Give "shortlog_len" parameter to the fmt_merge_msg(), remove its
"merge_summary" parameter, and remove fmt_merge_msg_shortlog() function.
In the updated API, shortlog_len == 0 means no shortlog is given.

The parameter "merge_title" controls if the title of the merge commit is
autogenerated (it reads something like "Merge branch ..."), and typically
it is set to true when the caller does not give its own message.
Signed-off-by: NRamkumar Ramachandra <artagnon@gmail.com>
Mentored-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Port v1.7.0-rc0~83^2 (Teach --[no-]rerere-autoupdate option to
merge, revert and friends, 2009-12-04) to the example merge script.

After this change, all tests pass for me with the scripted
merge.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Port v1.6.2-rc1~10^2 (Teach @{-1} to git merge, 2009-02-13) to
the old merge script.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In a --no-ff merge with conflicts, "git commit" used to forget the
--no-ff when used to complete the merge.  That was fixed by
v1.6.1-rc1~134^2 (builtin-commit: use reduce_heads() only when
appropriate, 2008-10-03) for the builtin merge.  Port the change to
the merge script in contrib/examples.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Port v1.6.6-rc0~62^2 (Teach 'git merge' and 'git pull' the option
--ff-only, 2009-10-29) to the old merge script.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Based on v1.6.0-rc0~51^2~5 (Build in merge, 2008-07-07).
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Based on v1.7.1.1~23^2 (merge: --log appends shortlog to message if
specified, 2010-05-11).  Without this change, the scripted
(non-builtin) merge does not pass t7604.

Cc: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Some git-merge-* commands are not merge strategies.  This is based on
v1.6.1-rc1~294^2~7 (builtin-merge: allow using a custom strategy,
2008-07-30) but it is less smart: we just use a hard-coded list of
forbidden strategy names.  It is okay if this falls out of date, since
the code is just an example.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The idea comes from v1.6.1-rc1~294^2~7 (builtin-merge: allow using a
custom strategy, 2008-07-30).
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Without this support, the scripted merge cannot pass t6037.

Based on v1.7.0-rc0~55^2~5 (git merge -X<option>, 2009-11-25).
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

 - point out remote-tracking branches as "remote branch
   'upstream/master'";
 - avoid misleading log messages when a tag and branch
   share a name.

This approximates the builtin merge command's behavior well
enough to pass the relevant tests.

Based roughly on v1.6.4.2~10^2 (merge: indicate remote tracking
branches in merge message, 2009-08-09) and v1.6.4.2~10^2~1 (merge: fix
incorrect merge message for ambiguous tag/branch, 2009-08-09).

Cc: Jeff King <peff@peff.net>
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Check MERGE_HEAD and bail out if it exists.  Based on v1.6.3.3~3^2
(refuse to merge during a merge, 2009-06-01).  Without this change,
the scripted merge does not pass t3030.

Cc: Clemens Buchacher <drizzd@aon.at>
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Before:

	You are in the middle of a conflicted merge.

After:

	Merge is not possible because you have unmerged files.

I prefer the old message, but the new one is more consistent with
other commands and tests expect it.  In particular, without this
change the scripted merge does not pass t3030.

Based on v1.7.0-rc0~66^2 (Be more user-friendly when refusing to do
something because of conflict., 2010-01-12).

Cc: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Port v1.6.1-rc1~319 (provide more errors for the "merge into empty
head" case, 2008-08-21) to the example merge script.

Noticed by comparison with builtin merge.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Since v1.7.1.1~23^2 (merge: --log appends shortlog to message if
specified, 2010-05-11), the fmt-merge-msg backend supports custom text
to override the merge title "Merge <foo> into <bar>".

Expose this functionality for scripted callers.  Example:

 git fmt-merge-msg --log -m \
	"$(printf '%s\n' \
	    "Merge branch 'api-cleanup' into feature" \
	    '' \
	    'This is to use a few functions refactored for this purpose.'
	)" <.git/FETCH_HEAD

Cc: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

While show-branch --independent does not support more than MAX_REVS
revs, git internally supports more with a different algorithm.
Expose that functionality as "git merge-base --independent".

This should help scripts to catch up with builtin merge in supporting
dodecapus.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

While show-branch --merge-base does not support more than MAX_REVS
revs, git supports more with a different algorithm
(v1.6.0-rc0~51^2~13, Introduce get_octopus_merge_bases() in commit.c,
2008-06-27).  Expose that functionality.

This should help scripts to catch up with builtin merge in supporting
dodecapus.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

For example, a person reading the merge-base man page might wonder
about the fastest way to check if one commit is an ancestor of
another (which would require rev-list).
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Guard setup with test_expect_success.  Use test_might_fail
instead of ignoring the exit code from git config --unset.
Point out setup commands that are shared by multiple tests,
to make it easy to write GIT_SKIP_TESTS specifications that
work.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Guard setup with test_expect_success, put the opening quote
starting each test on the same line as the test_expect_* invocation,
and combine related actions into single tests.

While at it:

 - use test_cmp instead of expr or test $foo = $bar, for more helpful
   output with -v when tests fail;

 - use test_commit for brevity.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Some people like to "git fetch origin && merge origin/master" from
the unborn branch provided when first initializing a repository.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The details of the reflog message are not important, but
including something sane in the reflog is.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Probably as a development aid, this test script runs gitk --all
to allow the driver to inspect history between tests when run
with --debug.  As a result, running all tests with --debug
requires closing a long series of gitk displays, one at a time.

Use git log --graph --oneline instead.  This way, the history is
available for viewing with "git show" but the test script finishes
without interaction.

Longer term, it would be nice to have an option to run a
user-specified command between tests.  This patch does not do
that.

Cc: Lars Hjemli <hjemli@gmail.com>
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Guard setup commands with test_expect_success, so they are easier
to visually skip over and get to the good part.  While at it:

 - use "printf '%s\n' a b ..." instead of "cat <<EOF" for test
   vectors with short lines;

 - use test_cmp instead of test foo = bar where possible, for
   better output with -v on failure;

 - do not go to extraordinary lengths to print a relevant message
   when test commands fail.  There is a patch in flight that could be
   used to restore the nice error messages in a cleaner way.

Cc: Lars Hjemli <hjemli@gmail.com>
Cc: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

If no branch 'foo' exists but a tag 'foo' does, then
git merge foo^ results in

	Merge branch 'foo' (early part)

as a commit message, because the relevant code path checks that
refs/heads/foo is a valid refname for writing rather than for
reading.
Signed-off-by: NJonathan Nieder <jrnieder@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

These two tests weren't about how "git reflog show <branch>" exits when
there is no reflog, but were about "checkout" and "branch" create or not
create reflog when creating a new <branch>.  Update the tests to check
what we are interested in, using "git rev-parse --verify".

Also lose tests based on "test -f .git/logs/refs/heads/<branch>" from
nearby, to avoid exposing this particular implementation detail
unnecessarily.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

See e2007832Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The correct advice should have been taken from c289c315 (t/t7006: ignore
return status of shell's unset builtin, 2010-06-02).  A real-life issue
we experienced was with "unset", not with "export" (exporting an
unset variable may have similar portability issues, though).
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The '!()' notation is interpreted as a pattern-list on Ksh.  The Ksh man
page describe it as follows:

   !(pattern-list)
      Matches anything except one of the given patterns.

Ksh performs a file glob using the pattern-list and then tries to execute
the first file in the list.  If a space is added between the '!' and the
open parens, then Ksh will not interpret it as a pattern list, but in this
case, it is preferred to use test_must_fail, so lets do so.
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

These two lines use the negation '!' operator to negate the result of a
simple command.  Since these commands do not contain any pipes or other
complexities, the test_must_fail function can be used and is preferred
since it will additionally detect termination due to a signal.

This was noticed because the second use of '!' does not include a space
between the '!' and the opening parens.  Ksh interprets this as follows:

   !(pattern-list)
      Matches anything except one of the given patterns.

Ksh performs a file glob using the pattern-list and then tries to execute
the first file in the list.  If a space is added between the '!' and the
open parens, then Ksh will not interpret it as a pattern list, but in this
case, it is preferred to use test_must_fail, so lets do so.
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* maint:
  t/README: clarify test_must_fail description
  Check size of path buffer before writing into it

Conflicts:
	t/README

Some have found the wording of the description to be somewhat ambiguous
with respect to when it is desirable to use test_must_fail instead of
"! <git-command>".  Tweak the wording somewhat to hopefully clarify that
it is _because_ test_must_fail can detect segmentation fault that it is
desirable to use it instead of "! <git-command>".
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This prevents a buffer overrun that could otherwise be triggered by
creating a file called '.git' with contents

  gitdir: (something really long)
Signed-off-by: NGreg Brockman <gdb@mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* maint:
  update-server-info: Shorten read_pack_info_file()
  Documentation: Explain git-mergetool's use of temporary files

The correct responses to a D and a T line in .git/objects/info/packs
are the same, so combine their case arms.  In both cases we already
‘goto’ out of the switch so while at it, remove a redundant ‘break’
to avoid yet another line of code.
Signed-off-by: NRalf Thielow <ralf.thielow@googlemail.com>
Reviewed-by: Jonathan Nieder <jrnieder <at> gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

All the git add options were listed in the synopsis until the
--ignore-missing option was added. Change that so that the git add
documentation now has the complete listing.
Signed-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

git submodule add no longer implicitly adds with --force. Remove
references to the old functionality in the documentation.
Signed-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: NJens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

To make the behavior of "git submodule add" more consistent with "git add"
ignored submodule paths should not be silently added when they match an
entry in a .gitignore file. To be able to override that default behavior
in the same way as we can do that for "git add", the new option "--force"
is introduced.
Signed-off-by: NJens Lehmann <Jens.Lehmann@web.de>
Acked-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* jl/add-n-ignore-missing:
  git add: Add the "--ignore-missing" option for the dry run