putty/settings.c

1335 строки
56 KiB
C
Исходник Обычный вид История

/*
* settings.c: read and write saved sessions. (platform-independent)
*/
#include <assert.h>
#include <stdio.h>
#include <stdlib.h>
#include "putty.h"
#include "storage.h"
Support GSS key exchange, for Kerberos 5 only. This is a heavily edited (by me) version of a patch originally due to Nico Williams and Viktor Dukhovni. Their comments: * Don't delegate credentials when rekeying unless there's a new TGT or the old service ticket is nearly expired. * Check for the above conditions more frequently (every two minutes by default) and rekey when we would delegate credentials. * Do not rekey with very short service ticket lifetimes; some GSSAPI libraries may lose the race to use an almost expired ticket. Adjust the timing of rekey checks to try to avoid this possibility. My further comments: The most interesting thing about this patch to me is that the use of GSS key exchange causes a switch over to a completely different model of what host keys are for. This comes from RFC 4462 section 2.1: the basic idea is that when your session is mostly bidirectionally authenticated by the GSSAPI exchanges happening in initial kex and every rekey, host keys become more or less vestigial, and their remaining purpose is to allow a rekey to happen if the requirements of the SSH protocol demand it at an awkward moment when the GSS credentials are not currently available (e.g. timed out and haven't been renewed yet). As such, there's no need for host keys to be _permanent_ or to be a reliable identifier of a particular host, and RFC 4462 allows for the possibility that they might be purely transient and only for this kind of emergency fallback purpose. Therefore, once PuTTY has done a GSS key exchange, it disconnects itself completely from the permanent host key cache functions in storage.h, and instead switches to a _transient_ host key cache stored in memory with the lifetime of just that SSH session. That cache is populated with keys received from the server as a side effect of GSS kex (via the optional SSH2_MSG_KEXGSS_HOSTKEY message), and used if later in the session we have to fall back to a non-GSS key exchange. However, in practice servers we've tested against do not send a host key in that way, so we also have a fallback method of populating the transient cache by triggering an immediate non-GSS rekey straight after userauth (reusing the code path we also use to turn on OpenSSH delayed encryption without the race condition).
2018-04-26 09:18:59 +03:00
#ifndef NO_GSSAPI
#include "sshgssc.h"
#include "sshgss.h"
#endif
/* The cipher order given here is the default order. */
static const struct keyvalwhere ciphernames[] = {
{ "aes", CIPHER_AES, -1, -1 },
{ "chacha20", CIPHER_CHACHA20, CIPHER_AES, +1 },
{ "3des", CIPHER_3DES, -1, -1 },
{ "WARN", CIPHER_WARN, -1, -1 },
{ "des", CIPHER_DES, -1, -1 },
{ "blowfish", CIPHER_BLOWFISH, -1, -1 },
{ "arcfour", CIPHER_ARCFOUR, -1, -1 },
};
/* The default order here is sometimes overridden by the backward-
* compatibility warts in load_open_settings(), and should be kept
* in sync with those. */
static const struct keyvalwhere kexnames[] = {
{ "ecdh", KEX_ECDH, -1, +1 },
/* This name is misleading: it covers both SHA-256 and SHA-1 variants */
{ "dh-gex-sha1", KEX_DHGEX, -1, -1 },
{ "dh-group14-sha1", KEX_DHGROUP14, -1, -1 },
{ "dh-group1-sha1", KEX_DHGROUP1, KEX_WARN, +1 },
{ "rsa", KEX_RSA, KEX_WARN, -1 },
{ "WARN", KEX_WARN, -1, -1 }
};
static const struct keyvalwhere hknames[] = {
{ "ed25519", HK_ED25519, -1, +1 },
{ "ecdsa", HK_ECDSA, -1, -1 },
{ "dsa", HK_DSA, -1, -1 },
{ "rsa", HK_RSA, -1, -1 },
{ "WARN", HK_WARN, -1, -1 },
};
/*
* All the terminal modes that we know about for the "TerminalModes"
* setting. (Also used by config.c for the drop-down list.)
* This is currently precisely the same as the set in ssh.c, but could
* in principle differ if other backends started to support tty modes
* (e.g., the pty backend).
* The set of modes in in this array is currently significant for
* settings migration from old versions; if they change, review the
* gppmap() invocation for "TerminalModes".
*/
const char *const ttymodes[] = {
"INTR", "QUIT", "ERASE", "KILL", "EOF",
"EOL", "EOL2", "START", "STOP", "SUSP",
"DSUSP", "REPRINT", "WERASE", "LNEXT", "FLUSH",
"SWTCH", "STATUS", "DISCARD", "IGNPAR", "PARMRK",
"INPCK", "ISTRIP", "INLCR", "IGNCR", "ICRNL",
"IUCLC", "IXON", "IXANY", "IXOFF", "IMAXBEL",
"IUTF8", "ISIG", "ICANON", "XCASE", "ECHO",
"ECHOE", "ECHOK", "ECHONL", "NOFLSH", "TOSTOP",
"IEXTEN", "ECHOCTL", "ECHOKE", "PENDIN", "OPOST",
"OLCUC", "ONLCR", "OCRNL", "ONOCR", "ONLRET",
"CS7", "CS8", "PARENB", "PARODD", NULL
};
/*
* Convenience functions to access the backends[] array
* (which is only present in tools that manage settings).
*/
const struct BackendVtable *backend_vt_from_name(const char *name)
{
const struct BackendVtable *const *p;
for (p = backends; *p != NULL; p++)
if (!strcmp((*p)->name, name))
return *p;
return NULL;
}
const struct BackendVtable *backend_vt_from_proto(int proto)
{
const struct BackendVtable *const *p;
for (p = backends; *p != NULL; p++)
if ((*p)->protocol == proto)
return *p;
return NULL;
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
char *get_remote_username(Conf *conf)
{
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
char *username = conf_get_str(conf, CONF_username);
if (*username) {
return dupstr(username);
} else if (conf_get_bool(conf, CONF_username_from_env)) {
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
/* Use local username. */
return get_username(); /* might still be NULL */
} else {
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
return NULL;
}
}
static char *gpps_raw(settings_r *sesskey, const char *name, const char *def)
{
char *ret = sesskey ? read_setting_s(sesskey, name) : NULL;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
if (!ret)
ret = platform_default_s(name);
if (!ret)
ret = def ? dupstr(def) : NULL; /* permit NULL as final fallback */
return ret;
}
static void gpps(settings_r *sesskey, const char *name, const char *def,
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
Conf *conf, int primary)
{
char *val = gpps_raw(sesskey, name, def);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_str(conf, primary, val);
sfree(val);
}
/*
* gppfont and gppfile cannot have local defaults, since the very
* format of a Filename or FontSpec is platform-dependent. So the
* platform-dependent functions MUST return some sort of value.
*/
static void gppfont(settings_r *sesskey, char *name,
Conf *conf, int primary)
{
FontSpec *result = sesskey ? read_setting_fontspec(sesskey, name) : NULL;
if (!result)
result = platform_default_fontspec(name);
conf_set_fontspec(conf, primary, result);
fontspec_free(result);
}
static void gppfile(settings_r *sesskey, const char *name,
Conf *conf, int primary)
{
Filename *result = sesskey ? read_setting_filename(sesskey, name) : NULL;
if (!result)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
result = platform_default_filename(name);
conf_set_filename(conf, primary, result);
filename_free(result);
}
static bool gppb_raw(settings_r *sesskey, const char *name, bool def)
{
def = platform_default_b(name, def);
return sesskey ? read_setting_i(sesskey, name, def) != 0 : def;
}
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
static void gppb(settings_r *sesskey, const char *name, bool def,
Conf *conf, int primary)
{
conf_set_bool(conf, primary, gppb_raw(sesskey, name, def));
}
static int gppi_raw(settings_r *sesskey, const char *name, int def)
{
def = platform_default_i(name, def);
return sesskey ? read_setting_i(sesskey, name, def) : def;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
}
static void gppi(settings_r *sesskey, const char *name, int def,
Conf *conf, int primary)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
{
conf_set_int(conf, primary, gppi_raw(sesskey, name, def));
}
/*
* Read a set of name-value pairs in the format we occasionally use:
* NAME\tVALUE\0NAME\tVALUE\0\0 in memory
* NAME=VALUE,NAME=VALUE, in storage
* If there's no "=VALUE" (e.g. just NAME,NAME,NAME) then those keys
* are mapped to the empty string.
*/
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
static bool gppmap(settings_r *sesskey, const char *name,
Conf *conf, int primary)
{
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
char *buf, *p, *q, *key, *val;
/*
* Start by clearing any existing subkeys of this key from conf.
*/
while ((key = conf_get_str_nthstrkey(conf, primary, 0)) != NULL)
conf_del_str_str(conf, primary, key);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
/*
* Now read a serialised list from the settings and unmarshal it
* into its components.
*/
buf = gpps_raw(sesskey, name, NULL);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
if (!buf)
return false;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
p = buf;
while (*p) {
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
q = buf;
val = NULL;
while (*p && *p != ',') {
int c = *p++;
if (c == '=')
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
c = '\0';
if (c == '\\')
c = *p++;
*q++ = c;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
if (!c)
val = q;
}
if (*p == ',')
p++;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
if (!val)
val = q;
*q = '\0';
if (primary == CONF_portfwd && strchr(buf, 'D') != NULL) {
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
/*
* Backwards-compatibility hack: dynamic forwardings are
* indexed in the data store as a third type letter in the
* key, 'D' alongside 'L' and 'R' - but really, they
* should be filed under 'L' with a special _value_,
* because local and dynamic forwardings both involve
* _listening_ on a local port, and are hence mutually
* exclusive on the same port number. So here we translate
* the legacy storage format into the sensible internal
* form, by finding the D and turning it into a L.
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
*/
char *newkey = dupstr(buf);
*strchr(newkey, 'D') = 'L';
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_str_str(conf, primary, newkey, "D");
sfree(newkey);
} else {
conf_set_str_str(conf, primary, buf, val);
}
}
sfree(buf);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
return true;
}
/*
* Write a set of name/value pairs in the above format, or just the
* names if include_values is false.
*/
static void wmap(settings_w *sesskey, char const *outkey, Conf *conf,
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
int primary, bool include_values)
{
char *buf, *p, *key, *realkey;
const char *val, *q;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
int len;
len = 1; /* allow for NUL */
for (val = conf_get_str_strs(conf, primary, NULL, &key);
val != NULL;
val = conf_get_str_strs(conf, primary, key, &key))
len += 2 + 2 * (strlen(key) + strlen(val)); /* allow for escaping */
buf = snewn(len, char);
p = buf;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
for (val = conf_get_str_strs(conf, primary, NULL, &key);
val != NULL;
val = conf_get_str_strs(conf, primary, key, &key)) {
if (primary == CONF_portfwd && !strcmp(val, "D")) {
/*
* Backwards-compatibility hack, as above: translate from
* the sensible internal representation of dynamic
* forwardings (key "L<port>", value "D") to the
* conceptually incoherent legacy storage format (key
* "D<port>", value empty).
*/
char *L;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
realkey = key; /* restore it at end of loop */
val = "";
key = dupstr(key);
L = strchr(key, 'L');
if (L) *L = 'D';
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
} else {
realkey = NULL;
}
if (p != buf)
*p++ = ',';
for (q = key; *q; q++) {
if (*q == '=' || *q == ',' || *q == '\\')
*p++ = '\\';
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
*p++ = *q;
}
if (include_values) {
*p++ = '=';
for (q = val; *q; q++) {
if (*q == '=' || *q == ',' || *q == '\\')
*p++ = '\\';
*p++ = *q;
}
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
if (realkey) {
free(key);
key = realkey;
}
}
*p = '\0';
write_setting_s(sesskey, outkey, buf);
sfree(buf);
}
static int key2val(const struct keyvalwhere *mapping,
int nmaps, char *key)
{
int i;
for (i = 0; i < nmaps; i++)
if (!strcmp(mapping[i].s, key)) return mapping[i].v;
return -1;
}
static const char *val2key(const struct keyvalwhere *mapping,
int nmaps, int val)
{
int i;
for (i = 0; i < nmaps; i++)
if (mapping[i].v == val) return mapping[i].s;
return NULL;
}
/*
* Helper function to parse a comma-separated list of strings into
* a preference list array of values. Any missing values are added
* to the end and duplicates are weeded.
* XXX: assumes vals in 'mapping' are small +ve integers
*/
static void gprefs_from_str(const char *str,
const struct keyvalwhere *mapping, int nvals,
Conf *conf, int primary)
{
char *commalist = dupstr(str);
char *p, *q;
int i, j, n, v, pos;
unsigned long seen = 0; /* bitmap for weeding dups etc */
/*
* Go through that list and convert it into values.
*/
n = 0;
p = commalist;
while (1) {
while (*p && *p == ',') p++;
if (!*p)
break; /* no more words */
q = p;
while (*p && *p != ',') p++;
if (*p) *p++ = '\0';
v = key2val(mapping, nvals, q);
if (v != -1 && !(seen & (1 << v))) {
seen |= (1 << v);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int_int(conf, primary, n, v);
n++;
}
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
sfree(commalist);
/*
* Now go through 'mapping' and add values that weren't mentioned
* in the list we fetched. We may have to loop over it multiple
* times so that we add values before other values whose default
* positions depend on them.
*/
while (n < nvals) {
for (i = 0; i < nvals; i++) {
assert(mapping[i].v >= 0);
assert(mapping[i].v < 32);
if (!(seen & (1 << mapping[i].v))) {
/*
* This element needs adding. But can we add it yet?
*/
if (mapping[i].vrel != -1 && !(seen & (1 << mapping[i].vrel)))
continue; /* nope */
/*
* OK, we can work out where to add this element, so
* do so.
*/
if (mapping[i].vrel == -1) {
pos = (mapping[i].where < 0 ? n : 0);
} else {
for (j = 0; j < n; j++)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
if (conf_get_int_int(conf, primary, j) ==
mapping[i].vrel)
break;
assert(j < n); /* implied by (seen & (1<<vrel)) */
pos = (mapping[i].where < 0 ? j : j+1);
}
/*
* And add it.
*/
for (j = n-1; j >= pos; j--)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int_int(conf, primary, j+1,
conf_get_int_int(conf, primary, j));
conf_set_int_int(conf, primary, pos, mapping[i].v);
seen |= (1 << mapping[i].v);
n++;
}
}
}
}
/*
* Read a preference list.
*/
static void gprefs(settings_r *sesskey, const char *name, const char *def,
const struct keyvalwhere *mapping, int nvals,
Conf *conf, int primary)
{
/*
* Fetch the string which we'll parse as a comma-separated list.
*/
char *value = gpps_raw(sesskey, name, def);
gprefs_from_str(value, mapping, nvals, conf, primary);
sfree(value);
}
/*
* Write out a preference list.
*/
static void wprefs(settings_w *sesskey, const char *name,
const struct keyvalwhere *mapping, int nvals,
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
Conf *conf, int primary)
{
char *buf, *p;
int i, maxlen;
for (maxlen = i = 0; i < nvals; i++) {
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
const char *s = val2key(mapping, nvals,
conf_get_int_int(conf, primary, i));
if (s) {
maxlen += (maxlen > 0 ? 1 : 0) + strlen(s);
}
}
buf = snewn(maxlen + 1, char);
p = buf;
for (i = 0; i < nvals; i++) {
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
const char *s = val2key(mapping, nvals,
conf_get_int_int(conf, primary, i));
if (s) {
p += sprintf(p, "%s%s", (p > buf ? "," : ""), s);
}
}
assert(p - buf == maxlen);
*p = '\0';
write_setting_s(sesskey, name, buf);
sfree(buf);
}
static void write_setting_b(settings_w *handle, const char *key, bool value)
{
write_setting_i(handle, key, value ? 1 : 0);
}
static void write_clip_setting(settings_w *sesskey, const char *savekey,
Conf *conf, int confkey, int strconfkey)
{
int val = conf_get_int(conf, confkey);
switch (val) {
case CLIPUI_NONE:
default:
write_setting_s(sesskey, savekey, "none");
break;
case CLIPUI_IMPLICIT:
write_setting_s(sesskey, savekey, "implicit");
break;
case CLIPUI_EXPLICIT:
write_setting_s(sesskey, savekey, "explicit");
break;
case CLIPUI_CUSTOM:
{
char *sval = dupcat("custom:", conf_get_str(conf, strconfkey),
(const char *)NULL);
write_setting_s(sesskey, savekey, sval);
sfree(sval);
}
break;
}
}
static void read_clip_setting(settings_r *sesskey, char *savekey,
int def, Conf *conf, int confkey, int strconfkey)
{
char *setting = read_setting_s(sesskey, savekey);
int val;
conf_set_str(conf, strconfkey, "");
if (!setting) {
val = def;
} else if (!strcmp(setting, "implicit")) {
val = CLIPUI_IMPLICIT;
} else if (!strcmp(setting, "explicit")) {
val = CLIPUI_EXPLICIT;
} else if (!strncmp(setting, "custom:", 7)) {
val = CLIPUI_CUSTOM;
conf_set_str(conf, strconfkey, setting + 7);
} else {
val = CLIPUI_NONE;
}
conf_set_int(conf, confkey, val);
sfree(setting);
}
char *save_settings(const char *section, Conf *conf)
{
struct settings_w *sesskey;
char *errmsg;
sesskey = open_settings_w(section, &errmsg);
if (!sesskey)
return errmsg;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
save_open_settings(sesskey, conf);
close_settings_w(sesskey);
return NULL;
}
void save_open_settings(settings_w *sesskey, Conf *conf)
{
int i;
const char *p;
write_setting_i(sesskey, "Present", 1);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "HostName", conf_get_str(conf, CONF_host));
write_setting_filename(sesskey, "LogFileName", conf_get_filename(conf, CONF_logfilename));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "LogType", conf_get_int(conf, CONF_logtype));
write_setting_i(sesskey, "LogFileClash", conf_get_int(conf, CONF_logxfovr));
write_setting_b(sesskey, "LogFlush", conf_get_bool(conf, CONF_logflush));
write_setting_b(sesskey, "LogHeader", conf_get_bool(conf, CONF_logheader));
write_setting_b(sesskey, "SSHLogOmitPasswords", conf_get_bool(conf, CONF_logomitpass));
write_setting_b(sesskey, "SSHLogOmitData", conf_get_bool(conf, CONF_logomitdata));
p = "raw";
{
const struct BackendVtable *vt =
backend_vt_from_proto(conf_get_int(conf, CONF_protocol));
if (vt)
p = vt->name;
}
write_setting_s(sesskey, "Protocol", p);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "PortNumber", conf_get_int(conf, CONF_port));
/* The CloseOnExit numbers are arranged in a different order from
* the standard FORCE_ON / FORCE_OFF / AUTO. */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "CloseOnExit", (conf_get_int(conf, CONF_close_on_exit)+2)%3);
write_setting_b(sesskey, "WarnOnClose", !!conf_get_bool(conf, CONF_warn_on_close));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "PingInterval", conf_get_int(conf, CONF_ping_interval) / 60); /* minutes */
write_setting_i(sesskey, "PingIntervalSecs", conf_get_int(conf, CONF_ping_interval) % 60); /* seconds */
write_setting_b(sesskey, "TCPNoDelay", conf_get_bool(conf, CONF_tcp_nodelay));
write_setting_b(sesskey, "TCPKeepalives", conf_get_bool(conf, CONF_tcp_keepalives));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "TerminalType", conf_get_str(conf, CONF_termtype));
write_setting_s(sesskey, "TerminalSpeed", conf_get_str(conf, CONF_termspeed));
wmap(sesskey, "TerminalModes", conf, CONF_ttymodes, true);
/* Address family selection */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "AddressFamily", conf_get_int(conf, CONF_addressfamily));
/* proxy settings */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "ProxyExcludeList", conf_get_str(conf, CONF_proxy_exclude_list));
write_setting_i(sesskey, "ProxyDNS", (conf_get_int(conf, CONF_proxy_dns)+2)%3);
write_setting_b(sesskey, "ProxyLocalhost", conf_get_bool(conf, CONF_even_proxy_localhost));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "ProxyMethod", conf_get_int(conf, CONF_proxy_type));
write_setting_s(sesskey, "ProxyHost", conf_get_str(conf, CONF_proxy_host));
write_setting_i(sesskey, "ProxyPort", conf_get_int(conf, CONF_proxy_port));
write_setting_s(sesskey, "ProxyUsername", conf_get_str(conf, CONF_proxy_username));
write_setting_s(sesskey, "ProxyPassword", conf_get_str(conf, CONF_proxy_password));
write_setting_s(sesskey, "ProxyTelnetCommand", conf_get_str(conf, CONF_proxy_telnet_command));
write_setting_i(sesskey, "ProxyLogToTerm", conf_get_int(conf, CONF_proxy_log_to_term));
wmap(sesskey, "Environment", conf, CONF_environmt, true);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "UserName", conf_get_str(conf, CONF_username));
write_setting_b(sesskey, "UserNameFromEnvironment", conf_get_bool(conf, CONF_username_from_env));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "LocalUserName", conf_get_str(conf, CONF_localusername));
write_setting_b(sesskey, "NoPTY", conf_get_bool(conf, CONF_nopty));
write_setting_b(sesskey, "Compression", conf_get_bool(conf, CONF_compression));
write_setting_b(sesskey, "TryAgent", conf_get_bool(conf, CONF_tryagent));
write_setting_b(sesskey, "AgentFwd", conf_get_bool(conf, CONF_agentfwd));
#ifndef NO_GSSAPI
write_setting_b(sesskey, "GssapiFwd", conf_get_bool(conf, CONF_gssapifwd));
#endif
write_setting_b(sesskey, "ChangeUsername", conf_get_bool(conf, CONF_change_username));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
wprefs(sesskey, "Cipher", ciphernames, CIPHER_MAX, conf, CONF_ssh_cipherlist);
wprefs(sesskey, "KEX", kexnames, KEX_MAX, conf, CONF_ssh_kexlist);
wprefs(sesskey, "HostKey", hknames, HK_MAX, conf, CONF_ssh_hklist);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "RekeyTime", conf_get_int(conf, CONF_ssh_rekey_time));
#ifndef NO_GSSAPI
Support GSS key exchange, for Kerberos 5 only. This is a heavily edited (by me) version of a patch originally due to Nico Williams and Viktor Dukhovni. Their comments: * Don't delegate credentials when rekeying unless there's a new TGT or the old service ticket is nearly expired. * Check for the above conditions more frequently (every two minutes by default) and rekey when we would delegate credentials. * Do not rekey with very short service ticket lifetimes; some GSSAPI libraries may lose the race to use an almost expired ticket. Adjust the timing of rekey checks to try to avoid this possibility. My further comments: The most interesting thing about this patch to me is that the use of GSS key exchange causes a switch over to a completely different model of what host keys are for. This comes from RFC 4462 section 2.1: the basic idea is that when your session is mostly bidirectionally authenticated by the GSSAPI exchanges happening in initial kex and every rekey, host keys become more or less vestigial, and their remaining purpose is to allow a rekey to happen if the requirements of the SSH protocol demand it at an awkward moment when the GSS credentials are not currently available (e.g. timed out and haven't been renewed yet). As such, there's no need for host keys to be _permanent_ or to be a reliable identifier of a particular host, and RFC 4462 allows for the possibility that they might be purely transient and only for this kind of emergency fallback purpose. Therefore, once PuTTY has done a GSS key exchange, it disconnects itself completely from the permanent host key cache functions in storage.h, and instead switches to a _transient_ host key cache stored in memory with the lifetime of just that SSH session. That cache is populated with keys received from the server as a side effect of GSS kex (via the optional SSH2_MSG_KEXGSS_HOSTKEY message), and used if later in the session we have to fall back to a non-GSS key exchange. However, in practice servers we've tested against do not send a host key in that way, so we also have a fallback method of populating the transient cache by triggering an immediate non-GSS rekey straight after userauth (reusing the code path we also use to turn on OpenSSH delayed encryption without the race condition).
2018-04-26 09:18:59 +03:00
write_setting_i(sesskey, "GssapiRekey", conf_get_int(conf, CONF_gssapirekey));
#endif
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "RekeyBytes", conf_get_str(conf, CONF_ssh_rekey_data));
write_setting_b(sesskey, "SshNoAuth", conf_get_bool(conf, CONF_ssh_no_userauth));
write_setting_b(sesskey, "SshBanner", conf_get_bool(conf, CONF_ssh_show_banner));
write_setting_b(sesskey, "AuthTIS", conf_get_bool(conf, CONF_try_tis_auth));
write_setting_b(sesskey, "AuthKI", conf_get_bool(conf, CONF_try_ki_auth));
#ifndef NO_GSSAPI
write_setting_b(sesskey, "AuthGSSAPI", conf_get_bool(conf, CONF_try_gssapi_auth));
write_setting_b(sesskey, "AuthGSSAPIKEX", conf_get_bool(conf, CONF_try_gssapi_kex));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
wprefs(sesskey, "GSSLibs", gsslibkeywords, ngsslibs, conf, CONF_ssh_gsslist);
write_setting_filename(sesskey, "GSSCustom", conf_get_filename(conf, CONF_ssh_gss_custom));
#endif
write_setting_b(sesskey, "SshNoShell", conf_get_bool(conf, CONF_ssh_no_shell));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "SshProt", conf_get_int(conf, CONF_sshprot));
write_setting_s(sesskey, "LogHost", conf_get_str(conf, CONF_loghost));
write_setting_b(sesskey, "SSH2DES", conf_get_bool(conf, CONF_ssh2_des_cbc));
write_setting_filename(sesskey, "PublicKeyFile", conf_get_filename(conf, CONF_keyfile));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "RemoteCommand", conf_get_str(conf, CONF_remote_cmd));
write_setting_b(sesskey, "RFCEnviron", conf_get_bool(conf, CONF_rfc_environ));
write_setting_b(sesskey, "PassiveTelnet", conf_get_bool(conf, CONF_passive_telnet));
write_setting_b(sesskey, "BackspaceIsDelete", conf_get_bool(conf, CONF_bksp_is_delete));
write_setting_b(sesskey, "RXVTHomeEnd", conf_get_bool(conf, CONF_rxvt_homeend));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "LinuxFunctionKeys", conf_get_int(conf, CONF_funky_type));
write_setting_b(sesskey, "NoApplicationKeys", conf_get_bool(conf, CONF_no_applic_k));
write_setting_b(sesskey, "NoApplicationCursors", conf_get_bool(conf, CONF_no_applic_c));
write_setting_b(sesskey, "NoMouseReporting", conf_get_bool(conf, CONF_no_mouse_rep));
write_setting_b(sesskey, "NoRemoteResize", conf_get_bool(conf, CONF_no_remote_resize));
write_setting_b(sesskey, "NoAltScreen", conf_get_bool(conf, CONF_no_alt_screen));
write_setting_b(sesskey, "NoRemoteWinTitle", conf_get_bool(conf, CONF_no_remote_wintitle));
write_setting_b(sesskey, "NoRemoteClearScroll", conf_get_bool(conf, CONF_no_remote_clearscroll));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "RemoteQTitleAction", conf_get_int(conf, CONF_remote_qtitle_action));
write_setting_b(sesskey, "NoDBackspace", conf_get_bool(conf, CONF_no_dbackspace));
write_setting_b(sesskey, "NoRemoteCharset", conf_get_bool(conf, CONF_no_remote_charset));
write_setting_b(sesskey, "ApplicationCursorKeys", conf_get_bool(conf, CONF_app_cursor));
write_setting_b(sesskey, "ApplicationKeypad", conf_get_bool(conf, CONF_app_keypad));
write_setting_b(sesskey, "NetHackKeypad", conf_get_bool(conf, CONF_nethack_keypad));
write_setting_b(sesskey, "AltF4", conf_get_bool(conf, CONF_alt_f4));
write_setting_b(sesskey, "AltSpace", conf_get_bool(conf, CONF_alt_space));
write_setting_b(sesskey, "AltOnly", conf_get_bool(conf, CONF_alt_only));
write_setting_b(sesskey, "ComposeKey", conf_get_bool(conf, CONF_compose_key));
write_setting_b(sesskey, "CtrlAltKeys", conf_get_bool(conf, CONF_ctrlaltkeys));
#ifdef OSX_META_KEY_CONFIG
write_setting_b(sesskey, "OSXOptionMeta", conf_get_bool(conf, CONF_osx_option_meta));
write_setting_b(sesskey, "OSXCommandMeta", conf_get_bool(conf, CONF_osx_command_meta));
#endif
write_setting_b(sesskey, "TelnetKey", conf_get_bool(conf, CONF_telnet_keyboard));
write_setting_b(sesskey, "TelnetRet", conf_get_bool(conf, CONF_telnet_newline));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "LocalEcho", conf_get_int(conf, CONF_localecho));
write_setting_i(sesskey, "LocalEdit", conf_get_int(conf, CONF_localedit));
write_setting_s(sesskey, "Answerback", conf_get_str(conf, CONF_answerback));
write_setting_b(sesskey, "AlwaysOnTop", conf_get_bool(conf, CONF_alwaysontop));
write_setting_b(sesskey, "FullScreenOnAltEnter", conf_get_bool(conf, CONF_fullscreenonaltenter));
write_setting_b(sesskey, "HideMousePtr", conf_get_bool(conf, CONF_hide_mouseptr));
write_setting_b(sesskey, "SunkenEdge", conf_get_bool(conf, CONF_sunken_edge));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "WindowBorder", conf_get_int(conf, CONF_window_border));
write_setting_i(sesskey, "CurType", conf_get_int(conf, CONF_cursor_type));
write_setting_b(sesskey, "BlinkCur", conf_get_bool(conf, CONF_blink_cur));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "Beep", conf_get_int(conf, CONF_beep));
write_setting_i(sesskey, "BeepInd", conf_get_int(conf, CONF_beep_ind));
write_setting_filename(sesskey, "BellWaveFile", conf_get_filename(conf, CONF_bell_wavefile));
write_setting_b(sesskey, "BellOverload", conf_get_bool(conf, CONF_bellovl));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "BellOverloadN", conf_get_int(conf, CONF_bellovl_n));
write_setting_i(sesskey, "BellOverloadT", conf_get_int(conf, CONF_bellovl_t)
#ifdef PUTTY_UNIX_H
* 1000
#endif
);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "BellOverloadS", conf_get_int(conf, CONF_bellovl_s)
#ifdef PUTTY_UNIX_H
* 1000
#endif
);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "ScrollbackLines", conf_get_int(conf, CONF_savelines));
write_setting_b(sesskey, "DECOriginMode", conf_get_bool(conf, CONF_dec_om));
write_setting_b(sesskey, "AutoWrapMode", conf_get_bool(conf, CONF_wrap_mode));
write_setting_b(sesskey, "LFImpliesCR", conf_get_bool(conf, CONF_lfhascr));
write_setting_b(sesskey, "CRImpliesLF", conf_get_bool(conf, CONF_crhaslf));
write_setting_b(sesskey, "DisableArabicShaping", conf_get_bool(conf, CONF_no_arabicshaping));
write_setting_b(sesskey, "DisableBidi", conf_get_bool(conf, CONF_no_bidi));
write_setting_b(sesskey, "WinNameAlways", conf_get_bool(conf, CONF_win_name_always));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "WinTitle", conf_get_str(conf, CONF_wintitle));
write_setting_i(sesskey, "TermWidth", conf_get_int(conf, CONF_width));
write_setting_i(sesskey, "TermHeight", conf_get_int(conf, CONF_height));
write_setting_fontspec(sesskey, "Font", conf_get_fontspec(conf, CONF_font));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "FontQuality", conf_get_int(conf, CONF_font_quality));
write_setting_i(sesskey, "FontVTMode", conf_get_int(conf, CONF_vtmode));
write_setting_b(sesskey, "UseSystemColours", conf_get_bool(conf, CONF_system_colour));
write_setting_b(sesskey, "TryPalette", conf_get_bool(conf, CONF_try_palette));
write_setting_b(sesskey, "ANSIColour", conf_get_bool(conf, CONF_ansi_colour));
write_setting_b(sesskey, "Xterm256Colour", conf_get_bool(conf, CONF_xterm_256_colour));
write_setting_b(sesskey, "TrueColour", conf_get_bool(conf, CONF_true_colour));
write_setting_i(sesskey, "BoldAsColour", conf_get_int(conf, CONF_bold_style)-1);
for (i = 0; i < 22; i++) {
char buf[20], buf2[30];
sprintf(buf, "Colour%d", i);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
sprintf(buf2, "%d,%d,%d",
conf_get_int_int(conf, CONF_colours, i*3+0),
conf_get_int_int(conf, CONF_colours, i*3+1),
conf_get_int_int(conf, CONF_colours, i*3+2));
write_setting_s(sesskey, buf, buf2);
}
write_setting_b(sesskey, "RawCNP", conf_get_bool(conf, CONF_rawcnp));
write_setting_b(sesskey, "UTF8linedraw", conf_get_bool(conf, CONF_utf8linedraw));
write_setting_b(sesskey, "PasteRTF", conf_get_bool(conf, CONF_rtf_paste));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "MouseIsXterm", conf_get_int(conf, CONF_mouse_is_xterm));
write_setting_b(sesskey, "RectSelect", conf_get_bool(conf, CONF_rect_select));
write_setting_b(sesskey, "PasteControls", conf_get_bool(conf, CONF_paste_controls));
write_setting_b(sesskey, "MouseOverride", conf_get_bool(conf, CONF_mouse_override));
for (i = 0; i < 256; i += 32) {
char buf[20], buf2[256];
int j;
sprintf(buf, "Wordness%d", i);
*buf2 = '\0';
for (j = i; j < i + 32; j++) {
sprintf(buf2 + strlen(buf2), "%s%d",
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
(*buf2 ? "," : ""),
conf_get_int_int(conf, CONF_wordness, j));
}
write_setting_s(sesskey, buf, buf2);
}
write_setting_b(sesskey, "MouseAutocopy",
conf_get_bool(conf, CONF_mouseautocopy));
write_clip_setting(sesskey, "MousePaste", conf,
CONF_mousepaste, CONF_mousepaste_custom);
write_clip_setting(sesskey, "CtrlShiftIns", conf,
CONF_ctrlshiftins, CONF_ctrlshiftins_custom);
write_clip_setting(sesskey, "CtrlShiftCV", conf,
CONF_ctrlshiftcv, CONF_ctrlshiftcv_custom);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "LineCodePage", conf_get_str(conf, CONF_line_codepage));
write_setting_b(sesskey, "CJKAmbigWide", conf_get_bool(conf, CONF_cjk_ambig_wide));
write_setting_b(sesskey, "UTF8Override", conf_get_bool(conf, CONF_utf8_override));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "Printer", conf_get_str(conf, CONF_printer));
write_setting_b(sesskey, "CapsLockCyr", conf_get_bool(conf, CONF_xlat_capslockcyr));
write_setting_b(sesskey, "ScrollBar", conf_get_bool(conf, CONF_scrollbar));
write_setting_b(sesskey, "ScrollBarFullScreen", conf_get_bool(conf, CONF_scrollbar_in_fullscreen));
write_setting_b(sesskey, "ScrollOnKey", conf_get_bool(conf, CONF_scroll_on_key));
write_setting_b(sesskey, "ScrollOnDisp", conf_get_bool(conf, CONF_scroll_on_disp));
write_setting_b(sesskey, "EraseToScrollback", conf_get_bool(conf, CONF_erase_to_scrollback));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "LockSize", conf_get_int(conf, CONF_resize_action));
write_setting_b(sesskey, "BCE", conf_get_bool(conf, CONF_bce));
write_setting_b(sesskey, "BlinkText", conf_get_bool(conf, CONF_blinktext));
write_setting_b(sesskey, "X11Forward", conf_get_bool(conf, CONF_x11_forward));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_s(sesskey, "X11Display", conf_get_str(conf, CONF_x11_display));
write_setting_i(sesskey, "X11AuthType", conf_get_int(conf, CONF_x11_auth));
write_setting_filename(sesskey, "X11AuthFile", conf_get_filename(conf, CONF_xauthfile));
write_setting_b(sesskey, "LocalPortAcceptAll", conf_get_bool(conf, CONF_lport_acceptall));
write_setting_b(sesskey, "RemotePortAcceptAll", conf_get_bool(conf, CONF_rport_acceptall));
wmap(sesskey, "PortForwardings", conf, CONF_portfwd, true);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "BugIgnore1", 2-conf_get_int(conf, CONF_sshbug_ignore1));
write_setting_i(sesskey, "BugPlainPW1", 2-conf_get_int(conf, CONF_sshbug_plainpw1));
write_setting_i(sesskey, "BugRSA1", 2-conf_get_int(conf, CONF_sshbug_rsa1));
write_setting_i(sesskey, "BugIgnore2", 2-conf_get_int(conf, CONF_sshbug_ignore2));
write_setting_i(sesskey, "BugHMAC2", 2-conf_get_int(conf, CONF_sshbug_hmac2));
write_setting_i(sesskey, "BugDeriveKey2", 2-conf_get_int(conf, CONF_sshbug_derivekey2));
write_setting_i(sesskey, "BugRSAPad2", 2-conf_get_int(conf, CONF_sshbug_rsapad2));
write_setting_i(sesskey, "BugPKSessID2", 2-conf_get_int(conf, CONF_sshbug_pksessid2));
write_setting_i(sesskey, "BugRekey2", 2-conf_get_int(conf, CONF_sshbug_rekey2));
write_setting_i(sesskey, "BugMaxPkt2", 2-conf_get_int(conf, CONF_sshbug_maxpkt2));
write_setting_i(sesskey, "BugOldGex2", 2-conf_get_int(conf, CONF_sshbug_oldgex2));
write_setting_i(sesskey, "BugWinadj", 2-conf_get_int(conf, CONF_sshbug_winadj));
write_setting_i(sesskey, "BugChanReq", 2-conf_get_int(conf, CONF_sshbug_chanreq));
write_setting_b(sesskey, "StampUtmp", conf_get_bool(conf, CONF_stamp_utmp));
write_setting_b(sesskey, "LoginShell", conf_get_bool(conf, CONF_login_shell));
write_setting_b(sesskey, "ScrollbarOnLeft", conf_get_bool(conf, CONF_scrollbar_on_left));
write_setting_fontspec(sesskey, "BoldFont", conf_get_fontspec(conf, CONF_boldfont));
write_setting_fontspec(sesskey, "WideFont", conf_get_fontspec(conf, CONF_widefont));
write_setting_fontspec(sesskey, "WideBoldFont", conf_get_fontspec(conf, CONF_wideboldfont));
write_setting_b(sesskey, "ShadowBold", conf_get_bool(conf, CONF_shadowbold));
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
write_setting_i(sesskey, "ShadowBoldOffset", conf_get_int(conf, CONF_shadowboldoffset));
write_setting_s(sesskey, "SerialLine", conf_get_str(conf, CONF_serline));
write_setting_i(sesskey, "SerialSpeed", conf_get_int(conf, CONF_serspeed));
write_setting_i(sesskey, "SerialDataBits", conf_get_int(conf, CONF_serdatabits));
write_setting_i(sesskey, "SerialStopHalfbits", conf_get_int(conf, CONF_serstopbits));
write_setting_i(sesskey, "SerialParity", conf_get_int(conf, CONF_serparity));
write_setting_i(sesskey, "SerialFlowControl", conf_get_int(conf, CONF_serflow));
write_setting_s(sesskey, "WindowClass", conf_get_str(conf, CONF_winclass));
write_setting_b(sesskey, "ConnectionSharing", conf_get_bool(conf, CONF_ssh_connection_sharing));
write_setting_b(sesskey, "ConnectionSharingUpstream", conf_get_bool(conf, CONF_ssh_connection_sharing_upstream));
write_setting_b(sesskey, "ConnectionSharingDownstream", conf_get_bool(conf, CONF_ssh_connection_sharing_downstream));
wmap(sesskey, "SSHManualHostKeys", conf, CONF_ssh_manual_hostkeys, false);
}
bool load_settings(const char *section, Conf *conf)
{
settings_r *sesskey;
sesskey = open_settings_r(section);
bool exists = (sesskey != NULL);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
load_open_settings(sesskey, conf);
close_settings_r(sesskey);
if (exists && conf_launchable(conf))
add_session_to_jumplist(section);
return exists;
}
void load_open_settings(settings_r *sesskey, Conf *conf)
{
int i;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
char *prot;
conf_set_bool(conf, CONF_ssh_subsys, false); /* FIXME: load this properly */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_str(conf, CONF_remote_cmd, "");
conf_set_str(conf, CONF_remote_cmd2, "");
conf_set_str(conf, CONF_ssh_nc_host, "");
gpps(sesskey, "HostName", "", conf, CONF_host);
gppfile(sesskey, "LogFileName", conf, CONF_logfilename);
gppi(sesskey, "LogType", 0, conf, CONF_logtype);
gppi(sesskey, "LogFileClash", LGXF_ASK, conf, CONF_logxfovr);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "LogFlush", true, conf, CONF_logflush);
gppb(sesskey, "LogHeader", true, conf, CONF_logheader);
gppb(sesskey, "SSHLogOmitPasswords", true, conf, CONF_logomitpass);
gppb(sesskey, "SSHLogOmitData", false, conf, CONF_logomitdata);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
prot = gpps_raw(sesskey, "Protocol", "default");
conf_set_int(conf, CONF_protocol, default_protocol);
conf_set_int(conf, CONF_port, default_port);
{
const struct BackendVtable *vt = backend_vt_from_name(prot);
if (vt) {
conf_set_int(conf, CONF_protocol, vt->protocol);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "PortNumber", default_port, conf, CONF_port);
}
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
sfree(prot);
/* Address family selection */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "AddressFamily", ADDRTYPE_UNSPEC, conf, CONF_addressfamily);
/* The CloseOnExit numbers are arranged in a different order from
* the standard FORCE_ON / FORCE_OFF / AUTO. */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
i = gppi_raw(sesskey, "CloseOnExit", 1); conf_set_int(conf, CONF_close_on_exit, (i+1)%3);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "WarnOnClose", true, conf, CONF_warn_on_close);
{
/* This is two values for backward compatibility with 0.50/0.51 */
int pingmin, pingsec;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
pingmin = gppi_raw(sesskey, "PingInterval", 0);
pingsec = gppi_raw(sesskey, "PingIntervalSecs", 0);
conf_set_int(conf, CONF_ping_interval, pingmin * 60 + pingsec);
}
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "TCPNoDelay", true, conf, CONF_tcp_nodelay);
gppb(sesskey, "TCPKeepalives", false, conf, CONF_tcp_keepalives);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "TerminalType", "xterm", conf, CONF_termtype);
gpps(sesskey, "TerminalSpeed", "38400,38400", conf, CONF_termspeed);
if (gppmap(sesskey, "TerminalModes", conf, CONF_ttymodes)) {
/*
* Backwards compatibility with old saved settings.
*
* From the invention of this setting through 0.67, the set of
* terminal modes was fixed, and absence of a mode from this
* setting meant the user had explicitly removed it from the
* UI and we shouldn't send it.
*
* In 0.68, the IUTF8 mode was added, and in handling old
* settings we inadvertently removed the ability to not send
* a mode. Any mode not mentioned was treated as if it was
* set to 'auto' (A).
*
* After 0.68, we added explicit notation to the setting format
* when the user removes a known terminal mode from the list.
*
* So: if any of the modes from the original set is missing, we
* assume this was an intentional removal by the user and add
* an explicit removal ('N'); but if IUTF8 (or any other mode
* added after 0.67) is missing, we assume that its absence is
* due to the setting being old rather than intentional, and
* add it with its default setting.
*
* (This does mean that if a 0.68 user explicitly removed IUTF8,
* we add it back; but removing IUTF8 had no effect in 0.68, so
* we're preserving behaviour, which is the best we can do.)
*/
for (i = 0; ttymodes[i]; i++) {
if (!conf_get_str_str_opt(conf, CONF_ttymodes, ttymodes[i])) {
/* Mode not mentioned in setting. */
const char *def;
if (!strcmp(ttymodes[i], "IUTF8")) {
/* Any new modes we add in future should be treated
* this way too. */
def = "A"; /* same as new-setting default below */
} else {
/* One of the original modes. Absence is probably
* deliberate. */
def = "N"; /* don't send */
}
conf_set_str_str(conf, CONF_ttymodes, ttymodes[i], def);
}
}
} else {
/* This hardcodes a big set of defaults in any new saved
* sessions. Let's hope we don't change our mind. */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
for (i = 0; ttymodes[i]; i++)
conf_set_str_str(conf, CONF_ttymodes, ttymodes[i], "A");
}
/* proxy settings */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "ProxyExcludeList", "", conf, CONF_proxy_exclude_list);
i = gppi_raw(sesskey, "ProxyDNS", 1); conf_set_int(conf, CONF_proxy_dns, (i+1)%3);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "ProxyLocalhost", false, conf, CONF_even_proxy_localhost);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "ProxyMethod", -1, conf, CONF_proxy_type);
if (conf_get_int(conf, CONF_proxy_type) == -1) {
int i;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
i = gppi_raw(sesskey, "ProxyType", 0);
if (i == 0)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int(conf, CONF_proxy_type, PROXY_NONE);
else if (i == 1)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int(conf, CONF_proxy_type, PROXY_HTTP);
else if (i == 3)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int(conf, CONF_proxy_type, PROXY_TELNET);
else if (i == 4)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int(conf, CONF_proxy_type, PROXY_CMD);
else {
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
i = gppi_raw(sesskey, "ProxySOCKSVersion", 5);
if (i == 5)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int(conf, CONF_proxy_type, PROXY_SOCKS5);
else
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int(conf, CONF_proxy_type, PROXY_SOCKS4);
}
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "ProxyHost", "proxy", conf, CONF_proxy_host);
gppi(sesskey, "ProxyPort", 80, conf, CONF_proxy_port);
gpps(sesskey, "ProxyUsername", "", conf, CONF_proxy_username);
gpps(sesskey, "ProxyPassword", "", conf, CONF_proxy_password);
gpps(sesskey, "ProxyTelnetCommand", "connect %host %port\\n",
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf, CONF_proxy_telnet_command);
gppi(sesskey, "ProxyLogToTerm", FORCE_OFF, conf, CONF_proxy_log_to_term);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppmap(sesskey, "Environment", conf, CONF_environmt);
gpps(sesskey, "UserName", "", conf, CONF_username);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "UserNameFromEnvironment", false,
conf, CONF_username_from_env);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "LocalUserName", "", conf, CONF_localusername);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "NoPTY", false, conf, CONF_nopty);
gppb(sesskey, "Compression", false, conf, CONF_compression);
gppb(sesskey, "TryAgent", true, conf, CONF_tryagent);
gppb(sesskey, "AgentFwd", false, conf, CONF_agentfwd);
gppb(sesskey, "ChangeUsername", false, conf, CONF_change_username);
#ifndef NO_GSSAPI
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "GssapiFwd", false, conf, CONF_gssapifwd);
#endif
gprefs(sesskey, "Cipher", "\0",
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
ciphernames, CIPHER_MAX, conf, CONF_ssh_cipherlist);
{
/* Backward-compatibility: before 0.58 (when the "KEX"
* preference was first added), we had an option to
* disable gex under the "bugs" panel after one report of
* a server which offered it then choked, but we never got
* a server version string or any other reports. */
const char *default_kexes,
*normal_default = "ecdh,dh-gex-sha1,dh-group14-sha1,rsa,"
"WARN,dh-group1-sha1",
*bugdhgex2_default = "ecdh,dh-group14-sha1,rsa,"
"WARN,dh-group1-sha1,dh-gex-sha1";
char *raw;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
i = 2 - gppi_raw(sesskey, "BugDHGEx2", 0);
if (i == FORCE_ON)
default_kexes = bugdhgex2_default;
else
default_kexes = normal_default;
/* Migration: after 0.67 we decided we didn't like
* dh-group1-sha1. If it looks like the user never changed
* the defaults, quietly upgrade their settings to demote it.
* (If they did, they're on their own.) */
raw = gpps_raw(sesskey, "KEX", default_kexes);
assert(raw != NULL);
/* Lack of 'ecdh' tells us this was saved by 0.58-0.67
* inclusive. If it was saved by a later version, we need
* to leave it alone. */
if (strcmp(raw, "dh-group14-sha1,dh-group1-sha1,rsa,"
"WARN,dh-gex-sha1") == 0) {
/* Previously migrated from BugDHGEx2. */
sfree(raw);
raw = dupstr(bugdhgex2_default);
} else if (strcmp(raw, "dh-gex-sha1,dh-group14-sha1,"
"dh-group1-sha1,rsa,WARN") == 0) {
/* Untouched old default setting. */
sfree(raw);
raw = dupstr(normal_default);
}
/* (For the record: after 0.70, the default algorithm list
* very briefly contained the string 'gss-sha1-krb5'; this was
* never used in any committed version of code, but was left
* over from a pre-commit version of GSS key exchange.
* Mentioned here as it is remotely possible that it will turn
* up in someone's saved settings in future.) */
gprefs_from_str(raw, kexnames, KEX_MAX, conf, CONF_ssh_kexlist);
sfree(raw);
}
gprefs(sesskey, "HostKey", "ed25519,ecdsa,rsa,dsa,WARN",
hknames, HK_MAX, conf, CONF_ssh_hklist);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "RekeyTime", 60, conf, CONF_ssh_rekey_time);
#ifndef NO_GSSAPI
gppi(sesskey, "GssapiRekey", GSS_DEF_REKEY_MINS, conf, CONF_gssapirekey);
#endif
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "RekeyBytes", "1G", conf, CONF_ssh_rekey_data);
{
/* SSH-2 only by default */
int sshprot = gppi_raw(sesskey, "SshProt", 3);
/* Old sessions may contain the values corresponding to the fallbacks
* we used to allow; migrate them */
if (sshprot == 1) sshprot = 0; /* => "SSH-1 only" */
else if (sshprot == 2) sshprot = 3; /* => "SSH-2 only" */
conf_set_int(conf, CONF_sshprot, sshprot);
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "LogHost", "", conf, CONF_loghost);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "SSH2DES", false, conf, CONF_ssh2_des_cbc);
gppb(sesskey, "SshNoAuth", false, conf, CONF_ssh_no_userauth);
gppb(sesskey, "SshBanner", true, conf, CONF_ssh_show_banner);
gppb(sesskey, "AuthTIS", false, conf, CONF_try_tis_auth);
gppb(sesskey, "AuthKI", true, conf, CONF_try_ki_auth);
#ifndef NO_GSSAPI
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "AuthGSSAPI", true, conf, CONF_try_gssapi_auth);
gppb(sesskey, "AuthGSSAPIKEX", true, conf, CONF_try_gssapi_kex);
gprefs(sesskey, "GSSLibs", "\0",
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gsslibkeywords, ngsslibs, conf, CONF_ssh_gsslist);
gppfile(sesskey, "GSSCustom", conf, CONF_ssh_gss_custom);
#endif
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "SshNoShell", false, conf, CONF_ssh_no_shell);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppfile(sesskey, "PublicKeyFile", conf, CONF_keyfile);
gpps(sesskey, "RemoteCommand", "", conf, CONF_remote_cmd);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "RFCEnviron", false, conf, CONF_rfc_environ);
gppb(sesskey, "PassiveTelnet", false, conf, CONF_passive_telnet);
gppb(sesskey, "BackspaceIsDelete", true, conf, CONF_bksp_is_delete);
gppb(sesskey, "RXVTHomeEnd", false, conf, CONF_rxvt_homeend);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "LinuxFunctionKeys", 0, conf, CONF_funky_type);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "NoApplicationKeys", false, conf, CONF_no_applic_k);
gppb(sesskey, "NoApplicationCursors", false, conf, CONF_no_applic_c);
gppb(sesskey, "NoMouseReporting", false, conf, CONF_no_mouse_rep);
gppb(sesskey, "NoRemoteResize", false, conf, CONF_no_remote_resize);
gppb(sesskey, "NoAltScreen", false, conf, CONF_no_alt_screen);
gppb(sesskey, "NoRemoteWinTitle", false, conf, CONF_no_remote_wintitle);
gppb(sesskey, "NoRemoteClearScroll", false,
conf, CONF_no_remote_clearscroll);
{
/* Backward compatibility */
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
int no_remote_qtitle = gppi_raw(sesskey, "NoRemoteQTitle", 1);
/* We deliberately interpret the old setting of "no response" as
* "empty string". This changes the behaviour, but hopefully for
* the better; the user can always recover the old behaviour. */
gppi(sesskey, "RemoteQTitleAction",
no_remote_qtitle ? TITLE_EMPTY : TITLE_REAL,
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf, CONF_remote_qtitle_action);
}
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "NoDBackspace", false, conf, CONF_no_dbackspace);
gppb(sesskey, "NoRemoteCharset", false, conf, CONF_no_remote_charset);
gppb(sesskey, "ApplicationCursorKeys", false, conf, CONF_app_cursor);
gppb(sesskey, "ApplicationKeypad", false, conf, CONF_app_keypad);
gppb(sesskey, "NetHackKeypad", false, conf, CONF_nethack_keypad);
gppb(sesskey, "AltF4", true, conf, CONF_alt_f4);
gppb(sesskey, "AltSpace", false, conf, CONF_alt_space);
gppb(sesskey, "AltOnly", false, conf, CONF_alt_only);
gppb(sesskey, "ComposeKey", false, conf, CONF_compose_key);
gppb(sesskey, "CtrlAltKeys", true, conf, CONF_ctrlaltkeys);
#ifdef OSX_META_KEY_CONFIG
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "OSXOptionMeta", true, conf, CONF_osx_option_meta);
gppb(sesskey, "OSXCommandMeta", false, conf, CONF_osx_command_meta);
#endif
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "TelnetKey", false, conf, CONF_telnet_keyboard);
gppb(sesskey, "TelnetRet", true, conf, CONF_telnet_newline);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "LocalEcho", AUTO, conf, CONF_localecho);
gppi(sesskey, "LocalEdit", AUTO, conf, CONF_localedit);
gpps(sesskey, "Answerback", "PuTTY", conf, CONF_answerback);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "AlwaysOnTop", false, conf, CONF_alwaysontop);
gppb(sesskey, "FullScreenOnAltEnter", false,
conf, CONF_fullscreenonaltenter);
gppb(sesskey, "HideMousePtr", false, conf, CONF_hide_mouseptr);
gppb(sesskey, "SunkenEdge", false, conf, CONF_sunken_edge);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "WindowBorder", 1, conf, CONF_window_border);
gppi(sesskey, "CurType", 0, conf, CONF_cursor_type);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "BlinkCur", false, conf, CONF_blink_cur);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
/* pedantic compiler tells me I can't use conf, CONF_beep as an int * :-) */
gppi(sesskey, "Beep", 1, conf, CONF_beep);
gppi(sesskey, "BeepInd", 0, conf, CONF_beep_ind);
gppfile(sesskey, "BellWaveFile", conf, CONF_bell_wavefile);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "BellOverload", true, conf, CONF_bellovl);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "BellOverloadN", 5, conf, CONF_bellovl_n);
i = gppi_raw(sesskey, "BellOverloadT", 2*TICKSPERSEC
#ifdef PUTTY_UNIX_H
*1000
#endif
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
);
conf_set_int(conf, CONF_bellovl_t, i
#ifdef PUTTY_UNIX_H
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
/ 1000
#endif
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
);
i = gppi_raw(sesskey, "BellOverloadS", 5*TICKSPERSEC
#ifdef PUTTY_UNIX_H
*1000
#endif
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
);
conf_set_int(conf, CONF_bellovl_s, i
#ifdef PUTTY_UNIX_H
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
/ 1000
#endif
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
);
gppi(sesskey, "ScrollbackLines", 2000, conf, CONF_savelines);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "DECOriginMode", false, conf, CONF_dec_om);
gppb(sesskey, "AutoWrapMode", true, conf, CONF_wrap_mode);
gppb(sesskey, "LFImpliesCR", false, conf, CONF_lfhascr);
gppb(sesskey, "CRImpliesLF", false, conf, CONF_crhaslf);
gppb(sesskey, "DisableArabicShaping", false, conf, CONF_no_arabicshaping);
gppb(sesskey, "DisableBidi", false, conf, CONF_no_bidi);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "WinNameAlways", true, conf, CONF_win_name_always);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "WinTitle", "", conf, CONF_wintitle);
gppi(sesskey, "TermWidth", 80, conf, CONF_width);
gppi(sesskey, "TermHeight", 24, conf, CONF_height);
gppfont(sesskey, "Font", conf, CONF_font);
gppi(sesskey, "FontQuality", FQ_DEFAULT, conf, CONF_font_quality);
gppi(sesskey, "FontVTMode", VT_UNICODE, conf, CONF_vtmode);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "UseSystemColours", false, conf, CONF_system_colour);
gppb(sesskey, "TryPalette", false, conf, CONF_try_palette);
gppb(sesskey, "ANSIColour", true, conf, CONF_ansi_colour);
gppb(sesskey, "Xterm256Colour", true, conf, CONF_xterm_256_colour);
gppb(sesskey, "TrueColour", true, conf, CONF_true_colour);
i = gppi_raw(sesskey, "BoldAsColour", 1); conf_set_int(conf, CONF_bold_style, i+1);
for (i = 0; i < 22; i++) {
static const char *const defaults[] = {
"187,187,187", "255,255,255", "0,0,0", "85,85,85", "0,0,0",
"0,255,0", "0,0,0", "85,85,85", "187,0,0", "255,85,85",
"0,187,0", "85,255,85", "187,187,0", "255,255,85", "0,0,187",
"85,85,255", "187,0,187", "255,85,255", "0,187,187",
"85,255,255", "187,187,187", "255,255,255"
};
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
char buf[20], *buf2;
int c0, c1, c2;
sprintf(buf, "Colour%d", i);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
buf2 = gpps_raw(sesskey, buf, defaults[i]);
if (sscanf(buf2, "%d,%d,%d", &c0, &c1, &c2) == 3) {
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int_int(conf, CONF_colours, i*3+0, c0);
conf_set_int_int(conf, CONF_colours, i*3+1, c1);
conf_set_int_int(conf, CONF_colours, i*3+2, c2);
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
sfree(buf2);
}
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "RawCNP", false, conf, CONF_rawcnp);
gppb(sesskey, "UTF8linedraw", false, conf, CONF_utf8linedraw);
gppb(sesskey, "PasteRTF", false, conf, CONF_rtf_paste);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "MouseIsXterm", 0, conf, CONF_mouse_is_xterm);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "RectSelect", false, conf, CONF_rect_select);
gppb(sesskey, "PasteControls", false, conf, CONF_paste_controls);
gppb(sesskey, "MouseOverride", true, conf, CONF_mouse_override);
for (i = 0; i < 256; i += 32) {
static const char *const defaults[] = {
"0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0",
"0,1,2,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1",
"1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,2",
"1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1",
"1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1",
"1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1",
"2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,1,2,2,2,2,2,2,2,2",
"2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,1,2,2,2,2,2,2,2,2"
};
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
char buf[20], *buf2, *p;
int j;
sprintf(buf, "Wordness%d", i);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
buf2 = gpps_raw(sesskey, buf, defaults[i / 32]);
p = buf2;
for (j = i; j < i + 32; j++) {
char *q = p;
while (*p && *p != ',')
p++;
if (*p == ',')
*p++ = '\0';
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int_int(conf, CONF_wordness, j, atoi(q));
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
sfree(buf2);
}
gppb(sesskey, "MouseAutocopy", CLIPUI_DEFAULT_AUTOCOPY,
conf, CONF_mouseautocopy);
read_clip_setting(sesskey, "MousePaste", CLIPUI_DEFAULT_MOUSE,
conf, CONF_mousepaste, CONF_mousepaste_custom);
read_clip_setting(sesskey, "CtrlShiftIns", CLIPUI_DEFAULT_INS,
conf, CONF_ctrlshiftins, CONF_ctrlshiftins_custom);
read_clip_setting(sesskey, "CtrlShiftCV", CLIPUI_NONE,
conf, CONF_ctrlshiftcv, CONF_ctrlshiftcv_custom);
/*
* The empty default for LineCodePage will be converted later
* into a plausible default for the locale.
*/
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "LineCodePage", "", conf, CONF_line_codepage);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "CJKAmbigWide", false, conf, CONF_cjk_ambig_wide);
gppb(sesskey, "UTF8Override", true, conf, CONF_utf8_override);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "Printer", "", conf, CONF_printer);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "CapsLockCyr", false, conf, CONF_xlat_capslockcyr);
gppb(sesskey, "ScrollBar", true, conf, CONF_scrollbar);
gppb(sesskey, "ScrollBarFullScreen", false,
conf, CONF_scrollbar_in_fullscreen);
gppb(sesskey, "ScrollOnKey", false, conf, CONF_scroll_on_key);
gppb(sesskey, "ScrollOnDisp", true, conf, CONF_scroll_on_disp);
gppb(sesskey, "EraseToScrollback", true, conf, CONF_erase_to_scrollback);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppi(sesskey, "LockSize", 0, conf, CONF_resize_action);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "BCE", true, conf, CONF_bce);
gppb(sesskey, "BlinkText", false, conf, CONF_blinktext);
gppb(sesskey, "X11Forward", false, conf, CONF_x11_forward);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gpps(sesskey, "X11Display", "", conf, CONF_x11_display);
gppi(sesskey, "X11AuthType", X11_MIT, conf, CONF_x11_auth);
gppfile(sesskey, "X11AuthFile", conf, CONF_xauthfile);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "LocalPortAcceptAll", false, conf, CONF_lport_acceptall);
gppb(sesskey, "RemotePortAcceptAll", false, conf, CONF_rport_acceptall);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppmap(sesskey, "PortForwardings", conf, CONF_portfwd);
i = gppi_raw(sesskey, "BugIgnore1", 0); conf_set_int(conf, CONF_sshbug_ignore1, 2-i);
i = gppi_raw(sesskey, "BugPlainPW1", 0); conf_set_int(conf, CONF_sshbug_plainpw1, 2-i);
i = gppi_raw(sesskey, "BugRSA1", 0); conf_set_int(conf, CONF_sshbug_rsa1, 2-i);
i = gppi_raw(sesskey, "BugIgnore2", 0); conf_set_int(conf, CONF_sshbug_ignore2, 2-i);
{
int i;
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
i = gppi_raw(sesskey, "BugHMAC2", 0); conf_set_int(conf, CONF_sshbug_hmac2, 2-i);
if (2-i == AUTO) {
i = gppi_raw(sesskey, "BuggyMAC", 0);
if (i == 1)
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
conf_set_int(conf, CONF_sshbug_hmac2, FORCE_ON);
}
}
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
i = gppi_raw(sesskey, "BugDeriveKey2", 0); conf_set_int(conf, CONF_sshbug_derivekey2, 2-i);
i = gppi_raw(sesskey, "BugRSAPad2", 0); conf_set_int(conf, CONF_sshbug_rsapad2, 2-i);
i = gppi_raw(sesskey, "BugPKSessID2", 0); conf_set_int(conf, CONF_sshbug_pksessid2, 2-i);
i = gppi_raw(sesskey, "BugRekey2", 0); conf_set_int(conf, CONF_sshbug_rekey2, 2-i);
i = gppi_raw(sesskey, "BugMaxPkt2", 0); conf_set_int(conf, CONF_sshbug_maxpkt2, 2-i);
i = gppi_raw(sesskey, "BugOldGex2", 0); conf_set_int(conf, CONF_sshbug_oldgex2, 2-i);
i = gppi_raw(sesskey, "BugWinadj", 0); conf_set_int(conf, CONF_sshbug_winadj, 2-i);
i = gppi_raw(sesskey, "BugChanReq", 0); conf_set_int(conf, CONF_sshbug_chanreq, 2-i);
conf_set_bool(conf, CONF_ssh_simple, false);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "StampUtmp", true, conf, CONF_stamp_utmp);
gppb(sesskey, "LoginShell", true, conf, CONF_login_shell);
gppb(sesskey, "ScrollbarOnLeft", false, conf, CONF_scrollbar_on_left);
gppb(sesskey, "ShadowBold", false, conf, CONF_shadowbold);
Post-release destabilisation! Completely remove the struct type 'Config' in putty.h, which stores all PuTTY's settings and includes an arbitrary length limit on every single one of those settings which is stored in string form. In place of it is 'Conf', an opaque data type everywhere outside the new file conf.c, which stores a list of (key, value) pairs in which every key contains an integer identifying a configuration setting, and for some of those integers the key also contains extra parts (so that, for instance, CONF_environmt is a string-to-string mapping). Everywhere that a Config was previously used, a Conf is now; everywhere there was a Config structure copy, conf_copy() is called; every lookup, adjustment, load and save operation on a Config has been rewritten; and there's a mechanism for serialising a Conf into a binary blob and back for use with Duplicate Session. User-visible effects of this change _should_ be minimal, though I don't doubt I've introduced one or two bugs here and there which will eventually be found. The _intended_ visible effects of this change are that all arbitrary limits on configuration strings and lists (e.g. limit on number of port forwardings) should now disappear; that list boxes in the configuration will now be displayed in a sorted order rather than the arbitrary order in which they were added to the list (since the underlying data structure is now a sorted tree234 rather than an ad-hoc comma-separated string); and one more specific change, which is that local and dynamic port forwardings on the same port number are now mutually exclusive in the configuration (putting 'D' in the key rather than the value was a mistake in the first place). One other reorganisation as a result of this is that I've moved all the dialog.c standard handlers (dlg_stdeditbox_handler and friends) out into config.c, because I can't really justify calling them generic any more. When they took a pointer to an arbitrary structure type and the offset of a field within that structure, they were independent of whether that structure was a Config or something completely different, but now they really do expect to talk to a Conf, which can _only_ be used for PuTTY configuration, so I've renamed them all things like conf_editbox_handler and moved them out of the nominally independent dialog-box management module into the PuTTY-specific config.c. [originally from svn r9214]
2011-07-14 22:52:21 +04:00
gppfont(sesskey, "BoldFont", conf, CONF_boldfont);
gppfont(sesskey, "WideFont", conf, CONF_widefont);
gppfont(sesskey, "WideBoldFont", conf, CONF_wideboldfont);
gppi(sesskey, "ShadowBoldOffset", 1, conf, CONF_shadowboldoffset);
gpps(sesskey, "SerialLine", "", conf, CONF_serline);
gppi(sesskey, "SerialSpeed", 9600, conf, CONF_serspeed);
gppi(sesskey, "SerialDataBits", 8, conf, CONF_serdatabits);
gppi(sesskey, "SerialStopHalfbits", 2, conf, CONF_serstopbits);
gppi(sesskey, "SerialParity", SER_PAR_NONE, conf, CONF_serparity);
gppi(sesskey, "SerialFlowControl", SER_FLOW_XONXOFF, conf, CONF_serflow);
gpps(sesskey, "WindowClass", "", conf, CONF_winclass);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
gppb(sesskey, "ConnectionSharing", false,
conf, CONF_ssh_connection_sharing);
gppb(sesskey, "ConnectionSharingUpstream", true,
conf, CONF_ssh_connection_sharing_upstream);
gppb(sesskey, "ConnectionSharingDownstream", true,
conf, CONF_ssh_connection_sharing_downstream);
gppmap(sesskey, "SSHManualHostKeys", conf, CONF_ssh_manual_hostkeys);
}
bool do_defaults(const char *session, Conf *conf)
{
return load_settings(session, conf);
}
static int sessioncmp(const void *av, const void *bv)
{
const char *a = *(const char *const *) av;
const char *b = *(const char *const *) bv;
/*
* Alphabetical order, except that "Default Settings" is a
* special case and comes first.
*/
if (!strcmp(a, "Default Settings"))
return -1; /* a comes first */
if (!strcmp(b, "Default Settings"))
return +1; /* b comes first */
/*
* FIXME: perhaps we should ignore the first & in determining
* sort order.
*/
return strcmp(a, b); /* otherwise, compare normally */
}
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
void get_sesslist(struct sesslist *list, bool allocate)
{
int i;
char *p;
settings_e *handle;
if (allocate) {
strbuf *sb = strbuf_new();
if ((handle = enum_settings_start()) != NULL) {
while (enum_settings_next(handle, sb))
put_byte(sb, '\0');
enum_settings_finish(handle);
}
put_byte(sb, '\0');
list->buffer = strbuf_to_str(sb);
/*
* Now set up the list of sessions. Note that "Default
* Settings" must always be claimed to exist, even if it
* doesn't really.
*/
p = list->buffer;
list->nsessions = 1; /* "Default Settings" counts as one */
while (*p) {
if (strcmp(p, "Default Settings"))
list->nsessions++;
while (*p)
p++;
p++;
}
list->sessions = snewn(list->nsessions + 1, const char *);
list->sessions[0] = "Default Settings";
p = list->buffer;
i = 1;
while (*p) {
if (strcmp(p, "Default Settings"))
list->sessions[i++] = p;
while (*p)
p++;
p++;
}
qsort(list->sessions, i, sizeof(const char *), sessioncmp);
} else {
sfree(list->buffer);
sfree(list->sessions);
list->buffer = NULL;
list->sessions = NULL;
}
}