git.git
14 years agoblame,cat-file: Demonstrate --textconv is wrongly running converter on symlinks
Kirill Smelkov [Wed, 29 Sep 2010 11:35:23 +0000 (15:35 +0400)]
blame,cat-file: Demonstrate --textconv is wrongly running converter on symlinks

git blame --textconv is wrongly calling the textconv filter on
symlinks: symlinks are stored as blobs whose content is the target of
the link, and blame calls the textconv filter on a temporary file
filled-in with the content of this blob.

For example:

    $ git blame -C -C regular-file.pdf
    Error: May not be a PDF file (continuing anyway)
    Error: PDF file is damaged - attempting to reconstruct xref table...
    Error: Couldn't find trailer dictionary
    Error: Couldn't read xref table
    Warning: program returned non-zero exit code #1
    fatal: unable to read files to diff

That errors come from pdftotext run on symlink.pdf being extracted to
/tmp/ with one-line plain-text content pointing to link destination.

So several failures are demonstrated here:

  - git cat-file --textconv :symlink.bin    # also HEAD:symlink.bin
  - git blame --textconv symlink.bin
  - git blame -C -C --textconv regular-file # but also looks on symlink.bin

At present they all fail with something like.

    E: /tmp/j3ELEs_symlink.bin is not "binary" file

NOTE: git diff doesn't try to textconv the pathnames, it runs the
textual diff without textconv, which is the expected behavior.

Signed-off-by: Kirill Smelkov <kirr@landau.phys.spbu.ru>
Reviewed-by: Matthieu Moy <Matthieu.Moy@grenoble-inp.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoblame,cat-file: Prepare --textconv tests for correctly-failing conversion program
Kirill Smelkov [Wed, 29 Sep 2010 11:35:22 +0000 (15:35 +0400)]
blame,cat-file: Prepare --textconv tests for correctly-failing conversion program

The textconv filter is sometimes incorrectly ran on a temporary file
whose content is the target of a symbolic link, instead of actual file
content. Prepare to test this by marking the content of the file to
convert with "bin:", and let the helper die if "bin:" is not found in
the file content.

NOTE: I've changed $@ to $1 in helper becase textconv program "should
take a single argument" (see Documentation/gitattributes.txt), so
making this more explicit makes sense and also helps to avoid
problems with feeding arguments to echo.

Signed-off-by: Kirill Smelkov <kirr@landau.phys.spbu.ru>
Reviewed-by: Matthieu Moy <Matthieu.Moy@grenoble-inp.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agodiff: don't use pathname-based diff drivers for symlinks
Jeff King [Tue, 21 Sep 2010 21:01:24 +0000 (17:01 -0400)]
diff: don't use pathname-based diff drivers for symlinks

When we're diffing symlinks, we consider the contents to be
the pathname that the symlink points to. When a user sets up
a userdiff driver like "*.pdf diff=pdf", their "diff.pdf.*"
config generally tells us what to do with the content of
pdf files.

With the current code, we will actually process a symlink
like "link.pdf" using a configured pdf driver, meaning we
are using contents which consist of a pathname with
configuration that is expecting contents that consist of an
actual pdf file.

The most noticeable example of this would have been
textconv; however, it was already protected in its own
textconv-specific code path. We can still see the breakage
with something like "diff.*.binary", though. You could
also see it with diff.*.funcname, though it is a bit harder
to trigger accidentally there.

This patch adds a check for S_ISREG lower in the callstack
than the textconv-specific check, which should block use of
any userdiff config for non-regular files. We can drop the
check in the textconv code, which is now redundant.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot3101: modernise style
Junio C Hamano [Sat, 11 Sep 2010 17:53:29 +0000 (10:53 -0700)]
t3101: modernise style

Also add a few " &&" cascade that were missing.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agocompat/nedmalloc: don't force NDEBUG on the rest of git
René Scharfe [Sat, 11 Sep 2010 09:59:18 +0000 (11:59 +0200)]
compat/nedmalloc: don't force NDEBUG on the rest of git

Define the nedmalloc feature configuration macros for nedmalloc.o, only.
This keeps assert(3) working for the rest of the git source; it was
turned off for nedmalloc users before by defining NDEBUG globally.

Also remove -DUSE_NED_ALLOCATOR as this macro isn't used anywhere.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoconfig.txt: fix placement of diff.noprefix
Mark Lodato [Fri, 10 Sep 2010 01:39:37 +0000 (21:39 -0400)]
config.txt: fix placement of diff.noprefix

In git-config(1), diff.noprefix was placed in between
diff.mnemonicprefix and the list of mnemonic prefixes, which is
obviously incorrect and very confusing to readers.  Now, it is located
after the end of the explanation of mnemonicprefix, which makes much
more sense.

Signed-off-by: Mark Lodato <lodatom@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoxdiff-interface.c: always trim trailing space from xfuncname matches
Brandon Casey [Thu, 9 Sep 2010 19:02:46 +0000 (14:02 -0500)]
xdiff-interface.c: always trim trailing space from xfuncname matches

Generally, trailing space is removed from the string matched by the
xfuncname patterns.  The exception is when the matched string exceeds the
length of the fixed-size buffer that it will be copied in to.  But, a
string that exceeds the buffer can still contain trailing space in the
portion of the string that will be copied into the buffer.  So, simplify
this code slightly, and just perform the trailing space removal always.

Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agodiff.c: call regfree to free memory allocated by regcomp when necessary
Brandon Casey [Thu, 9 Sep 2010 19:02:45 +0000 (14:02 -0500)]
diff.c: call regfree to free memory allocated by regcomp when necessary

Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agorevert: Fix trivial comment style issue
Elijah Newren [Mon, 6 Sep 2010 21:53:24 +0000 (15:53 -0600)]
revert: Fix trivial comment style issue

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agocache_tree_free: Fix small memory leak
Elijah Newren [Mon, 6 Sep 2010 21:40:16 +0000 (15:40 -0600)]
cache_tree_free: Fix small memory leak

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'xx/trivial' into maint
Junio C Hamano [Mon, 6 Sep 2010 07:11:59 +0000 (00:11 -0700)]
Merge branch 'xx/trivial' into maint

* xx/trivial:
  tag.c: whitespace breakages fix
  Fix whitespace issue in object.c
  t5505: add missing &&

14 years agotag.c: whitespace breakages fix
Junio C Hamano [Mon, 6 Sep 2010 05:32:05 +0000 (22:32 -0700)]
tag.c: whitespace breakages fix

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoFix whitespace issue in object.c
Jared Hance [Sun, 5 Sep 2010 19:36:33 +0000 (15:36 -0400)]
Fix whitespace issue in object.c

Change some expanded tabs (spaces) to tabs in object.c.

Signed-off-by: Jared Hance <jaredhance@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot5505: add missing &&
Jens Lehmann [Sun, 5 Sep 2010 12:56:11 +0000 (14:56 +0200)]
t5505: add missing &&

Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoGit 1.7.2.3 v1.7.2.3
Junio C Hamano [Fri, 3 Sep 2010 16:38:53 +0000 (09:38 -0700)]
Git 1.7.2.3

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agobuiltin/merge_recursive.c: Add an usage string and make use of it.
Thiago Farina [Mon, 30 Aug 2010 03:30:22 +0000 (00:30 -0300)]
builtin/merge_recursive.c: Add an usage string and make use of it.

This improves the usage output by adding builtin_merge_recursive_usage string
that follows the same pattern used by the other builtin commands.

The previous output for git merger-recursive was:
usage: merge-recursive <base>... -- <head> <remote> ...

Now the output is:
usage: git merge-recursive <base>... -- <head> <remote> ...

Since cmd_merge_recursive is used to handle four different commands we need
the %s in the usage string, so the following example:

$ git merge-subtree -h

Will output:
usage: git merge-subtree <base>... -- <head> <remote> ...

Signed-off-by: Thiago Farina <tfransosi@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoPrepare for 1.7.2.3
Junio C Hamano [Wed, 1 Sep 2010 21:15:35 +0000 (14:15 -0700)]
Prepare for 1.7.2.3

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'gp/pack-refs-remove-empty-dirs' into maint
Junio C Hamano [Wed, 1 Sep 2010 21:05:05 +0000 (14:05 -0700)]
Merge branch 'gp/pack-refs-remove-empty-dirs' into maint

* gp/pack-refs-remove-empty-dirs:
  pack-refs: remove newly empty directories

14 years agoMerge branch 'sg/rerere-gc-old-still-used' into maint
Junio C Hamano [Wed, 1 Sep 2010 20:57:23 +0000 (13:57 -0700)]
Merge branch 'sg/rerere-gc-old-still-used' into maint

* sg/rerere-gc-old-still-used:
  rerere: fix overeager gc
  mingw_utime(): handle NULL times parameter

14 years agoMerge branch 'np/maint-huge-delta-generation' into maint
Junio C Hamano [Wed, 1 Sep 2010 20:56:10 +0000 (13:56 -0700)]
Merge branch 'np/maint-huge-delta-generation' into maint

* np/maint-huge-delta-generation:
  fix >4GiB source delta assertion failure

14 years agoMerge branch 'dj/fetch-tagopt' into maint
Junio C Hamano [Wed, 1 Sep 2010 20:52:10 +0000 (13:52 -0700)]
Merge branch 'dj/fetch-tagopt' into maint

* dj/fetch-tagopt:
  fetch: allow command line --tags to override config

14 years agoMerge branch 'da/fix-submodule-sync-superproject-config' into maint
Junio C Hamano [Wed, 1 Sep 2010 20:50:46 +0000 (13:50 -0700)]
Merge branch 'da/fix-submodule-sync-superproject-config' into maint

* da/fix-submodule-sync-superproject-config:
  submodule sync: Update "submodule.<name>.url"

14 years agoMerge branch 'en/rebase-against-rebase-fix' into maint
Junio C Hamano [Wed, 1 Sep 2010 20:43:55 +0000 (13:43 -0700)]
Merge branch 'en/rebase-against-rebase-fix' into maint

* en/rebase-against-rebase-fix:
  pull --rebase: Avoid spurious conflicts and reapplying unnecessary patches
  t5520-pull: Add testcases showing spurious conflicts from git pull --rebase

14 years agoMerge branch 'maint-1.7.1' into maint
Junio C Hamano [Mon, 30 Aug 2010 06:36:57 +0000 (23:36 -0700)]
Merge branch 'maint-1.7.1' into maint

* maint-1.7.1:
  t0003: add missing && at end of lines

14 years agot0003: add missing && at end of lines
Matthieu Moy [Sat, 28 Aug 2010 18:18:36 +0000 (20:18 +0200)]
t0003: add missing && at end of lines

Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agofor-each-ref: fix objectname:short bug
Jay Soffian [Thu, 26 Aug 2010 21:34:29 +0000 (17:34 -0400)]
for-each-ref: fix objectname:short bug

When objectname:short was introduced, it forgot to copy the result of
find_unique_abbrev. Because the result of find_unique_abbrev is a
pointer to static buffer, this resulted in the same value being
substituted in for each ref.

Signed-off-by: Jay Soffian <jaysoffian@gmail.com>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agotree-walk: Correct bitrotted comment about tree_entry()
Elijah Newren [Wed, 25 Aug 2010 02:53:11 +0000 (20:53 -0600)]
tree-walk: Correct bitrotted comment about tree_entry()

There was a code comment that referred to the "above two functions" but
over time the functions immediately preceding the comment have changed.
Just mention the relevant functions by name.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoFix 'git log' early pager startup error case
Linus Torvalds [Tue, 24 Aug 2010 17:33:59 +0000 (10:33 -0700)]
Fix 'git log' early pager startup error case

We start the pager too early for several git commands, which results in
the errors sometimes going to the pager rather than show up as errors.

This is often hidden by the fact that we pass in '-X' to less by default,
which causes 'less' to exit for small output, but if you do

  export LESS=-S

you can then clearly see the problem by doing

  git log --prretty

which shows the error message ("fatal: unrecognized argument: --prretty")
being sent to the pager.

This happens for pretty much all git commands that use USE_PAGER, and then
check arguments separately. But "git diff" does it too early too (even
though it does an explicit setup_pager() call)

This only fixes it for the trivial "git log" family case.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoparse-options: clarify PARSE_OPT_NOARG description
Jonathan Nieder [Sun, 22 Aug 2010 16:26:38 +0000 (21:56 +0530)]
parse-options: clarify PARSE_OPT_NOARG description

Here "takes no argument" means "does not take an argument".  The
latter phrasing might make it clearer that PARSE_OPT_NOARG does not
make an option with an argument that can optionally be left off.

Noticed-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot3302 (notes): Port to Solaris
Jonathan Nieder [Tue, 24 Aug 2010 07:14:53 +0000 (02:14 -0500)]
t3302 (notes): Port to Solaris

The time_notes script, which uses POSIX shell features, is
currently sometimes run with a non-POSIX /bin/sh.

Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoTypos in code comments, an error message, documentation
Ralf Wildenhues [Sun, 22 Aug 2010 11:12:12 +0000 (13:12 +0200)]
Typos in code comments, an error message, documentation

Signed-off-by: Ralf Wildenhues <Ralf.Wildenhues@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agofix >4GiB source delta assertion failure
Nicolas Pitre [Sat, 21 Aug 2010 05:00:13 +0000 (01:00 -0400)]
fix >4GiB source delta assertion failure

When people try insane things such as delta-compressing 4GiB files, they
get this assertion:

diff-delta.c:285: create_delta_index: Assertion `packed_entry - (struct index_entry *)mem == entries' failed.

This happens because:

1) the 'entries' variable is an unsigned int

2) it is assigned with entries = (bufsize - 1) / RABIN_WINDOW
   (that itself is not a problem unless bufsize > 4G * RABIN_WINDOW)

3) the buffer is indexed from top to bottom starting at
   "data = buffer + entries * RABIN_WINDOW" and the multiplication
   here does indeed overflows, making the resulting top of the buffer
   much lower than expected.

This makes the number of actually produced index entries smaller than
what was computed initially, hence the assertion.

Furthermore, the current delta encoding format cannot represent offsets
into a reference buffer with more than 32 bits anyway.  So let's just
limit the number of entries to what the delta format can encode.

Reported-by: Ilari Liusvaara <ilari.liusvaara@elisanet.fi>
Signed-off-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot7403: add missing &&'s
Jens Lehmann [Wed, 18 Aug 2010 21:20:33 +0000 (23:20 +0200)]
t7403: add missing &&'s

Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoTell ignore file about generate files in /gitweb/static
Mark Rada [Sat, 21 Aug 2010 16:18:37 +0000 (12:18 -0400)]
Tell ignore file about generate files in /gitweb/static

Signed-off-by: Mark Rada <marada@uwaterloo.ca>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'jc/maint-follow-rename-fix' into maint
Junio C Hamano [Fri, 20 Aug 2010 19:53:09 +0000 (12:53 -0700)]
Merge branch 'jc/maint-follow-rename-fix' into maint

* jc/maint-follow-rename-fix:
  log: test for regression introduced in v1.7.2-rc0~103^2~2
  diff --follow: do call diffcore_std() as necessary
  diff --follow: do not waste cycles while recursing

14 years agoMerge branch 'jn/maint-plug-leak' into maint
Junio C Hamano [Fri, 20 Aug 2010 19:53:09 +0000 (12:53 -0700)]
Merge branch 'jn/maint-plug-leak' into maint

* jn/maint-plug-leak:
  write-tree: Avoid leak when index refers to an invalid object
  read-tree: stop leaking tree objects
  core: Stop leaking ondisk_cache_entrys

14 years agoMerge branch 'jn/fix-abbrev' into maint
Junio C Hamano [Fri, 20 Aug 2010 19:53:09 +0000 (12:53 -0700)]
Merge branch 'jn/fix-abbrev' into maint

* jn/fix-abbrev:
  examples/commit: use --abbrev for commit summary
  checkout, commit: remove confusing assignments to rev.abbrev
  archive: abbreviate substituted commit ids again

14 years agoMerge branch 'vs/doc-spell' into maint
Junio C Hamano [Fri, 20 Aug 2010 19:53:08 +0000 (12:53 -0700)]
Merge branch 'vs/doc-spell' into maint

* vs/doc-spell:
  Documentation: spelling fixes

14 years agoMerge branch 'jn/rebase-rename-am' into maint
Junio C Hamano [Fri, 20 Aug 2010 19:53:08 +0000 (12:53 -0700)]
Merge branch 'jn/rebase-rename-am' into maint

* jn/rebase-rename-am:
  rebase: protect against diff.renames configuration
  t3400 (rebase): whitespace cleanup
  Teach "apply --index-info" to handle rename patches
  t4150 (am): futureproof against failing tests
  t4150 (am): style fix

14 years agoMerge branch 'jn/doc-pull' into maint
Junio C Hamano [Fri, 20 Aug 2010 19:53:08 +0000 (12:53 -0700)]
Merge branch 'jn/doc-pull' into maint

* jn/doc-pull:
  Documentation: flesh out “git pull” description

14 years agoMerge branch 'bc/use-more-hardlinks-in-install' into maint
Junio C Hamano [Fri, 20 Aug 2010 19:53:07 +0000 (12:53 -0700)]
Merge branch 'bc/use-more-hardlinks-in-install' into maint

* bc/use-more-hardlinks-in-install:
  Makefile: make hard/symbolic links for non-builtins too
  Makefile: link builtins residing in bin directory to main git binary too

14 years agoMerge branch 'tr/rfc-reset-doc' into maint
Junio C Hamano [Fri, 20 Aug 2010 19:53:07 +0000 (12:53 -0700)]
Merge branch 'tr/rfc-reset-doc' into maint

* tr/rfc-reset-doc:
  Documentation/reset: move "undo permanently" example behind "make topic"
  Documentation/reset: reorder examples to match description
  Documentation/reset: promote 'examples' one section up
  Documentation/reset: separate options by mode
  Documentation/git-reset: reorder modes for soft-mixed-hard progression

14 years agoxmalloc: include size in the failure message
Matthieu Moy [Fri, 20 Aug 2010 15:09:51 +0000 (17:09 +0200)]
xmalloc: include size in the failure message

Out-of-memory errors can either be actual lack of memory, or bugs (like
code trying to call xmalloc(-1) by mistake). A little more information
may help tracking bugs reported by users.

Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoGit 1.7.2.2 v1.7.2.2
Junio C Hamano [Thu, 19 Aug 2010 23:03:18 +0000 (16:03 -0700)]
Git 1.7.2.2

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'tr/xsize-bits' into maint
Junio C Hamano [Thu, 19 Aug 2010 22:49:54 +0000 (15:49 -0700)]
Merge branch 'tr/xsize-bits' into maint

* tr/xsize-bits:
  xsize_t: check whether we lose bits

14 years agoMerge branch 'jc/sha1-name-find-fix' into maint
Junio C Hamano [Thu, 19 Aug 2010 22:49:24 +0000 (15:49 -0700)]
Merge branch 'jc/sha1-name-find-fix' into maint

* jc/sha1-name-find-fix:
  sha1_name.c: fix parsing of ":/token" syntax

Conflicts:
sha1_name.c

14 years agoMerge branch 'sp/fix-smart-http-deadlock-on-error' into maint
Junio C Hamano [Wed, 18 Aug 2010 23:30:11 +0000 (16:30 -0700)]
Merge branch 'sp/fix-smart-http-deadlock-on-error' into maint

* sp/fix-smart-http-deadlock-on-error:
  smart-http: Don't deadlock on server failure

14 years agosubmodule sync: Update "submodule.<name>.url"
David Aguilar [Wed, 18 Aug 2010 15:58:33 +0000 (08:58 -0700)]
submodule sync: Update "submodule.<name>.url"

When "git submodule sync" synchronizes the repository URLs
it only updates submodules' .git/config.  However, the old
URLs still exist in the super-project's .git/config.

Update the super-project's configuration so that commands
such as "git submodule update" use the URLs from .gitmodules.

Signed-off-by: David Aguilar <davvid@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agolog: test for regression introduced in v1.7.2-rc0~103^2~2
Ævar Arnfjörð Bjarmason [Sun, 15 Aug 2010 10:16:25 +0000 (10:16 +0000)]
log: test for regression introduced in v1.7.2-rc0~103^2~2

Add a regression test for the git log -M --follow $diff_option bug
introduced in v1.7.2-rc0~103^2~2, $diff_option being diff related
options like -p, --stat, --name-only etc.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agofetch: allow command line --tags to override config
Daniel Johnson [Wed, 11 Aug 2010 22:57:20 +0000 (18:57 -0400)]
fetch: allow command line --tags to override config

Originally, if remote.<name>.tagopt was set, the --tags and option would
have no effect when given to git fetch. So if
tagopt="--no-tags"

git fetch --tags

would not actually fetch tags.

This patch changes this behavior to only follow what is written in the
config if there is no option passed by the command line.

Signed-off-by: Daniel Johnson <ComputerDruid@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agodiff --follow: do call diffcore_std() as necessary
Junio C Hamano [Fri, 13 Aug 2010 19:17:45 +0000 (12:17 -0700)]
diff --follow: do call diffcore_std() as necessary

Usually, diff frontends populate the output queue with filepairs without
any rename information and call diffcore_std() to sort the renames out.
When --follow is in effect, however, diff-tree family of frontend has a
hack that looks like this:

    diff-tree frontend
    -> diff_tree_sha1()
       . populate diff_queued_diff
       . if --follow is in effect and there is only one change that
         creates the target path, then
       -> try_to_follow_renames()
  -> diff_tree_sha1() with no pathspec but with -C
  -> diffcore_std() to find renames
  . if rename is found, tweak diff_queued_diff and put a
    single filepair that records the found rename there
    -> diffcore_std()
       . tweak elements on diff_queued_diff by
       - rename detection
       - path ordering
       - pickaxe filtering

We need to skip parts of the second call to diffcore_std() that is related
to rename detection, and do so only when try_to_follow_renames() did find
a rename.  Earlier 1da6175 (Make diffcore_std only can run once before a
diff_flush, 2010-05-06) tried to deal with this issue incorrectly; it
unconditionally disabled any second call to diffcore_std().

This hopefully fixes the breakage.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agodiff --follow: do not waste cycles while recursing
Junio C Hamano [Fri, 13 Aug 2010 17:36:01 +0000 (10:36 -0700)]
diff --follow: do not waste cycles while recursing

The "--follow" logic is called from diff_tree_sha1() function, but the
input trees to diff_tree_sha1() are not necessarily the top-level trees
(compare_tree_entry() calls it while it recursively descends into
subtrees).  When a newly created path lives in somewhere deep in the
source hierarchy, e.g. "platform/", but the rename source is in a totally
different place in the destination hierarchy, e.g. "lang-api/src/com/...",
running "try_to_find_renames()" while base is set to "platform/" is a
wasted call.

We only need to run the rename following at the very top level.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agopull --rebase: Avoid spurious conflicts and reapplying unnecessary patches
Elijah Newren [Fri, 13 Aug 2010 01:50:50 +0000 (19:50 -0600)]
pull --rebase: Avoid spurious conflicts and reapplying unnecessary patches

Prior to c85c792 (pull --rebase: be cleverer with rebased upstream
branches, 2008-01-26), pull --rebase would run

  git rebase $merge_head

which resulted in a call to

  git format-patch ... --ignore-if-in-upstream $merge_head..$cur_branch

This resulted in patches from $merge_head..$cur_branch being applied, as
long as they did not already exist in $cur_branch..$merge_head.

Unfortunately, when upstream is rebased, $merge_head..$cur_branch also
refers to "old" commits that have already been rebased upstream, meaning
that many patches that were already fixed upstream would be reapplied.
This could result in many spurious conflicts, as well as reintroduce
patches that were intentionally dropped upstream.

So the algorithm was changed in c85c792 (pull --rebase: be cleverer with
rebased upstream branches, 2008-01-26) and d44e712 (pull: support rebased
upstream + fetch + pull --rebase, 2009-07-19).  Defining $old_remote_ref to
be the most recent entry in the reflog for @{upstream} that is an ancestor
of $cur_branch, pull --rebase was changed to run

  git rebase --onto $merge_head $old_remote_ref

which results in a call to

  git format-patch ... --ignore-if-in-upstream $old_remote_ref..$cur_branch

The whole point of this change was to reduce the number of commits being
reapplied, by avoiding commits that upstream already has or had.

In the rebased upstream case, this change achieved that purpose.  It is
worth noting, though, that since $old_remote_ref is always an ancestor of
$cur_branch (by its definition), format-patch will not know what upstream
is and thus will not be able to determine if any patches are already
upstream; they will all be reapplied.

In the non-rebased upstream case, this new form is usually the same as the
original code but in some cases $old_remote_ref can be an ancestor of

   $(git merge-base $merge_head $cur_branch)

meaning that instead of avoiding reapplying commits that upstream already
has, it actually includes more such commits.  Combined with the fact that
format-patch can no longer detect commits that are already upstream (since
it is no longer told what upstream is), results in lots of confusion for
users (e.g. "git is giving me lots of conflicts in stuff I didn't even
change since my last push.")

Cases where additional commits could be reapplied include forking from a
commit other than the tracking branch, or amending/rebasing after pushing.
Cases where the inability to detect upstreamed commits cause problems
include independent discovery of a fix and having your patches get
upstreamed by some alternative route (e.g. pulling your changes to a third
machine, pushing from there, and then going back to your original machine
and trying to pull --rebase).

Fix the non-rebased upstream case by ignoring $old_remote_ref whenever it
is contained in $(git merge-base $merge_head $cur_branch).  This should
have no affect on the rebased upstream case.

Acked-by: Santi Béjar <santi@agolina.net>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot5520-pull: Add testcases showing spurious conflicts from git pull --rebase
Elijah Newren [Fri, 13 Aug 2010 01:50:49 +0000 (19:50 -0600)]
t5520-pull: Add testcases showing spurious conflicts from git pull --rebase

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agopush: mention "git pull" in error message for non-fast forwards
Matthieu Moy [Mon, 2 Aug 2010 13:42:44 +0000 (15:42 +0200)]
push: mention "git pull" in error message for non-fast forwards

The message remains fuzzy to include "git pull", "git pull --rebase" and
others, but directs the user to the simplest solution in the vast
majority of cases.

Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoStandardize do { ... } while (0) style
Jonathan Nieder [Thu, 12 Aug 2010 22:11:15 +0000 (17:11 -0500)]
Standardize do { ... } while (0) style

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot/t7003: replace \t with literal tab in sed expression
Brandon Casey [Thu, 12 Aug 2010 20:55:54 +0000 (15:55 -0500)]
t/t7003: replace \t with literal tab in sed expression

The sed utilities on IRIX and Solaris do not interpret the sequence '\t'
to mean a tab character;  they read a literal character 't'.  So, use a
literal tab instead.

Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoindex-pack: Don't follow replace refs.
Nelson Elhage [Thu, 12 Aug 2010 14:18:12 +0000 (10:18 -0400)]
index-pack: Don't follow replace refs.

Without this, attempting to index a pack containing objects that have been
replaced results in a fatal error that looks like:

fatal: SHA1 COLLISION FOUND WITH <replaced-object> !

Signed-off-by: Nelson Elhage <nelhage@ksplice.com>
Acked-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agopost-receive-email: remove spurious commas in email subject
Matthieu Moy [Wed, 11 Aug 2010 08:14:13 +0000 (10:14 +0200)]
post-receive-email: remove spurious commas in email subject

The previous form produced subjects like

  [SCM] project.git branch, foo, updated. ...

The new one will produce the lighter

  [SCM] project.git branch foo updated. ...

Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agofast-import: export correctly marks larger than 2^20-1
Raja R Harinath [Tue, 13 Jul 2010 11:51:48 +0000 (17:21 +0530)]
fast-import: export correctly marks larger than 2^20-1

dump_marks_helper() has a bug when dumping marks larger than 2^20-1,
i.e., when the sparse array has more than two levels.  The bug was
that the 'base' counter was being shifted by 20 bits at level 3, and
then again by 10 bits at level 2, rather than a total shift of 20 bits
in this argument to the recursive call:

  (base + k) << m->shift

There are two ways to fix this correctly, the elegant:

  (base + k) << 10

and the one I chose due to edit distance:

  base + (k << m->shift)

Signed-off-by: Raja R Harinath <harinath@hurrynot.org>
Acked-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot/lib-git-svn.sh: use $PERL_PATH for perl, not perl from $PATH
Ævar Arnfjörð Bjarmason [Tue, 10 Aug 2010 12:14:25 +0000 (12:14 +0000)]
t/lib-git-svn.sh: use $PERL_PATH for perl, not perl from $PATH

Change the git-svn tests to use $PERL_PATH, not the "perl" in $PATH.

Using perl in $PATH was added by Sam Vilain in v1.6.6-rc0~95^2~3,
Philippe Bruhat introduced $PERL_PATH to the test suite in
v1.6.6-rc0~9^2, but the lib-git-svn.sh tests weren't updated to use
the new convention.

This resulted in the git-svn tests always being skipped on my
system. My /usr/bin/perl has access to SVN::Core and SVN::Repos, but
the perl in my $PATH does not.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agowrite-tree: Avoid leak when index refers to an invalid object
Jonathan Nieder [Tue, 10 Aug 2010 03:32:11 +0000 (22:32 -0500)]
write-tree: Avoid leak when index refers to an invalid object

Noticed by valgrind during test t0000.35 “writing this tree without
--missing-ok”.

Even in the cherry-pick foo..bar code path, such an error is the
end of the line.  But maybe some day an interactive porcelain will
want to link to libgit, making this matter.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoread-tree: stop leaking tree objects
Jonathan Nieder [Tue, 10 Aug 2010 03:33:44 +0000 (22:33 -0500)]
read-tree: stop leaking tree objects

The underlying problem is that the fill_tree_descriptor()
API is easy to misuse, and this patch does not fix that.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agocore: Stop leaking ondisk_cache_entrys
Jonathan Nieder [Tue, 10 Aug 2010 03:28:07 +0000 (22:28 -0500)]
core: Stop leaking ondisk_cache_entrys

Noticed with valgrind.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agodiff: strip extra "/" when stripping prefix
Jakub Narebski [Mon, 9 Aug 2010 14:50:53 +0000 (10:50 -0400)]
diff: strip extra "/" when stripping prefix

There are two ways a user might want to use "diff --relative":

  1. For a file in a directory, like "subdir/file", the user
     can use "--relative=subdir/" to strip the directory.

  2. To strip part of a filename, like "foo-10", they can
     use "--relative=foo-".

We currently handle both of those situations. However, if the user passes
"--relative=subdir" (without the trailing slash), we produce inconsistent
results. For the unified diff format, we collapse the double-slash of
"a//file" correctly into "a/file". But for other formats (raw, stat,
name-status), we end up with "/file".

We can do what the user means here and strip the extra "/" (and only a
slash).  We are not hurting any existing users of (2) above with this
behavior change because the existing output for this case was nonsensical.

Patch by Jakub, tests and commit message by Jeff King.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agogitweb: clarify search results page when no matching commit found
Jonathan Nieder [Sat, 7 Aug 2010 21:56:47 +0000 (16:56 -0500)]
gitweb: clarify search results page when no matching commit found

When searching commits for a string that never occurs, the results
page looks something like this:

projects / foo.git / search                                 \o/
summary | ... | tree          [commit] search: [ kfjdkas ] [ ]re
first ⋅ prev ⋅ next

Merge branch 'maint'

Foo: a demonstration project

Without a list of hits to compare it to, the header describing the
commit named by the hash parameter (usually HEAD) may itself look
like a hit.  Add some text (“No match.”) to replace the empty
list of hits and avoid this confusion.

While at it, remove some nearby dead code, left behind from a
simplification a few years ago (v1.5.4-rc0~276^2~4, 2007-11-01).

Noticed-by: Erick Mattos <erick.mattos@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Acked-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoDocumentation: add a FILES section for show-ref
Jonathan Nieder [Mon, 9 Aug 2010 01:32:43 +0000 (20:32 -0500)]
Documentation: add a FILES section for show-ref

A peek at where the refs are kept might help understanding, even if,
as the DESCRIPTION section suggests, direct access is not part of the
public API.

Balance that out with a pointer to update-ref.

Suggested-by: Geoff Russell <geoffrey.russell@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMakefile: add missing dependency on http.h
Jonathan Nieder [Sun, 8 Aug 2010 21:48:59 +0000 (16:48 -0500)]
Makefile: add missing dependency on http.h

v1.7.1-rc0~65^2~2 (http: init and cleanup separately from
http-walker, 2010-03-02) introduced a direct dependency from
http-fetch on the HTTP request library.  Declare it.

Detected with "make CHECK_HEADER_DEPENDENCIES=1".

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Acked-by: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMakefile: add missing dependencies on url.h
Jonathan Nieder [Sun, 8 Aug 2010 21:25:11 +0000 (16:25 -0500)]
Makefile: add missing dependencies on url.h

v1.7.2-rc0~56^2 and its parent (decode file:// and ssh://
URLs, 2010-05-23) introduced a new url library.  Update the
Makefile with the relevant dependencies.

Detected with "make CHECK_HEADER_DEPENDENCIES=1".

Cc: Jeff King <peff@peff.net>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoDocumentation/git-log: Clarify --full-diff
Michael J Gruber [Sun, 8 Aug 2010 15:31:34 +0000 (17:31 +0200)]
Documentation/git-log: Clarify --full-diff

The current description gives the impression that "--full-diff" affects
"log -p" only.

Make it clearer that it affects all diff-based output types.

Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agogit-rebase: fix typo when parsing --force-rebase
Willy Tarreau [Sun, 8 Aug 2010 05:13:32 +0000 (07:13 +0200)]
git-rebase: fix typo when parsing --force-rebase

Due to two missing hyphens, The "force" keyword on the command line
would be taken as an alias for the --force-rebase option.

Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoimap-send: Fix sprintf usage
Ævar Arnfjörð Bjarmason [Sat, 7 Aug 2010 23:09:45 +0000 (18:09 -0500)]
imap-send: Fix sprintf usage

When composing a command for the imap server, imap-send uses a single
nfsnprintf() invocation for brevity instead of dealing separately with
the case when there is a message to be sent and the case when there
isn’t.  The unused argument in the second case, while valid, is
confusing for static analyzers and human readers.

v1.6.4-rc0~117 (imap-send: add support for IPv6, 2009-05-25)
mistakenly used %hu as the format for an int “port”, by analogy with
existing usage for the unsigned short “addr.sin_port”.  Use %d
instead.

Noticed with clang.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoprune: allow --dry-run for -n and --verbose for -v
René Scharfe [Fri, 6 Aug 2010 20:28:05 +0000 (22:28 +0200)]
prune: allow --dry-run for -n and --verbose for -v

For consistency with other git commands, let git prune accept the long
options --dry-run and --verbose for the respective short ones -n and -v.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agonotes: allow --dry-run for -n and --verbose for -v
René Scharfe [Fri, 6 Aug 2010 20:28:09 +0000 (22:28 +0200)]
notes: allow --dry-run for -n and --verbose for -v

For consistency with other git commands, let the prune subcommand of
git notes accept the long options --dry-run and --verbose for the
respective short ones -n and -v.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoDocument -B<n>[/<m>], -M<n> and -C<n> variants of -B, -M and -C
Matthieu Moy [Thu, 5 Aug 2010 16:14:25 +0000 (18:14 +0200)]
Document -B<n>[/<m>], -M<n> and -C<n> variants of -B, -M and -C

These options take an optional argument, but this optional argument was
not documented.

Original patch by Matthieu Moy, but documentation for -B mostly copied
from the explanations of Junio C Hamano.

While we're there, fix a typo in a comment in diffcore.h.

Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoDocumentation: cite git-am from git-apply
Brad King [Tue, 3 Aug 2010 18:41:50 +0000 (14:41 -0400)]
Documentation: cite git-am from git-apply

Users reading git-apply documentation may also be interested in git-am,
especially after receiving an email created with git-format-patch.  The
documentation for git-am already references git-apply.  Add the reverse.

Signed-off-by: Brad King <brad.king@kitware.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot7003: fix subdirectory-filter test
Thomas Rast [Thu, 29 Jul 2010 15:10:22 +0000 (17:10 +0200)]
t7003: fix subdirectory-filter test

The test would not fail if the filtering failed to do anything, since
in

  test -z "$(git diff HEAD directorymoved:newsubdir)"'

'directorymoved:newsubdir' is not valid, so git-diff fails without
printing anything on stdout.  But then the exit status of git-diff is
lost, whereas test -z "" succeeds.

Use 'git diff --exit-code' instead, which does the right thing and has
the added bonus of showing the differences if there are any.

Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agosmart-http: Don't deadlock on server failure
Shawn O. Pearce [Fri, 6 Aug 2010 21:19:44 +0000 (14:19 -0700)]
smart-http: Don't deadlock on server failure

If the remote HTTP server fails (e.g. returns 404 or 500) when we
posted the RPC to it, we won't have sent anything to the background
Git process that is supposed to handle the stream.  Because we
didn't send anything, its waiting for input from remote-curl, and
remote-curl cannot read its response payload because doing so would
lead to a deadlock.

Send the background task EOF on its input before we try to read
its response back, that way it will break out of its read loop
and terminate.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoAllow "check-ref-format --branch" from subdirectory
Jonathan Nieder [Fri, 6 Aug 2010 03:39:22 +0000 (22:39 -0500)]
Allow "check-ref-format --branch" from subdirectory

check-ref-format --branch requires access to the repository
to resolve refs like @{-1}.

Noticed by Nguyễn Thái Ngọc Duy.

Cc: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agocheck-ref-format: handle subcommands in separate functions
Jonathan Nieder [Fri, 6 Aug 2010 03:36:07 +0000 (22:36 -0500)]
check-ref-format: handle subcommands in separate functions

The code for each subcommand should be easier to read and manipulate
this way.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agopretty-options.txt: match --format's documentation with implementation.
Matthieu Moy [Thu, 5 Aug 2010 13:08:05 +0000 (15:08 +0200)]
pretty-options.txt: match --format's documentation with implementation.

Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'sv/maint-diff-q-clear-fix' into maint
Junio C Hamano [Tue, 3 Aug 2010 22:17:34 +0000 (15:17 -0700)]
Merge branch 'sv/maint-diff-q-clear-fix' into maint

* sv/maint-diff-q-clear-fix:
  Fix DIFF_QUEUE_CLEAR refactoring

14 years agoMerge branch 'rr/svn-fe' into maint
Junio C Hamano [Tue, 3 Aug 2010 22:17:11 +0000 (15:17 -0700)]
Merge branch 'rr/svn-fe' into maint

* rr/svn-fe:
  contrib/svn-fe: Add the svn-fe target to .gitignore
  contrib/svn-fe: Fix IncludePath

14 years agoMerge branch 'pt/git-gui' into maint
Junio C Hamano [Tue, 3 Aug 2010 22:16:38 +0000 (15:16 -0700)]
Merge branch 'pt/git-gui' into maint

* pt/git-gui:
  git-gui: fix size and position of window panes on startup
  git-gui: mc cannot be used before msgcat has been loaded
  git-gui: use textconv filter for diff and blame
  git-gui: Avoid using the <<Copy>> binding as a menu accelerator on win32
  git-gui: fix shortcut creation on cygwin
  git-gui: fix PATH environment for mingw development environment
  git-gui: fix usage of _gitworktree when creating shortcut for windows
  git-gui: fix "Explore Working Copy" for Windows again
  git-gui: fix usage of themed widgets variable
  git-gui: Handle failure of core.worktree to identify the working directory.
  git-gui: check whether systems nice command works or disable it

14 years agocontrib/svn-fe: Add the svn-fe target to .gitignore
Ramkumar Ramachandra [Thu, 29 Jul 2010 05:36:20 +0000 (11:06 +0530)]
contrib/svn-fe: Add the svn-fe target to .gitignore

Add the svn-fe target to .gitignroe.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agocontrib/svn-fe: Fix IncludePath
Ramkumar Ramachandra [Thu, 29 Jul 2010 05:36:19 +0000 (11:06 +0530)]
contrib/svn-fe: Fix IncludePath

Include the path "../../vcs-svn" while compiling it in the Makefile
and change svn-fe.c to include svndump.h.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoDocumentation: flesh out “git pull” description
Jonathan Nieder [Mon, 2 Aug 2010 21:39:30 +0000 (16:39 -0500)]
Documentation: flesh out “git pull” description

The current description in the pull man page does not say much more
than that “git pull” is fetch + merge.  Though that is all a person
needs to know in the end, it would be useful to summarize a bit about
what those commands do for new readers.

Most of this description is taken from the “git merge” docs.

Now that we explain how to back out of a failed merge (reset --merge),
we can tone down the warning against that a bit.

Except, as Thomas noticed, there’s a risk with that because people
might read this version of the manpage online and then conclude that
it is safe to try a merge with uncommitted changes, only to find that
their “git reset” doesn't support --merge yet.  Or worse, verify that
their git-reset has --merge by a quick test (1b5b465 is in 1.6.2) but
then find that it does not help with backing out of a merge (e11d7b5
is only in 1.7.0!).  So keep the warning.

With clarifications from Ævar, Thomas, and Junio.

Noticed-by: Geoff Russell <geoffrey.russell@gmail.com>
Cc: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Cc: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agosha1_name.c: fix parsing of ":/token" syntax
Junio C Hamano [Mon, 2 Aug 2010 21:37:06 +0000 (14:37 -0700)]
sha1_name.c: fix parsing of ":/token" syntax

The parser tried to clean up the object flags it used while finding
commits with matching string, but was not doing a very good job at it.

This caused "checkout -b new ':/token'", which internally tries to parse
':/token' twice as an object name, to fail when the commit in question
was reachable from only one ref.

The mask bits given to pop_most_recent_commit(&list, MASK) means "I have
already been on the list to be processed, so please do not place me again
even if I am found to be a parent of some other commit on the list."  So
mark them when we add them to the list at the beginning.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'ab/tap' into maint
Junio C Hamano [Mon, 2 Aug 2010 18:53:58 +0000 (11:53 -0700)]
Merge branch 'ab/tap' into maint

* ab/tap:
  test-lib: Remove 3 year old no-op --no-python option
  test-lib: Ignore --quiet under a TAP harness

14 years agotest-lib: Remove 3 year old no-op --no-python option
Ævar Arnfjörð Bjarmason [Sat, 31 Jul 2010 16:49:53 +0000 (16:49 +0000)]
test-lib: Remove 3 year old no-op --no-python option

The --no-python option was added to test-lib.sh by Johannes Schindelin
in early 2006 in abb7c7b3. It was later turned into a no-op by Junio C
Hamano in 7cdbff14 the same year.

Over three years is long enough before removing this old wart which
was retained for backwards compatibility. Our tests have been using
NO_PYTHON and "test_have_prereq PYTHON" for a long time now.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agotest-lib: Ignore --quiet under a TAP harness
Ævar Arnfjörð Bjarmason [Sat, 31 Jul 2010 16:40:05 +0000 (16:40 +0000)]
test-lib: Ignore --quiet under a TAP harness

Running the tests with --quiet under a TAP harness will always fail,
since a TAP harness always needs actual test output to go along with
the plan that's being emitted.

Change the test-lib.sh to ignore the --quiet option under
HARNESS_ACTIVE to work around this. Then users that have --quiet in
their GIT_TEST_OPTS can run tests under prove(1) without everything
breaking.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge git://repo.or.cz/git-gui into pt/git-gui
Junio C Hamano [Mon, 2 Aug 2010 16:28:30 +0000 (09:28 -0700)]
Merge git://repo.or.cz/git-gui into pt/git-gui

* git://repo.or.cz/git-gui:
  git-gui: fix size and position of window panes on startup
  git-gui: mc cannot be used before msgcat has been loaded
  git-gui: use textconv filter for diff and blame
  git-gui: Avoid using the <<Copy>> binding as a menu accelerator on win32
  git-gui: fix shortcut creation on cygwin
  git-gui: fix PATH environment for mingw development environment
  git-gui: fix usage of _gitworktree when creating shortcut for windows
  git-gui: fix "Explore Working Copy" for Windows again
  git-gui: fix usage of themed widgets variable
  git-gui: Handle failure of core.worktree to identify the working directory.
  git-gui: check whether systems nice command works or disable it

14 years agoDocumentation/rev-parse: quoting is required with --parseopt
Thomas Rast [Fri, 30 Jul 2010 15:01:50 +0000 (17:01 +0200)]
Documentation/rev-parse: quoting is required with --parseopt

When calling rev-parse --parseopt, as in the (now fixed) documented
example

  eval "$(echo "$OPTS_SPEC" | git rev-parse --parseopt -- "$@" || echo exit $?)"

the outermost quoting is required, as otherwise all runs of arbitrary
whitespace inside the resulting 'set -- ...' call would be collapsed
into a single space.

This was exposed as a result of our new use of cat <<\EOF since
47e9cd2 (parseopt: wrap rev-parse --parseopt usage for eval
consumption, 2010-06-12), but has always been a problem when handling
arguments containing e.g. newlines.

Point this out in the documentation, and in particular correct the
example that did not have the quotes.

Noticed-by: Joshua Jensen <jjensen@workspacewhiz.com>
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoDocumentation: reporting bugs
Junio C Hamano [Mon, 2 Aug 2010 16:07:39 +0000 (09:07 -0700)]
Documentation: reporting bugs

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoFix DIFF_QUEUE_CLEAR refactoring
Junio C Hamano [Mon, 2 Aug 2010 15:29:42 +0000 (08:29 -0700)]
Fix DIFF_QUEUE_CLEAR refactoring

It introduced a macro to reduce repeated assignments to three fields,
but an unrelated and incorrect change snuck in by mistake, which broke
commands like "git diff-files -p --submodule".

Noticed by Sven Verdoolaege.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agogit-gui: fix size and position of window panes on startup
Pat Thoyts [Mon, 2 Aug 2010 12:42:45 +0000 (13:42 +0100)]
git-gui: fix size and position of window panes on startup

The themed panedwindow needs to have the sash position set after the
widget has been mapped therefore apply this setting in the Map event
binding. To avoid visible redraws as the application is constructed
the main window should be withdrawn until all the widgets have been added

Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
14 years agogit-gui: mc cannot be used before msgcat has been loaded
Pat Thoyts [Mon, 2 Aug 2010 11:13:05 +0000 (12:13 +0100)]
git-gui: mc cannot be used before msgcat has been loaded

If someone attempts to use an older version that Tk 8.4 the error was
masked by the lack of a mc command.

Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
14 years agogit-gui: use textconv filter for diff and blame
Clément Poulain [Fri, 30 Jul 2010 08:11:02 +0000 (09:11 +0100)]
git-gui: use textconv filter for diff and blame

Create a checkbox "Use Textconv For Diffs and Blame" in git-gui options.
If checked and if the driver for the concerned file exists, git-gui calls diff
and blame with --textconv option

Signed-off-by: Clément Poulain <clement.poulain@ensimag.imag.fr>
Signed-off-by: Diane Gasselin <diane.gasselin@ensimag.imag.fr>
Signed-off-by: Axel Bonnet <axel.bonnet@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
14 years agogit-gui: Avoid using the <<Copy>> binding as a menu accelerator on win32
Pat Thoyts [Mon, 8 Mar 2010 12:43:27 +0000 (12:43 +0000)]
git-gui: Avoid using the <<Copy>> binding as a menu accelerator on win32

On Windows the Control-C binding is used to copy and is mapped to the Tk
virtual event <<Copy>>. In the initial git-gui dialog this is also bound
as an accelerator for the Clone menu item. The effect is that both bindings
run, copying the text but resetting the clone page or switching to the clone
page when the user tries to copy text from one of the entry fields.
This patch avoids this by using Control-L instead for Windows only.

Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
14 years agogit-gui: fix shortcut creation on cygwin
Heiko Voigt [Sat, 27 Feb 2010 20:48:48 +0000 (21:48 +0100)]
git-gui: fix shortcut creation on cygwin

When the user tried to create a desktop icon with git gui on cygwin
wscript was complaining about an unknown option and displaying the
non-native path as such.

Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>