git.git
18 years agoupload-pack: stop the other side when they have more roots than we do.
Junio C Hamano [Thu, 6 Jul 2006 04:28:20 +0000 (21:28 -0700)]
upload-pack: stop the other side when they have more roots than we do.

When the downloader has more roots than we do, we will see many
"have" that leads to the root we do not have and let the other
side walk all the way to that root.  Tell them to stop sending the
branch'es ancestors by sending a fake "ACK" when we already have
common ancestor for the wanted refs.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoupload-pack: minor clean-up in multi-ack logic
Junio C Hamano [Thu, 6 Jul 2006 01:12:12 +0000 (18:12 -0700)]
upload-pack: minor clean-up in multi-ack logic

No changes to what it does, but separating the codepath clearly
with if ... else if ... chain makes it easier to follow.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'jc/upload-pack'
Junio C Hamano [Sun, 13 Aug 2006 05:16:51 +0000 (22:16 -0700)]
Merge branch 'jc/upload-pack'

18 years agoMerge branch 'jc/racy-delay'
Junio C Hamano [Sun, 13 Aug 2006 02:41:44 +0000 (19:41 -0700)]
Merge branch 'jc/racy-delay'

18 years agoMerge branch 'th/diff-extra'
Junio C Hamano [Sun, 13 Aug 2006 02:34:41 +0000 (19:34 -0700)]
Merge branch 'th/diff-extra'

18 years agoMerge branch 'jc/pack-objects'
Junio C Hamano [Sun, 13 Aug 2006 02:33:16 +0000 (19:33 -0700)]
Merge branch 'jc/pack-objects'

18 years agoMerge branch 'js/read-tree'
Junio C Hamano [Sun, 13 Aug 2006 02:29:11 +0000 (19:29 -0700)]
Merge branch 'js/read-tree'

18 years agoMerge branch 'js/http-mb'
Junio C Hamano [Sun, 13 Aug 2006 02:24:51 +0000 (19:24 -0700)]
Merge branch 'js/http-mb'

18 years agoMerge branch 'js/color-diff'
Junio C Hamano [Sun, 13 Aug 2006 02:24:47 +0000 (19:24 -0700)]
Merge branch 'js/color-diff'

18 years agoMerge branch 'jn/web'
Junio C Hamano [Sun, 13 Aug 2006 02:24:15 +0000 (19:24 -0700)]
Merge branch 'jn/web'

18 years agoMerge branch 'lt/web'
Junio C Hamano [Sun, 13 Aug 2006 02:24:09 +0000 (19:24 -0700)]
Merge branch 'lt/web'

18 years agoMerge branch 'jn/conf'
Junio C Hamano [Sun, 13 Aug 2006 02:23:09 +0000 (19:23 -0700)]
Merge branch 'jn/conf'

18 years agoMerge branch 'jc/grep'
Junio C Hamano [Sun, 13 Aug 2006 02:16:33 +0000 (19:16 -0700)]
Merge branch 'jc/grep'

18 years agoMerge branch 'mk/rename'
Junio C Hamano [Sun, 13 Aug 2006 02:13:31 +0000 (19:13 -0700)]
Merge branch 'mk/rename'

18 years agoMerge branch 'ml/pager'
Junio C Hamano [Sun, 13 Aug 2006 02:13:25 +0000 (19:13 -0700)]
Merge branch 'ml/pager'

18 years agoGIT 1.4.2 v1.4.2
Junio C Hamano [Sun, 13 Aug 2006 01:32:31 +0000 (18:32 -0700)]
GIT 1.4.2

18 years agoMerge git://git.kernel.org/pub/scm/gitk/gitk
Junio C Hamano [Sun, 13 Aug 2006 01:32:17 +0000 (18:32 -0700)]
Merge git://git.kernel.org/pub/scm/gitk/gitk

* git://git.kernel.org/pub/scm/gitk/gitk:
  gitk: Show the currently checked-out head in bold font
  gitk: Allow the user to set some colors

18 years agot/t4013: fix futzing with the version string.
Junio C Hamano [Sun, 13 Aug 2006 01:04:07 +0000 (18:04 -0700)]
t/t4013: fix futzing with the version string.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoBetter error message when we are unable to lock the index file
Junio C Hamano [Sat, 12 Aug 2006 08:03:47 +0000 (01:03 -0700)]
Better error message when we are unable to lock the index file

Most of the callers except the one in refs.c use the function to
update the index file.  Among the index writers, everybody
except write-tree dies if they cannot open it for writing.

This gives the function an extra argument, to tell it to die
when it cannot create a new file as the lockfile.

The only caller that does not have to die is write-tree, because
updating the index for the cache-tree part is optional and not
being able to do so does not affect the correctness.  I think we
do not have to be so careful and make the failure into die() the
same way as other callers, but that would be a different patch.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-am: give better diagnostics when the patch does not apply during --3way
Junio C Hamano [Sat, 12 Aug 2006 23:16:47 +0000 (16:16 -0700)]
git-am: give better diagnostics when the patch does not apply during --3way

If the user tries to apply a patch that was hand-edited in such
a way that it does not apply to the original file recorded on
its "index" line anymore, we did detect the situation but did
not issue an error message that is specific enough.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'rj/header'
Junio C Hamano [Sat, 12 Aug 2006 23:09:19 +0000 (16:09 -0700)]
Merge branch 'rj/header'

* rj/header:
  Fix header breakage with _XOPEN_SOURCE.

18 years agogit-svn: split the path from the url correctly with limited perms
Eric Wong [Sat, 12 Aug 2006 06:21:41 +0000 (23:21 -0700)]
git-svn: split the path from the url correctly with limited perms

This version of the splitter (that only affects SVN:: library
users) works when one only has limited read-permissions to
the repository they're fetching from.

Updated from the original patch to workaround some SVN bug
somewhere, which only seems to happen against file://
repositories...  Here's the diff against the original patch I
submitted:

@@ -1159,8 +1159,8 @@ sub repo_path_split {
  }

  if ($_use_lib) {
- $SVN = libsvn_connect($full_url);
- my $url = $SVN->get_repos_root;
+ my $tmp = libsvn_connect($full_url);
+ my $url = $tmp->get_repos_root;
  $full_url =~ s#^\Q$url\E/*##;
  push @repo_path_split_cache, qr/^(\Q$url\E)/;
  return ($url, $full_url);

Somehow connecting to a repository with the full url makes the
returned SVN::Ra object act strangely and break things, so now
we just drop the SVN::Ra object that we made our initial
connection with.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-grep: show pathnames relative to the current directory
Junio C Hamano [Fri, 11 Aug 2006 07:44:42 +0000 (00:44 -0700)]
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>
18 years agogit-sh-setup: do not use repo-config to test the git directory
Junio C Hamano [Sat, 12 Aug 2006 01:47:50 +0000 (18:47 -0700)]
git-sh-setup: do not use repo-config to test the git directory

Since repo-config does not fail in non-git directory, it is not
a good command to use to test the git-ness nor validate the
repository revision of $GIT_DIR.

Original patch by Robert Shearman but with minor fixes.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-svn: bugfix: allow SVN:: lib users to track the root of the repository
Eric Wong [Fri, 11 Aug 2006 18:11:29 +0000 (11:11 -0700)]
git-svn: bugfix: allow SVN:: lib users to track the root of the repository

I'm not sure if anybody has hit this (besides me), but this
fixes the problem where I ran into while attempting to import a
small repo at the root level:  I ended up with all the commits, but
with no file/tree changes at all throughout the entire history.

Also, fix a warning if the commit message is not defined for revision 0.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-svn: correctly kill keyword expansion without munging EOLs
Eric Wong [Fri, 11 Aug 2006 11:34:07 +0000 (04:34 -0700)]
git-svn: correctly kill keyword expansion without munging EOLs

This bugfix applies to users of the svn command-line client only.

We no longer muck with newlines when killing keyword expansion.
This tended to generate unintended diffs in commits because svn
revert -R would destroy the manual EOL changes we were doing. Of
course, we didn't need the EOL munging in the first place, as
svn seems to do it for us even in the text-base files.

Now we set the mtime and atime the files changed by keyword
expansion killing to avoid triggering a change on svn revert,
which svn still seems to want to do.

Thanks to Seth Falcon for reporting this bug.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agodrop length argument of has_extension
Rene Scharfe [Fri, 11 Aug 2006 12:01:45 +0000 (14:01 +0200)]
drop length argument of has_extension

As Fredrik points out the current interface of has_extension() is
potentially confusing.  Its parameters include both a nul-terminated
string and a length-limited string.

This patch drops the length argument, requiring two nul-terminated
strings; all callsites are updated.  I checked that all of them indeed
provide nul-terminated strings.  Filenames need to be nul-terminated
anyway if they are to be passed to open() etc.  The performance penalty
of the additional strlen() is negligible compared to the system calls
which inevitably surround has_extension() calls.

Additionally, change has_extension() to use size_t inside instead of
int, as that is the exact type strlen() returns and memcmp() expects.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'rn/push-dav'
Junio C Hamano [Fri, 11 Aug 2006 06:13:50 +0000 (23:13 -0700)]
Merge branch 'rn/push-dav'

* rn/push-dav:
  http-push: Make WebDAV work with (broken?) default apache2 WebDAV module

18 years agoAdd the --color-words option to the diff options family
Johannes Schindelin [Fri, 28 Jul 2006 21:56:15 +0000 (23:56 +0200)]
Add the --color-words option to the diff options family

With this option, the changed words are shown inline. For example,
if a file containing "This is foo" is changed to "This is bar", the diff
will now show "This is " in plain text, "foo" in red, and "bar" in green.

Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAdd Documentation/howto/setup-git-server-over-http.txt
Rutger Nijlunsing [Thu, 10 Aug 2006 20:00:26 +0000 (22:00 +0200)]
Add Documentation/howto/setup-git-server-over-http.txt

A small howto on how to setup GIT over HTTP transport protocol by
setting up WebDAV access on apache2.

[jc: minimum ispell fixes applied]

Signed-off-by: Rutger Nijlunsing <git@tux.tmfweb.nl>
Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Whitespace cleanup - tabs are for indent, spaces are for align
Jakub Narebski [Thu, 10 Aug 2006 10:38:47 +0000 (12:38 +0200)]
gitweb: Whitespace cleanup - tabs are for indent, spaces are for align

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-verify-pack: make builtin
Rene Scharfe [Thu, 10 Aug 2006 15:02:38 +0000 (17:02 +0200)]
git-verify-pack: make builtin

Convert git-verify-pack to a builtin command.  Also rename ac to argc
and av to argv for consistancy.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'master' into mk/rename
Junio C Hamano [Thu, 10 Aug 2006 21:17:49 +0000 (14:17 -0700)]
Merge branch 'master' into mk/rename

* master:
  git-verify-pack: no need to count errors
  git-verify-pack: buffer overrun paranoia
  git-verify-pack: free pack after use and a cleanup
  git-verify-pack: get rid of while loop
  git-verify-pack: insist on .idx extension
  git-verify-pack: more careful path handling
  git-verify-pack: show usage when no pack was specified
  Add has_extension()
  builtin-apply: remove unused increment
  Fix git-diff A...B
  combine-diff: use color
  git-apply: applying a patch to make a symlink shorter.
  allow diff.renamelimit to be set regardless of -M/-C
  make --find-copies-harder imply -C
  find_unique_abbrev() with len=0 should not abbreviate
  check return value from diff_setup_done()
  Fix tutorial-2.html
  Documentation: git-status takes the same options as git-commit
  Update git-init-db(1) and documentation of core.sharedRepository

18 years agogit-verify-pack: no need to count errors
Rene Scharfe [Thu, 10 Aug 2006 15:02:37 +0000 (17:02 +0200)]
git-verify-pack: no need to count errors

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-verify-pack: buffer overrun paranoia
Rene Scharfe [Thu, 10 Aug 2006 15:02:36 +0000 (17:02 +0200)]
git-verify-pack: buffer overrun paranoia

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-verify-pack: free pack after use and a cleanup
Rene Scharfe [Thu, 10 Aug 2006 15:02:35 +0000 (17:02 +0200)]
git-verify-pack: free pack after use and a cleanup

Plug memory leak in verify_one_pack() by freeing the struct packed_git
we got from add_packed_git().  Also rename g to pack and pull an
assignment out of an if statement while we're at it.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-verify-pack: get rid of while loop
Rene Scharfe [Thu, 10 Aug 2006 15:02:34 +0000 (17:02 +0200)]
git-verify-pack: get rid of while loop

Get rid of that while loop which was apparently used as a way to avoid
goto's (why?).  It's easy now because there is only one break left at
the end of it.  Also make the comment clearer.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-verify-pack: insist on .idx extension
Rene Scharfe [Thu, 10 Aug 2006 15:02:33 +0000 (17:02 +0200)]
git-verify-pack: insist on .idx extension

git-verify-pack can be called with a filename without .idx extension.
add_packed_git() on the other hand depends on its presence.  So
instead of trying to call it with whatever the user gave us check for
that extension and add it if it's missing.

That means that you can't name your index file "blah" and your pack
file ".pack" anymore ("git-verify-pack blah" currently works in that
case).  I think this regression is a good change. ;-)

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-verify-pack: more careful path handling
Rene Scharfe [Thu, 10 Aug 2006 15:02:32 +0000 (17:02 +0200)]
git-verify-pack: more careful path handling

Use strlcpy() to copy the filename into a buffer and complain if it
doesn't fit.  Also move the path buffer into verify_one_pack(); it is
used only there.  Now we can const'ify the first argument of this
function.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-verify-pack: show usage when no pack was specified
Rene Scharfe [Thu, 10 Aug 2006 15:02:31 +0000 (17:02 +0200)]
git-verify-pack: show usage when no pack was specified

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAdd has_extension()
Rene Scharfe [Thu, 10 Aug 2006 15:02:30 +0000 (17:02 +0200)]
Add has_extension()

The little helper has_extension() documents through its name what we are
trying to do and makes sure we don't forget the underrun check.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-apply: remove unused increment
Junio C Hamano [Thu, 10 Aug 2006 07:56:40 +0000 (00:56 -0700)]
builtin-apply: remove unused increment

We do not use desc.alloc after assigning desc.buffer to patch->result;
do not bother to increment it.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix git-diff A...B
Junio C Hamano [Thu, 10 Aug 2006 07:50:15 +0000 (00:50 -0700)]
Fix git-diff A...B

Commit 9919f41 meant to make git-diff A...B to (usually) mean
"git-diff `git-merge-base A B` B", but it got the parameters wrong
and ended up showing "git-diff `git-merge-base A B` A" by mistake.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agocombine-diff: use color
Junio C Hamano [Thu, 10 Aug 2006 07:30:33 +0000 (00:30 -0700)]
combine-diff: use color

Using the same mechanism as the regular diffs, color combined diff
output.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agohttp-push: Make WebDAV work with (broken?) default apache2 WebDAV module
Rutger Nijlunsing [Wed, 9 Aug 2006 18:54:23 +0000 (20:54 +0200)]
http-push: Make WebDAV work with (broken?) default apache2 WebDAV module

WebDAV on Debian unstable cannot handle renames on WebDAV from
file.ext to newfile (without ext) when newfile* already
exists. Normally, git creates a file like 'objects/xx/sha1.token',
which is renamed to 'objects/xx/sha1' when transferred completely.

Just use '_' instead of '.' so WebDAV doesn't see it as an extension
change.

Signed-off-by: Rutger Nijlunsing <git@tux.tmfweb.nl>
Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-apply: applying a patch to make a symlink shorter.
Junio C Hamano [Thu, 10 Aug 2006 05:47:25 +0000 (22:47 -0700)]
git-apply: applying a patch to make a symlink shorter.

The internal representation of the result is counted string
(i.e. char *buf and ulong size), which is fine for writing out
to regular file, but throwing the buf at symlink(2) was a
no-no.

Reported by Willy Tarreau.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoallow diff.renamelimit to be set regardless of -M/-C v1.4.2-rc4
Junio C Hamano [Wed, 9 Aug 2006 21:05:23 +0000 (14:05 -0700)]
allow diff.renamelimit to be set regardless of -M/-C

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agomake --find-copies-harder imply -C
Junio C Hamano [Wed, 9 Aug 2006 20:17:19 +0000 (13:17 -0700)]
make --find-copies-harder imply -C

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agofind_unique_abbrev() with len=0 should not abbreviate
Junio C Hamano [Wed, 9 Aug 2006 20:17:04 +0000 (13:17 -0700)]
find_unique_abbrev() with len=0 should not abbreviate

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agocheck return value from diff_setup_done()
Junio C Hamano [Wed, 9 Aug 2006 19:45:27 +0000 (12:45 -0700)]
check return value from diff_setup_done()

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix tutorial-2.html
Junio C Hamano [Wed, 9 Aug 2006 07:57:01 +0000 (00:57 -0700)]
Fix tutorial-2.html

Honza Pazdziora noticed that one example did not match reality.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: Add support for setting CURLDIR, OPENSSLDIR, EXPATDIR
Jakub Narebski [Wed, 9 Aug 2006 00:15:10 +0000 (02:15 +0200)]
autoconf: Add support for setting CURLDIR, OPENSSLDIR, EXPATDIR

Add support for --with-openssl=PATH and --without-openssl,
--with-curl=PATH and --without-curl, --with-expat=PATH and
--without-expat ./configure options, each setting or unsetting
appropriate NO_PACKAGE and if called with argument also PACKAGEDIR
(of which only CURLDIR is documented in Makefile)

All above options are supported as override to autodetection; more to
come in the same style (override to autodetection), so moved the bulk
of comments for site configuration down.

Needs review by someone well versed in autoconf and m4.

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoDocumentation: git-status takes the same options as git-commit
Junio C Hamano [Tue, 8 Aug 2006 05:41:32 +0000 (22:41 -0700)]
Documentation: git-status takes the same options as git-commit

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: It is --without-python, not --no-python
Jakub Narebski [Wed, 9 Aug 2006 00:19:22 +0000 (02:19 +0200)]
autoconf: It is --without-python, not --no-python

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoUpdate git-init-db(1) and documentation of core.sharedRepository
Jonas Fonseca [Wed, 9 Aug 2006 00:26:23 +0000 (02:26 +0200)]
Update git-init-db(1) and documentation of core.sharedRepository

Combine option descriptions in git-init-db(1). Reflect the changes to
additionally allow all users to read the created git repository.

Signed-off-by: Jonas Fonseca <fonseca@diku.dk>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: clean temporary file mak.append
Junio C Hamano [Tue, 8 Aug 2006 20:42:35 +0000 (13:42 -0700)]
autoconf: clean temporary file mak.append

When configure is interrupted in the middle it leaves
config.mak.append behind.  Add it to .gitignore and make sure
$(MAKE) clean removes it.

Also earlier .gitignore listed config.mak.in which is a tracked
file.  Fix it.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: Move variables which we always set to config.mak.in
Jakub Narebski [Tue, 8 Aug 2006 16:39:06 +0000 (18:39 +0200)]
autoconf: Move variables which we always set to config.mak.in

Move detected NO_STH and NEED_STH variables, which we always output,
either setting or unsetting (setting to empty string) to config.mak.in
and use setting appropriately named variables and doing AC_SUBST
instead of adding them via GIT_CONF_APPEND_LINE macro and
config.mak.append temporary file.

Variables which might and might not be set are still added via
config.mak.append; this include all STH_PATH variables.

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoread-cache: tweak racy-git delay logic
Junio C Hamano [Tue, 8 Aug 2006 21:47:32 +0000 (14:47 -0700)]
read-cache: tweak racy-git delay logic

Instead of looping over the entries and writing out, use a
separate loop after all entries have been written out to check
how many entries are racily clean.  Make sure that the newly
created index file gets the right timestamp when we check by
flushing the buffered data by ce_write().

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'master' into mk/rename
Junio C Hamano [Tue, 8 Aug 2006 22:42:20 +0000 (15:42 -0700)]
Merge branch 'master' into mk/rename

18 years agoautoconf: Improvements in NO_PYTHON/PYTHON_PATH handling
Jakub Narebski [Tue, 8 Aug 2006 16:38:06 +0000 (18:38 +0200)]
autoconf: Improvements in NO_PYTHON/PYTHON_PATH handling

Unset NO_PYTHON for --with-python without arguments, and when
PYTHON_PATH is set.  Do not check for PYTHON_PATH if it is set
via --with-python=PYTHON_PATH.  Prefer "python" over version
specific names such as "python2.4".

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: Error out on --without-shell and --without-perl
Jakub Narebski [Tue, 8 Aug 2006 16:36:21 +0000 (18:36 +0200)]
autoconf: Error out on --without-shell and --without-perl

Error out on --without-shell/--with-shell=no and
--without-perl/--with-perl=no instead of just warning
and continuing.

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: Add configure target to main Makefile
Jakub Narebski [Tue, 8 Aug 2006 16:35:23 +0000 (18:35 +0200)]
autoconf: Add configure target to main Makefile

While at it fill git version information in configure.ac
configure target needs autoconf, of course.

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-push: allow pushing from subdirectories
Jeff King [Tue, 8 Aug 2006 20:01:32 +0000 (16:01 -0400)]
git-push: allow pushing from subdirectories

The semantics are equivalent to pushing from the root; we just try harder to
find the .git directory.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoGIT-VERSION-GEN: adjust for ancient git
Junio C Hamano [Tue, 8 Aug 2006 20:11:16 +0000 (13:11 -0700)]
GIT-VERSION-GEN: adjust for ancient git

When an ancient "git" that does not understand "describe"
command is on the $PATH, "git describe" emitted a Usage message
without exiting non-zero status (which is a mistake we cannot
fix retroactively).  Catch this case to make sure we do not try
using phoney multi-line string as a version number.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-mv: fix use of uninitialized memory.
Junio C Hamano [Tue, 8 Aug 2006 19:21:33 +0000 (12:21 -0700)]
builtin-mv: fix use of uninitialized memory.

Juergen Ruehle noticed that add_slash() tries to strcat()
into uninitialized memory and fails.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agodebugging: XMALLOC_POISON
Junio C Hamano [Tue, 8 Aug 2006 19:23:47 +0000 (12:23 -0700)]
debugging: XMALLOC_POISON

Compile with -DXMALLOC_POISON=1 to catch errors from using uninitialized
memory returned by xmalloc.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: blame table row no highlight fix
Luben Tuikov [Fri, 4 Aug 2006 22:09:59 +0000 (15:09 -0700)]
gitweb: blame table row no highlight fix

Until now blame just used the commit/tree/tags/etc style of
highlight-able table rows, which have alternating light/dark rows that
flash when mouse pointer passes over them. This is very annoying in
blame, since the text is static and it interferes with the
per-revision block highlighting.

Signed-off-by: Luben Tuikov <ltuikov@yahoo.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: bugfix: git_commit and git_commitdiff parents
Luben Tuikov [Fri, 4 Aug 2006 22:11:47 +0000 (15:11 -0700)]
gitweb: bugfix: git_commit and git_commitdiff parents

In git_commit() the hash base of $from_id is $parent, not
$hash:
 - If status is "D", then action blob for $from_id wants
   $parent, not $hash.  History needs $parent too.
 - If status is "R", then action blob for $from_id wants
   $parent, not $hash.

Similarly in git_commitdiff() the hash base of $from_id is
$hash_parent, not $hash.

Signed-off-by: Luben Tuikov <ltuikov@yahoo.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Remove unused variables in git_shortlog_body and git_heads
Jakub Narebski [Mon, 7 Aug 2006 23:15:05 +0000 (01:15 +0200)]
gitweb: Remove unused variables in git_shortlog_body and git_heads

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoannotate: Fix bug when parsing merges with differing real and logical parents.
Ryan Anderson [Mon, 7 Aug 2006 12:11:24 +0000 (05:11 -0700)]
annotate: Fix bug when parsing merges with differing real and logical parents.

Signed-off-by: Ryan Anderson <ryan@michonline.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agolog-tree: show_log() should respect the setting of diffopt->line_termination
Ryan Anderson [Mon, 7 Aug 2006 12:11:23 +0000 (05:11 -0700)]
log-tree: show_log() should respect the setting of diffopt->line_termination

Signed-off-by: Ryan Anderson <ryan@michonline.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agohandle https:// protocol in git-clone
Michael Krelin [Mon, 7 Aug 2006 15:10:21 +0000 (17:10 +0200)]
handle https:// protocol in git-clone

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoRacy git: avoid having to be always too careful
Junio C Hamano [Sat, 5 Aug 2006 11:16:02 +0000 (04:16 -0700)]
Racy git: avoid having to be always too careful

Immediately after a bulk checkout, most of the paths in the
working tree would have the same timestamp as the index file,
and this would force ce_match_stat() to take slow path for all
of them.  When writing an index file out, if many of the paths
have very new (read: the same timestamp as the index file being
written out) timestamp, we are better off delaying the return
from the command, to make sure that later command to touch the
working tree files will leave newer timestamps than recorded in
the index, thereby avoiding to take the slow path.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: Unset NO_STH and NEED_STH when it is detected not needed
Jakub Narebski [Fri, 4 Aug 2006 21:43:59 +0000 (23:43 +0200)]
autoconf: Unset NO_STH and NEED_STH when it is detected not needed

When configure detects some NO_XXX or NEEDS_XXX is unneeded, unset
this variable (actually set it to empty string).  This allow
autodetection to override the default set in Makefile.

[jc: while at it fixed a thinko in IPv6 detection.]

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: fix $project usage
Matthias Lederhofer [Sun, 6 Aug 2006 11:25:41 +0000 (13:25 +0200)]
gitweb: fix $project usage

There were some places where $project was used even if it was not
defined.

Signed-off-by: Matthias Lederhofer <matled@gmx.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: fix commitdiff_plain for root commits
Matthias Lederhofer [Sun, 6 Aug 2006 17:24:47 +0000 (19:24 +0200)]
gitweb: fix commitdiff_plain for root commits

Signed-off-by: Matthias Lederhofer <matled@gmx.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Skip nonmatching lines in difftree output, consistently
Jakub Narebski [Sun, 6 Aug 2006 15:59:52 +0000 (17:59 +0200)]
gitweb: Skip nonmatching lines in difftree output, consistently

This fixes error for commitdiff on root commit (without parents).

Noticed-by: Matthias Lederhofer (matled)
Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: fix commitdiff for root commits
Jakub Narebski [Sun, 6 Aug 2006 14:14:25 +0000 (16:14 +0200)]
gitweb: fix commitdiff for root commits

After changing all "-|" open invocations to list form, commitdiff for
initial commit (without parent) got broken; it returned incorrectly
empty patch earlier.  Use '--root' option to git-diff-tree for initial
(root) commit.

No checking for empty $hash_parent in git_commitdiff_plain -- we rely
on gitweb to give correct parameters for commitdiff_plain action.

Noticed by Matthias Lederhofer (matled).

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: check if HTTP_ACCEPT is really set
Matthias Lederhofer [Sun, 6 Aug 2006 13:55:02 +0000 (15:55 +0200)]
gitweb: check if HTTP_ACCEPT is really set

Signed-off-by: Matthias Lederhofer <matled@gmx.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAdd gitweb.cgi to .gitignore
Matthias Kestenholz [Sun, 6 Aug 2006 10:47:33 +0000 (12:47 +0200)]
Add gitweb.cgi to .gitignore

Signed-off-by: Matthias Kestenholz <matthias@spinlock.ch>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Refactor untabifying - converting tabs to spaces
Jakub Narebski [Sun, 6 Aug 2006 00:08:31 +0000 (02:08 +0200)]
gitweb: Refactor untabifying - converting tabs to spaces

Add untabify subroutine and use it.  It also fixes git_diff_print
which used to get the tabstop wrong.

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Inline $rss_link
Jakub Narebski [Sat, 5 Aug 2006 11:18:58 +0000 (13:18 +0200)]
gitweb: Inline $rss_link

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: PATH_INFO=/ means no project
Jakub Narebski [Sat, 5 Aug 2006 11:16:03 +0000 (13:16 +0200)]
gitweb: PATH_INFO=/ means no project

Prepared for refactoring input validation.

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: No error messages with unescaped/unprotected user input
Jakub Narebski [Sat, 5 Aug 2006 11:15:24 +0000 (13:15 +0200)]
gitweb: No error messages with unescaped/unprotected user input

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: No periods for error messages
Jakub Narebski [Sat, 5 Aug 2006 11:13:53 +0000 (13:13 +0200)]
gitweb: No periods for error messages

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Acked-by: Luben Tuikov <ltuikov@yahoo.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Cleanup and uniquify error messages
Jakub Narebski [Sat, 5 Aug 2006 11:12:51 +0000 (13:12 +0200)]
gitweb: Cleanup and uniquify error messages

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Don't undefine query parameter related variables before die_error
Jakub Narebski [Sat, 5 Aug 2006 10:58:06 +0000 (12:58 +0200)]
gitweb: Don't undefine query parameter related variables before die_error

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Use undef for die_error to use default first (status) parameter value
Jakub Narebski [Sat, 5 Aug 2006 10:56:42 +0000 (12:56 +0200)]
gitweb: Use undef for die_error to use default first (status) parameter value

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: die_error first (optional) parameter is HTTP status
Jakub Narebski [Sat, 5 Aug 2006 10:56:04 +0000 (12:56 +0200)]
gitweb: die_error first (optional) parameter is HTTP status

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: Separate input validation and dispatch, add comment about opml action
Jakub Narebski [Sat, 5 Aug 2006 10:55:20 +0000 (12:55 +0200)]
gitweb: Separate input validation and dispatch, add comment about opml action

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix "grep -w"
Junio C Hamano [Sat, 5 Aug 2006 05:16:42 +0000 (22:16 -0700)]
Fix "grep -w"

We used to find the first match of the pattern and then if the
match is not for the entire word, declared that the whole line
does not match.

But that is wrong.  The command "git grep -w -e mmap" should
find that a line "foo_mmap bar mmap baz" matches, by tring the
second instance of pattern "mmap" on the same line.

Problems an earlier round of "fix" had were pointed out by Morten
Welinder, which have been incorporated in the t7002 tests.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMakefile: Cygwin does not seem to need NO_STRLCPY
Junio C Hamano [Sat, 5 Aug 2006 07:20:51 +0000 (00:20 -0700)]
Makefile: Cygwin does not seem to need NO_STRLCPY

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agocommit walkers: setup_ident() to record correct committer in ref-log.
Ramsay Jones [Fri, 4 Aug 2006 21:01:34 +0000 (22:01 +0100)]
commit walkers: setup_ident() to record correct committer in ref-log.

The function pull() in fetch.c calls write_ref_sha1(), which may
need committer identity to update the ref-log, so they need to
call setup_ident() before calling git_config() function.

Acked-by: Shawn Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agohttp-push: avoid fork() by calling merge_bases() directly
Johannes Schindelin [Fri, 4 Aug 2006 15:50:41 +0000 (17:50 +0200)]
http-push: avoid fork() by calling merge_bases() directly

Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAllow config file to specify Signed-off-by identity in format-patch.
Ramsay Jones [Fri, 4 Aug 2006 21:01:30 +0000 (22:01 +0100)]
Allow config file to specify Signed-off-by identity in format-patch.

Unlike git-commit, git-format-patch was not picking up and using the
user.email config variable for the email part of the committer info.
I was forced to use the GIT_COMMITTER_EMAIL environment variable to
override the default <user@localhost.localdomain>. The fix was to
simply move the call to setup_ident() to come before the git_config()
call.

Signed-off-by: Ramsay Allan Jones <ramsay@ramsay1.demon.co.uk>
Acked-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitweb: git_tree displays blame based on repository config
Luben Tuikov [Fri, 4 Aug 2006 22:14:27 +0000 (15:14 -0700)]
gitweb: git_tree displays blame based on repository config

git_tree() will now conditionally display "blame"
depending on how "gitweb.blame" variable is configured
using "git-repo-config".

Signed-off-by: Luben Tuikov <ltuikov@yahoo.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: NO_IPV6
Junio C Hamano [Fri, 4 Aug 2006 23:46:16 +0000 (16:46 -0700)]
autoconf: NO_IPV6

We would need both "struct addrinfo" and getaddrinfo()
available.  Check them and set NO_IPV6 otherwise.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: Set NEEDS_LIBICONV unconditionally if there is no iconv in libc
Jakub Narebski [Fri, 4 Aug 2006 21:28:11 +0000 (23:28 +0200)]
autoconf: Set NEEDS_LIBICONV unconditionally if there is no iconv in libc

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoautoconf: fix NEEDS_SSL_WITH_CRYPTO
Junio C Hamano [Fri, 4 Aug 2006 23:33:18 +0000 (16:33 -0700)]
autoconf: fix NEEDS_SSL_WITH_CRYPTO

NEEDS_SSL_WITH_CRYPTO means you cannot just say "-lcrypto" to
use SHA1 stuff, but need to say "-lcrypto -lssl".

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'jc/c99'
Junio C Hamano [Fri, 4 Aug 2006 19:09:53 +0000 (12:09 -0700)]
Merge branch 'jc/c99'

* jc/c99:
  Cygwin needs NO_C99_FORMAT???