One of the ways that locking might fail is that there is a
DF conflict between two refs (e.g., you want to lock
"foo/bar" but "foo" already exists). In this case, we return
an error, but there is no way for the caller to know the
specific problem.
This patch sets errno to ENOTDIR, which is the most sensible
code. It's what we would see if the refs were stored purely
in the filesystem (but these days we must check the
namespace manually due to packed refs).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Avoid "diff-index --cached" optimization under --find-copies-harder
When find-copies-harder is in effect, the diff frontends are expected to
feed all paths, not just changed paths, to the diffcore, so that copy
sources can be picked up. In such a case, not descending into subtrees
using the cache-tree information is simply wrong.
When running "diff-index --cached" after making a change to only a small
portion of the index, there is no point unpacking unchanged subtrees into
the index recursively, only to find that all entries match anyway. Tweak
unpack_trees() logic that is used to read in the tree object to catch the
case where the tree entry we are looking at matches the index as a whole
by looking at the cache-tree.
As an exercise, after modifying a few paths in the kernel tree, here are
a few numbers on my Athlon 64X2 3800+:
(without patch, hot cache)
$ /usr/bin/time git diff --cached --raw
:100644 100644 b57e1f5... e69de29... M Makefile
:100644 000000 8c86b72... 0000000... D arch/x86/Makefile
:000000 100644 0000000... e69de29... A arche
0.07user 0.02system 0:00.09elapsed 102%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+9407minor)pagefaults 0swaps
(with patch, hot cache)
$ /usr/bin/time ../git.git/git-diff --cached --raw
:100644 100644 b57e1f5... e69de29... M Makefile
:100644 000000 8c86b72... 0000000... D arch/x86/Makefile
:000000 100644 0000000... e69de29... A arche
0.02user 0.00system 0:00.02elapsed 103%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2446minor)pagefaults 0swaps
Cold cache numbers are very impressive, but it does not matter very much
in practice:
(without patch, cold cache)
$ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches'
$ /usr/bin/time git diff --cached --raw
:100644 100644 b57e1f5... e69de29... M Makefile
:100644 000000 8c86b72... 0000000... D arch/x86/Makefile
:000000 100644 0000000... e69de29... A arche
0.06user 0.17system 0:10.26elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
247032inputs+0outputs (1172major+8237minor)pagefaults 0swaps
(with patch, cold cache)
$ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches'
$ /usr/bin/time ../git.git/git-diff --cached --raw
:100644 100644 b57e1f5... e69de29... M Makefile
:100644 000000 8c86b72... 0000000... D arch/x86/Makefile
:000000 100644 0000000... e69de29... A arche
0.02user 0.01system 0:01.01elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
18440inputs+0outputs (79major+2369minor)pagefaults 0swaps
This is one of the oldest scripts; update it to match more modern style.
Notably, we should:
- Put the test title on the same line as the "test_expect_success", and
end the line with a single-quote to begin the body of the test which is
one multi-line string; and
- Run as many commands inside test_expect_success, not outside, to catch
unexpected breakages.
cache-tree.c::cache_tree_find(): simplify internal API
Earlier cache_tree_find() needs to be called with a valid cache_tree,
but repeated look-up may find an invalid or missing cache_tree in between.
Help simplify the callers by returning NULL to mean "nothing appropriate
found" when the input is NULL.
This particular readlink call never NUL-terminated its
result, making it a potential source of bugs (though there
is no bug now, as it currently always respects the length
field). Let's just switch it to strbuf_readlink which is
shorter and less error-prone.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t3701: ensure correctly set up repository after skipped tests
There are two tests that are skipped if file modes are not obeyed by the
file system. In this case, the subsequent test failed because the
repository was in an unexpected state. This corrects it.
Signed-off-by: Johannes Sixt <j6t@kdbg.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
parse-opts: add OPT_FILENAME and transition builtins
Commit dbd0f5c (Files given on the command line are relative to $cwd,
2008-08-06) introduced parse_options_fix_filename() as a minimal fix.
OPT_FILENAME is intended to be a more robust fix for the same issue.
OPT_FILENAME and its associated enum OPTION_FILENAME are used to
represent filename options within the parse options API.
This option is similar to OPTION_STRING. If --no is prefixed to the
option the filename is unset. If no argument is given and the default
value is set, the filename is set to the default value. The difference
is that the filename is prefixed with the prefix passed to
parse_options() (or parse_options_start()).
Update git-apply, git-commit, git-fmt-merge-msg, and git-tag to use
OPT_FILENAME with their filename options. Also, rename
parse_options_fix_filename() to fix_filename() as it is no longer
extern.
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
To give OPT_FILENAME the prefix, we pass the prefix to parse_options()
which passes the prefix to parse_options_start() which sets the prefix
member of parse_opts_ctx accordingly. If there isn't a prefix in the
calling context, passing NULL will suffice.
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jc/mktree:
mktree: validate entry type in input
mktree --batch: build more than one tree object
mktree --missing: updated usage message and man page
mktree --missing: allow missing objects
t1010: add mktree test
mktree: do not barf on a submodule commit
builtin-mktree.c: use a helper function to handle one line of input
mktree: use parse-options
build-in git-mktree
* master: (654 commits)
http-push.c::remove_locks(): fix use after free
t/t3400-rebase.sh: add more tests to help migrating git-rebase.sh to C
post-receive-email: hooks.showrev: show how to include both web link and patch
MinGW: Fix compiler warning in merge-recursive
MinGW: Add a simple getpass()
MinGW: use POSIX signature of waitpid()
MinGW: the path separator to split GITPERLLIB is ';' on Win32
MinGW: Scan for \r in addition to \n when reading shbang lines
gitweb: Sanitize title attribute in format_subject_html
Terminate argv with NULL before calling setup_revisions()
doc/git-rebase.txt: remove mention of multiple strategies
git-send-email: Handle quotes when parsing .mailrc files
git-svn: add --authors-prog option
git-svn: Set svn.authorsfile if it is passed to git svn clone
git-svn: Correctly report max revision when following deleted paths
git-svn: Fix for svn paths removed > log-window-size revisions ago
git-svn testsuite: use standard configuration for Subversion tools
grep: fix word-regexp colouring
completion: use git rev-parse to detect bare repos
Cope better with a _lot_ of packs
...
Change xdl_merge to generate output even for null merges
xdl_merge used to have a check to ensure that there was at least
some change in one or other side being merged but this suppressed
output for the degenerate case when base, local and remote
contents were all identical.
Removing this check enables correct output in the degenerate case
and xdl_free_script handles freeing NULL scripts so there is no
need to have the check for these calls.
Signed-off-by: Charles Bailey <charles@hashpling.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
gitk: Add option 'Simple history' to the options menu
When clicked, the option --simplify-by-decoration is added to gitk/git log.
This yields to a simplified history where only decorated commits are shown,
i.e. those with a yellow tag or a green branch flag.
Signed-off-by: Dirk Suesserott <newsletter@dirk.my1.cc> Signed-off-by: Paul Mackerras <paulus@samba.org>
After bol is forwarded, it doesn't represent the beginning of the line
any more. This means that the beginning-of-line marker (^) mustn't match,
i.e. the regex flag REG_NOTBOL needs to be set.
Commit dbd0f5c7 (Files given on the command line are relative to $cwd,
2008-08-06) only fixed git-commit and git-tag. But, git-apply and
git-fmt-merge-msg didn't get the update and exhibit the same behavior.
Fix them and add tests for "apply --build-fake-ancestor" and
"fmt-merge-msg -F".
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit dbd0f5c7 (Files given on the command line are relative to $cwd,
2008-08-06) introduced parse_options_fix_filename() as a quick fix for
filename arguments used in the parse options API.
git-commit was still broken. This means
git commit -F log -t temp
in a subdirectory would make git think the log message should be taken
from temp instead of log.
This is because parse_options_fix_filename() calls prefix_filename()
which uses a single static char buffer to do its work. Making two calls
with two char pointers causes the pointers to alias. To prevent
aliasing, we duplicate the string returned by
parse_options_fix_filename().
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
post-receive-email: hooks.showrev: show how to include both web link and patch
Add a comment showing how to include a web link (i.e. gitweb/cgit)
and a patch in the email that is sent for each pushed commit.
The quoting was tricky enough that it's worth documenting. To add
two blank lines (i.e. put \n\n in the printf), you would need to
say \\\\n\\\\n, and in the end, the pair of "echo" statements seemed
better. This is used in glibc.git repository:
GCC 4.4.0 on Windows does not like the format %zu. It is quite unlikely,
though, that we need more merge bases than a %d can display, so replace
the %zu by a %d.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Steffen Prohaska <prohaska@zib.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We need getpass() to activate curl on MinGW. Although the default
Makefile currently has 'NO_CURL = YesPlease', msysgit releases do
provide curl support, so getpass() is used.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Steffen Prohaska <prohaska@zib.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Git's source code expects waitpid() to return a signed int status.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Steffen Prohaska <prohaska@zib.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
MinGW: the path separator to split GITPERLLIB is ';' on Win32
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Steffen Prohaska <prohaska@zib.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jn/gitweb-cleanup:
gitweb: Remove unused $hash_base parameter from normalize_link_target
gitweb: Simplify snapshot format detection logic in evaluate_path_info
gitweb: Use capturing parentheses only when you intend to capture
gitweb: Replace wrongly added tabs with spaces
gitweb: Use block form of map/grep in a few cases more
gitweb: Always use three argument form of open
gitweb: Always use three argument form of open
gitweb: Do not use bareword filehandles
* 'cc/bisect' (early part):
bisect: make "git bisect" use new "--next-all" bisect-helper function
bisect: add "check_good_are_ancestors_of_bad" function
bisect: implement the "check_merge_bases" function
bisect: automatically sort sha1_array if needed when looking it up
bisect: make skipped array functions more generic
bisect: remove too much function nesting
bisect: use new "struct argv_array" to prepare argv for "setup_revisions"
bisect: store good revisions in a "sha1_array"
bisect: implement "rev_argv_push" to fill an argv with revs
bisect: use "sha1_array" to store skipped revisions
am: simplify "sq" function by using "git rev-parse --sq-quote"
bisect: use "git rev-parse --sq-quote" instead of a custom "sq" function
rev-parse: add --sq-quote to shell quote arguments
rev-list: remove stringed output flag from "show_bisect_vars"
bisect--helper: remove "--next-vars" option as it is now useless
bisect: use "git bisect--helper --next-exit" in "git-bisect.sh"
bisect--helper: add "--next-exit" to output bisect results
bisect: move common bisect functionality to "bisect_common"
rev-list: refactor printing bisect vars
rev-list: make "estimate_bisect_steps" non static
Later in the same header, compilation with -c99 is made to fail if _XPG6 is
not set, like this:
#if defined(_STDC_C99) && (defined(__XOPEN_OR_POSIX) && !defined(_XPG6))
#error "Compiler or options invalid for pre-UNIX 03 X/Open applications \
and pre-2001 POSIX applications"
#elif ...
The problem is that they check things in an order that is inconvenient for
us. When they see _XOPEN_SOURCE_EXTENDED, they declare that we are XPG4v2,
regardless of the value of _XOPEN_SOURCE.
To work around this problem, do not define _XOPEN_SOURCE_EXTENDED on
Sun's.
Note that "-g" no longer uses an equals '=' sign for its optional
arguments, but "--reflog" still does. This is normal behavior for parse
options, as arguments to "-g" are put immediately after the option with
no space.
For example
git show-branch -g=4
is now
git show-branch -g4
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
parse-options: add PARSE_OPT_LITERAL_ARGHELP for complicated argh's
Usually, the argh element in struct option points at a placeholder value
(e.g. "val"), and is shown in the usage message as
--option=<val>
by enclosing the string inside of angle brackets.
When the option is more complex (e.g. optional arguments separated by a
comma), you would want to produce a usage message that looks like
--option=<val1>[,<val2>]
In such a case, the caller can pass a string to argh with placeholders
already enclosed in necessary angle brackets (e.g. "<val1>[,<val2>]")
and set this flag.
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
On IRIX 6.5 CP1251 is unknown, but WIN1251 (which seems to be a
non-standard name) is known. On Solaris 10, the opposite is true. Solaris
also knows CP1251 as WINDOWS-1251, but this too is not recognized on IRIX.
I could not find a name that both platforms recognized for this character
set.
An alternative character set which covers the same alphabet seems to be the
ISO8859-5 character set. Both platforms support this character set, so use
it instead.
This allows t8005.4 to pass on Solaris 7, and part of the test to pass on
IRIX. (My IRIX can't convert SJIS to UTF-8 :(
Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil> Signed-off-by: Junio C Hamano <gitster@pobox.com>
gitweb: Sanitize title attribute in format_subject_html
Replace control characters with question mark '?' (like in
chop_and_esc_str).
A little background: some web browsers turn on strict (and
unforgiving) XML validating mode for XHTML documents served using
application/xhtml+xml content type. This means among others that
control characters are forbidden to appear in gitweb output.
CGI.pm does by default slight escaping (using simple_escape subroutine
from CGI::Util) of all _attribute_ values (depending on the value of
autoEscape, by default on). This escaping, at least in CGI.pm version
3.10 (most current version at CPAN is 3.43), is minimal: only '"',
'&', '<' and '>' are escaped using named HTML entity references
(", &, < and > respectively). But simple_escape does
not do escaping of control characters such as ^X which are invalid in
XHTML (in strict mode).
If by some accident commit message do contain some control character
in first 50 characters (more or less) of first line of commit message,
and this line is longer than 50 characters (so gitweb shortens it for
display), then gitweb would put this control character in title
attribute (and CGI.pm would not remove them). The tag _contents_ is
safe because it is escaped using esc_html() explicitly, and it
replaces control characters by their printable representation.
While at it: chop_and_escape_str doesn't need capturing group.
Noticed-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Terminate argv with NULL before calling setup_revisions()
It is convention that argv should be terminated with NULL, even if
argc is used to specify the size of argv. setup_revisions() requires
this and may segfault otherwise.
This patch makes sure that all argv (that I can find) is NULL terminated.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* git://git.bogomips.org/git-svn:
git-svn: add --authors-prog option
git-svn: Set svn.authorsfile if it is passed to git svn clone
git-svn: Correctly report max revision when following deleted paths
git-svn: Fix for svn paths removed > log-window-size revisions ago
git-svn testsuite: use standard configuration for Subversion tools
git-send-email: Handle quotes when parsing .mailrc files
It is legal and not uncommon to use quotes in a .mailrc file so
you can include a persons fullname as well as their email alias.
Handle this by using quotewords instead of split when parsing
.mailrc files.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a new option, --authors-prog, to git-svn that allows a more flexible
alternative (or supplement) to --authors-file. This allows more
advanced username operations than the authors file will allow. For
example, one may look up Subversion users via LDAP, or may generate the
name and email address from the Subversion username.
Notes:
* If both --authors-name and --authors-prog are given, the former is
tried first, falling back to the later.
* The program is called once per unique SVN username, and the result is
cached.
* The command-line argument must be the path to a program, not a generic
shell command line. The absolute path to this program is taken at
startup since the git-svn script changes directory during operation.
* The option is not enabled for `git svn log'.
[ew: fixed case where neither --authors-(name|prog) were defined] Signed-off-by: Mark Lodato <lodatom@gmail.com> Acked-by: Eric Wong <normalperson@yhbt.net>
git-svn: Fix for svn paths removed > log-window-size revisions ago
Instead of trying to find the end of the commit history only in the
last window, track if we have seen commits yet, and use that to judge
if we need to backtrack and look for a tail. Otherwise, conversion
can silently lose up to 100 revisions of a branch if it was deleted
>100 revisions ago.
Signed-off-by: Alex Vandiver <alexmv@mit.edu> Acked-by: Eric Wong <normalperson@yhbt.net>
git-svn testsuite: use standard configuration for Subversion tools
I have tweaked configuration in my ~/.subversion directory, namely I am
running auto-properties and automatically adding '$Id$' expansion to
every file. This choke the last test named 'proplist' from
t9101-git-svn-props.sh, because one more property, svn:keywords is
automatically added.
I had just wrapped svn invocation with the svn_cmd that specifies empty
directory via --config-dir argument. Since the latter is the global
option, it should be recognized by all svn subcommands, so no
regressions will be introduced.
Now svn_cmd is used everywhere, not just in the failed test module: this
should guard us from the future clashes with user-defined configuration
tweaks.
Signed-off-by: Eygene Ryabinkin <rea-git@codelabs.ru> Acked-by: Eric Wong <normalperson@yhbt.net>
* maint:
grep: fix word-regexp colouring
completion: use git rev-parse to detect bare repos
Cope better with a _lot_ of packs
for-each-ref: fix segfault in copy_email
As noticed by Dmitry Gryazin: When a pattern is found but it doesn't
start and end at word boundaries, bol is forwarded to after the match and
the pattern is searched again. When a pattern is finally found between
word boundaries, the match offsets are off by the number of characters
that have been skipped.
This patch corrects the offsets to be relative to the value of bol as
passed to match_one_pattern() by its caller.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx> Signed-off-by: Junio C Hamano <gitster@pobox.com>
You might end up with a situation where you have tons of pack files, e.g.
when using hg2git. In this situation, all kinds of operations may
end up with a "too many files open" error. Let's recover gracefully from
that.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Looks-right-to-me-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
This allows you to discard the cache-tree information before writing the
tree out of the index (i.e. it always recomputes the tree object names for
all the subtrees).
This is only useful as a debug option, so I did not bother documenting it.
The v0.99 tag is special in that it contains no "tagger"
header.
The bug is obvious in copy_email, which carefully checks to
make sure the result of a strchr is non-NULL, but only after
already having used it to perform other work. The fix is to
move the check up.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Extend sample pre-commit hook to check for non ascii filenames
At the moment non-ascii encodings of filenames are not portably
converted between different filesystems by git. This will most likely
change in the future but to allow repositories to be portable among
different file/operating systems this check is enabled by default.
Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t5100: use ancient encoding syntax for backwards compatibility
Some ancient platforms do not have an extensive list of alternate names for
character encodings. For example, Solaris 7 does not know that ISO-8859-1
is the same as ISO8859-1. Modern platforms do know this, so use the older
names.
Some ancient platforms do not have an extensive list of alternate names for
character encodings. For example, Solaris 7 does not know that ISO-8859-1
is the same as ISO8859-1. Modern platforms do know this, so use the older
name.
Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t3901: Use ISO8859-1 instead of ISO-8859-1 for backward compatibility
Some ancient platforms do not have an extensive list of alternate names for
character encodings. For example, Solaris 7 does not know that ISO-8859-1
is the same as ISO8859-1. Modern platforms do know this, so use the older
name.
Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil> Signed-off-by: Junio C Hamano <gitster@pobox.com>
builtin-mailinfo.c: use "ISO8859-1" instead of "latin1" as fallback encoding
Some platforms do not understand the character encoding "latin1" which is
another name for "ISO8859-1". So use "ISO8859-1" instead which all tested
platforms understand.
Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil> Signed-off-by: Junio C Hamano <gitster@pobox.com>
builtin-mailinfo.c: compare character encodings case insensitively
When converting between character encodings, git tests whether the "from"
encoding and the "to" encoding have the same name. git should perform this
test case insensitively so that e.g. utf-8 is not seen as a different
encoding than UTF-8.
Additionally, it is not necessary to call tolower() anymore on the encodings
extracted from the mail message.
Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t3900: use ancient iconv names for backward compatibility
Some old iconv implementations do not have many alternate names and/or
do not match character encoding names case insensitively. These
implementations can not tell that utf-8 and UTF-8 are the same encoding
and fail when trying to do the conversion. So use the old names, which
modern implementations still support.
gitk: Handle msysGit version during version comparisons
msysGit generates version strings with text appended which cannot
be used with vcompare; trying to use them generates a Tcl error.
Limit git_version to the first three digits which are the real git
version to avoid this error.
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net> Signed-off-by: Paul Mackerras <paulus@samba.org>
* ar/unlink-err:
print unlink(2) errno in copy_or_link_directory
replace direct calls to unlink(2) with unlink_or_warn
Introduce an unlink(2) wrapper which gives warning if unlink failed
bisect: check ancestors without forking a "git rev-list" process
We must save the pending commits that will be used during revision
walking and unparse them after, because we want to leave a clean
state for the next revision walking that will try to find the best
bisection point.
As we don't fork a process anymore to call "git rev-list", we need
to remove the use of GIT_TRACE to check how "git rev-list" is
called from the t6030 test that uses it.
Signed-off-by: Christian Couder <chriscool@tuxfamily.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>