* es/rebase-i-no-abbrev:
rebase -i: fix short SHA-1 collision
t3404: rebase -i: demonstrate short SHA-1 collision
t3404: make tests more self-contained
Conflicts:
t/t3404-rebase-interactive.sh
- Don't start tests with 'test $? = 0' to catch preparation done
outside the test_expect_success block.
- Move writing the bogus patch and the expected output into the
appropriate test_expect_success blocks.
- Use the test_must_fail helper instead of manually checking for
non-zero exit code.
- Use the debug-friendly test_path_is_file helper instead of 'test -f'.
- No space after '>'.
Signed-off-by: SZEDER Gábor <szeder@ira.uka.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The test 'choking "git rm" should not let it die with cruft' is
supposed to check 'git rm's behavior when interrupted by provoking a
SIGPIPE while 'git rm' is busily deleting files from a specially
crafted index.
This test is silently broken for the following reasons:
- The test crafts a special index by feeding a large number of index
entries with null shas to 'git update-index --index-info'. It was
OK back then when this test was introduced in commit 0693f9ddad
(Make sure lockfiles are unlocked when dying on SIGPIPE,
2008-12-18), but since commit 4337b5856f (do not write null sha1s to
on-disk index, 2012-07-28) null shas are not allowed in the on-disk
index causing 'git update-index' to error out.
- The barfing 'git update-index --index-info' should fail the test,
but it remains unnoticed because of the severely broken && chain:
the test's result depends solely on whether there is a stale lock
file left behind, but after 'git update-index' errors out 'git rm'
won't be executed at all.
To fix this test feed only non-null shas to 'git update-index' and
restore the && chain (partly by adding a missing && and by using the
test_when_finished helper instead of manual cleanup).
Signed-off-by: SZEDER Gábor <szeder@ira.uka.de>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit a15d069 taught git to use curl's SOCKOPTFUNCTION hook
to turn on TCP keepalives. However, modern versions of curl
have a TCP_KEEPALIVE option, which can do this for us. As an
added bonus, the curl code knows how to turn on keepalive
for a much wider variety of platforms. The only downside to
using this option is that not everybody has a new enough curl.
Let's split our keepalive options into three conditionals:
1. With curl 7.25.0 and newer, we rely on curl to do it
right.
2. With older curl that still knows SOCKOPTFUNCTION, we
use the code from a15d069.
3. Otherwise, we are out of luck, and the call is a no-op.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* git://git.bogomips.org/git-svn:
git-svn: Warn about changing default for --prefix in Git v2.0
Documentation/git-svn: Promote the use of --prefix in docs + examples
git-svn.txt: elaborate on rev_map files
git-svn.txt: replace .git with $GIT_DIR
git-svn.txt: reword description of gc command
git-svn.txt: fix AsciiDoc formatting error
git-svn: fix signed commit parsing
Correct all hits from
git grep -e '\(&&\|||\)[^ ]' -e '[^ ]\(&&\|||\)' -- '*.c'
i.e. && or || operators that are followed by anything but a SP,
or that follow something other than a SP or a HT, so that these
operators have a SP around it when necessary.
We usually refrain from making this kind of a tree-wide change in
order to avoid unnecessary conflicts with other "real work" patches,
but in this case, the end result does not have a potentially
cumbersome tree-wide impact, while this is a tree-wide cleanup.
Fixes to compat/regex/regcomp.c and xdiff/xemit.c are to replace a
HT immediately after && with a SP.
This is based on Felipe's patch to bultin/symbolic-ref.c; I did all
the finding out what other files in the whole tree need to be fixed
and did the fix and also the log message while reviewing that single
liner, so any screw-ups in this version are mine.
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The gnome-keyring lib (0.4) distributed with RHEL 4.X is really ancient
and does not provide most of the synchronous functions that even ancient
releases do. Thankfully, we're only using one function that is missing.
Let's emulate gnome_keyring_item_delete_sync() by calling the asynchronous
function and then triggering the event loop processing until our
callback is called.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The gnome-keyring lib distributed with RHEL 5.X is ancient and does
not provide a few of the functions/defines that more recent versions
do, but mostly the API is the same. Let's provide the missing bits
via macro definitions and function implementation.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Produce an error message when we fail to store a password to the keyring.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rather than roll our own, let's use the messaging functions provided
by glib.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rather than roll our own, let's use the memory allocation/free routines
provided by glib.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
gnome-keyring provides functions to allocate non-pageable memory (if
possible). Let's use them to allocate memory that may be used to hold
secure data read from the keyring.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
gnome-keyring provides functions for allocating non-pageable memory (if
possible) intended to be used for storing passwords. Let's use them.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rather than carefully allocating memory for sprintf() to write into,
let's make use of the glib helper function g_strdup_printf(), which
makes things a lot easier and less error-prone.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since this is a Gnome application, let's set the application name to
something reasonable. This will be displayed in Gnome dialog boxes
e.g. the one that prompts for the user's keyring password.
We add an include statement for glib.h and add the glib-2.0 cflags and
libs to the compilation arguments, but both of these are really noops
since glib is already a dependency of gnome-keyring.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Ensure buffer length is non-zero before attempting to access the last
element.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Also, initialization is not necessary since it is assigned before it is
used.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If the correct arguments were not specified, this program should exit
non-zero. Let's do so.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A range notation "A..B" means exactly the same thing as what "^A B"
means, i.e. the set of commits that are reachable from B but not
from A. But the internal representation after the revision parser
parsed these two notations are subtly different.
- "rev-list ^A B" leaves A and B in the revs->pending.objects[]
array, with the former marked as UNINTERESTING and the revision
traversal machinery propagates the mark to underlying commit
objects A^0 and B^0.
- "rev-list A..B" peels tags and leaves A^0 (marked as
UNINTERESTING) and B^0 in revs->pending.objects[] array before
the traversal machinery kicks in.
This difference usually does not matter, but starts to matter when
the --objects option is used. For example, we see this:
$ git rev-list --objects v1.8.4^1..v1.8.4 | grep $(git rev-parse v1.8.4)
$ git rev-list --objects v1.8.4 ^v1.8.4^1 | grep $(git rev-parse v1.8.4)
04f013dc38 v1.8.4
With the former invocation, the revision traversal machinery never
hears about the tag v1.8.4 (it only sees the result of peeling it,
i.e. the commit v1.8.4^0), and the tag itself does not appear in the
output. The latter does send the tag object itself to the output.
Make the range notation keep the unpeeled objects and feed them to
the traversal machinery to fix this inconsistency.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
git-prune-packed operates on GIT_OBJECT_DIRECTORY, not
GIT_OBJECT_DIR.
Signed-off-by: Steffen Prohaska <prohaska@zib.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Labeled lists require a double colon.
[jc] I eyeballed the output from
git grep '[^:]:$' Documentation/\*.txt
and the patch fixes all breakages of this kind.
Signed-off-by: Steffen Prohaska <prohaska@zib.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The way to spell the current repository with a '.' dot is
independent from how the pathspec allows globs expanded by Git.
Make them two separate bullet items in the enumeration.
Signed-off-by: Philip Oakley <philipoakley@iee.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Our default_remote_name starts at "origin", but may be
overridden by the config file. In the former case, we
allocate a new string, but in the latter case, we point to
the remote name in an existing "struct branch".
This gives the variable inconsistent free() semantics (we
are sometimes responsible for freeing the string and
sometimes pointing to somebody else's storage), and causes a
small leak when the allocated string is overridden by
config.
We can fix both by simply dropping the extra copy and
pointing to the string literal.
Noticed-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The '+=' operator is not supported by old Bash versions (3.0) we still
care about.
Signed-off-by: SZEDER Gábor <szeder@ira.uka.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Split_ident currently parses left to right. Given this
input:
Your Name <email@example.com> 123456789 -0500\n
We assume the name starts the line and runs until the first
"<". That starts the email address, which runs until the
first ">". Everything after that is assumed to be the
timestamp.
This works fine in the normal case, but is easily broken by
corrupted ident lines that contain an extra ">". Some
examples seen in the wild are:
1. Name <email>-<> 123456789 -0500\n
2. Name <email> <Name<email>> 123456789 -0500\n
3. Name1 <email1>, Name2 <email2> 123456789 -0500\n
Currently each of these produces some email address (which
is not necessarily the one the user intended) and end up
with a NULL date (which is generally interpreted as the
epoch by "git log" and friends).
But in each case we could get the correct timestamp simply
by parsing from the right-hand side, looking backwards for
the final ">", and then reading the timestamp from there.
In general, it's a losing battle to try to automatically
guess what the user meant with their broken crud. But this
particular workaround is probably worth doing. One, it's
dirt simple, and can't impact non-broken cases. Two, it
doesn't catch a single breakage we've seen, but rather a
large class of errors (i.e., any breakage inside the email
angle brackets may affect the email, but won't spill over
into the timestamp parsing). And three, the timestamp is
arguably more valuable to get right, because it can affect
correctness (e.g., in --until cutoffs).
This patch implements the right-to-left scheme described
above. We adjust the tests in t4212, which generate a commit
with such a broken ident, and now gets the timestamp right.
We also add a test that fsck continues to detect the
breakage.
For reference, here are pointers to the breakages seen (as
numbered above):
[1] http://article.gmane.org/gmane.comp.version-control.git/221441
[2] http://article.gmane.org/gmane.comp.version-control.git/222362
[3] http://perl5.git.perl.org/perl.git/commit/13b79730adea97e660de84bbe67f9d7cbe344302
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
For efficiency and security reasons, an earlier commit in
this series taught http_get_* to re-write the base url based
on redirections we saw while making a specific request.
This commit wires that option into the info/refs request,
meaning that a redirect from
http://example.com/foo.git/info/refs
to
https://example.com/bar.git/info/refs
will behave as if "https://example.com/bar.git" had been
provided to git in the first place.
The tests bear some explanation. We introduce two new
hierearchies into the httpd test config:
1. Requests to /smart-redir-limited will work only for the
initial info/refs request, but not any subsequent
requests. As a result, we can confirm whether the
client is re-rooting its requests after the initial
contact, since otherwise it will fail (it will ask for
"repo.git/git-upload-pack", which is not redirected).
2. Requests to smart-redir-auth will redirect, and require
auth after the redirection. Since we are using the
redirected base for further requests, we also update
the credential struct, in order not to mislead the user
(or credential helpers) about which credential is
needed. We can therefore check the GIT_ASKPASS prompts
to make sure we are prompting for the new location.
Because we have neither multiple servers nor https
support in our test setup, we can only redirect between
paths, meaning we need to turn on
credential.useHttpPath to see the difference.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
We use a strbuf to generate the string containing the remote
URL, but then detach it to a bare pointer. This makes it
harder to later manipulate the URL, as we have forgotten the
length (and the allocation semantics are not as clear).
Let's instead keep the strbuf around. As a bonus, this
eliminates a confusing double-use of the "buf" strbuf in
main(). Prior to this, it was used both for constructing the
url, and for reading commands from stdin.
The downside is that we have to update each call site to
refer to "url.buf" rather than just "url" when they want the
C string.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
In the discover_refs function, we use a strbuf named
"buffer" for multiple purposes. First we build the info/refs
URL in it, and then detach that to a bare pointer. Then, we
use the same strbuf to store the result of fetching the
refs.
Let's instead keep a separate refs_url strbuf. This is less
confusing, as the "buffer" strbuf is now used for only one
thing.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
If a caller asks the http_get_* functions to go to a
particular URL and we end up elsewhere due to a redirect,
the effective_url field can tell us where we went.
It would be nice to remember this redirect and short-cut
further requests for two reasons:
1. It's more efficient. Otherwise we spend an extra http
round-trip to the server for each subsequent request,
just to get redirected.
2. If we end up with an http 401 and are going to ask for
credentials, it is to feed them to the redirect target.
If the redirect is an http->https upgrade, this means
our credentials may be provided on the http leg, just
to end up redirected to https. And if the redirect
crosses server boundaries, then curl will drop the
credentials entirely as it follows the redirect.
However, it, it is not enough to simply record the effective
URL we saw and use that for subsequent requests. We were
originally fed a "base" url like:
http://example.com/foo.git
and we want to figure out what the new base is, even though
the URLs we see may be:
original: http://example.com/foo.git/info/refs
effective: http://example.com/bar.git/info/refs
Subsequent requests will not be for "info/refs", but for
other paths relative to the base. We must ask the caller to
pass in the original base, and we must pass the redirected
base back to the caller (so that it can generate more URLs
from it). Furthermore, we need to feed the new base to the
credential code, so that requests to credential helpers (or
to the user) match the URL we will be requesting.
This patch teaches http_request_reauth to do this munging.
Since it is the caller who cares about making more URLs, it
seems at first glance that callers could simply check
effective_url themselves and handle it. However, since we
need to update the credential struct before the second
re-auth request, we have to do it inside http_request_reauth.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
When we ask curl to access a URL, it may follow one or more
redirects to reach the final location. We have no idea
this has happened, as curl takes care of the details and
simply returns the final content to us.
The final URL that we ended up with can be accessed via
CURLINFO_EFFECTIVE_URL. Let's make that optionally available
to callers of http_get_*, so that they can make further
decisions based on the redirection.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
When we are handling a curl response code in http_request or
in the remote-curl RPC code, we use the handle_curl_result
helper to translate curl's response into an easy-to-use
code. When we see an HTTP 401, we do one of two things:
1. If we already had a filled-in credential, we mark it as
rejected, and then return HTTP_NOAUTH to indicate to
the caller that we failed.
2. If we didn't, then we ask for a new credential and tell
the caller HTTP_REAUTH to indicate that they may want
to try again.
Rejecting in the first case makes sense; it is the natural
result of the request we just made. However, prompting for
more credentials in the second step does not always make
sense. We do not know for sure that the caller is going to
make a second request, and nor are we sure that it will be
to the same URL. Logically, the prompt belongs not to the
request we just finished, but to the request we are (maybe)
about to make.
In practice, it is very hard to trigger any bad behavior.
Currently, if we make a second request, it will always be to
the same URL (even in the face of redirects, because curl
handles the redirects internally). And we almost always
retry on HTTP_REAUTH these days. The one exception is if we
are streaming a large RPC request to the server (e.g., a
pushed packfile), in which case we cannot restart. It's
extremely unlikely to see a 401 response at this stage,
though, as we would typically have seen it when we sent a
probe request, before streaming the data.
This patch drops the automatic prompt out of case 2, and
instead requires the caller to do it. This is a few extra
lines of code, and the bug it fixes is unlikely to come up
in practice. But it is conceptually cleaner, and paves the
way for better handling of credentials across redirects.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
The help text for the `tool` flag should mention:
--tool=<tool>
instead of:
--tool-<tool>
Signed-off-by: Stefan Saasen <ssaasen@atlassian.com>
Reviewed-by: David Aguilar <davvid@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Sparse issues an "using sizeof on a function" warning for each
call to curl_easy_setopt() which sets an option that takes a
function pointer parameter. (currently 12 such warnings over 4
files.)
The warnings relate to the use of the "typecheck-gcc.h" header
file which adds a layer of type-checking macros to the curl
function invocations (for gcc >= 4.3 and !__cplusplus). As part
of the type-checking layer, 'sizeof' is applied to the function
parameter of curl_easy_setopt(). Note that, in the context of
sizeof, the function to function pointer conversion is not
performed and that sizeof(f) != sizeof(&f).
A simple solution, therefore, would be to replace the function
name in each such call to curl_easy_setopt() with an explicit
function pointer expression (i.e. replace f with &f).
However, the "typecheck-gcc.h" header file is only conditionally
included, in addition to the gcc and C++ checks mentioned above,
depending on the CURL_DISABLE_TYPECHECK preprocessor variable.
In order to suppress the warnings, we use target-specific variable
assignments to add -DCURL_DISABLE_TYPECHECK to SPARSE_FLAGS for
each file affected (http-push.c, http.c, http-walker.c and
remote-curl.c).
Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
The Thunderbird section of the 'MUA-specific hints' contains three
different approaches to setting up the mail client to leave patch
emails unmolested. The second approach (configuration) has a step
missing when configuring the composition window not to wrap. In
particular, the "mailnews.wraplength" configuration variable needs
to be set to zero. Update the documentation to add the missing
setting.
Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Junio C Hamano
SZEDER Gábor
Jeff King
Felipe Contreras
Brandon Casey
Steffen Prohaska
Philip Oakley
Stefan Saasen
Ramsay Jones