so that an ugly commit message like this can be
handled sanely.

Currently, --pretty=oneline and --pretty=email (hence
format-patch) take and use only the first line of the commit log
message.  This changes them to:

 - Take the first paragraph, where the definition of the first
   paragraph is "skip all blank lines from the beginning, and
   then grab everything up to the next empty line".

 - Replace all line breaks with a whitespace.

This change would not affect a well-behaved commit message that
adheres to the convention of "single line summary, a blank line,
and then body of message", as its first paragraph always
consists of a single line.  Commit messages from different
culture, such as the ones imported from CVS/SVN, can however get
chomped with the existing behaviour at the first linebreak in
the middle of sentence right now, which would become much easier
to see with this change.

The Subject: and --pretty=oneline output would become very long
and unsightly for non-conforming commits, but their messages are
already ugly anyway, and thischange at least avoids the loss of
information.

The Subject: line from a multi-line paragraph is folded using
RFC2822 line folding rules at the places where line breaks were
in the original.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Traditionally we had 16kB limit when formatting log messages for
output, because it was easier to arrange for the caller to have
a reasonably big buffer and pass it down without ever worrying
about reallocating.

This changes the calling convention of pretty_print_commit() to
lift this limit.  Instead of the buffer and remaining length, it
now takes a pointer to the pointer that points at the allocated
buffer, and another pointer to the location that stores the
allocated length, and reallocates the buffer as necessary.

To support the user format, the error return of interpolate()
needed to be changed.  It used to return a bool telling "Ok the
result fits", or "Sorry, I had to truncate it".  Now it returns
0 on success, and returns the size of the buffer it wants in
order to fit the whole result.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* 'jc/blame' (early part):
  git-blame -w: ignore whitespace
  git-blame: do not indent with spaces.

This is in preparation for keeping two entry lists in the
dir object.

This patch adds and uses the ALLOC_GROW() macro, which
implements the commonly used idiom of growing a dynamic
array using the alloc_nr function (not just in dir.c, but
everywhere).

We also move creation of a dir_entry to dir_entry_new.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

'git-remote show remote-name' lists the refs that are pushed to the remote
by showing the 'Push' line from the config file. But before showing it,
it shortened 'refs/heads/here:refs/heads/there' to 'here:there'. However,
if the Push line is prefixed with a plus, the ref was not shortened.
Signed-off-by: NJohannes Sixt <johannes.sixt@telecom.at>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

/.git or .git is removed from the project name and the
basename of the remaining path is used as the beginning of
the filename and as the directory in the archive.

The regexp will actually not strip off /.git or .git if there
wouldn't be anything left after removing it.

Currently the full project name is used as directory in the
archive and the basename is used as filename.  For example a
repository named foo/bar/.git will have a archive named
.git-<version>.* and extract to foo/bar/.git.  With this patch
the file is named bar-<version>.* and extracts to bar.
Signed-off-by: NMatthias Lederhofer <matled@gmx.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

There are only a dozen or so uses of strdup in all of git.
Of those, most seem ok, but this one isn't:
Signed-off-by: NJim Meyering <jim@meyering.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* maint:

* aw/cvs:
  cvsimport: add <remote>/HEAD reference in separate remotes more
  cvsimport: update documentation to include separate remotes option
  cvsimport: add support for new style remote layout

* ep/cvstag:
  Use git-tag in git-cvsimport

* ar/clone:
  Fix clone to setup the origin if its name ends with .git

* sv/objfixes:
  Don't assume tree entries that are not dirs are blobs

Earlier, a second "-C" on the command line had no effect.
But "--find-copies-harder" is so long to type, let's make doubled -C
enable that option.  It is in line with how "git blame" handles such
doubled options to mean "work harder".
Signed-off-by: NJohannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

All the other directory location variables do not have the trailing
slash.
Signed-off-by: NJohannes Sixt <johannes.sixt@telecom.at>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Recent git-gui has the ability to determine the location of its library
files relative to the --exec-dir. Its Makefile enables this capability
depending on the install paths that are specified. However, without this
fix there is an extra slash in a path specification, so that the Makefile
does not recognize the equivalence of two paths that it compares.

A side-effect is that all "standard" builds (which do not set $(sharedir)
explicitly) now exploit above mentioned gut-gui feature.

Another side-effect is that an ugly compiled-in double-slash in
$(template_dir) is avoided.
Signed-off-by: NJohannes Sixt <johannes.sixt@telecom.at>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* lh/submodule:
  git-submodule: clone during update, not during init
  git-submodule: move cloning into a separate function

* maint:
  Unquote From line from patch before comparing with given from address.
  git-cherry: Document 'limit' command-line option

* 'master' of git://repo.or.cz/git-gui:
  git-gui: Save geometry before the window layout is damaged
  git-gui: Give amend precedence to HEAD over MERGE_MSG
  git-gui: Include 'war on whitespace' fixes from git.git

This makes --suppress-from actually work when you're unfortunate enough
to have non-ASCII in your name.  Also, if there's a match use the optionally
RFC2047 quoted version from the email.
Signed-off-by: NKristian Høgsberg <krh@redhat.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NLuiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* 'maint' of git://repo.or.cz/git-gui:
  git-gui: Save geometry before the window layout is damaged
  git-gui: Give amend precedence to HEAD over MERGE_MSG
  git-gui: Include 'war on whitespace' fixes from git.git

* maint:
  git-gui: Save geometry before the window layout is damaged
  git-gui: Give amend precedence to HEAD over MERGE_MSG

Because Tk does not assure us the order that it will process
children in before it destroys the main toplevel we cannot safely
save our geometry data during a "bind . <Destroy>" event binding.
The geometry may have already changed as a result of a one or
more children being removed from the layout.  This was pointed
out in gitk by Mark Levedahl, and patched over there by commit
b6047c5a.

So we now also use "wm protocol . WM_DELETE_WINDOW" to detect when
the window is closed by the user, and forward that close event to
our main do_quit routine.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Apparently git-commit.sh (the command line commit user interface in
core Git) always gives precedence to the prior commit's message if
`commit --amend` is used and a $GIT_DIR/MERGE_MSG file also exists.

We actually were doing the same here in git-gui, but the amended
message got lost if $GIT_DIR/MERGE_MSG already existed because
we started a rescan immediately after loading the prior commit's
body into the edit buffer.  When that happened the rescan found
MERGE_MSG existed and replaced the commit message buffer with the
contents of that file.  This meant the user never saw us pick up
the commit message of the prior commit we are about to replace.

Johannes Sixt <J.Sixt@eudaptics.com> found this bug in git-gui by
running `git cherry-pick -n $someid` and then trying to amend the
prior commit in git-gui, thus combining the contents of $someid
with the contents of HEAD, and reusing the commit message of HEAD,
not $someid.  With the recent changes to make cherry-pick use the
$GIT_DIR/MERGE_MSG file Johannes saw git-gui pick up the message
of $someid, not HEAD.  Now we always use HEAD if we are amending.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

* maint:
  git-gui: Include 'war on whitespace' fixes from git.git

Earlier git.git applied a large "war on whitespace" patch that was
created using 'apply --whitespace=strip'.  Unfortunately a few of
git-gui's own files got caught in the mix and were also cleaned up.
That was a6080a0a.

This patch is needed in git-gui.git to reapply those exact same
changes here, otherwise our version generator script is unable to
obtain our version number from git-describe when we are hosted in
the git.git repository.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

* 'master' of git://repo.or.cz/git-gui:
  git-gui: Internalize symbolic-ref HEAD reading logic
  git-gui: Expose the merge.diffstat configuration option
  git-gui: Allow users to delete remote branches
  git-gui: Allow users to rename branches through 'branch -m'
  git-gui: Disable tearoff menus on Windows, Mac OS X
  git-gui: Provide fatal error if library is unavailable
  git-gui: Enable verbose Tcl loading earlier
  git-gui: Show the git-gui library path in 'About git-gui'
  git-gui: GUI support for running 'git remote prune <name>'
  git gui 0.8.0

* maint: (46 commits)
  git-gui: Changed blame header bar background to match main window
  git-gui: Favor the original annotations over the recent ones
  git-gui: Improve our labeling of blame annotation types
  git-gui: Use three colors for the blame viewer background
  git-gui: Jump to original line in blame viewer
  git-gui: Display both commits in our tooltips
  git-gui: Run blame twice on the same file and display both outputs
  git-gui: Display the "Loading annotation..." message in italic
  git-gui: Rename fields in blame viewer to better descriptions
  git-gui: Label the uncommitted blame history entry
  git-gui: Switch internal blame structure to Tcl lists
  git-gui: Cleanup redundant column management in blame viewer
  git-gui: Better document our blame variables
  git-gui: Remove unused commit_list from blame viewer
  git-gui: Automatically expand the line number column as needed
  git-gui: Make the line number column slightly wider in blame
  git-gui: Use lighter colors in blame view
  git-gui: Remove unnecessary space between columns in blame viewer
  git-gui: Remove the loaded column from the blame viewer
  git-gui: Clip the commit summaries in the blame history menu
  ...

* 'maint' of git://repo.or.cz/git-gui: (46 commits)
  git-gui: Changed blame header bar background to match main window
  git-gui: Favor the original annotations over the recent ones
  git-gui: Improve our labeling of blame annotation types
  git-gui: Use three colors for the blame viewer background
  git-gui: Jump to original line in blame viewer
  git-gui: Display both commits in our tooltips
  git-gui: Run blame twice on the same file and display both outputs
  git-gui: Display the "Loading annotation..." message in italic
  git-gui: Rename fields in blame viewer to better descriptions
  git-gui: Label the uncommitted blame history entry
  git-gui: Switch internal blame structure to Tcl lists
  git-gui: Cleanup redundant column management in blame viewer
  git-gui: Better document our blame variables
  git-gui: Remove unused commit_list from blame viewer
  git-gui: Automatically expand the line number column as needed
  git-gui: Make the line number column slightly wider in blame
  git-gui: Use lighter colors in blame view
  git-gui: Remove unnecessary space between columns in blame viewer
  git-gui: Remove the loaded column from the blame viewer
  git-gui: Clip the commit summaries in the blame history menu
  ...

* maint: (38 commits)
  git-gui: Changed blame header bar background to match main window
  git-gui: Favor the original annotations over the recent ones
  git-gui: Improve our labeling of blame annotation types
  git-gui: Use three colors for the blame viewer background
  git-gui: Jump to original line in blame viewer
  git-gui: Display both commits in our tooltips
  git-gui: Run blame twice on the same file and display both outputs
  git-gui: Display the "Loading annotation..." message in italic
  git-gui: Rename fields in blame viewer to better descriptions
  git-gui: Label the uncommitted blame history entry
  git-gui: Switch internal blame structure to Tcl lists
  git-gui: Cleanup redundant column management in blame viewer
  git-gui: Better document our blame variables
  git-gui: Remove unused commit_list from blame viewer
  git-gui: Automatically expand the line number column as needed
  git-gui: Make the line number column slightly wider in blame
  git-gui: Use lighter colors in blame view
  git-gui: Remove unnecessary space between columns in blame viewer
  git-gui: Remove the loaded column from the blame viewer
  git-gui: Clip the commit summaries in the blame history menu
  ...

Allow choosing between '-c' (combined diff) and '--cc' (compact
combined) diff format in 'commitdiff' view for merge (multiparent)
commits.  Default is now '--cc'.

In the bottom part of navigation bar there is link allowing to change
diff format: "combined" for '-c' (when using '--cc') and "compact" for
'--cc' (when using '-c'), just on the right of "raw" link to
'commitdiff_plain" view.

About patchset part of diff --cc output: the difftree (whatchanged
table) has "patch" links to anchors to individual patches (on the same
page). The --cc option further compresses the patch output by
omitting some hunks; when this optimization makes all hunks disappear,
the patch is not shown (like in any other "empty diff" case). But the
fact that patch has been simplified out is not reflected in the raw
(difftree) part of diff output; the raw part is the same for '-c' and
'--cc' options. As correcting difftree is rather out of the question,
as it would require scanning patchset part before writing out
difftree, we add "Simple merge" empty diffs as a place to have anchor
to in place of those simplified out and removed patches.
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Add links to diff to file ('blobdiff' view) for each of individual
versions of the file in a merge commit to the from-file/to-file header
in the patch part of combined 'commitdiff' view for merges.

The from-file/to-file header for combined diff now looks like:

  --- _1_/_git-gui/git-gui.sh_
  --- _2_/_git-gui.sh_
  +++ b/_git-gui/git-gui.sh_

where _<filename>_ link is link to appropriate version of a file
('blob' view), and _<n>_ is link to respective diff to mentioned
version of a file ('blobdiff' view). There is even hint provided in
the form of title attribute.
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Instead of using default, diff(1) like from-file/to-file header for
combined diff (for a merge commit), which looks like:

  --- a/git-gui/git-gui.sh
  +++ b/_git-gui/git-gui.sh_

(where _link_ denotes [hidden] hyperlink), create from-file(n)/to-file
header, using "--- <n>/_<filename>_" for each of parents, e.g.:

  --- 1/_git-gui/git-gui.sh_
  --- 2/_git-gui.sh_
  +++ b/_git-gui/git-gui.sh_

Test it on one of merge commits involving rename, e.g.
  95f97567c1887d77f3a46b42d8622c76414d964d (rename at top)
  5bac4a67 (file from one branch)

This is mainly meant to easier see renames in a merge commit.
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This commit makes git_patchset_body easier to read, and reduces level of
nesting and indent level. It adds more lines that it removes because of
extra parameter passing in subroutines, and subroutine calls in
git_patchset_body. Also because there are few added comments.

Below there are descriptions of all split-off subroutines:

Separate formatting "git diff" header into format_git_diff_header_line.
While at it fix it so it always escapes pathname. It would be even more
useful if we decide to use `--cc' for merges, and need to generate by
hand empty patches for anchors.

Separate formatting extended (git) diff header lines into
format_extended_diff_header_line. This one is copied without changes.

Separate formatting two-lines from-file/to-file diff header into
format_diff_from_to_header subroutine. While at it fix it so it always
escapes pathname. Beware calling convention: it takes _two_ lines.

Separate generating %from and %to hashes (with info used among others to
generate hyperlinks) into parse_from_to_diffinfo subroutine. This one is
copied without changes.

Separate checking if file was deleted (and among others therefore does
not have link to the result file) into is_deleted subroutine. This would
allow us to easily change the algotithm to find if file is_deleted in
the result.
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Check if 'hp' (hash_parent) parameter to 'commitdiff' view is one of
'h' (hash) commit parents, i.e. if commitdiff is of the form
"<commit>^<n> <commit>", and mark it as such in the bottom part of
navigation bar. The "next" link in commitdiff view was introduced
in commit 151602df

If 'hb' is n-th parent of 'h', show the following at the bottom
of navigation bar:
  (from parent n: _commit_)
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Since commit-fb1dde4a we show combined diff for merges in 'commitdiff'
view, and since commit-208ecb2e also in 'commit' view. Sometimes
though one would want to see diff to one of merge commit parents. It
is easy in 'commit' view: in the commit header part there are "diff"
links for each of parent header. This commit adds such links also for
'commitdiff' view.

Add to difftree / whatchanged table row with "1", "2", ... links to
'commitdiff' view for diff with n-th parent for merge commits, as a
table header.  This is visible only in 'comitdiff' view, and only for
a merge commit (comit with more than one parent).

To save space links are shown as "n", where "n" is number of a parent,
and not as for example shortened (to 7 characters) sha1 of a parent
commit.  To make it easier to discover what links is for, each link
has 'title' attribute explaining the link.

Note that one would need to remember that difftree table in 'commit'
view has one less column (it doesn't have "patch" link column), if one
would want to add such table header also in 'commit' view.

Example output:
                          1       2       3
  Makefile      patch | diff1 | diff2 | diff3 | blob | history
  cache.h       patch | diff1 | diff2 | diff3 | blob | history
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* maint:
  tutorial: use "project history" instead of "changelog" in header
  Documentation: user-manual todo
  user-manual: add a missing section ID
  Fix typo in remote branch example in git user manual
  user-manual: quick-start updates

It lacked a description for the (historically) optional tagger header line.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In the kernel we have a rule for *.c -> *.s files exactly because
it's nice to be able to easily say "ok, what does that generate".

Here's a patch to add such a rule to git too, in case anybody is
interested. It makes it much simpler to just do

	make sha1_file.s

and look at the compiler-generated output that way, rather than having to
fire up gdb on the resulting binary.

(Add -fverbose-asm or something if you want to, it can make the result
even more readable)

[jc: add *.s to .gitignore]
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The word "changelog" seems a little too much like jargon to me, and beginners
must understand section headers so they know where to look for help.
Signed-off-by: N"J. Bruce Fields" <bfields@citi.umich.edu>