git.git
17 years agocvsimport: introduce _fetchfile() method and used a 1M buffer to read()
Martin Langhoff [Tue, 23 May 2006 08:08:58 +0000 (20:08 +1200)]
cvsimport: introduce _fetchfile() method and used a 1M buffer to read()

File retrieval from the socket is now moved to _fetchfile() and we now
cap reads at 1MB. This should limit the memory growth of the cvsimport
process.

Signed-off-by: Martin Langhoff <martin@catalyst.net.nz>
Signed-off-by: Junio C Hamano <junkio@cox.net>
17 years agocvsimport: cleanup commit function
Jeff King [Tue, 23 May 2006 07:27:46 +0000 (03:27 -0400)]
cvsimport: cleanup commit function

This change attempts to clean up the commit function to make it a bit
easier to read (or at least the first half of it). It also improves
robustness and performance. Specifically:
  - report get_headref errors on opening ref unless the error is ENOENT
  - use regex to check for sha1 instead of length
  - use lexically scoped filehandles which get cleaned up automagically
  - check for error on both 'print' and 'close' (since output is buffered)
  - avoid "fork, do some perl, then exec" in commit(). It's not necessary,
    and we probably end up COW'ing parts of the perl process. Plus the code
    is much smaller because we can use open2()
  - avoid calling strftime over and over (mainly a readability cleanup)

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
17 years agocvsimport: use git-update-index --index-info
Jeff King [Tue, 23 May 2006 07:27:45 +0000 (03:27 -0400)]
cvsimport: use git-update-index --index-info

This should reduce the number of git-update-index forks required per
commit. We now do adds/removes in one call, and we are no longer forced to
deal with argv limitations.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
17 years agogit status: skip empty directories, and add -u to show all untracked files
Matthias Lederhofer [Mon, 22 May 2006 21:02:06 +0000 (23:02 +0200)]
git status: skip empty directories, and add -u to show all untracked files

By default, we use --others --directory to show uninteresting
directories (to get user's attention) without their contents (to
unclutter output).  Showing empty directories do not make sense,
so pass --no-empty-directory when we do so.

Giving -u (or --untracked) disables this uncluttering to let the
user get all untracked files.

Signed-off-by: Junio C Hamano <junkio@cox.net>
17 years agocvsimport: repack every kilo-commits.
Linus Torvalds [Tue, 23 May 2006 02:28:37 +0000 (19:28 -0700)]
cvsimport: repack every kilo-commits.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Acked-by: Martin Langhoff <martin@catalyst.net.nz>
Signed-off-by: Junio C Hamano <junkio@cox.net>
17 years agocvsimport: introduce -L<imit> option to workaround memory leaks
Martin Langhoff [Mon, 22 May 2006 11:38:08 +0000 (23:38 +1200)]
cvsimport: introduce -L<imit> option to workaround memory leaks

Signed-off-by: Martin Langhoff <martin@catalyst.net.nz>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoRemove possible segfault in http-fetch.
Sean [Sat, 20 May 2006 22:46:33 +0000 (18:46 -0400)]
Remove possible segfault in http-fetch.

Free the curl string lists after running http_cleanup to
avoid an occasional segfault in the curl library.  Seems
to only occur if the website returns a 405 error.

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoInstall git builtins into gitexecdir rather than bindir.
Sean [Mon, 22 May 2006 04:42:59 +0000 (00:42 -0400)]
Install git builtins into gitexecdir rather than bindir.

Moving "git-cmd" commands out of the path and into a special
git exec path, should include the builtins.

[jc: fixed the case where bindir == gitexecdir - ln -f fails
 with a complaint that src and dst are the same, likewise for
 the fallback cp.]

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoChange GIT-VERSION-GEN to call git commands with "git" not "git-".
Sean [Mon, 22 May 2006 04:39:52 +0000 (00:39 -0400)]
Change GIT-VERSION-GEN to call git commands with "git" not "git-".

GIT-VERSION-GEN can incorrectly return a default version of
"v1.3.GIT" because it tries to execute git commands using the
"git-cmd" format that expects all git commands to be in the $PATH.
Convert these to  "git cmd" format so that a proper answer is
returned even when the git commands have been moved out of the
$PATH and into a $gitexecdir.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agocvsimport: replace anonymous sub ref with a normal sub
Martin Langhoff [Mon, 22 May 2006 12:45:39 +0000 (00:45 +1200)]
cvsimport: replace anonymous sub ref with a normal sub

commit() does not need to be an anonymous subreference. Keep it simple.

Signed-off-by: Martin Langhoff <martin@catalyst.net.nz>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agocvsimport: minor fixups
Martin Langhoff [Mon, 22 May 2006 12:45:47 +0000 (00:45 +1200)]
cvsimport: minor fixups

Cleanup @skipped after it's used. Close a fhandle.
Removing suspects one at a time.

Signed-off-by: Martin Langhoff <martin@catalyst.net.nz>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoProblem: 'trap...exit' causes error message when /bin/sh is ash.
Yakov Lerner [Mon, 22 May 2006 19:34:00 +0000 (22:34 +0300)]
Problem: 'trap...exit' causes error message when /bin/sh is ash.

Problem: 'trap...exit' causes error message when /bin/sh is ash.
Fix: Change 'trap...exit' to 'trap...0' like in other scripts.

Signed-off-by: Yakov Lerner <iler.ml@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAvoid segfault in diff --stat rename output.
Sean [Tue, 23 May 2006 00:36:34 +0000 (20:36 -0400)]
Avoid segfault in diff --stat rename output.

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoTutorial #2: broken link fix.
Junio C Hamano [Mon, 22 May 2006 22:34:40 +0000 (15:34 -0700)]
Tutorial #2: broken link fix.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit help: remove whatchanged from list of common commands
Martin Waitz [Mon, 22 May 2006 10:09:14 +0000 (12:09 +0200)]
git help: remove whatchanged from list of common commands

whatchanged is replaced by git log now.

Signed-off-by: Martin Waitz
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agofetch-pack: output refs in the order they were given on the command line.
Junio C Hamano [Thu, 11 May 2006 22:28:44 +0000 (15:28 -0700)]
fetch-pack: output refs in the order they were given on the command line.

Currently, fetched refs are output in the order the remote side
happened to send them.  This changes the order to match the
order of refs that were given on the command line.  To the
existing core callers (git-fetch and git-clone) this does not
make any difference, but for other Porcelain use, it would be
more intuitive.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agov266
Kay Sievers [Mon, 22 May 2006 12:31:09 +0000 (14:31 +0200)]
v266

18 years agoremove Christian's email address
Kay Sievers [Mon, 22 May 2006 12:30:47 +0000 (14:30 +0200)]
remove Christian's email address

18 years agodiff: minor option combination fix.
Junio C Hamano [Mon, 22 May 2006 07:31:02 +0000 (00:31 -0700)]
diff: minor option combination fix.

output_format == DIFFSTAT and with_stat == true does not make sense, and
the way the code is structured it causes trouble.  Avoid it.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin format-patch: squelch content-type for 7-bit ASCII
Junio C Hamano [Mon, 22 May 2006 06:55:00 +0000 (23:55 -0700)]
builtin format-patch: squelch content-type for 7-bit ASCII

When --attach is not used, usually we do not say Content-Type:
and fluff, but if the commit message is not 7-bit ASCII, mark
it as "text/plain; charset=UTF-8".  This unclutters output
somewhat.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoCMIT_FMT_EMAIL: Q-encode Subject: and display-name part of From: fields.
Junio C Hamano [Tue, 16 May 2006 09:29:42 +0000 (02:29 -0700)]
CMIT_FMT_EMAIL: Q-encode Subject: and display-name part of From: fields.

By convention, the commit message and the author/committer names
in the commit objects are UTF-8 encoded.  When formatting for
e-mails, Q-encode them according to RFC 2047.

While we are at it, generate the content-type and
content-transfer-encoding headers as well.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agomailinfo: skip bogus UNIX From line inside body
Junio C Hamano [Mon, 22 May 2006 00:15:06 +0000 (17:15 -0700)]
mailinfo: skip bogus UNIX From line inside body

Sometimes people just include the whole format-patch output in
the commit e-mail.  Detect it and skip the bogus ">From " line.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agotutorial-2: typofix in examples.
Junio C Hamano [Mon, 22 May 2006 00:37:25 +0000 (17:37 -0700)]
tutorial-2: typofix in examples.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agotutorial: add discussion of index file, object database
J. Bruce Fields [Sun, 21 May 2006 23:49:34 +0000 (19:49 -0400)]
tutorial: add discussion of index file, object database

Add a sequel to tutorial.txt which discusses the index file and
the object database.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agotutorial: expanded discussion of commit history
J. Bruce Fields [Sun, 21 May 2006 20:54:05 +0000 (16:54 -0400)]
tutorial: expanded discussion of commit history

Expand the history-browsing section of the tutorial a bit, in part to
address Junio's suggestion that we mention "git grep" and Linus's
complaint that people are missing the flexibility of the commandline
interfaces for selecting commits.

This reads a little more like a collection of examples than a
"tutorial", but maybe that's what people need at this point.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agotutorial: replace "whatchanged" by "log"
J. Bruce Fields [Sun, 21 May 2006 20:52:34 +0000 (16:52 -0400)]
tutorial: replace "whatchanged" by "log"

Junio suggested changing references to git-whatchanged to git-log.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoNO_INET_NTOP and compat/inet_ntop.c for some systems (e.g. old Cygwin).
Yakov Lerner [Sun, 21 May 2006 21:37:00 +0000 (00:37 +0300)]
NO_INET_NTOP and compat/inet_ntop.c for some systems (e.g. old Cygwin).

For systems which lack inet_ntop(), this adds compat/inet_ntop.c,
and related build constant, NO_INET_NTOP. Older Cygwin(s) lack
inet_ntop().

Signed-off-by: Yakov Lerner <iler.ml@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoremove superflous "const"
Alex Riesen [Sun, 21 May 2006 20:45:46 +0000 (22:45 +0200)]
remove superflous "const"

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-rebase: use canonical A..B syntax to format-patch
Junio C Hamano [Sun, 21 May 2006 10:16:38 +0000 (03:16 -0700)]
git-rebase: use canonical A..B syntax to format-patch

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agocheckdiff_consume: strtol parameter fix.
Junio C Hamano [Sun, 21 May 2006 10:01:59 +0000 (03:01 -0700)]
checkdiff_consume: strtol parameter fix.

The second parameter is not the end of string input; it is
the optional return value to retrieve where the parser stopped.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit-format-patch: now built-in.
Junio C Hamano [Sun, 21 May 2006 09:48:21 +0000 (02:48 -0700)]
git-format-patch: now built-in.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoElaborate on why ':' is a bad idea in a ref name.
Shawn Pearce [Sun, 21 May 2006 02:03:14 +0000 (22:03 -0400)]
Elaborate on why ':' is a bad idea in a ref name.

With the new cat-file syntax of 'v1.3.3:refs.c' we should mention
it as part of the reason why ':' is not permitted in a ref name.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoReference git-check-ref-format in git-branch.
Shawn Pearce [Sun, 21 May 2006 01:54:46 +0000 (21:54 -0400)]
Reference git-check-ref-format in git-branch.

Its nice to have git-check-ref-format actually get mentioned in
git-branch's documentation as the syntax of a ref name must conform
to what is described in git-check-ref-format.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agofmt-patch: Support --attach
Johannes Schindelin [Sat, 20 May 2006 13:40:29 +0000 (15:40 +0200)]
fmt-patch: Support --attach

This patch touches a couple of files, because it adds options to print a
custom text just after the subject of a commit, and just after the
diffstat.

[jc: made "many dashes" used as the boundary leader into a single
 variable, to reduce the possibility of later tweaks to miscount the
 number of dashes to break it.]

Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'master' into js/fmt-patch
Junio C Hamano [Sun, 21 May 2006 08:34:54 +0000 (01:34 -0700)]
Merge branch 'master' into js/fmt-patch

* master: (119 commits)
  diff family: add --check option
  Document that "git add" only adds non-ignored files.
  Add a conversion tool to migrate remote information into the config
  fetch, pull: ask config for remote information
  Fix build procedure for builtin-init-db
  read-tree -m -u: do not overwrite or remove untracked working tree files.
  apply --cached: do not check newly added file in the working tree
  Implement a --dry-run option to git-quiltimport
  Implement git-quiltimport
  Revert "builtin-grep: workaround for non GNU grep."
  builtin-grep: workaround for non GNU grep.
  builtin-grep: workaround for non GNU grep.
  git-am: use apply --cached
  apply --cached: apply a patch without using working tree.
  apply --numstat: show new name, not old name.
  Documentation/Makefile: create tarballs for the man pages and html files
  Allow pickaxe and diff-filter options to be used by git log.
  Libify the index refresh logic
  Builtin git-init-db
  Remove unnecessary local in get_ref_sha1.
  ...

18 years agodiff family: add --check option
Johannes Schindelin [Sat, 20 May 2006 21:43:13 +0000 (23:43 +0200)]
diff family: add --check option

Actually, it is a diff option now, so you can say

git diff --check

to ask if what you are about to commit is a good patch.

[jc: this also would work for fmt-patch, but the point is that
 the check is done before making a commit.  format-patch is run
 from an already created commit, and that is too late to catch
 whitespace damaged change.]

Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoDocument that "git add" only adds non-ignored files.
Santi [Fri, 19 May 2006 21:02:34 +0000 (23:02 +0200)]
Document that "git add" only adds non-ignored files.

Signed-off-by: Santi Béjar <sbejar@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogitk: Highlight paths of interest in tree view as well
Paul Mackerras [Sat, 20 May 2006 12:56:27 +0000 (22:56 +1000)]
gitk: Highlight paths of interest in tree view as well

With this, when the file list window is in tree view mode, we highlight
the paths of interest.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agogitk: Highlight entries in the file list as well
Paul Mackerras [Sat, 20 May 2006 11:31:52 +0000 (21:31 +1000)]
gitk: Highlight entries in the file list as well

This applies a bold highlight to entries in the file list pane in the
bottom right corner when it is displaying the list of changed files.
This doesn't yet highlight file list entries when it is in tree view
mode.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] gitk: Display commit messages with word wrap
Sergey Vlasov [Mon, 15 May 2006 15:13:14 +0000 (19:13 +0400)]
[PATCH] gitk: Display commit messages with word wrap

Some people put very long strings into commit messages, which then
become invisible in gitk (word wrapping in the commit details window is
turned off, and there is no horizontal scroll bar).  Enabling word wrap
for just the commit message looks much better.

Wrapping is controlled by the "wrapcomment" option in ~/.gitk.  By
default this option is set to "none", which disables wrapping; setting
it to "word" enables word wrap for commit messages.

Signed-off-by: Sergey Vlasov <vsu@altlinux.ru>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge branch 'lt/dirwalk' into jc/dirwalk-n-cache-tree
Junio C Hamano [Sat, 20 May 2006 08:28:49 +0000 (01:28 -0700)]
Merge branch 'lt/dirwalk' into jc/dirwalk-n-cache-tree

This commit is what this branch is all about.  It records the
evil merge needed to adjust built-in git-add and git-rm for
the cache-tree extension.

* lt/dirwalk:
  Add builtin "git rm" command
  Move pathspec matching from builtin-add.c into dir.c
  Prevent bogus paths from being added to the index.
  builtin-add: fix unmatched pathspec warnings.
  Remove old "git-add.sh" remnants
  builtin-add: warn on unmatched pathspecs
  Do "git add" as a builtin
  Clean up git-ls-file directory walking library interface
  libify git-ls-files directory traversal

Conflicts:

Makefile
builtin.h
git.c
update-index.c

18 years agoMerge branch 'jc/cache-tree' into jc/dirwalk-n-cache-tree
Junio C Hamano [Sat, 20 May 2006 07:56:11 +0000 (00:56 -0700)]
Merge branch 'jc/cache-tree' into jc/dirwalk-n-cache-tree

* jc/cache-tree: (24 commits)
  Fix crash when reading the empty tree
  fsck-objects: do not segfault on missing tree in cache-tree
  cache-tree: a bit more debugging support.
  read-tree: invalidate cache-tree entry when a new index entry is added.
  Fix test-dump-cache-tree in one-tree disappeared case.
  fsck-objects: mark objects reachable from cache-tree
  cache-tree: replace a sscanf() by two strtol() calls
  cache-tree.c: typefix
  test-dump-cache-tree: validate the cached data as well.
  cache_tree_update: give an option to update cache-tree only.
  read-tree: teach 1-way merege and plain read to prime cache-tree.
  read-tree: teach 1 and 2 way merges about cache-tree.
  update-index: when --unresolve, smudge the relevant cache-tree entries.
  test-dump-cache-tree: report number of subtrees.
  cache-tree: sort the subtree entries.
  Teach fsck-objects about cache-tree.
  index: make the index file format extensible.
  cache-tree: protect against "git prune".
  Add test-dump-cache-tree
  Use cache-tree in update-index.
  ...

18 years agobuilt-in tar-tree and remote tar-tree
Junio C Hamano [Fri, 19 May 2006 01:46:44 +0000 (18:46 -0700)]
built-in tar-tree and remote tar-tree

This makes tar-tree a built-in.  As an added bonus, you can now
say:

git tar-tree --remote=remote-repository <ent> [<base>]

This does not work with git-daemon yet, but should work with
localhost and git over ssh transports.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'js/fetchconfig'
Junio C Hamano [Sat, 20 May 2006 00:44:07 +0000 (17:44 -0700)]
Merge branch 'js/fetchconfig'

* js/fetchconfig:
  Add a conversion tool to migrate remote information into the config
  fetch, pull: ask config for remote information

18 years agoAdd builtin "git rm" command
Linus Torvalds [Fri, 19 May 2006 23:19:34 +0000 (16:19 -0700)]
Add builtin "git rm" command

This changes semantics very subtly, because it adds a new atomicity
guarantee.

In particular, if you "git rm" several files, it will now do all or
nothing. The old shell-script really looped over the removed files one by
one, and would basically randomly fail in the middle if "-f" was used and
one of the files didn't exist in the working directory.

This C builtin one will not re-write the index after each remove, but
instead remove all files at once. However, that means that if "-f" is used
(to also force removal of the file from the working directory), and some
files have already been removed from the workspace, it won't stop in the
middle in some half-way state like the old one did.

So what happens is that if the _first_ file fails to be removed with "-f",
we abort the whole "git rm". But once we've started removing, we don't
leave anything half done. If some of the other files don't exist, we'll
just ignore errors of removal from the working tree.

This is only an issue with "-f", of course.

I think the new behaviour is strictly an improvement, but perhaps more
importantly, it is _different_. As a special case, the semantics are
identical for the single-file case (which is the only one our test-suite
seems to test).

The other question is what to do with leading directories. The old "git
rm" script didn't do anything, which is somewhat inconsistent. This one
will actually clean up directories that have become empty as a result of
removing the last file, but maybe we want to have a flag to decide the
behaviour?

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix build procedure for builtin-init-db
Junio C Hamano [Sat, 20 May 2006 00:23:07 +0000 (17:23 -0700)]
Fix build procedure for builtin-init-db

c3c8835fbb182d971d71939b9a3ec7c8b86d6caf broke the default template
location which is in builtin-init-db.o, by not supplying the
compilation-time constant to the right build commands.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'jc/read-tree-safety'
Junio C Hamano [Sat, 20 May 2006 00:02:22 +0000 (17:02 -0700)]
Merge branch 'jc/read-tree-safety'

* jc/read-tree-safety:
  read-tree -m -u: do not overwrite or remove untracked working tree files.

18 years agoMerge branch 'jc/apply'
Junio C Hamano [Fri, 19 May 2006 23:59:29 +0000 (16:59 -0700)]
Merge branch 'jc/apply'

* jc/apply:
  apply --cached: do not check newly added file in the working tree

18 years agoMerge branch 'master' into new
Paul Mackerras [Fri, 19 May 2006 23:59:10 +0000 (09:59 +1000)]
Merge branch 'master' into new

18 years agogitk: Fix bug where page-up/down wouldn't always work properly
Paul Mackerras [Fri, 19 May 2006 23:58:49 +0000 (09:58 +1000)]
gitk: Fix bug where page-up/down wouldn't always work properly

If the user pressed page up or page down and the new page wasn't
already drawn, we failed to select the line we wanted in the new
page.  This fixes it.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge branch 'eb/quilt'
Junio C Hamano [Fri, 19 May 2006 23:57:11 +0000 (16:57 -0700)]
Merge branch 'eb/quilt'

* eb/quilt:
  Implement a --dry-run option to git-quiltimport
  Implement git-quiltimport

18 years agoMerge branch 'jc/grep'
Junio C Hamano [Fri, 19 May 2006 23:55:33 +0000 (16:55 -0700)]
Merge branch 'jc/grep'

* jc/grep:
  Revert "builtin-grep: workaround for non GNU grep."
  builtin-grep: workaround for non GNU grep.

18 years agogitk: Make a row of controls for controlling highlighting
Paul Mackerras [Fri, 19 May 2006 23:38:11 +0000 (09:38 +1000)]
gitk: Make a row of controls for controlling highlighting

Now there is a bar across the middle (just below the bar containing
the sha1 ID, find string etc.) which controls highlighting.  There are
three ways to highlight: the user can highlight commits affecting
a list of paths, commits in a view, or commits where the author or
committer matches any of a list of strings (case-insensitive).  The
elements of the list of paths and list of names are delimited by
whitespace with shell quoting rules.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge branch 'lt/grep'
Junio C Hamano [Fri, 19 May 2006 23:26:45 +0000 (16:26 -0700)]
Merge branch 'lt/grep'

* lt/grep:
  builtin-grep: workaround for non GNU grep.
  git-am: use apply --cached
  apply --cached: apply a patch without using working tree.
  apply --numstat: show new name, not old name.

18 years agoMerge branch 'ts/doctar'
Junio C Hamano [Fri, 19 May 2006 23:26:01 +0000 (16:26 -0700)]
Merge branch 'ts/doctar'

* ts/doctar:
  Documentation/Makefile: create tarballs for the man pages and html files

18 years agoAllow pickaxe and diff-filter options to be used by git log.
Sean [Fri, 19 May 2006 04:19:20 +0000 (00:19 -0400)]
Allow pickaxe and diff-filter options to be used by git log.

Handle the -S option when passed to git log such that only the
appropriate commits are displayed.  Also per Junio's comments, do
the same for "--diff-filter", so that it too can be used as an option
to git log.  By default no patch or diff information is displayed.

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMove pathspec matching from builtin-add.c into dir.c
Linus Torvalds [Fri, 19 May 2006 23:07:51 +0000 (16:07 -0700)]
Move pathspec matching from builtin-add.c into dir.c

I'll use it for builtin-rm.c too.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoLibify the index refresh logic
Linus Torvalds [Fri, 19 May 2006 16:56:35 +0000 (09:56 -0700)]
Libify the index refresh logic

This cleans up and libifies the "git update-index --[really-]refresh"
functionality. This will be eventually required for eventually doing the
"commit" and "status" commands as built-ins.

It really just moves "refresh_index()" from update-index.c to
read-cache.c, but it also has to change the calling convention so that the
function uses a "unsigned int flags" argument instead of various static
flags variables for passing down the information about whether to be quiet
or not, and allow unmerged entries etc.

That actually cleans up update-index.c too, since it turns out that all
those flags were really specific to that one function of the index update,
so they shouldn't have had file-scope visibility even before.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoBuiltin git-init-db
Timo Hirvonen [Fri, 19 May 2006 10:03:57 +0000 (13:03 +0300)]
Builtin git-init-db

Basically this just renames init-db.c to builtin-init-db.c and makes
some strings const.

Signed-off-by: Timo Hirvonen <tihirvon@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge part of 'js/fmt-patch' for RFC2822 dates into 'sp/reflog'
Junio C Hamano [Fri, 19 May 2006 22:25:57 +0000 (15:25 -0700)]
Merge part of 'js/fmt-patch' for RFC2822 dates into 'sp/reflog'

An earlier patch from Shawn Pearce dependes on a change that is
only in "next".  I do not want to make this series hostage to
the yet-to-graduate js/fmt-patch branch, but let's try fixing it
by merging the early parts of the branch to see what happens.

Right now, 'sp/reflog' will not be in "next" for now, so I won't
have to regret this -- if this merge causes problem down the road
merging I can always rebuild the topic branch ;-).

18 years agoEnable ref log creation in git checkout -b.
Shawn Pearce [Fri, 19 May 2006 09:17:16 +0000 (05:17 -0400)]
Enable ref log creation in git checkout -b.

Switch git checkout -b to use git-update-ref rather than echo and
a shell I/O redirection.  This is more in line with typical GIT
commands and allows -b to be logged according to the normal ref
logging rules.

Added -l option to allow users to create the ref log at the same
time as creating a branch.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoCreate/delete branch ref logs.
Shawn Pearce [Fri, 19 May 2006 09:16:58 +0000 (05:16 -0400)]
Create/delete branch ref logs.

When crating a new branch offer '-l' as a way for the user to
quickly enable ref logging for the new branch.

When deleting a branch also delete its ref log.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoInclude ref log detail in commit, reset, etc.
Shawn Pearce [Fri, 19 May 2006 09:16:18 +0000 (05:16 -0400)]
Include ref log detail in commit, reset, etc.

When updating a ref at the direction of the user include a reason why
head was changed as part of the ref log (assuming it was enabled).

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoChange order of -m option to update-ref.
Shawn Pearce [Fri, 19 May 2006 09:15:50 +0000 (05:15 -0400)]
Change order of -m option to update-ref.

The actual position doesn't matter but most people prefer to see
options appear before the arguments.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoCorrect force_write bug in refs.c
Shawn Pearce [Fri, 19 May 2006 09:15:28 +0000 (05:15 -0400)]
Correct force_write bug in refs.c

My earlier attempt at forcing a write for non-existant refs worked;
it forced a write for pretty much all refs.  This corrects the
condition to only force a write for refs which don't exist yet.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoChange 'master@noon' syntax to 'master@{noon}'.
Shawn Pearce [Fri, 19 May 2006 07:29:43 +0000 (03:29 -0400)]
Change 'master@noon' syntax to 'master@{noon}'.

Its ambiguous to parse "master@2006-05-17 18:30:foo" when foo is
meant as a file name and ":30" is meant as 30 minutes past 6 pm.
Therefore all date specifications in a sha1 expression must now
appear within brackets and the ':' splitter used for the path name
in a sha1 expression ignores ':' appearing within brackets.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoLog ref updates made by fetch.
Shawn Pearce [Fri, 19 May 2006 07:29:26 +0000 (03:29 -0400)]
Log ref updates made by fetch.

If a ref is changed by http-fetch, local-fetch or ssh-fetch
record the change and the remote URL/name in the log for the ref.
This requires loading the config file to check logAllRefUpdates.

Also fixed a bug in the ref lock generation; the log file name was
not being produced right due to a bad prefix length.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoForce writing ref if it doesn't exist.
Shawn Pearce [Fri, 19 May 2006 07:29:05 +0000 (03:29 -0400)]
Force writing ref if it doesn't exist.

Normally we try to skip writing a ref if its value hasn't changed
but in the special case that the ref doesn't exist but the new
value is going to be 0{40} then force writing the ref anyway.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAdded logs/ directory to repository layout.
Shawn Pearce [Fri, 19 May 2006 07:28:46 +0000 (03:28 -0400)]
Added logs/ directory to repository layout.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoGeneral ref log reading improvements.
Shawn Pearce [Fri, 19 May 2006 07:28:19 +0000 (03:28 -0400)]
General ref log reading improvements.

Corrected the log starting time displayed in the error message
(as it was always showing the epoch due to a bad input to strtoul).

Improved the log parser so we only scan backwards towards the
'\n' from the end of the prior log; during this scan the last '>'
is remembered to improve performance (rather than scanning forward
to it).

If the log record matched is the last log record in the file only
use its new sha1 value if the date matches exactly; otherwise we
leave the passed in sha1 alone as it already contains the current
value of the ref.  This way lookups of dates later than the log
end to stick with the current ref value in case the ref was updated
without logging.

If it looks like someone changed the ref without logging it and we
are going to return the sha1 which should have been valid during
the missing period then warn the user that there might be log data
missing and thus their query result may not be accurate.  The check
isn't perfect as its just based on comparing the old and new sha1
values between the two log records but its better than not checking
at all.

Implemented test cases based on git-rev-parse for most of the
boundary conditions.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge early part of 'sp/reflog' branch
Junio C Hamano [Fri, 19 May 2006 16:10:38 +0000 (09:10 -0700)]
Merge early part of 'sp/reflog' branch

18 years agoImplement a --dry-run option to git-quiltimport
Eric W. Biederman [Wed, 17 May 2006 20:10:25 +0000 (14:10 -0600)]
Implement a --dry-run option to git-quiltimport

Since large quilt trees like -mm can easily have patches
without clear authorship information, add a --dry-run
option to make the problem patches easy to find.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoImplement git-quiltimport
Eric W. Biederman [Wed, 17 May 2006 18:44:40 +0000 (12:44 -0600)]
Implement git-quiltimport

Importing a quilt patch series into git is not very difficult
but parsing the patch descriptions and all of the other
minutia take a bit of effort to get right, so this automates it.

Since git and quilt complement each other it makes sense
to make it easy to go back and forth between the two.

If a patch is encountered that it cannot derive the author
from the user is asked.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoDocumentation/Makefile: create tarballs for the man pages and html files
Tilman Sauerbeck [Thu, 18 May 2006 10:57:04 +0000 (12:57 +0200)]
Documentation/Makefile: create tarballs for the man pages and html files

[jc: rewrote by stealing from what I run to update html and
 man branches automatically]

Signed-off-by: Tilman Sauerbeck <tilman@code-monkey.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoSubmittingPatches: The download location of External Editor has moved
Lukas Sandström [Thu, 18 May 2006 12:23:59 +0000 (14:23 +0200)]
SubmittingPatches: The download location of External Editor has moved

Signed-off-by: Lukas Sandström <lukass@etek.chalmers.se>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMake git-check-format-ref a builtin.
Lukas Sandström [Thu, 18 May 2006 12:15:55 +0000 (14:15 +0200)]
Make git-check-format-ref a builtin.

Signed-off-by: Lukas Sandström <lukass@etek.chalmers.se>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMake "git rev-list" be a builtin
Linus Torvalds [Thu, 18 May 2006 21:19:20 +0000 (14:19 -0700)]
Make "git rev-list" be a builtin

This was surprisingly easy. The diff is truly minimal: rename "main()" to
"cmd_rev_list()" in rev-list.c, and rename the whole file to reflect its
new built-in status.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-diff: do not say files are renamed when blob and file are given
Junio C Hamano [Thu, 18 May 2006 21:35:37 +0000 (14:35 -0700)]
builtin-diff: do not say files are renamed when blob and file are given

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoPrevent bogus paths from being added to the index.
Linus Torvalds [Thu, 18 May 2006 19:07:31 +0000 (12:07 -0700)]
Prevent bogus paths from being added to the index.

With this one, it's now a fatal error to try to add a pathname
that cannot be added with "git add", i.e.

[torvalds@g5 git]$ git add .git/config
fatal: unable to add .git/config to index

and

[torvalds@g5 git]$ git add foo/../bar
fatal: unable to add foo/../bar to index

instead of the old "Ignoring path xyz" warning that would end up
silently succeeding on any other paths.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-add: fix unmatched pathspec warnings.
Junio C Hamano [Thu, 18 May 2006 08:29:36 +0000 (01:29 -0700)]
builtin-add: fix unmatched pathspec warnings.

"git add Documentation/" when Documentation directory exists
does not barf (as it should not), but "git add ." barfed when it
did not add anything.  This was because we checked for the path
prefix ("Documentation/" in the former case, and an empty string
in the latter case) for existence, and lstat("", &st) would say
"Huh?".

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoProvide a way to flush git-diff-tree's output
Paul Mackerras [Thu, 18 May 2006 06:58:51 +0000 (16:58 +1000)]
Provide a way to flush git-diff-tree's output

Gitk wants to use git-diff-tree as a filter to tell it which ids from
a given list affect a set of files or directories.  We don't want to
fork and exec a new git-diff-tree process for each batch of ids, since
there could be a lot of relatively small batches.  For example, a
batch could contain as many ids as fit in gitk's headline display
window, i.e. 20 or so, and we would be processing a new batch every
time the user scrolls that window.

The --stdin flag to git-diff-tree is suitable for this, but the main
difficulty is that the output of git-diff-tree gets buffered and
doesn't get sent until the buffer is full.

This provides a way to get git-diff-tree to flush its output buffers.
If a blank line is supplied on git-diff-tree's standard input, it will
flush its output buffers and then accept further input.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'maint'
Junio C Hamano [Thu, 18 May 2006 05:33:06 +0000 (22:33 -0700)]
Merge branch 'maint'

* maint:
  git-cvsimport: Handle "Removed" from pserver

18 years agogit-cvsimport: Handle "Removed" from pserver maint
Elrond [Wed, 10 May 2006 17:37:04 +0000 (19:37 +0200)]
git-cvsimport: Handle "Removed" from pserver

Sometimes the pserver says "Removed" instead of "Remove-entry".

Signed-off-by: Elrond <elrond+kernel.org@samba-tng.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoRemove old "git-add.sh" remnants
Linus Torvalds [Thu, 18 May 2006 04:21:04 +0000 (21:21 -0700)]
Remove old "git-add.sh" remnants

Repeat after me: "It's now a built-in"

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix ref log parsing so it works properly.
Shawn Pearce [Wed, 17 May 2006 22:34:48 +0000 (18:34 -0400)]
Fix ref log parsing so it works properly.

The log parser was only ever matching the last log record due to
calling strtoul on "> 1136091609" rather than " 1136091609".  Also
once a match for '@' has been found after the name of the ref there
is no point in looking for another '@' within the remaining text.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoSupport 'master@2 hours ago' syntax
Shawn Pearce [Wed, 17 May 2006 09:56:09 +0000 (05:56 -0400)]
Support 'master@2 hours ago' syntax

Extended sha1 expressions may now include date specifications
which indicate a point in time within the local repository's
history.  If the ref indicated to the left of '@' has a log in
$GIT_DIR/logs/<ref> then the value of the ref at the time indicated
by the specification is obtained from the ref's log.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoLog ref updates to logs/refs/<ref>
Shawn Pearce [Wed, 17 May 2006 09:55:40 +0000 (05:55 -0400)]
Log ref updates to logs/refs/<ref>

If config parameter core.logAllRefUpdates is true or the log
file already exists then append a line to ".git/logs/refs/<ref>"
whenever git-update-ref <ref> is executed.  Each log line contains
the following information:

  oldsha1 <SP> newsha1 <SP> committer <LF>

where committer is the current user, date, time and timezone in
the standard GIT ident format.  If the caller is unable to append
to the log file then git-update-ref will fail without updating <ref>.

An optional message may be included in the log line with the -m flag.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoConvert update-ref to use ref_lock API.
Shawn Pearce [Wed, 17 May 2006 09:55:19 +0000 (05:55 -0400)]
Convert update-ref to use ref_lock API.

This conversion also adds the '-m' switch to update-ref allowing
the caller to record why the ref is changing.  At present this is
merely copied down into the ref_lock API.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoImprove abstraction of ref lock/write.
Shawn Pearce [Wed, 17 May 2006 09:55:02 +0000 (05:55 -0400)]
Improve abstraction of ref lock/write.

Created 'struct ref_lock' to contain the data necessary to perform
a ref update.  This change improves writing a ref as the file names
are generated only once (rather than twice) and supports following
symrefs (up to the maximum depth).  Further the ref_lock structure
provides room to extend the update API with ref logging.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoRemove unnecessary local in get_ref_sha1.
Shawn Pearce [Wed, 17 May 2006 09:54:46 +0000 (05:54 -0400)]
Remove unnecessary local in get_ref_sha1.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoapply --cached: do not check newly added file in the working tree
Junio C Hamano [Wed, 17 May 2006 23:56:13 +0000 (16:56 -0700)]
apply --cached: do not check newly added file in the working tree

The --cached mode does not deal with the working tree, so we
should not check it with lstat.  An earlier code omitted the
call to lstat but forgot to omit the check for the errno.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-grep: workaround for non GNU grep.
Linus Torvalds [Wed, 17 May 2006 18:12:22 +0000 (11:12 -0700)]
builtin-grep: workaround for non GNU grep.

Of course, it still ignores the fact that not all grep's support some of
the flags like -F/-L/-A/-C etc, but for those cases, the external grep
itself will happily just say "unrecognized option -F" or similar.

So with this change, "git grep" should handle all the flags the native
grep handles, which is really quite fine. We don't _need_ to expose
anything more, and if you do want our extensions, you can get them with
"--uncached" and an up-to-date index.

No configuration necessary, and we automatically take advantage of any
native grep we have, if possible.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'ew/pretty-fmt'
Junio C Hamano [Wed, 17 May 2006 22:49:23 +0000 (15:49 -0700)]
Merge branch 'ew/pretty-fmt'

* ew/pretty-fmt:
  commit: allow --pretty= args to be abbreviated

18 years agoMerge branch 'jc/apply'
Junio C Hamano [Wed, 17 May 2006 22:47:33 +0000 (15:47 -0700)]
Merge branch 'jc/apply'

* jc/apply:
  git-am: use apply --cached
  apply --cached: apply a patch without using working tree.
  apply --numstat: show new name, not old name.

18 years agobuiltin-add: warn on unmatched pathspecs
Linus Torvalds [Wed, 17 May 2006 20:23:19 +0000 (13:23 -0700)]
builtin-add: warn on unmatched pathspecs

This is in the same spirit as what bba319b5 and 45e48120 tried
to do to help users.  A command such as "git add Documentaiton"
with misspelled pathspecs would give a friendly reminder with
this.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoDo "git add" as a builtin
Linus Torvalds [Wed, 17 May 2006 16:33:32 +0000 (09:33 -0700)]
Do "git add" as a builtin

First try. Let's see how well this works.

In many ways, the hard parts of "git commit" are not so different from
this, and a builtin commit would share a lot of the code, I think.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoRevert "builtin-grep: workaround for non GNU grep."
Junio C Hamano [Wed, 17 May 2006 18:42:14 +0000 (11:42 -0700)]
Revert "builtin-grep: workaround for non GNU grep."

This reverts 518920b764ee9150781e68217181b24d0712748e commit.
Linus has a more portable alternative.

18 years agocommit: allow --pretty= args to be abbreviated
Eric Wong [Mon, 15 May 2006 00:20:46 +0000 (17:20 -0700)]
commit: allow --pretty= args to be abbreviated

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-grep: workaround for non GNU grep.
Junio C Hamano [Wed, 17 May 2006 09:48:13 +0000 (02:48 -0700)]
builtin-grep: workaround for non GNU grep.

Some implementations do not know what to do with -H; define
NO_H_OPTION_IN_GREP when you build git if your grep lacks -H.

Most of the time, it can be worked around by prepending
/dev/null to the argument list, but that causes -L and -c to
slightly misbehave (they both expose /dev/null is given), so
when these options are given, do not run external grep that does
not understand -H.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoClean up git-ls-file directory walking library interface
Linus Torvalds [Wed, 17 May 2006 02:46:16 +0000 (19:46 -0700)]
Clean up git-ls-file directory walking library interface

This moves the code to add the per-directory ignore files for the base
directory into the library routine.

That not only allows us to turn the function push_exclude_per_directory()
static again, it also simplifies the library interface a lot (the caller
no longer needs to worry about any of the per-directory exclude files at
all).

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>