microsoft
/
git
зеркало из https://github.com/microsoft/git.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
git/builtin-grep.c

798 строки
18 KiB
C
Исходник Обычный вид История

built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
/*
* Builtin "git grep"
*
* Copyright (c) 2006 Junio C Hamano
*/
#include "cache.h"
#include "blob.h"
#include "tree.h"
#include "commit.h"
#include "tag.h"
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
#include "tree-walk.h"
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
#include "builtin.h"
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
#include "grep.h"
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
add NO_EXTERNAL_GREP build option Previously, we just chose whether to allow external grep based on the __unix__ define. However, there are systems which define this macro but which have an inferior group (e.g., one that does not support all options used by t7002). This allows users to accept the potential speed penalty to get a more consistent grep experience (and to pass the testsuite). Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-03-13 00:39:16 +03:00
#ifndef NO_EXTERNAL_GREP
#ifdef __unix__
#define NO_EXTERNAL_GREP 0
#else
#define NO_EXTERNAL_GREP 1
#endif
#endif
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
/*
* git grep pathspecs are somewhat different from diff-tree pathspecs;
* pathname wildcards are allowed.
*/
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
static int pathspec_matches(const char **paths, const char *name)
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
{
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
int namelen, i;
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
if (!paths || !*paths)
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
return 1;
namelen = strlen(name);
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
for (i = 0; paths[i]; i++) {
const char *match = paths[i];
int matchlen = strlen(match);
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
const char *cp, *meta;
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
Fix possible out-of-bounds array access If match is "", match[-1] is accessed. Let pathspec_matches return 1 in that case indicating that "" matches everything. Incidently this fixes git-grep'ing in ".". Signed-off-by: Uwe Zeisberger <Uwe_Zeisberger@digi.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-21 13:04:12 +04:00
if (!matchlen ||
((matchlen <= namelen) &&
!strncmp(name, match, matchlen) &&
(match[matchlen-1] == '/' ||
name[matchlen] == '\0' || name[matchlen] == '/')))
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
return 1;
if (!fnmatch(match, name, 0))
return 1;
if (name[namelen-1] != '/')
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
continue;
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
/* We are being asked if the directory ("name") is worth
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
* descending into.
*
* Find the longest leading directory name that does
* not have metacharacter in the pathspec; the name
* we are looking at must overlap with that directory.
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
*/
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
for (cp = match, meta = NULL; cp - match < matchlen; cp++) {
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
char ch = *cp;
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
if (ch == '*' || ch == '[' || ch == '?') {
meta = cp;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
break;
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
}
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
}
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
if (!meta)
meta = cp; /* fully literal */
if (namelen <= meta - match) {
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
/* Looking at "Documentation/" and
* the pattern says "Documentation/howto/", or
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
* "Documentation/diff*.txt". The name we
* have should match prefix.
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
*/
if (!memcmp(match, name, namelen))
return 1;
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
continue;
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
}
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
if (meta - match < namelen) {
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
/* Looking at "Documentation/howto/" and
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
* the pattern says "Documentation/h*";
* match up to "Do.../h"; this avoids descending
* into "Documentation/technical/".
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
*/
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
if (!memcmp(match, name, meta - match))
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
return 1;
builtin-grep: tighten path wildcard vs tree traversal. The earlier code descended into Documentation/technical when given "Documentation/how*" as the pattern, which was too loose. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 04:27:07 +04:00
continue;
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
}
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
}
return 0;
}
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
static int grep_sha1(struct grep_opt *opt, const unsigned char *sha1, const char *name, int tree_name_len)
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
{
unsigned long size;
char *data;
convert object type handling from a string to a number We currently have two parallel notation for dealing with object types in the code: a string and a numerical value. One of them is obviously redundent, and the most used one requires more stack space and a bunch of strcmp() all over the place. This is an initial step for the removal of the version using a char array found in object reading code paths. The patch is unfortunately large but there is no sane way to split it in smaller parts without breaking the system. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-02-26 22:55:59 +03:00
enum object_type type;
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
char *to_free = NULL;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
int hit;
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
convert object type handling from a string to a number We currently have two parallel notation for dealing with object types in the code: a string and a numerical value. One of them is obviously redundent, and the most used one requires more stack space and a bunch of strcmp() all over the place. This is an initial step for the removal of the version using a char array found in object reading code paths. The patch is unfortunately large but there is no sane way to split it in smaller parts without breaking the system. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-02-26 22:55:59 +03:00
data = read_sha1_file(sha1, &type, &size);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
if (!data) {
error("'%s': unable to read %s", name, sha1_to_hex(sha1));
return 0;
}
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
if (opt->relative && opt->prefix_length) {
static char name_buf[PATH_MAX];
char *cp;
int name_len = strlen(name) - opt->prefix_length + 1;
if (!tree_name_len)
name += opt->prefix_length;
else {
if (ARRAY_SIZE(name_buf) <= name_len)
cp = to_free = xmalloc(name_len);
else
cp = name_buf;
memcpy(cp, name, tree_name_len);
strcpy(cp + tree_name_len,
name + tree_name_len + opt->prefix_length);
name = cp;
}
}
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
hit = grep_buffer(opt, name, data, size);
free(data);
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
free(to_free);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
return hit;
}
static int grep_file(struct grep_opt *opt, const char *filename)
{
struct stat st;
int i;
char *data;
Cast 64 bit off_t to 32 bit size_t Some systems have sizeof(off_t) == 8 while sizeof(size_t) == 4. This implies that we are able to access and work on files whose maximum length is around 2^63-1 bytes, but we can only malloc or mmap somewhat less than 2^32-1 bytes of memory. On such a system an implicit conversion of off_t to size_t can cause the size_t to wrap, resulting in unexpected and exciting behavior. Right now we are working around all gcc warnings generated by the -Wshorten-64-to-32 option by passing the off_t through xsize_t(). In the future we should make xsize_t on such problematic platforms detect the wrapping and die if such a file is accessed. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-07 04:44:37 +03:00
size_t sz;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
if (lstat(filename, &st) < 0) {
err_ret:
if (errno != ENOENT)
error("'%s': %s", filename, strerror(errno));
return 0;
}
if (!st.st_size)
return 0; /* empty file -- no grep hit */
if (!S_ISREG(st.st_mode))
return 0;
Cast 64 bit off_t to 32 bit size_t Some systems have sizeof(off_t) == 8 while sizeof(size_t) == 4. This implies that we are able to access and work on files whose maximum length is around 2^63-1 bytes, but we can only malloc or mmap somewhat less than 2^32-1 bytes of memory. On such a system an implicit conversion of off_t to size_t can cause the size_t to wrap, resulting in unexpected and exciting behavior. Right now we are working around all gcc warnings generated by the -Wshorten-64-to-32 option by passing the off_t through xsize_t(). In the future we should make xsize_t on such problematic platforms detect the wrapping and die if such a file is accessed. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-07 04:44:37 +03:00
sz = xsize_t(st.st_size);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
i = open(filename, O_RDONLY);
if (i < 0)
goto err_ret;
Cast 64 bit off_t to 32 bit size_t Some systems have sizeof(off_t) == 8 while sizeof(size_t) == 4. This implies that we are able to access and work on files whose maximum length is around 2^63-1 bytes, but we can only malloc or mmap somewhat less than 2^32-1 bytes of memory. On such a system an implicit conversion of off_t to size_t can cause the size_t to wrap, resulting in unexpected and exciting behavior. Right now we are working around all gcc warnings generated by the -Wshorten-64-to-32 option by passing the off_t through xsize_t(). In the future we should make xsize_t on such problematic platforms detect the wrapping and die if such a file is accessed. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-07 04:44:37 +03:00
data = xmalloc(sz + 1);
if (st.st_size != read_in_full(i, data, sz)) {
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
error("'%s': short read %s", filename, strerror(errno));
close(i);
free(data);
return 0;
}
close(i);
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
if (opt->relative && opt->prefix_length)
filename += opt->prefix_length;
Cast 64 bit off_t to 32 bit size_t Some systems have sizeof(off_t) == 8 while sizeof(size_t) == 4. This implies that we are able to access and work on files whose maximum length is around 2^63-1 bytes, but we can only malloc or mmap somewhat less than 2^32-1 bytes of memory. On such a system an implicit conversion of off_t to size_t can cause the size_t to wrap, resulting in unexpected and exciting behavior. Right now we are working around all gcc warnings generated by the -Wshorten-64-to-32 option by passing the off_t through xsize_t(). In the future we should make xsize_t on such problematic platforms detect the wrapping and die if such a file is accessed. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-07 04:44:37 +03:00
i = grep_buffer(opt, filename, data, sz);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
free(data);
return i;
}
add NO_EXTERNAL_GREP build option Previously, we just chose whether to allow external grep based on the __unix__ define. However, there are systems which define this macro but which have an inferior group (e.g., one that does not support all options used by t7002). This allows users to accept the potential speed penalty to get a more consistent grep experience (and to pass the testsuite). Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-03-13 00:39:16 +03:00
#if !NO_EXTERNAL_GREP
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
static int exec_grep(int argc, const char **argv)
{
pid_t pid;
int status;
argv[argc] = NULL;
pid = fork();
if (pid < 0)
return pid;
if (!pid) {
execvp("grep", (char **) argv);
exit(255);
}
while (waitpid(pid, &status, 0) < 0) {
if (errno == EINTR)
continue;
return -1;
}
if (WIFEXITED(status)) {
if (!WEXITSTATUS(status))
return 1;
return 0;
}
return -1;
}
#define MAXARGS 1000
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
#define ARGBUF 4096
#define push_arg(a) do { \
if (nr < MAXARGS) argv[nr++] = (a); \
else die("maximum number of args exceeded"); \
} while (0)
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
Split grep arguments in a way that does not requires to add /dev/null. In order to (almost) always show the name of the file without relying on "-H" option of GNU grep, we used to add /dev/null to the argument list unless we are doing -l or -L. This caused "/dev/null:0" to show up when -c is given in the output. It is not enough to add -c to the set of options we do not pass /dev/null for. When we have too many files, we invoke grep multiple times and we need to avoid giving a widow filename to the last invocation -- otherwise we will not see the name. This keeps two filenames when the argv[] buffer is about to overflow and we have not finished iterating over the index, so that the last round will always have at least two paths to work with (and not require /dev/null). An obvious and the only exception is when there is only 1 file that is given to the underlying grep, and in that case we avoid passing /dev/null and let the external "grep -c" report only the number of matches. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-09-14 11:31:00 +04:00
/*
* If you send a singleton filename to grep, it does not give
* the name of the file. GNU grep has "-H" but we would want
* that behaviour in a portable way.
*
* So we keep two pathnames in argv buffer unsent to grep in
* the main loop if we need to do more than one grep.
*/
static int flush_grep(struct grep_opt *opt,
int argc, int arg0, const char **argv, int *kept)
{
int status;
int count = argc - arg0;
const char *kept_0 = NULL;
if (count <= 2) {
/*
* Because we keep at least 2 paths in the call from
* the main loop (i.e. kept != NULL), and MAXARGS is
* far greater than 2, this usually is a call to
* conclude the grep. However, the user could attempt
* to overflow the argv buffer by giving too many
* options to leave very small number of real
* arguments even for the call in the main loop.
*/
if (kept)
die("insanely many options to grep");
/*
* If we have two or more paths, we do not have to do
* anything special, but we need to push /dev/null to
* get "-H" behaviour of GNU grep portably but when we
* are not doing "-l" nor "-L" nor "-c".
*/
if (count == 1 &&
!opt->name_only &&
!opt->unmatch_name_only &&
!opt->count) {
argv[argc++] = "/dev/null";
argv[argc] = NULL;
}
}
else if (kept) {
/*
* Called because we found many paths and haven't finished
* iterating over the cache yet. We keep two paths
* for the concluding call. argv[argc-2] and argv[argc-1]
* has the last two paths, so save the first one away,
* replace it with NULL while sending the list to grep,
* and recover them after we are done.
*/
*kept = 2;
kept_0 = argv[argc-2];
argv[argc-2] = NULL;
argc -= 2;
}
status = exec_grep(argc, argv);
if (kept_0) {
/*
* Then recover them. Now the last arg is beyond the
* terminating NULL which is at argc, and the second
* from the last is what we saved away in kept_0
*/
argv[arg0++] = kept_0;
argv[arg0] = argv[argc+1];
}
return status;
}
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
static int external_grep(struct grep_opt *opt, const char **paths, int cached)
{
git-grep: fix exit code when we use external grep. Upon hit, we should exit with status 0. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:43:40 +04:00
int i, nr, argc, hit, len, status;
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
const char *argv[MAXARGS+1];
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
char randarg[ARGBUF];
char *argptr = randarg;
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
struct grep_pat *p;
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
if (opt->extended || (opt->relative && opt->prefix_length))
git-grep: boolean expression on pattern matching. This extends the behaviour of git-grep when multiple -e options are given. So far, we allowed multiple -e to behave just like regular grep with multiple -e, i.e. the patterns are OR'ed together. With this change, you can also have multiple patterns AND'ed together, or form boolean expressions, like this (the parentheses are quoted from the shell in this example): $ git grep -e _PATTERN --and \( -e atom -e token \) Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-30 14:04:05 +04:00
return -1;
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
len = nr = 0;
push_arg("grep");
if (opt->fixed)
Fix silly typo in new builtin grep The "-F" flag apparently got mis-translated due to some over-eager copy-paste work into a duplicate "-H" when using the external grep. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 04:54:01 +04:00
push_arg("-F");
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
if (opt->linenum)
push_arg("-n");
Add "-h/-H" parsing to "git grep" It turns out that I actually wanted to avoid the filenames (because I didn't care - I just wanted to see the context in which something was used) when doing a grep. But since "git grep" didn't take the "-h" parameter, I ended up having to do "grep -5 -h *.c" instead. So here's a trivial patch that adds "-h" (and thus has to enable -H too) to "git grep" parsing. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-14 21:45:12 +04:00
if (!opt->pathname)
push_arg("-h");
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
if (opt->regflags & REG_EXTENDED)
push_arg("-E");
builtin-grep: pass ignore case option to external grep Don't just read the --ignore-case/-i option, pass the flag on to the external grep program. Signed-off-by: Robert Fitzsimons <robfitz@273k.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-07 03:15:16 +04:00
if (opt->regflags & REG_ICASE)
push_arg("-i");
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
if (opt->word_regexp)
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
push_arg("-w");
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
if (opt->name_only)
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
push_arg("-l");
if (opt->unmatch_name_only)
push_arg("-L");
if (opt->count)
push_arg("-c");
if (opt->post_context || opt->pre_context) {
if (opt->post_context != opt->pre_context) {
if (opt->pre_context) {
push_arg("-B");
len += snprintf(argptr, sizeof(randarg)-len,
grep -An -Bm: fix invocation of external grep command When building command line to invoke external grep, the arguments to -A/-B/-C options were placd in randarg[] buffer, but the code forgot that snprintf() does not count terminating NUL in its return value. This caused "git grep -A1 -B2" to invoke external grep with "-B21 -A1". Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-11-18 08:18:14 +03:00
"%u", opt->pre_context) + 1;
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
if (sizeof(randarg) <= len)
die("maximum length of args exceeded");
push_arg(argptr);
argptr += len;
}
if (opt->post_context) {
push_arg("-A");
len += snprintf(argptr, sizeof(randarg)-len,
grep -An -Bm: fix invocation of external grep command When building command line to invoke external grep, the arguments to -A/-B/-C options were placd in randarg[] buffer, but the code forgot that snprintf() does not count terminating NUL in its return value. This caused "git grep -A1 -B2" to invoke external grep with "-B21 -A1". Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-11-18 08:18:14 +03:00
"%u", opt->post_context) + 1;
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
if (sizeof(randarg) <= len)
die("maximum length of args exceeded");
push_arg(argptr);
argptr += len;
}
}
else {
push_arg("-C");
len += snprintf(argptr, sizeof(randarg)-len,
grep -An -Bm: fix invocation of external grep command When building command line to invoke external grep, the arguments to -A/-B/-C options were placd in randarg[] buffer, but the code forgot that snprintf() does not count terminating NUL in its return value. This caused "git grep -A1 -B2" to invoke external grep with "-B21 -A1". Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-11-18 08:18:14 +03:00
"%u", opt->post_context) + 1;
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
if (sizeof(randarg) <= len)
die("maximum length of args exceeded");
push_arg(argptr);
argptr += len;
}
}
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
for (p = opt->pattern_list; p; p = p->next) {
builtin-grep: unparse more command line options. The earlier one to use external grep missed some often used options. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-16 00:28:01 +04:00
push_arg("-e");
push_arg(p->pattern);
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
}
builtin-grep: workaround for non GNU grep. Of course, it still ignores the fact that not all grep's support some of the flags like -F/-L/-A/-C etc, but for those cases, the external grep itself will happily just say "unrecognized option -F" or similar. So with this change, "git grep" should handle all the flags the native grep handles, which is really quite fine. We don't _need_ to expose anything more, and if you do want our extensions, you can get them with "--uncached" and an up-to-date index. No configuration necessary, and we automatically take advantage of any native grep we have, if possible. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-17 22:12:22 +04:00
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
hit = 0;
argc = nr;
for (i = 0; i < active_nr; i++) {
struct cache_entry *ce = active_cache[i];
remove superflous "const" Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-22 00:45:46 +04:00
char *name;
Split grep arguments in a way that does not requires to add /dev/null. In order to (almost) always show the name of the file without relying on "-H" option of GNU grep, we used to add /dev/null to the argument list unless we are doing -l or -L. This caused "/dev/null:0" to show up when -c is given in the output. It is not enough to add -c to the set of options we do not pass /dev/null for. When we have too many files, we invoke grep multiple times and we need to avoid giving a widow filename to the last invocation -- otherwise we will not see the name. This keeps two filenames when the argv[] buffer is about to overflow and we have not finished iterating over the index, so that the last round will always have at least two paths to work with (and not require /dev/null). An obvious and the only exception is when there is only 1 file that is given to the underlying grep, and in that case we avoid passing /dev/null and let the external "grep -c" report only the number of matches. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-09-14 11:31:00 +04:00
int kept;
Make on-disk index representation separate from in-core one This converts the index explicitly on read and write to its on-disk format, allowing the in-core format to contain more flags, and be simpler. In particular, the in-core format is now host-endian (as opposed to the on-disk one that is network endian in order to be able to be shared across machines) and as a result we can dispense with all the htonl/ntohl on accesses to the cache_entry fields. This will make it easier to make use of various temporary flags that do not exist in the on-disk format. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-01-15 03:03:17 +03:00
if (!S_ISREG(ce->ce_mode))
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
continue;
if (!pathspec_matches(paths, ce->name))
continue;
builtin-grep: workaround for non GNU grep. Of course, it still ignores the fact that not all grep's support some of the flags like -F/-L/-A/-C etc, but for those cases, the external grep itself will happily just say "unrecognized option -F" or similar. So with this change, "git grep" should handle all the flags the native grep handles, which is really quite fine. We don't _need_ to expose anything more, and if you do want our extensions, you can get them with "--uncached" and an up-to-date index. No configuration necessary, and we automatically take advantage of any native grep we have, if possible. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-17 22:12:22 +04:00
name = ce->name;
if (name[0] == '-') {
int len = ce_namelen(ce);
name = xmalloc(len + 3);
memcpy(name, "./", 2);
memcpy(name + 2, ce->name, len + 1);
}
argv[argc++] = name;
git grep shows the same hit repeatedly for unmerged paths When the index is unmerged, e.g. $ git ls-files -u 100644 faf413748eb6ccb15161a212156c5e348302b1b6 1 setup.c 100644 145eca50f41d811c4c8fcb21ed2604e6b2971aba 2 setup.c 100644 cb9558c49b6027bf225ba2a6154c4d2a52bcdbe2 3 setup.c running "git grep" for work tree files repeats hits for each unmerged stage. $ git grep -n -e setup_work_tree -- '*.[ch]' setup.c:209:void setup_work_tree(void) setup.c:209:void setup_work_tree(void) setup.c:209:void setup_work_tree(void) This should fix it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-12-06 03:13:08 +03:00
if (MAXARGS <= argc) {
status = flush_grep(opt, argc, nr, argv, &kept);
if (0 < status)
hit = 1;
argc = nr + kept;
}
grep: do not skip unmerged entries when grepping in the working tree. We used to skip unmerged entries, which made sense for grepping in the cached copies, but not for grepping in the working tree. Noticed by Johannes Sixt. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-11-26 23:47:52 +03:00
if (ce_stage(ce)) {
do {
i++;
} while (i < active_nr &&
!strcmp(ce->name, active_cache[i]->name));
i--; /* compensate for loop control */
}
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
}
git-grep: fix exit code when we use external grep. Upon hit, we should exit with status 0. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:43:40 +04:00
if (argc > nr) {
Split grep arguments in a way that does not requires to add /dev/null. In order to (almost) always show the name of the file without relying on "-H" option of GNU grep, we used to add /dev/null to the argument list unless we are doing -l or -L. This caused "/dev/null:0" to show up when -c is given in the output. It is not enough to add -c to the set of options we do not pass /dev/null for. When we have too many files, we invoke grep multiple times and we need to avoid giving a widow filename to the last invocation -- otherwise we will not see the name. This keeps two filenames when the argv[] buffer is about to overflow and we have not finished iterating over the index, so that the last round will always have at least two paths to work with (and not require /dev/null). An obvious and the only exception is when there is only 1 file that is given to the underlying grep, and in that case we avoid passing /dev/null and let the external "grep -c" report only the number of matches. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-09-14 11:31:00 +04:00
status = flush_grep(opt, argc, nr, argv, NULL);
git-grep: fix exit code when we use external grep. Upon hit, we should exit with status 0. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:43:40 +04:00
if (0 < status)
hit = 1;
}
return hit;
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
}
Don't build external_grep if its not used Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-07 04:44:14 +03:00
#endif
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
static int grep_cache(struct grep_opt *opt, const char **paths, int cached)
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
{
int hit = 0;
int nr;
read_cache();
add NO_EXTERNAL_GREP build option Previously, we just chose whether to allow external grep based on the __unix__ define. However, there are systems which define this macro but which have an inferior group (e.g., one that does not support all options used by t7002). This allows users to accept the potential speed penalty to get a more consistent grep experience (and to pass the testsuite). Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-03-13 00:39:16 +03:00
#if !NO_EXTERNAL_GREP
builtin-grep: use external grep when we can take advantage of it It's not perfect, but it gets the "git grep some-random-string" down to the good old half-a-second range for the kernel. It should convert more of the argument flags for "grep", that should be trivial to expand (I did a few just as an example). It should also bother to try to return the right "hit" value (which it doesn't, right now - the code is kind of there, but I didn't actually bother to do it _right_). Also, right now it _just_ limits by number of arguments, but it should also strictly speaking limit by total argument size (ie add up the length of the filenames, and do the "exec_grep()" flush call if it's bigger than some random value like 32kB). But I think that it's _conceptually_ doing all the right things, and it seems to work. So maybe somebody else can do some of the final polish. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-15 07:49:15 +04:00
/*
* Use the external "grep" command for the case where
* we grep through the checked-out files. It tends to
* be a lot more optimized
*/
if (!cached) {
hit = external_grep(opt, paths, cached);
if (hit >= 0)
return hit;
}
#endif
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
for (nr = 0; nr < active_nr; nr++) {
struct cache_entry *ce = active_cache[nr];
Make on-disk index representation separate from in-core one This converts the index explicitly on read and write to its on-disk format, allowing the in-core format to contain more flags, and be simpler. In particular, the in-core format is now host-endian (as opposed to the on-disk one that is network endian in order to be able to be shared across machines) and as a result we can dispense with all the htonl/ntohl on accesses to the cache_entry fields. This will make it easier to make use of various temporary flags that do not exist in the on-disk format. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-01-15 03:03:17 +03:00
if (!S_ISREG(ce->ce_mode))
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
continue;
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
if (!pathspec_matches(paths, ce->name))
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
continue;
grep: do not skip unmerged entries when grepping in the working tree. We used to skip unmerged entries, which made sense for grepping in the cached copies, but not for grepping in the working tree. Noticed by Johannes Sixt. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-11-26 23:47:52 +03:00
if (cached) {
if (ce_stage(ce))
continue;
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
hit |= grep_sha1(opt, ce->sha1, ce->name, 0);
grep: do not skip unmerged entries when grepping in the working tree. We used to skip unmerged entries, which made sense for grepping in the cached copies, but not for grepping in the working tree. Noticed by Johannes Sixt. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-11-26 23:47:52 +03:00
}
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
else
hit |= grep_file(opt, ce->name);
grep: do not skip unmerged entries when grepping in the working tree. We used to skip unmerged entries, which made sense for grepping in the cached copies, but not for grepping in the working tree. Noticed by Johannes Sixt. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-11-26 23:47:52 +03:00
if (ce_stage(ce)) {
do {
nr++;
} while (nr < active_nr &&
!strcmp(ce->name, active_cache[nr]->name));
nr--; /* compensate for loop control */
}
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
}
grep: free expressions and patterns when done. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-28 03:27:10 +04:00
free_grep_patterns(opt);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
return hit;
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
static int grep_tree(struct grep_opt *opt, const char **paths,
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
struct tree_desc *tree,
const char *tree_name, const char *base)
{
int len;
int hit = 0;
tree_entry(): new tree-walking helper function This adds a "tree_entry()" function that combines the common operation of doing a "tree_entry_extract()" + "update_tree_entry()". It also has a simplified calling convention, designed for simple loops that traverse over a whole tree: the arguments are pointers to the tree descriptor and a name_entry structure to fill in, and it returns a boolean "true" if there was an entry left to be gotten in the tree. This allows tree traversal with struct tree_desc desc; struct name_entry entry; desc.buf = tree->buffer; desc.size = tree->size; while (tree_entry(&desc, &entry) { ... use "entry.{path, sha1, mode, pathlen}" ... } which is not only shorter than writing it out in full, it's hopefully less error prone too. [ It's actually a tad faster too - we don't need to recalculate the entry pathlength in both extract and update, but need to do it only once. Also, some callers can avoid doing a "strlen()" on the result, since it's returned as part of the name_entry structure. However, by now we're talking just 1% speedup on "git-rev-list --objects --all", and we're definitely at the point where tree walking is no longer the issue any more. ] NOTE! Not everybody wants to use this new helper function, since some of the tree walkers very much on purpose do the descriptor update separately from the entry extraction. So the "extract + update" sequence still remains as the core sequence, this is just a simplified interface. We should probably add a silly two-line inline helper function for initializing the descriptor from the "struct tree" too, just to cut down on the noise from that common "desc" initializer. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-30 20:45:45 +04:00
struct name_entry entry;
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
char *down;
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
int tn_len = strlen(tree_name);
char *path_buf = xmalloc(PATH_MAX + tn_len + 100);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
if (tn_len) {
tn_len = sprintf(path_buf, "%s:", tree_name);
down = path_buf + tn_len;
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
strcat(down, base);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
}
else {
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
down = path_buf;
strcpy(down, base);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
}
len = strlen(path_buf);
tree_entry(): new tree-walking helper function This adds a "tree_entry()" function that combines the common operation of doing a "tree_entry_extract()" + "update_tree_entry()". It also has a simplified calling convention, designed for simple loops that traverse over a whole tree: the arguments are pointers to the tree descriptor and a name_entry structure to fill in, and it returns a boolean "true" if there was an entry left to be gotten in the tree. This allows tree traversal with struct tree_desc desc; struct name_entry entry; desc.buf = tree->buffer; desc.size = tree->size; while (tree_entry(&desc, &entry) { ... use "entry.{path, sha1, mode, pathlen}" ... } which is not only shorter than writing it out in full, it's hopefully less error prone too. [ It's actually a tad faster too - we don't need to recalculate the entry pathlength in both extract and update, but need to do it only once. Also, some callers can avoid doing a "strlen()" on the result, since it's returned as part of the name_entry structure. However, by now we're talking just 1% speedup on "git-rev-list --objects --all", and we're definitely at the point where tree walking is no longer the issue any more. ] NOTE! Not everybody wants to use this new helper function, since some of the tree walkers very much on purpose do the descriptor update separately from the entry extraction. So the "extract + update" sequence still remains as the core sequence, this is just a simplified interface. We should probably add a silly two-line inline helper function for initializing the descriptor from the "struct tree" too, just to cut down on the noise from that common "desc" initializer. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-30 20:45:45 +04:00
while (tree_entry(tree, &entry)) {
strcpy(path_buf + len, entry.path);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
tree_entry(): new tree-walking helper function This adds a "tree_entry()" function that combines the common operation of doing a "tree_entry_extract()" + "update_tree_entry()". It also has a simplified calling convention, designed for simple loops that traverse over a whole tree: the arguments are pointers to the tree descriptor and a name_entry structure to fill in, and it returns a boolean "true" if there was an entry left to be gotten in the tree. This allows tree traversal with struct tree_desc desc; struct name_entry entry; desc.buf = tree->buffer; desc.size = tree->size; while (tree_entry(&desc, &entry) { ... use "entry.{path, sha1, mode, pathlen}" ... } which is not only shorter than writing it out in full, it's hopefully less error prone too. [ It's actually a tad faster too - we don't need to recalculate the entry pathlength in both extract and update, but need to do it only once. Also, some callers can avoid doing a "strlen()" on the result, since it's returned as part of the name_entry structure. However, by now we're talking just 1% speedup on "git-rev-list --objects --all", and we're definitely at the point where tree walking is no longer the issue any more. ] NOTE! Not everybody wants to use this new helper function, since some of the tree walkers very much on purpose do the descriptor update separately from the entry extraction. So the "extract + update" sequence still remains as the core sequence, this is just a simplified interface. We should probably add a silly two-line inline helper function for initializing the descriptor from the "struct tree" too, just to cut down on the noise from that common "desc" initializer. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-30 20:45:45 +04:00
if (S_ISDIR(entry.mode))
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
/* Match "abc/" against pathspec to
* decide if we want to descend into "abc"
* directory.
*/
Remove "pathlen" from "struct name_entry" Since we have the "tree_entry_len()" helper function these days, and don't need to do a full strlen(), there's no point in saving the path length - it's just redundant information. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-21 20:07:46 +03:00
strcpy(path_buf + len + tree_entry_len(entry.path, entry.sha1), "/");
builtin-grep: wildcard pathspec fixes This tweaks the pathspec wildcard used in builtin-grep to match that of ls-files. With this: git grep -e DEBUG -- '*/Kconfig*' would work like the shell script version, and you could even do: git grep -e DEBUG --cached -- '*/Kconfig*' ;# from index git grep -e DEBUG v2.6.12 -- '*/Kconfig*' ;# from rev Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 23:27:56 +04:00
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
if (!pathspec_matches(paths, down))
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
;
tree_entry(): new tree-walking helper function This adds a "tree_entry()" function that combines the common operation of doing a "tree_entry_extract()" + "update_tree_entry()". It also has a simplified calling convention, designed for simple loops that traverse over a whole tree: the arguments are pointers to the tree descriptor and a name_entry structure to fill in, and it returns a boolean "true" if there was an entry left to be gotten in the tree. This allows tree traversal with struct tree_desc desc; struct name_entry entry; desc.buf = tree->buffer; desc.size = tree->size; while (tree_entry(&desc, &entry) { ... use "entry.{path, sha1, mode, pathlen}" ... } which is not only shorter than writing it out in full, it's hopefully less error prone too. [ It's actually a tad faster too - we don't need to recalculate the entry pathlength in both extract and update, but need to do it only once. Also, some callers can avoid doing a "strlen()" on the result, since it's returned as part of the name_entry structure. However, by now we're talking just 1% speedup on "git-rev-list --objects --all", and we're definitely at the point where tree walking is no longer the issue any more. ] NOTE! Not everybody wants to use this new helper function, since some of the tree walkers very much on purpose do the descriptor update separately from the entry extraction. So the "extract + update" sequence still remains as the core sequence, this is just a simplified interface. We should probably add a silly two-line inline helper function for initializing the descriptor from the "struct tree" too, just to cut down on the noise from that common "desc" initializer. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-30 20:45:45 +04:00
else if (S_ISREG(entry.mode))
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
hit |= grep_sha1(opt, entry.sha1, path_buf, tn_len);
tree_entry(): new tree-walking helper function This adds a "tree_entry()" function that combines the common operation of doing a "tree_entry_extract()" + "update_tree_entry()". It also has a simplified calling convention, designed for simple loops that traverse over a whole tree: the arguments are pointers to the tree descriptor and a name_entry structure to fill in, and it returns a boolean "true" if there was an entry left to be gotten in the tree. This allows tree traversal with struct tree_desc desc; struct name_entry entry; desc.buf = tree->buffer; desc.size = tree->size; while (tree_entry(&desc, &entry) { ... use "entry.{path, sha1, mode, pathlen}" ... } which is not only shorter than writing it out in full, it's hopefully less error prone too. [ It's actually a tad faster too - we don't need to recalculate the entry pathlength in both extract and update, but need to do it only once. Also, some callers can avoid doing a "strlen()" on the result, since it's returned as part of the name_entry structure. However, by now we're talking just 1% speedup on "git-rev-list --objects --all", and we're definitely at the point where tree walking is no longer the issue any more. ] NOTE! Not everybody wants to use this new helper function, since some of the tree walkers very much on purpose do the descriptor update separately from the entry extraction. So the "extract + update" sequence still remains as the core sequence, this is just a simplified interface. We should probably add a silly two-line inline helper function for initializing the descriptor from the "struct tree" too, just to cut down on the noise from that common "desc" initializer. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-30 20:45:45 +04:00
else if (S_ISDIR(entry.mode)) {
convert object type handling from a string to a number We currently have two parallel notation for dealing with object types in the code: a string and a numerical value. One of them is obviously redundent, and the most used one requires more stack space and a bunch of strcmp() all over the place. This is an initial step for the removal of the version using a char array found in object reading code paths. The patch is unfortunately large but there is no sane way to split it in smaller parts without breaking the system. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-02-26 22:55:59 +03:00
enum object_type type;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
struct tree_desc sub;
void *data;
Initialize tree descriptors with a helper function rather than by hand. This removes slightly more lines than it adds, but the real reason for doing this is that future optimizations will require more setup of the tree descriptor, and so we want to do it in one place. Also renamed the "desc.buf" field to "desc.buffer" just to trigger compiler errors for old-style manual initializations, making sure I didn't miss anything. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-21 20:08:25 +03:00
unsigned long size;
data = read_sha1_file(entry.sha1, &type, &size);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
if (!data)
die("unable to read tree (%s)",
tree_entry(): new tree-walking helper function This adds a "tree_entry()" function that combines the common operation of doing a "tree_entry_extract()" + "update_tree_entry()". It also has a simplified calling convention, designed for simple loops that traverse over a whole tree: the arguments are pointers to the tree descriptor and a name_entry structure to fill in, and it returns a boolean "true" if there was an entry left to be gotten in the tree. This allows tree traversal with struct tree_desc desc; struct name_entry entry; desc.buf = tree->buffer; desc.size = tree->size; while (tree_entry(&desc, &entry) { ... use "entry.{path, sha1, mode, pathlen}" ... } which is not only shorter than writing it out in full, it's hopefully less error prone too. [ It's actually a tad faster too - we don't need to recalculate the entry pathlength in both extract and update, but need to do it only once. Also, some callers can avoid doing a "strlen()" on the result, since it's returned as part of the name_entry structure. However, by now we're talking just 1% speedup on "git-rev-list --objects --all", and we're definitely at the point where tree walking is no longer the issue any more. ] NOTE! Not everybody wants to use this new helper function, since some of the tree walkers very much on purpose do the descriptor update separately from the entry extraction. So the "extract + update" sequence still remains as the core sequence, this is just a simplified interface. We should probably add a silly two-line inline helper function for initializing the descriptor from the "struct tree" too, just to cut down on the noise from that common "desc" initializer. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-30 20:45:45 +04:00
sha1_to_hex(entry.sha1));
Initialize tree descriptors with a helper function rather than by hand. This removes slightly more lines than it adds, but the real reason for doing this is that future optimizations will require more setup of the tree descriptor, and so we want to do it in one place. Also renamed the "desc.buf" field to "desc.buffer" just to trigger compiler errors for old-style manual initializations, making sure I didn't miss anything. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-21 20:08:25 +03:00
init_tree_desc(&sub, data, size);
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
hit |= grep_tree(opt, paths, &sub, tree_name, down);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
free(data);
}
}
return hit;
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
static int grep_object(struct grep_opt *opt, const char **paths,
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
struct object *obj, const char *name)
{
Remove TYPE_* constant macros and use object_type enums consistently. This updates the type-enumeration constants introduced to reduce the memory footprint of "struct object" to match the type bits already used in the packfile format, by removing the former (i.e. TYPE_* constant macros) and using the latter (i.e. enum object_type) throughout the code for consistency. Eventually we can stop passing around the "type strings" entirely, and this will help - no confusion about two different integer enumeration. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-12 07:45:31 +04:00
if (obj->type == OBJ_BLOB)
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
return grep_sha1(opt, obj->sha1, name, 0);
Remove TYPE_* constant macros and use object_type enums consistently. This updates the type-enumeration constants introduced to reduce the memory footprint of "struct object" to match the type bits already used in the packfile format, by removing the former (i.e. TYPE_* constant macros) and using the latter (i.e. enum object_type) throughout the code for consistency. Eventually we can stop passing around the "type strings" entirely, and this will help - no confusion about two different integer enumeration. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-12 07:45:31 +04:00
if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
struct tree_desc tree;
void *data;
Initialize tree descriptors with a helper function rather than by hand. This removes slightly more lines than it adds, but the real reason for doing this is that future optimizations will require more setup of the tree descriptor, and so we want to do it in one place. Also renamed the "desc.buf" field to "desc.buffer" just to trigger compiler errors for old-style manual initializations, making sure I didn't miss anything. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-21 20:08:25 +03:00
unsigned long size;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
int hit;
data = read_object_with_reference(obj->sha1, tree_type,
Initialize tree descriptors with a helper function rather than by hand. This removes slightly more lines than it adds, but the real reason for doing this is that future optimizations will require more setup of the tree descriptor, and so we want to do it in one place. Also renamed the "desc.buf" field to "desc.buffer" just to trigger compiler errors for old-style manual initializations, making sure I didn't miss anything. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-21 20:08:25 +03:00
&size, NULL);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
if (!data)
die("unable to read tree (%s)", sha1_to_hex(obj->sha1));
Initialize tree descriptors with a helper function rather than by hand. This removes slightly more lines than it adds, but the real reason for doing this is that future optimizations will require more setup of the tree descriptor, and so we want to do it in one place. Also renamed the "desc.buf" field to "desc.buffer" just to trigger compiler errors for old-style manual initializations, making sure I didn't miss anything. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-21 20:08:25 +03:00
init_tree_desc(&tree, data, size);
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
hit = grep_tree(opt, paths, &tree, name, "");
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
free(data);
return hit;
}
Shrink "struct object" a bit This shrinks "struct object" by a small amount, by getting rid of the "struct type *" pointer and replacing it with a 3-bit bitfield instead. In addition, we merge the bitfields and the "flags" field, which incidentally should also remove a useless 4-byte padding from the object when in 64-bit mode. Now, our "struct object" is still too damn large, but it's now less obviously bloated, and of the remaining fields, only the "util" (which is not used by most things) is clearly something that should be eventually discarded. This shrinks the "git-rev-list --all" memory use by about 2.5% on the kernel archive (and, perhaps more importantly, on the larger mozilla archive). That may not sound like much, but I suspect it's more on a 64-bit platform. There are other remaining inefficiencies (the parent lists, for example, probably have horrible malloc overhead), but this was pretty obvious. Most of the patch is just changing the comparison of the "type" pointer from one of the constant string pointers to the appropriate new TYPE_xxx small integer constant. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-15 03:45:13 +04:00
die("unable to grep from object of type %s", typename(obj->type));
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
}
static const char builtin_grep_usage[] =
"git-grep <option>* <rev>* [-e] <pattern> [<path>...]";
git-grep: use a bit more specific error messages. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:44:48 +04:00
static const char emsg_invalid_context_len[] =
"%s: invalid context length argument";
static const char emsg_missing_context_len[] =
"missing context length argument";
static const char emsg_missing_argument[] =
"option requires an argument -%s";
Call setup_git_directory() much earlier This changes the calling convention of built-in commands and passes the "prefix" (i.e. pathname of $PWD relative to the project root level) down to them. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-29 09:44:25 +04:00
int cmd_grep(int argc, const char **argv, const char *prefix)
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
{
int hit = 0;
int cached = 0;
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
int seen_dashdash = 0;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
struct grep_opt opt;
Add "named object array" concept We've had this notion of a "object_list" for a long time, which eventually grew a "name" member because some users (notably git-rev-list) wanted to name each object as it is generated. That object_list is great for some things, but it isn't all that wonderful for others, and the "name" member is generally not used by everybody. This patch splits the users of the object_list array up into two: the traditional list users, who want the list-like format, and who don't actually use or want the name. And another class of users that really used the list as an extensible array, and generally wanted to name the objects. The patch is fairly straightforward, but it's also biggish. Most of it really just cleans things up: switching the revision parsing and listing over to the array makes things like the builtin-diff usage much simpler (we now see exactly how many members the array has, and we don't get the objects reversed from the order they were on the command line). One of the main reasons for doing this at all is that the malloc overhead of the simple object list was actually pretty high, and the array is just a lot denser. So this patch brings down memory usage by git-rev-list by just under 3% (on top of all the other memory use optimizations) on the mozilla archive. It does add more lines than it removes, and more importantly, it adds a whole new infrastructure for maintaining lists of objects, but on the other hand, the new dynamic array code is pretty obvious. The change to builtin-diff-tree.c shows a fairly good example of why an array interface is sometimes more natural, and just much simpler for everybody. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-20 04:42:35 +04:00
struct object_array list = { 0, 0, NULL };
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
const char **paths = NULL;
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
int i;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
memset(&opt, 0, sizeof(opt));
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
opt.prefix_length = (prefix && *prefix) ? strlen(prefix) : 0;
opt.relative = 1;
Add "-h/-H" parsing to "git grep" It turns out that I actually wanted to avoid the filenames (because I didn't care - I just wanted to see the context in which something was used) when doing a grep. But since "git grep" didn't take the "-h" parameter, I ended up having to do "grep -5 -h *.c" instead. So here's a trivial patch that adds "-h" (and thus has to enable -H too) to "git grep" parsing. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-14 21:45:12 +04:00
opt.pathname = 1;
builtin-grep: allow more than one patterns. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:40:49 +04:00
opt.pattern_tail = &opt.pattern_list;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
opt.regflags = REG_NEWLINE;
/*
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
* If there is no -- then the paths must exist in the working
* tree. If there is no explicit pattern specified with -e or
* -f, we take the first unrecognized non option to be the
* pattern, but then what follows it must be zero or more
* valid refs up to the -- (if exists), and then existing
* paths. If there is an explicit pattern, then the first
Fix more typos, primarily in the code The only visible change is that git-blame doesn't understand "--compability" anymore, but it does accept "--compatibility" instead, which is already documented. Signed-off-by: Pavel Roskin <proski@gnu.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-10 09:50:18 +04:00
* unrecognized non option is the beginning of the refs list
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
* that continues up to the -- (if exists), and then paths.
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
*/
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
while (1 < argc) {
const char *arg = argv[1];
argc--; argv++;
if (!strcmp("--cached", arg)) {
cached = 1;
continue;
}
builtin-grep: binary files -a and -I Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 08:05:29 +04:00
if (!strcmp("-a", arg) ||
!strcmp("--text", arg)) {
opt.binary = GREP_BINARY_TEXT;
continue;
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
if (!strcmp("-i", arg) ||
!strcmp("--ignore-case", arg)) {
opt.regflags |= REG_ICASE;
continue;
}
builtin-grep: binary files -a and -I Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 08:05:29 +04:00
if (!strcmp("-I", arg)) {
opt.binary = GREP_BINARY_NOMATCH;
continue;
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
if (!strcmp("-v", arg) ||
!strcmp("--invert-match", arg)) {
opt.invert = 1;
continue;
}
if (!strcmp("-E", arg) ||
!strcmp("--extended-regexp", arg)) {
opt.regflags |= REG_EXTENDED;
continue;
}
builtin-grep: -F (--fixed-strings) Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-10 05:28:41 +04:00
if (!strcmp("-F", arg) ||
!strcmp("--fixed-strings", arg)) {
opt.fixed = 1;
continue;
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
if (!strcmp("-G", arg) ||
!strcmp("--basic-regexp", arg)) {
opt.regflags &= ~REG_EXTENDED;
continue;
}
if (!strcmp("-n", arg)) {
opt.linenum = 1;
continue;
}
Add "-h/-H" parsing to "git grep" It turns out that I actually wanted to avoid the filenames (because I didn't care - I just wanted to see the context in which something was used) when doing a grep. But since "git grep" didn't take the "-h" parameter, I ended up having to do "grep -5 -h *.c" instead. So here's a trivial patch that adds "-h" (and thus has to enable -H too) to "git grep" parsing. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-14 21:45:12 +04:00
if (!strcmp("-h", arg)) {
opt.pathname = 0;
continue;
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
if (!strcmp("-H", arg)) {
Add "-h/-H" parsing to "git grep" It turns out that I actually wanted to avoid the filenames (because I didn't care - I just wanted to see the context in which something was used) when doing a grep. But since "git grep" didn't take the "-h" parameter, I ended up having to do "grep -5 -h *.c" instead. So here's a trivial patch that adds "-h" (and thus has to enable -H too) to "git grep" parsing. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-14 21:45:12 +04:00
opt.pathname = 1;
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
continue;
}
if (!strcmp("-l", arg) ||
Teach git-grep --name-only as synonym for -l I expected git grep --name-only to give me only the file names, much as git diff --name-only only generates filenames. Alas the option is -l, which matches common external greps but doesn't match other parts of the git UI. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-21 07:28:07 +03:00
!strcmp("--name-only", arg) ||
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
!strcmp("--files-with-matches", arg)) {
opt.name_only = 1;
continue;
}
builtin-grep: -L (--files-without-match). Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 08:46:29 +04:00
if (!strcmp("-L", arg) ||
!strcmp("--files-without-match", arg)) {
opt.unmatch_name_only = 1;
continue;
}
builtin-grep: support -c (--count). Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:40:49 +04:00
if (!strcmp("-c", arg) ||
!strcmp("--count", arg)) {
opt.count = 1;
continue;
}
builtin-grep: support -w (--word-regexp). Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:40:49 +04:00
if (!strcmp("-w", arg) ||
!strcmp("--word-regexp", arg)) {
opt.word_regexp = 1;
continue;
}
prefixcmp(): fix-up mechanical conversion. Previous step converted use of strncmp() with literal string mechanically even when the result is only used as a boolean: if (!strncmp("foo", arg, 3)) ==> if (!(-prefixcmp(arg, "foo"))) This step manually cleans them up to read: if (!prefixcmp(arg, "foo")) Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-02-20 12:54:00 +03:00
if (!prefixcmp(arg, "-A") ||
!prefixcmp(arg, "-B") ||
!prefixcmp(arg, "-C") ||
builtin-grep: allow -<n> and -[ABC]<n> notation for context lines. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:17:05 +04:00
(arg[0] == '-' && '1' <= arg[1] && arg[1] <= '9')) {
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
unsigned num;
builtin-grep: allow -<n> and -[ABC]<n> notation for context lines. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:17:05 +04:00
const char *scan;
switch (arg[1]) {
case 'A': case 'B': case 'C':
if (!arg[2]) {
if (argc <= 1)
git-grep: use a bit more specific error messages. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:44:48 +04:00
die(emsg_missing_context_len);
builtin-grep: allow -<n> and -[ABC]<n> notation for context lines. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:17:05 +04:00
scan = *++argv;
argc--;
}
else
scan = arg + 2;
break;
default:
scan = arg + 1;
break;
}
sscanf/strtoul: parse integers robustly * builtin-grep.c (strtoul_ui): Move function definition from here, to... * git-compat-util.h (strtoul_ui): ...here, with an added "base" parameter. * builtin-grep.c (cmd_grep): Update use of strtoul_ui to include base, "10". * builtin-update-index.c (read_index_info): Diagnose an invalid mode integer that is out of range or merely larger than INT_MAX. (cmd_update_index): Use strtoul_ui, not sscanf. * convert-objects.c (write_subdirectory): Likewise. Signed-off-by: Jim Meyering <jim@meyering.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-04-10 03:01:44 +04:00
if (strtoul_ui(scan, 10, &num))
git-grep: use a bit more specific error messages. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:44:48 +04:00
die(emsg_invalid_context_len, scan);
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
switch (arg[1]) {
case 'A':
opt.post_context = num;
break;
builtin-grep: allow -<n> and -[ABC]<n> notation for context lines. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:17:05 +04:00
default:
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
case 'C':
opt.post_context = num;
case 'B':
opt.pre_context = num;
break;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
continue;
}
Teach -f <file> option to builtin-grep. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 00:28:27 +04:00
if (!strcmp("-f", arg)) {
FILE *patterns;
int lno = 0;
char buf[1024];
if (argc <= 1)
git-grep: use a bit more specific error messages. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:44:48 +04:00
die(emsg_missing_argument, arg);
Teach -f <file> option to builtin-grep. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 00:28:27 +04:00
patterns = fopen(argv[1], "r");
if (!patterns)
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
die("'%s': %s", argv[1], strerror(errno));
Teach -f <file> option to builtin-grep. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 00:28:27 +04:00
while (fgets(buf, sizeof(buf), patterns)) {
int len = strlen(buf);
Don't access line[-1] for a zero-length "line" from fgets. A NUL byte at beginning of file, or just after a newline would provoke an invalid buf[-1] access in a few places. * builtin-grep.c (cmd_grep): Don't access buf[-1]. * builtin-pack-objects.c (get_object_list): Likewise. * builtin-rev-list.c (read_revisions_from_stdin): Likewise. * bundle.c (read_bundle_header): Likewise. * server-info.c (read_pack_info_file): Likewise. * transport.c (insert_packed_refs): Likewise. Signed-off-by: Jim Meyering <meyering@redhat.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-01-04 20:37:41 +03:00
if (len && buf[len-1] == '\n')
Teach -f <file> option to builtin-grep. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 00:28:27 +04:00
buf[len-1] = 0;
/* ignore empty line like grep does */
if (!buf[0])
continue;
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
append_grep_pattern(&opt, xstrdup(buf),
argv[1], ++lno,
GREP_PATTERN);
Teach -f <file> option to builtin-grep. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 00:28:27 +04:00
}
fclose(patterns);
argv++;
argc--;
continue;
}
git-grep: boolean expression on pattern matching. This extends the behaviour of git-grep when multiple -e options are given. So far, we allowed multiple -e to behave just like regular grep with multiple -e, i.e. the patterns are OR'ed together. With this change, you can also have multiple patterns AND'ed together, or form boolean expressions, like this (the parentheses are quoted from the shell in this example): $ git grep -e _PATTERN --and \( -e atom -e token \) Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-30 14:04:05 +04:00
if (!strcmp("--not", arg)) {
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
append_grep_pattern(&opt, arg, "command line", 0,
GREP_NOT);
git-grep: boolean expression on pattern matching. This extends the behaviour of git-grep when multiple -e options are given. So far, we allowed multiple -e to behave just like regular grep with multiple -e, i.e. the patterns are OR'ed together. With this change, you can also have multiple patterns AND'ed together, or form boolean expressions, like this (the parentheses are quoted from the shell in this example): $ git grep -e _PATTERN --and \( -e atom -e token \) Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-30 14:04:05 +04:00
continue;
}
if (!strcmp("--and", arg)) {
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
append_grep_pattern(&opt, arg, "command line", 0,
GREP_AND);
git-grep: boolean expression on pattern matching. This extends the behaviour of git-grep when multiple -e options are given. So far, we allowed multiple -e to behave just like regular grep with multiple -e, i.e. the patterns are OR'ed together. With this change, you can also have multiple patterns AND'ed together, or form boolean expressions, like this (the parentheses are quoted from the shell in this example): $ git grep -e _PATTERN --and \( -e atom -e token \) Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-30 14:04:05 +04:00
continue;
}
if (!strcmp("--or", arg))
continue; /* no-op */
if (!strcmp("(", arg)) {
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
append_grep_pattern(&opt, arg, "command line", 0,
GREP_OPEN_PAREN);
git-grep: boolean expression on pattern matching. This extends the behaviour of git-grep when multiple -e options are given. So far, we allowed multiple -e to behave just like regular grep with multiple -e, i.e. the patterns are OR'ed together. With this change, you can also have multiple patterns AND'ed together, or form boolean expressions, like this (the parentheses are quoted from the shell in this example): $ git grep -e _PATTERN --and \( -e atom -e token \) Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-30 14:04:05 +04:00
continue;
}
if (!strcmp(")", arg)) {
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
append_grep_pattern(&opt, arg, "command line", 0,
GREP_CLOSE_PAREN);
git-grep: boolean expression on pattern matching. This extends the behaviour of git-grep when multiple -e options are given. So far, we allowed multiple -e to behave just like regular grep with multiple -e, i.e. the patterns are OR'ed together. With this change, you can also have multiple patterns AND'ed together, or form boolean expressions, like this (the parentheses are quoted from the shell in this example): $ git grep -e _PATTERN --and \( -e atom -e token \) Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-30 14:04:05 +04:00
continue;
}
grep --all-match This lets you say: git grep --all-match -e A -e B -e C to find lines that match A or B or C but limit the matches from the files that have all of A, B and C. This is different from git grep -e A --and -e B --and -e C in that the latter looks for a single line that has all of these at the same time. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-28 04:50:52 +04:00
if (!strcmp("--all-match", arg)) {
opt.all_match = 1;
continue;
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
if (!strcmp("-e", arg)) {
if (1 < argc) {
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
append_grep_pattern(&opt, argv[1],
"-e option", 0,
GREP_PATTERN);
builtin-grep: allow more than one patterns. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:40:49 +04:00
argv++;
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
argc--;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
continue;
}
git-grep: use a bit more specific error messages. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:44:48 +04:00
die(emsg_missing_argument, arg);
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
}
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
if (!strcmp("--full-name", arg)) {
opt.relative = 0;
continue;
}
git-grep: fix parsing of pathspec separator '--' We used to misparse git grep -e foo -- '*.sh' Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:31:50 +04:00
if (!strcmp("--", arg)) {
/* later processing wants to have this at argv[1] */
argv--;
argc++;
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
break;
git-grep: fix parsing of pathspec separator '--' We used to misparse git grep -e foo -- '*.sh' Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-04 13:31:50 +04:00
}
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
if (*arg == '-')
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
usage(builtin_grep_usage);
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
/* First unrecognized non-option token */
builtin-grep: allow more than one patterns. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:40:49 +04:00
if (!opt.pattern_list) {
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
append_grep_pattern(&opt, arg, "command line", 0,
GREP_PATTERN);
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
break;
}
else {
/* We are looking at the first path or rev;
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
* it is found at argv[1] after leaving the
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
* loop.
*/
argc++; argv--;
break;
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
}
}
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
builtin-grep: allow more than one patterns. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-03 02:40:49 +04:00
if (!opt.pattern_list)
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
die("no pattern given.");
builtin-grep: -F (--fixed-strings) Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-10 05:28:41 +04:00
if ((opt.regflags != REG_NEWLINE) && opt.fixed)
die("cannot mix --fixed-strings and regexp");
builtin-grep: make pieces of it available as library. This makes three functions and associated option structures from builtin-grep available from other parts of the system. * options to drive built-in grep engine is stored in struct grep_opt; * pattern strings and extended grep expressions are added to struct grep_opt with append_grep_pattern(); * when finished calling append_grep_pattern(), call compile_grep_patterns() to prepare for execution; * call grep_buffer() to find matches in the in-core buffer. This also adds an internal option "status_only" to grep_opt, which suppresses any output from grep_buffer(). Callers of the function as library can use it to check if there is a match without producing any output. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-18 03:02:52 +04:00
compile_grep_patterns(&opt);
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
/* Check revs and then paths */
for (i = 1; i < argc; i++) {
const char *arg = argv[i];
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
unsigned char sha1[20];
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
/* Is it a rev? */
if (!get_sha1(arg, sha1)) {
struct object *object = parse_object(sha1);
if (!object)
die("bad object %s", arg);
Add "named object array" concept We've had this notion of a "object_list" for a long time, which eventually grew a "name" member because some users (notably git-rev-list) wanted to name each object as it is generated. That object_list is great for some things, but it isn't all that wonderful for others, and the "name" member is generally not used by everybody. This patch splits the users of the object_list array up into two: the traditional list users, who want the list-like format, and who don't actually use or want the name. And another class of users that really used the list as an extensible array, and generally wanted to name the objects. The patch is fairly straightforward, but it's also biggish. Most of it really just cleans things up: switching the revision parsing and listing over to the array makes things like the builtin-diff usage much simpler (we now see exactly how many members the array has, and we don't get the objects reversed from the order they were on the command line). One of the main reasons for doing this at all is that the malloc overhead of the simple object list was actually pretty high, and the array is just a lot denser. So this patch brings down memory usage by git-rev-list by just under 3% (on top of all the other memory use optimizations) on the mozilla archive. It does add more lines than it removes, and more importantly, it adds a whole new infrastructure for maintaining lists of objects, but on the other hand, the new dynamic array code is pretty obvious. The change to builtin-diff-tree.c shows a fairly good example of why an array interface is sometimes more natural, and just much simpler for everybody. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-20 04:42:35 +04:00
add_object_array(object, arg, &list);
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
continue;
}
if (!strcmp(arg, "--")) {
i++;
seen_dashdash = 1;
}
break;
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
}
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
/* The rest are paths */
if (!seen_dashdash) {
int j;
builtin-grep: typofix Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-10 05:15:21 +04:00
for (j = i; j < argc; j++)
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
verify_filename(prefix, argv[j]);
}
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
if (i < argc) {
builtin-grep: tighten argument parsing. I mistyped git grep next -e '"^@"' '*.c' and got many hits that contain "next" without complaint. Obviously what I meant to say was: git grep -e '"^@"' next -- '*.c' This tightens the argument parsing rule a bit: - All "grep" parameters should come first; - If there is no -e nor -f to specify pattern, the first non option string is the parameter; - After that, zero or more revs can follow. - An optional '--' can be present, and is skipped. - All the rest are pathspecs. If '--' was not there, they must be paths that exist in the working tree. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 10:55:47 +04:00
paths = get_pathspec(prefix, argv + i);
git-grep: show pathnames relative to the current directory By default, the command shows pathnames relative to the current directory. Use --full-name (the same flag to do so in ls-files) if you want to see the full pathname relative to the project root. This makes it very pleasant to run in Emacs compilation (or "grep-find") buffer. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-08-11 11:44:42 +04:00
if (opt.prefix_length && opt.relative) {
/* Make sure we do not get outside of paths */
for (i = 0; paths[i]; i++)
if (strncmp(prefix, paths[i], opt.prefix_length))
die("git-grep: cannot generate relative filenames containing '..'");
}
}
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
else if (prefix) {
paths = xcalloc(2, sizeof(const char *));
paths[0] = prefix;
paths[1] = NULL;
}
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
Add "named object array" concept We've had this notion of a "object_list" for a long time, which eventually grew a "name" member because some users (notably git-rev-list) wanted to name each object as it is generated. That object_list is great for some things, but it isn't all that wonderful for others, and the "name" member is generally not used by everybody. This patch splits the users of the object_list array up into two: the traditional list users, who want the list-like format, and who don't actually use or want the name. And another class of users that really used the list as an extensible array, and generally wanted to name the objects. The patch is fairly straightforward, but it's also biggish. Most of it really just cleans things up: switching the revision parsing and listing over to the array makes things like the builtin-diff usage much simpler (we now see exactly how many members the array has, and we don't get the objects reversed from the order they were on the command line). One of the main reasons for doing this at all is that the malloc overhead of the simple object list was actually pretty high, and the array is just a lot denser. So this patch brings down memory usage by git-rev-list by just under 3% (on top of all the other memory use optimizations) on the mozilla archive. It does add more lines than it removes, and more importantly, it adds a whole new infrastructure for maintaining lists of objects, but on the other hand, the new dynamic array code is pretty obvious. The change to builtin-diff-tree.c shows a fairly good example of why an array interface is sometimes more natural, and just much simpler for everybody. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-20 04:42:35 +04:00
if (!list.nr)
builtin-grep: do not use setup_revisions() Grep may want to grok multiple revisions, but it does not make much sense to walk revisions while doing so. This stops calling the code to parse parameters for the revision walker. The parameter parsing for the optional "-e" option becomes a lot simpler with it as well. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-02 02:58:29 +04:00
return !grep_cache(&opt, paths, cached);
Teach -f <file> option to builtin-grep. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 00:28:27 +04:00
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
if (cached)
Teach -f <file> option to builtin-grep. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-09 00:28:27 +04:00
die("both --cached and trees are given.");
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
Add "named object array" concept We've had this notion of a "object_list" for a long time, which eventually grew a "name" member because some users (notably git-rev-list) wanted to name each object as it is generated. That object_list is great for some things, but it isn't all that wonderful for others, and the "name" member is generally not used by everybody. This patch splits the users of the object_list array up into two: the traditional list users, who want the list-like format, and who don't actually use or want the name. And another class of users that really used the list as an extensible array, and generally wanted to name the objects. The patch is fairly straightforward, but it's also biggish. Most of it really just cleans things up: switching the revision parsing and listing over to the array makes things like the builtin-diff usage much simpler (we now see exactly how many members the array has, and we don't get the objects reversed from the order they were on the command line). One of the main reasons for doing this at all is that the malloc overhead of the simple object list was actually pretty high, and the array is just a lot denser. So this patch brings down memory usage by git-rev-list by just under 3% (on top of all the other memory use optimizations) on the mozilla archive. It does add more lines than it removes, and more importantly, it adds a whole new infrastructure for maintaining lists of objects, but on the other hand, the new dynamic array code is pretty obvious. The change to builtin-diff-tree.c shows a fairly good example of why an array interface is sometimes more natural, and just much simpler for everybody. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-20 04:42:35 +04:00
for (i = 0; i < list.nr; i++) {
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
struct object *real_obj;
Add "named object array" concept We've had this notion of a "object_list" for a long time, which eventually grew a "name" member because some users (notably git-rev-list) wanted to name each object as it is generated. That object_list is great for some things, but it isn't all that wonderful for others, and the "name" member is generally not used by everybody. This patch splits the users of the object_list array up into two: the traditional list users, who want the list-like format, and who don't actually use or want the name. And another class of users that really used the list as an extensible array, and generally wanted to name the objects. The patch is fairly straightforward, but it's also biggish. Most of it really just cleans things up: switching the revision parsing and listing over to the array makes things like the builtin-diff usage much simpler (we now see exactly how many members the array has, and we don't get the objects reversed from the order they were on the command line). One of the main reasons for doing this at all is that the malloc overhead of the simple object list was actually pretty high, and the array is just a lot denser. So this patch brings down memory usage by git-rev-list by just under 3% (on top of all the other memory use optimizations) on the mozilla archive. It does add more lines than it removes, and more importantly, it adds a whole new infrastructure for maintaining lists of objects, but on the other hand, the new dynamic array code is pretty obvious. The change to builtin-diff-tree.c shows a fairly good example of why an array interface is sometimes more natural, and just much simpler for everybody. Signed-off-by: Linus Torvalds <torvalds@osdl.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-20 04:42:35 +04:00
real_obj = deref_tag(list.objects[i].item, NULL, 0);
if (grep_object(&opt, paths, real_obj, list.objects[i].name))
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
hit = 1;
}
grep: free expressions and patterns when done. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-09-28 03:27:10 +04:00
free_grep_patterns(&opt);
built-in "git grep" This attempts to set up built-in "git grep" to further reduce our dependence on the shell, while at the same time optionally allowing to run grep against object database. You could do funky things like these: git grep --cached -e pattern ;# grep from index git grep -e pattern master ;# or in a rev git grep -e pattern master next ;# or in multiple revs git grep -e pattern pu^@ ;# even like this with an ;# extension from another topic ;-) git grep -e pattern master..next ;# or even from rev ranges git grep -e pattern master~20:Documentation ;# or an arbitrary tree git grep -e pattern next:git-commit.sh ;# or an arbitrary blob Right now, it does not understand and/or obey many options grep should accept, and the pattern must be given with -e option due to the way the parameter parser is structured, both of which obviously need to be fixed for usability. But this is going in the right direction. The shell script version is one of the worst Portability offender in the git barebone Porcelainish; it uses xargs -0 to pass paths around and shell arrays to sift flags and parameters. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-01 10:28:15 +04:00
return !hit;
}