Code

git.git
13 years agoMerge branch 'jk/fast-export-quote-path'
Junio C Hamano [Thu, 11 Aug 2011 18:03:16 +0000 (11:03 -0700)]
Merge branch 'jk/fast-export-quote-path'

* jk/fast-export-quote-path:
  fast-export: quote paths in output

13 years agoMerge branch 'rc/maint-http-wrong-free'
Junio C Hamano [Thu, 11 Aug 2011 18:03:13 +0000 (11:03 -0700)]
Merge branch 'rc/maint-http-wrong-free'

* rc/maint-http-wrong-free:
  Makefile: some changes for http-related flag documentation
  http.c: fix an invalid free()

Conflicts:
Makefile

13 years agoMerge branch 'rs/grep-function-context'
Junio C Hamano [Thu, 11 Aug 2011 18:03:09 +0000 (11:03 -0700)]
Merge branch 'rs/grep-function-context'

* rs/grep-function-context:
  grep: long context options
  grep: add option to show whole function as context

13 years agoMerge branch 'ef/ipv4-connect-error-report'
Junio C Hamano [Thu, 11 Aug 2011 18:03:06 +0000 (11:03 -0700)]
Merge branch 'ef/ipv4-connect-error-report'

* ef/ipv4-connect-error-report:
  connect: only log if all attempts failed (ipv4)

13 years agoam: Document new --exclude=<path> option
maximilian attems [Tue, 9 Aug 2011 15:21:08 +0000 (17:21 +0200)]
am: Document new --exclude=<path> option

The --exclude=<path> option is passed down to apply; document it as such.

Signed-off-by: maximilian attems <max@stro.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: add support for bisecting bare repositories
Jon Seymour [Tue, 9 Aug 2011 02:11:54 +0000 (12:11 +1000)]
bisect: add support for bisecting bare repositories

This enhances the support for bisecting history in bare repositories.

The "git bisect" command no longer needs to be run inside a repository
with a working tree; it defaults to --no-checkout when run in a bare
repository.

Two tests are included to demonstrate this behaviour.

Suggested-by: Junio C Hamano <gitster@pobox.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agohelping smart-http/stateless-rpc fetch race
Junio C Hamano [Fri, 5 Aug 2011 20:54:06 +0000 (13:54 -0700)]
helping smart-http/stateless-rpc fetch race

A request to fetch from a client over smart HTTP protocol is served in
multiple steps. In the first round, the server side shows the set of refs
it has and their values, and the client picks from them and sends "I want
to fetch the history leading to these commits".

When the server tries to respond to this second request, its refs may have
progressed by a push from elsewhere. By design, we do not allow fetching
objects that are not at the tip of an advertised ref, and the server
rejects such a request. The client needs to try again, which is not ideal
especially for a busy server.

Teach upload-pack (which is the workhorse driven by git-daemon and smart
http server interface) that it is OK for a smart-http client to ask for
commits that are not at the tip of any advertised ref, as long as they are
reachable from advertised refs.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoxdiff/xhistogram: drop need for additional variable
Tay Ray Chuan [Mon, 1 Aug 2011 04:20:10 +0000 (12:20 +0800)]
xdiff/xhistogram: drop need for additional variable

Having an additional variable (ptr) instead of changing line(1|2) and
count(1|2) was for debugging purposes.

Signed-off-by: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoxdiff/xhistogram: rely on xdl_trim_ends()
Tay Ray Chuan [Mon, 1 Aug 2011 04:20:09 +0000 (12:20 +0800)]
xdiff/xhistogram: rely on xdl_trim_ends()

Do away with reduce_common_start_end() and use xdf->dstart and xdf->dend
set by xdl_trim_ends() that similarly tells us where the first unmatched
line from the start and end occurs.

Signed-off-by: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoxdiff/xhistogram: rework handling of recursed results
Tay Ray Chuan [Mon, 1 Aug 2011 04:20:08 +0000 (12:20 +0800)]
xdiff/xhistogram: rework handling of recursed results

Previously we were over-complicating matters by trying to combine the
recursed results. Now, terminate immediately if a recursive call failed
and return its result.

Signed-off-by: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoMerge branch 'oa/pull-reflog'
Junio C Hamano [Mon, 8 Aug 2011 19:33:36 +0000 (12:33 -0700)]
Merge branch 'oa/pull-reflog'

* oa/pull-reflog:
  pull: remove extra space from reflog message

Conflicts:
git-pull.sh

13 years agoMerge branch 'ms/reflog-show-is-default'
Junio C Hamano [Mon, 8 Aug 2011 19:33:35 +0000 (12:33 -0700)]
Merge branch 'ms/reflog-show-is-default'

* ms/reflog-show-is-default:
  reflog: actually default to subcommand 'show'

13 years agoMerge branch 'jl/submodule-status-summary-doc'
Junio C Hamano [Mon, 8 Aug 2011 19:33:35 +0000 (12:33 -0700)]
Merge branch 'jl/submodule-status-summary-doc'

* jl/submodule-status-summary-doc:
  Documentation/submodule: add command references and update options

13 years agoMerge branch 'jn/gitweb-config-list-case'
Junio C Hamano [Mon, 8 Aug 2011 19:33:35 +0000 (12:33 -0700)]
Merge branch 'jn/gitweb-config-list-case'

* jn/gitweb-config-list-case:
  gitweb: Git config keys are case insensitive, make config search too

13 years agoMerge branch 'jl/submodule-update-quiet'
Junio C Hamano [Mon, 8 Aug 2011 19:33:34 +0000 (12:33 -0700)]
Merge branch 'jl/submodule-update-quiet'

* jl/submodule-update-quiet:
  submodule: update and add must honor --quiet flag

13 years agoMerge branch 'js/ls-tree-error'
Junio C Hamano [Mon, 8 Aug 2011 19:33:34 +0000 (12:33 -0700)]
Merge branch 'js/ls-tree-error'

* js/ls-tree-error:
  Ensure git ls-tree exits with a non-zero exit code if read_tree_recursive fails.
  Add a test to check that git ls-tree sets non-zero exit code on error.

13 years agoMerge branch 'jn/gitweb-system-config'
Junio C Hamano [Mon, 8 Aug 2011 19:33:34 +0000 (12:33 -0700)]
Merge branch 'jn/gitweb-system-config'

* jn/gitweb-system-config:
  gitweb: Introduce common system-wide settings for convenience

13 years agoMerge branch 'jk/reset-reflog-message-fix'
Junio C Hamano [Mon, 8 Aug 2011 19:33:33 +0000 (12:33 -0700)]
Merge branch 'jk/reset-reflog-message-fix'

* jk/reset-reflog-message-fix:
  reset: give better reflog messages

13 years agoMerge branch 'jc/diff-index-refactor'
Junio C Hamano [Mon, 8 Aug 2011 19:33:33 +0000 (12:33 -0700)]
Merge branch 'jc/diff-index-refactor'

* jc/diff-index-refactor:
  diff-lib: refactor run_diff_index() and do_diff_cache()
  diff-lib: simplify do_diff_cache()

13 years agoreceive-pack: do not overstep command line argument array
Junio C Hamano [Mon, 8 Aug 2011 19:31:01 +0000 (12:31 -0700)]
receive-pack: do not overstep command line argument array

Previous commit added one element to the command line, without
making sure the result fits there.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoMerge branch 'maint'
Junio C Hamano [Mon, 8 Aug 2011 19:24:16 +0000 (12:24 -0700)]
Merge branch 'maint'

* maint:
  filter-branch: Export variable `workdir' for --commit-filter
  Documentation/Makefile: add *.pdf to `clean' target
  Documentation: ignore *.pdf files

13 years agofilter-branch: Export variable `workdir' for --commit-filter
Michael Witten [Sun, 7 Aug 2011 02:44:43 +0000 (02:44 +0000)]
filter-branch: Export variable `workdir' for --commit-filter

According to `git help filter-branch':

       --commit-filter <command>
           ...
           You can use the _map_ convenience function in this filter,
           and other convenience functions, too...
           ...

However, it turns out that `map' hasn't been usable because it depends
on the variable `workdir', which is not propogated to the environment
of the shell that runs the commit-filter <command> because the
shell is created via a simple-command rather than a compound-command
subshell:

 @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
                 $(git write-tree) $parentstr < ../message > ../map/$commit ||
                         die "could not write rewritten commit"

One solution is simply to export `workdir'. However, it seems rather
heavy-handed to export `workdir' to the environments of all commands,
so instead this commit exports `workdir' for only the duration of the
shell command in question:

 workdir=$workdir @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
                 $(git write-tree) $parentstr < ../message > ../map/$commit ||
                         die "could not write rewritten commit"

Signed-off-by: Michael Witten <mfwitten@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agosubmodule: take advantage of gettextln and eval_gettextln.
Jon Seymour [Sun, 7 Aug 2011 11:58:17 +0000 (21:58 +1000)]
submodule: take advantage of gettextln and eval_gettextln.

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agostash: take advantage of eval_gettextln
Jon Seymour [Sun, 7 Aug 2011 11:58:16 +0000 (21:58 +1000)]
stash: take advantage of eval_gettextln

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agopull: take advantage of eval_gettextln
Jon Seymour [Sun, 7 Aug 2011 11:58:15 +0000 (21:58 +1000)]
pull: take advantage of eval_gettextln

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-am: take advantage of gettextln and eval_gettextln.
Jon Seymour [Sun, 7 Aug 2011 11:58:14 +0000 (21:58 +1000)]
git-am: take advantage of gettextln and eval_gettextln.

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogettext: add gettextln, eval_gettextln to encode common idiom
Jon Seymour [Sun, 7 Aug 2011 11:58:13 +0000 (21:58 +1000)]
gettext: add gettextln, eval_gettextln to encode common idiom

Currently, if you want to use gettext or eval_gettext to format a message
you may have to add a separate echo statement and a surrounding subshell
in order to interpolate the required trailing new line.

This patch introduces two new helper functions, gettextln and eval_gettextln
which append a trailing newline to the gettext output.

This allows constructions of the form:

if test -s "$GIT_DIR/BISECT_START"
then
(
gettext "You need to give me at least one good and one bad revisions.
(You can use \"git bisect bad\" and \"git bisect good\" for that.)" &&
echo
) >&2
else
...

to be expressed more concisely as:

if test -s "$GIT_DIR/BISECT_START"
then
gettextln "You need to give me at least one good and one bad revisions.
(You can use \"git bisect bad\" and \"git bisect good\" for that.)" >&2
else
...

Acked-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agotest: cope better with use of return for errors
Jonathan Nieder [Mon, 8 Aug 2011 01:17:09 +0000 (03:17 +0200)]
test: cope better with use of return for errors

In olden times, tests would quietly exit the script when they failed
at an inconvenient moment, which was a little disconcerting.
Therefore v0.99.5~24^2~4 (Trapping exit in tests, using return for
errors, 2005-08-10) switched to an idiom of using "return" instead,
wrapping evaluation of test code in a function to make that safe:

test_run_ () {
eval >&3 2>&4 "$1"
eval_ret="$?"
return 0
}

Years later, the implementation of test_when_finished (v1.7.1.1~95,
2010-05-02) and v1.7.2-rc2~1^2~13 (test-lib: output a newline before
"ok" under a TAP harness, 2010-06-24) took advantage of test_run_ as a
place to put code shared by all test assertion functions, without
paying attention to the function's former purpose:

test_run_ () {
...
eval >&3 2>&4 "$1"
eval_ret=$?

if should run cleanup
then
eval >&3 2>&4 "$test_cleanup"
fi
if TAP format requires a newline here
then
echo
fi
return 0
}

That means cleanup commands and the newline to put TAP output at
column 0 are skipped when tests use "return" to fail early.  Fix it by
introducing a test_eval_ function to catch the "return", with a
comment explaining the new function's purpose for the next person who
might touch this code.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agotest: simplify return value of test_run_
Jonathan Nieder [Mon, 8 Aug 2011 01:15:34 +0000 (03:15 +0200)]
test: simplify return value of test_run_

As v0.99.5~24^2~4 (Trapping exit in tests, using return for errors,
2005-08-10) explains, callers to test_run_ (such as test_expect_code)
used to check the result from eval and the return value separately so
tests that fail early could be distinguished from tests that completed
normally with successful (nonzero) status.  Eventually tests that
succeed with nonzero status were phased out (see v1.7.4-rc0~65^2~19,
2010-10-03 and especially v1.5.5-rc0~271, 2008-02-01) but the weird
two-return-value calling convention lives on.

Let's get rid of it.  The new rule: test_run_ succeeds (returns 0)
if and only if the test succeeded.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoDocumentation/Makefile: add *.pdf to `clean' target
Emilio G. Cota [Mon, 8 Aug 2011 08:33:05 +0000 (04:33 -0400)]
Documentation/Makefile: add *.pdf to `clean' target

user-manual.pdf is not removed by `make clean'; fix it.

Signed-off-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoDocumentation: ignore *.pdf files
Emilio G. Cota [Mon, 8 Aug 2011 08:33:04 +0000 (04:33 -0400)]
Documentation: ignore *.pdf files

user-manual.pdf is generated by the build and therefore
should be ignored by git.

Signed-off-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Propagate errors upwards from do_pick_commit
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:16 +0000 (16:09 +0530)]
revert: Propagate errors upwards from do_pick_commit

Currently, revert_or_cherry_pick can fail in two ways.  If it
encounters a conflict, it returns a positive number indicating the
intended exit status for the git wrapper to pass on; for all other
errors, it calls die().  The latter behavior is inconsiderate towards
callers, as it denies them the opportunity to recover from errors and
do other things.

After this patch, revert_or_cherry_pick will still return a positive
return value to indicate an exit status for conflicts as before, while
for some other errors, it will print an error message and return -1
instead of die()-ing.  The cmd_revert and cmd_cherry_pick are adjusted
to handle the fatal errors by die()-ing themselves.

While the full benefits of this patch will only be seen once all the
"die" calls are replaced with calls to "error", its immediate impact
is to change some "fatal:" messages to say "error:" and to add a new
"fatal: cherry-pick failed" message at the end when the operation
fails.

Inspired-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Jonathan Nieder <jrnieder@gmail.com>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Introduce --continue to continue the operation
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:15 +0000 (16:09 +0530)]
revert: Introduce --continue to continue the operation

Introduce a new "git cherry-pick --continue" command which uses the
information in ".git/sequencer" to continue a cherry-pick that stopped
because of a conflict or other error.  It works by dropping the first
instruction from .git/sequencer/todo and performing the remaining
cherry-picks listed there, with options (think "-s" and "-X") from the
initial command listed in ".git/sequencer/opts".

So now you can do:

  $ git cherry-pick -Xpatience foo..bar
  ... description conflict in commit moo ...
  $ git cherry-pick --continue
  error: 'cherry-pick' is not possible because you have unmerged files.
  fatal: failed to resume cherry-pick
  $ echo resolved >conflictingfile
  $ git add conflictingfile && git commit
  $ git cherry-pick --continue; # resumes with the commit after "moo"

During the "git commit" stage, CHERRY_PICK_HEAD will aid by providing
the commit message from the conflicting "moo" commit.  Note that the
cherry-pick mechanism has no control at this stage, so the user is
free to violate anything that was specified during the first
cherry-pick invocation.  For example, if "-x" was specified during the
first cherry-pick invocation, the user is free to edit out the message
during commit time.  Note that the "--signoff" option specified at
cherry-pick invocation time is not reflected in the commit message
provided by CHERRY_PICK_HEAD; the user must take care to add
"--signoff" during the "git commit" invocation.

Helped-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Don't implicitly stomp pending sequencer operation
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:14 +0000 (16:09 +0530)]
revert: Don't implicitly stomp pending sequencer operation

Protect the user from forgetting about a pending sequencer operation
by immediately erroring out when an existing cherry-pick or revert
operation is in progress like:

  $ git cherry-pick foo
  ... conflict ...
  $ git cherry-pick moo
  error: .git/sequencer already exists
  hint: A cherry-pick or revert is in progress
  hint: Use --reset to forget about it
  fatal: cherry-pick failed

A naive version of this would break the following established ways of
working:

  $ git cherry-pick foo
  ... conflict ...
  $ git reset --hard  # I actually meant "moo" when I said "foo"
  $ git cherry-pick moo

  $ git cherry-pick foo
  ... conflict ...
  $ git commit # commit the resolution
  $ git cherry-pick moo # New operation

However, the previous patches "reset: Make reset remove the sequencer
state" and "revert: Remove sequencer state when no commits are
pending" make sure that this does not happen.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Remove sequencer state when no commits are pending
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:13 +0000 (16:09 +0530)]
revert: Remove sequencer state when no commits are pending

When cherry-pick or revert is called on a list of commits, and a
conflict encountered somewhere in the middle, the data in
".git/sequencer" is required to continue the operation.  However, when
a conflict is encountered in the very last commit, the user will have
to "continue" after resolving the conflict and committing just so that
the sequencer state is removed.  This is how the current "rebase -i"
script works as well.

  $ git cherry-pick foo..bar
  ... conflict encountered while picking "bar" ...
  $ echo "resolved" >problematicfile
  $ git add problematicfile
  $ git commit
  $ git cherry-pick --continue # This would be a no-op

Change this so that the sequencer state is cleared when a conflict is
encountered in the last commit.  Incidentally, this patch makes sure
that some existing tests don't break when features like "--reset" and
"--continue" are implemented later in the series.

A better way to implement this feature is to get the last "git commit"
to remove the sequencer state.  However, that requires tighter
coupling between "git commit" and the sequencer, a goal that can be
pursued once the sequencer is made more general.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Acked-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoreset: Make reset remove the sequencer state
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:12 +0000 (16:09 +0530)]
reset: Make reset remove the sequencer state

Years of muscle memory have trained users to use "git reset --hard" to
remove the branch state after any sort operation.  Make it also remove
the sequencer state to facilitate this established workflow:

  $ git cherry-pick foo..bar
  ... conflict encountered ...
  $ git reset --hard # Oops, I didn't mean that
  $ git cherry-pick quux..bar
  ... cherry-pick succeeded ...

Guard against accidental removal of the sequencer state by providing
one level of "undo".  In the first "reset" invocation,
".git/sequencer" is moved to ".git/sequencer-old"; it is completely
removed only in the second invocation.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agofast-export: quote paths in output
Jeff King [Fri, 5 Aug 2011 22:36:22 +0000 (16:36 -0600)]
fast-export: quote paths in output

Many pathnames in a fast-import stream need to be quoted. In
particular:

  1. Pathnames at the end of an "M" or "D" line need quoting
     if they contain a LF or start with double-quote.

  2. Pathnames on a "C" or "R" line need quoting as above,
     but also if they contain spaces.

For (1), we weren't quoting at all. For (2), we put
double-quotes around the paths to handle spaces, but ignored
the possibility that they would need further quoting.

This patch checks whether each pathname needs c-style
quoting, and uses it. This is slightly overkill for (1),
which doesn't actually need to quote many characters that
vanilla c-style quoting does. However, it shouldn't hurt, as
any implementation needs to be ready to handle quoted
strings anyway.

In addition to adding a test, we have to tweak a test which
blindly assumed that case (2) would always use
double-quotes, whether it needed to or not.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agofilter-branch.sh: de-dent usage string
Junio C Hamano [Fri, 5 Aug 2011 22:06:21 +0000 (15:06 -0700)]
filter-branch.sh: de-dent usage string

"Usage: git filter-branch " that is prefixed to the first line is 25
columns long, so the "[--index-filter ..." on the second line would not
align with "[--env-filter ..." on the first line to begin with. If the
second and subsequent lines do not aim to align with anything on the
first line, it is just fine to indent them with a single HT.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agomisc-sh: fix up whitespace in some other .sh files.
Jon Seymour [Fri, 5 Aug 2011 13:31:29 +0000 (23:31 +1000)]
misc-sh: fix up whitespace in some other .sh files.

I found that the patched 4 files were different when this
filter is applied.

expand -i | unexpand --first-only

This patch contains the corrected files.

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoMerge branch 'tc/minix'
Junio C Hamano [Fri, 5 Aug 2011 21:55:00 +0000 (14:55 -0700)]
Merge branch 'tc/minix'

* tc/minix:
  Makefile: add Minix configuration options.

13 years agoMerge branch 'jc/pack-order-tweak'
Junio C Hamano [Fri, 5 Aug 2011 21:54:57 +0000 (14:54 -0700)]
Merge branch 'jc/pack-order-tweak'

* jc/pack-order-tweak:
  pack-objects: optimize "recency order"
  core: log offset pack data accesses happened

13 years agobisect: further style nitpicks
Junio C Hamano [Fri, 5 Aug 2011 17:09:23 +0000 (10:09 -0700)]
bisect: further style nitpicks

Fix a few remaining lines that indented with spaces.

Also simplify the logic of checking out the original branch and reporting
error during "bisect reset".

Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: replace "; then" with "\n<tab>*then"
Jon Seymour [Fri, 5 Aug 2011 11:31:31 +0000 (21:31 +1000)]
bisect: replace "; then" with "\n<tab>*then"

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: cleanup whitespace errors in git-bisect.sh.
Jon Seymour [Fri, 5 Aug 2011 11:31:30 +0000 (21:31 +1000)]
bisect: cleanup whitespace errors in git-bisect.sh.

All leading whitespace is now encoded with tabs.

After this patch, the following is true:

RAW=$(cat git-bisect.sh | md5sum) &&
ROUNDTRIP=$(cat git-bisect.sh | expand -i - | unexpand --first-only - | md5sum) &&
LEADING=$(sed -n "/^  */p" < git-bisect.sh | wc -l) &&
test $RAW = $ROUNDTRIP &&
test $LEADING = 0 &&
test -z "$(git diff -w HEAD~1 HEAD)"

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agotest-path-utils: Add subcommand "prefix_path"
Michael Haggerty [Thu, 4 Aug 2011 04:47:48 +0000 (06:47 +0200)]
test-path-utils: Add subcommand "prefix_path"

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agotest-path-utils: Add subcommand "absolute_path"
Michael Haggerty [Thu, 4 Aug 2011 04:47:47 +0000 (06:47 +0200)]
test-path-utils: Add subcommand "absolute_path"

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Normalize paths
Michael Haggerty [Thu, 4 Aug 2011 04:47:46 +0000 (06:47 +0200)]
git-check-attr: Normalize paths

Normalize the path arguments (relative to the working tree root, if
applicable) before looking up their attributes.  This requires passing
the prefix down the call chain.

This fixes two test cases for different reasons:

* "unnormalized paths" is fixed because the .gitattribute-file-seeking
  code is not confused into reading the top-level file twice.

* "relative paths" is fixed because the canonical pathnames are passed
  to get_check_attr() or get_all_attrs(), allowing them to match the
  pathname patterns as expected.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Demonstrate problems with relative paths
Michael Haggerty [Thu, 4 Aug 2011 04:47:45 +0000 (06:47 +0200)]
git-check-attr: Demonstrate problems with relative paths

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Demonstrate problems with unnormalized paths
Michael Haggerty [Thu, 4 Aug 2011 04:47:44 +0000 (06:47 +0200)]
git-check-attr: Demonstrate problems with unnormalized paths

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: test that no output is written to stderr
Michael Haggerty [Thu, 4 Aug 2011 04:47:43 +0000 (06:47 +0200)]
git-check-attr: test that no output is written to stderr

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoRename git_checkattr() to git_check_attr()
Michael Haggerty [Thu, 4 Aug 2011 04:36:33 +0000 (06:36 +0200)]
Rename git_checkattr() to git_check_attr()

Suggested by: Junio Hamano <gitster@pobox.com>

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Fix command-line handling to match docs
Michael Haggerty [Thu, 4 Aug 2011 04:36:32 +0000 (06:36 +0200)]
git-check-attr: Fix command-line handling to match docs

According to the git-check-attr synopsis, if the '--stdin' option is
used then no pathnames are expected on the command line.  Change the
behavior to match this description; namely, if '--stdin' is used but
not '--', then treat all command-line arguments as attribute names.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Drive two tests using the same raw data
Michael Haggerty [Thu, 4 Aug 2011 04:36:31 +0000 (06:36 +0200)]
git-check-attr: Drive two tests using the same raw data

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Add an --all option to show all attributes
Michael Haggerty [Thu, 4 Aug 2011 04:36:30 +0000 (06:36 +0200)]
git-check-attr: Add an --all option to show all attributes

Add new usage patterns

    git check-attr [-a | --all] [--] pathname...
    git check-attr --stdin [-a | --all] < <list-of-paths>

which display all attributes associated with the specified file(s).

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Error out if no pathnames are specified
Michael Haggerty [Thu, 4 Aug 2011 04:36:29 +0000 (06:36 +0200)]
git-check-attr: Error out if no pathnames are specified

If no pathnames are passed as command-line arguments and the --stdin
option is not specified, fail with the error message "No file
specified".  Add tests of this behavior.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Process command-line args more systematically
Michael Haggerty [Thu, 4 Aug 2011 04:36:28 +0000 (06:36 +0200)]
git-check-attr: Process command-line args more systematically

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Handle each error separately
Michael Haggerty [Thu, 4 Aug 2011 04:36:27 +0000 (06:36 +0200)]
git-check-attr: Handle each error separately

This will make the code easier to refactor.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Extract a function error_with_usage()
Michael Haggerty [Thu, 4 Aug 2011 04:36:26 +0000 (06:36 +0200)]
git-check-attr: Extract a function error_with_usage()

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Introduce a new variable
Michael Haggerty [Thu, 4 Aug 2011 04:36:25 +0000 (06:36 +0200)]
git-check-attr: Introduce a new variable

Avoid reusing variable "doubledash" to mean something other than the
expected "position of a double-dash, if any".

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Extract a function output_attr()
Michael Haggerty [Thu, 4 Aug 2011 04:36:24 +0000 (06:36 +0200)]
git-check-attr: Extract a function output_attr()

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoAllow querying all attributes on a file
Michael Haggerty [Thu, 4 Aug 2011 04:36:23 +0000 (06:36 +0200)]
Allow querying all attributes on a file

Add a function, git_all_attrs(), that reports on all attributes that
are set on a path.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoRemove redundant check
Michael Haggerty [Thu, 4 Aug 2011 04:36:22 +0000 (06:36 +0200)]
Remove redundant check

bootstrap_attr_stack() also checks whether attr_stack is already set.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoRemove redundant call to bootstrap_attr_stack()
Michael Haggerty [Thu, 4 Aug 2011 04:36:21 +0000 (06:36 +0200)]
Remove redundant call to bootstrap_attr_stack()

prepare_attr_stack() does the same thing.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoExtract a function collect_all_attrs()
Michael Haggerty [Thu, 4 Aug 2011 04:36:20 +0000 (06:36 +0200)]
Extract a function collect_all_attrs()

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoTeach prepare_attr_stack() to figure out dirlen itself
Michael Haggerty [Thu, 4 Aug 2011 04:36:19 +0000 (06:36 +0200)]
Teach prepare_attr_stack() to figure out dirlen itself

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Use git_attr_name()
Michael Haggerty [Thu, 4 Aug 2011 04:36:18 +0000 (06:36 +0200)]
git-check-attr: Use git_attr_name()

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoProvide access to the name attribute of git_attr
Michael Haggerty [Thu, 4 Aug 2011 04:36:17 +0000 (06:36 +0200)]
Provide access to the name attribute of git_attr

It will be present in any likely future reimplementation, and its
availability simplifies other code.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Add tests of command-line parsing
Michael Haggerty [Thu, 4 Aug 2011 04:36:16 +0000 (06:36 +0200)]
git-check-attr: Add tests of command-line parsing

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogit-check-attr: Add missing "&&"
Michael Haggerty [Thu, 4 Aug 2011 04:36:15 +0000 (06:36 +0200)]
git-check-attr: Add missing "&&"

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoDisallow the empty string as an attribute name
Michael Haggerty [Thu, 4 Aug 2011 04:36:14 +0000 (06:36 +0200)]
Disallow the empty string as an attribute name

Previously, it was possible to have a line like "file.txt =foo" in a
.gitattribute file, after which an invocation like "git check-attr ''
-- file.txt" would succeed.  This patch disallows both constructs.

Please note that any existing .gitattributes file that tries to set an
empty attribute will now trigger the error message "error: : not a
valid attribute name" whereas previously the nonsense was allowed
through.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoRemove anachronism from comment
Michael Haggerty [Thu, 4 Aug 2011 04:36:13 +0000 (06:36 +0200)]
Remove anachronism from comment

Setting attributes to arbitrary values ("attribute=value") is now
supported, so it is no longer necessary for this comment to justify
prohibiting '=' in an attribute name.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agodoc: Correct git_attr() calls in example code
Michael Haggerty [Thu, 4 Aug 2011 04:36:12 +0000 (06:36 +0200)]
doc: Correct git_attr() calls in example code

Commit 7fb0eaa2 (2010-01-17) changed git_attr() to take a string
instead of a string and a length.  Update the documentation
accordingly.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agodoc: Add a link from gitattributes(5) to git-check-attr(1)
Michael Haggerty [Thu, 4 Aug 2011 04:36:11 +0000 (06:36 +0200)]
doc: Add a link from gitattributes(5) to git-check-attr(1)

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agodocs: put listed example commands in backticks
Jeff King [Thu, 4 Aug 2011 02:13:29 +0000 (20:13 -0600)]
docs: put listed example commands in backticks

Many examples of git command invocation are given in asciidoc listing
blocks, which makes them monospaced and avoids further interpretation of
special characters.  Some manpages make a list of examples, like:

  git foo::
    Run git foo.

  git foo -q::
    Use the "-q" option.

to quickly show many variants. However, they can sometimes be hard to
read, because they are shown in a proportional-width font (so, for
example, seeing the difference between "-- foo" and "--foo" can be
difficult).

This patch puts all such examples into backticks, which gives the
equivalent formatting to a listing block (i.e., monospaced and without
character interpretation).

As a bonus, this also fixes an example in the git-push manpage, in which
"git push origin :::" was accidentally considered a newly-indented list,
and not a list item with "git push origin :" in it.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Introduce --reset to remove sequencer state
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:11 +0000 (16:09 +0530)]
revert: Introduce --reset to remove sequencer state

To explicitly remove the sequencer state for a fresh cherry-pick or
revert invocation, introduce a new subcommand called "--reset" to
remove the sequencer state.

Take the opportunity to publicly expose the sequencer paths, and a
generic function called "remove_sequencer_state" that various git
programs can use to remove the sequencer state in a uniform manner;
"git reset" uses it later in this series.  Introducing this public API
is also in line with our long-term goal of eventually factoring out
functions from revert.c into a generic commit sequencer.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Make pick_commits functionally act on a commit list
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:10 +0000 (16:09 +0530)]
revert: Make pick_commits functionally act on a commit list

Apart from its central objective of calling into the picking
mechanism, pick_commits creates a sequencer directory, prepares a todo
list, and even acts upon the "--reset" subcommand.  This makes for a
bad API since the central worry of callers is to figure out whether or
not any conflicts were encountered during the cherry picking.  The
current API is like:

  if (pick_commits(opts) < 0)
     print "Something failed, we're not sure what"

So, change pick_commits so that it's only responsible for picking
commits in a loop and reporting any errors, leaving the rest to a new
function called pick_revisions.  Consequently, the API of pick_commits
becomes much clearer:

  act_on_subcommand(opts->subcommand);
  todo_list = prepare_todo_list();
  if (pick_commits(todo_list, opts) < 0)
     print "Error encountered while picking commits"

Now, callers can easily call-in to the cherry-picking machinery by
constructing an arbitrary todo list along with some options.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Save command-line options for continuing operation
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:09 +0000 (16:09 +0530)]
revert: Save command-line options for continuing operation

In the same spirit as ".git/sequencer/head" and ".git/sequencer/todo",
introduce ".git/sequencer/opts" to persist the replay_opts structure
for continuing after a conflict resolution.  Use the gitconfig format
for this file so that it looks like:

  [options]
  signoff = true
  record-origin = true
  mainline = 1
  strategy = recursive
  strategy-option = patience
  strategy-option = ours

Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Helped-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Save data for continuing after conflict resolution
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:08 +0000 (16:09 +0530)]
revert: Save data for continuing after conflict resolution

Ever since v1.7.2-rc1~4^2~7 (revert: allow cherry-picking more than
one commit, 2010-06-02), a single invocation of "git cherry-pick" or
"git revert" can perform picks of several individual commits.  To
implement features like "--continue" to continue the whole operation,
we will need to store some information about the state and the plan at
the beginning.  Introduce a ".git/sequencer/head" file to store this
state, and ".git/sequencer/todo" file to store the plan.  The head
file contains the SHA-1 of the HEAD before the start of the operation,
and the todo file contains an instruction sheet whose format is
inspired by the format of the "rebase -i" instruction sheet.  As a
result, a typical todo file looks like:

  pick 8537f0e submodule add: test failure when url is not configured
  pick 4d68932 submodule add: allow relative repository path
  pick f22a17e submodule add: clean up duplicated code
  pick 59a5775 make copy_ref globally available

Since SHA-1 hex is abbreviated using an find_unique_abbrev(), it is
unambiguous.  This does not guarantee that there will be no ambiguity
when more objects are added to the repository.

These two files alone are not enough to implement a "--continue" that
remembers the command-line options specified; later patches in the
series save them too.

These new files are unrelated to the existing .git/CHERRY_PICK_HEAD,
which will still be useful while committing after a conflict
resolution.

Inspired-by: Christian Couder <chriscool@tuxfamily.org>
Helped-by: Junio C Hamano <gitster@pobox.com>
Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Don't create invalid replay_opts in parse_args
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:07 +0000 (16:09 +0530)]
revert: Don't create invalid replay_opts in parse_args

The "--ff" command-line option cannot be used with some other
command-line options.  However, parse_args still parses these
incompatible options into a replay_opts structure for use by the rest
of the program.  Although pick_commits, the current gatekeeper to the
cherry-pick machinery, checks the validity of the replay_opts
structure before before starting its operation, there will be multiple
entry points to the cherry-pick machinery in future.  To futureproof
the code and catch these errors in one place, make sure that an
invalid replay_opts structure is not created by parse_args in the
first place.  We still check the replay_opts structure for validity in
pick_commits, but this is an assert() now to emphasize that it's the
caller's responsibility to get it right.

Inspired-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Jonathan Nieder <jrnieder@gmail.com>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Separate cmdline parsing from functional code
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:06 +0000 (16:09 +0530)]
revert: Separate cmdline parsing from functional code

Currently, revert_or_cherry_pick sets up a default git config, parses
command-line arguments, before preparing to pick commits.  This makes
for a bad API as the central worry of callers is to assert whether or
not a conflict occured while cherry picking.  The current API is like:

  if (revert_or_cherry_pick(argc, argv, opts) < 0)
     print "Something failed, we're not sure what"

Simplify and rename revert_or_cherry_pick to pick_commits so that it
only has the responsibility of setting up the revision walker and
picking commits in a loop.  Transfer the remaining work to its
callers.  Now, the API is simplified as:

  if (parse_args(argc, argv, opts) < 0)
     print "Can't parse arguments"
  if (pick_commits(opts) < 0)
     print "Error encountered in picking machinery"

Later in the series, pick_commits will also serve as the starting
point for continuing a cherry-pick or revert.

Inspired-by: Christian Couder <chriscool@tuxfamily.org>
Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Introduce struct to keep command-line options
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:05 +0000 (16:09 +0530)]
revert: Introduce struct to keep command-line options

The current code uses a set of file-scope static variables to tell the
cherry-pick/ revert machinery how to replay the changes, and
initializes them by parsing the command-line arguments.  In later
steps in this series, we would like to introduce an API function that
calls into this machinery directly and have a way to tell it what to
do.  Hence, introduce a structure to group these variables, so that
the API can take them as a single replay_options parameter.  The only
exception is the variable "me" -- remove it since it not an
independent option, and can be inferred from the action.

Unfortunately, this patch introduces a minor regression.  Parsing
strategy-option violates a C89 rule: Initializers cannot refer to
variables whose address is not known at compile time.  Currently, this
rule is violated by some other parts of Git as well, and it is
possible to get GCC to report these instances using the "-std=c89
-pedantic" option.

Inspired-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Jonathan Nieder <jrnieder@gmail.com>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Eliminate global "commit" variable
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:04 +0000 (16:09 +0530)]
revert: Eliminate global "commit" variable

Functions which act on commits currently rely on a file-scope static
variable to be set before they're called.  Consequently, the API and
corresponding callsites are ugly and unclear.  Remove this variable
and change their API to accept the commit to act on as additional
argument so that the callsites change from looking like

  commit = prepare_a_commit();
  act_on_commit();

to looking like

  commit = prepare_a_commit();
  act_on_commit(commit);

This change is also in line with our long-term goal of exposing some
of these functions through a public API.

Inspired-by: Christian Couder <chriscool@tuxfamily.org>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Rename no_replay to record_origin
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:03 +0000 (16:09 +0530)]
revert: Rename no_replay to record_origin

The "-x" command-line option is used to record the name of the
original commits being picked in the commit message.  The variable
corresponding to this option is named "no_replay" for historical
reasons; the name is especially confusing because the term "replay" is
used to describe what cherry-pick does (for example, in the
documentation of the "--mainline" option).  So, give the variable
corresponding to the "-x" command-line option a better name:
"record_origin".

Mentored-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Don't check lone argument in get_encoding
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:02 +0000 (16:09 +0530)]
revert: Don't check lone argument in get_encoding

The only place get_encoding uses the global "commit" variable is when
writing an error message explaining that its lone argument was NULL.
Since the function's only caller ensures that a NULL argument isn't
passed, we can remove this check with two beneficial consequences:

1. Since the function doesn't use the global "commit" variable any
   more, it won't need to change when we eliminate the global variable
   later in the series.
2. Translators no longer need to localize an error message that will
   never be shown.

Suggested-by: Junio C Hamano <gitster@pobox.com>
Mentored-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agorevert: Simplify and inline add_message_to_msg
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:01 +0000 (16:09 +0530)]
revert: Simplify and inline add_message_to_msg

The add_message_to_msg function has some dead code, an unclear API,
only one callsite.  While it originally intended fill up an empty
commit message with the commit object name while picking, it really
doesn't do this -- a bug introduced in v1.5.1-rc1~65^2~2 (Make
git-revert & git-cherry-pick a builtin, 2007-03-01).  Today, tests in
t3505-cherry-pick-empty.sh indicate that not filling up an empty
commit message is the desired behavior.  Re-implement and inline the
function accordingly, with a beneficial side-effect: don't dereference
a NULL pointer when the commit doesn't have a delimeter after the
header.

Helped-by: Junio C Hamano <gitster@pobox.com>
Mentored-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoconfig: Introduce functions to write non-standard file
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:39:00 +0000 (16:09 +0530)]
config: Introduce functions to write non-standard file

Introduce two new functions corresponding to "git_config_set" and
"git_config_set_multivar" to write a non-standard configuration file.
Expose these new functions in cache.h for other git programs to use.

Helped-by: Jeff King <peff@peff.net>
Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoadvice: Introduce error_resolve_conflict
Ramkumar Ramachandra [Thu, 4 Aug 2011 10:38:59 +0000 (16:08 +0530)]
advice: Introduce error_resolve_conflict

Enable future callers to report a conflict and not die immediately by
introducing a new function called error_resolve_conflict.
Re-implement die_resolve_conflict as a call to error_resolve_conflict
followed by a call to die.  Consequently, the message printed by
die_resolve_conflict changes from

  fatal: 'commit' is not possible because you have unmerged files.
         Please, fix them up in the work tree ...
         ...

to

  error: 'commit' is not possible because you have unmerged files.
  hint: Fix them up in the work tree ...
  hint: ...
  fatal: Exiting because of an unresolved conflict.

Hints are printed using the same advise function introduced in
v1.7.3-rc0~26^2~3 (Introduce advise() to print hints, 2010-08-11).

Inspired-by: Christian Couder <chistian.couder@gmail.com>
Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: add documentation for --no-checkout option.
Jon Seymour [Thu, 4 Aug 2011 12:01:03 +0000 (22:01 +1000)]
bisect: add documentation for --no-checkout option.

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: add tests for the --no-checkout option.
Jon Seymour [Thu, 4 Aug 2011 12:01:02 +0000 (22:01 +1000)]
bisect: add tests for the --no-checkout option.

These tests verify that git-bisect --no-checkout can successfully
bisect commit histories that reference damaged trees.

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: introduce --no-checkout support into porcelain.
Jon Seymour [Thu, 4 Aug 2011 12:01:01 +0000 (22:01 +1000)]
bisect: introduce --no-checkout support into porcelain.

git-bisect can now perform bisection of a history without performing
a checkout at each stage of the bisection process. Instead, HEAD is updated.

One use-case for this function is allow git bisect to be used with
damaged repositories where git checkout would fail because the tree
referenced by the commit is damaged.

It can also be used in other cases where actual checkout of the tree
is not required to progress the bisection.

Improved-by: Christian Couder <chriscool@tuxfamily.org>
Improved-by: Junio C Hamano <gitster@pobox.com>
Improved-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: introduce support for --no-checkout option.
Jon Seymour [Thu, 4 Aug 2011 12:01:00 +0000 (22:01 +1000)]
bisect: introduce support for --no-checkout option.

If --no-checkout is specified, then the bisection process uses:

git update-ref --no-deref HEAD <trial>

at each trial instead of:

git checkout <trial>

Improved-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: add tests to document expected behaviour in presence of broken trees.
Jon Seymour [Thu, 4 Aug 2011 12:00:59 +0000 (22:00 +1000)]
bisect: add tests to document expected behaviour in presence of broken trees.

If the repo is broken, we expect bisect to fail.

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: use && to connect statements that are deferred with eval.
Jon Seymour [Thu, 4 Aug 2011 12:00:58 +0000 (22:00 +1000)]
bisect: use && to connect statements that are deferred with eval.

Christian Couder pointed out that the existing eval strategy
swallows an initial non-zero return. Using && to connect
the statements should fix this.

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agobisect: move argument parsing before state modification.
Jon Seymour [Thu, 4 Aug 2011 12:00:57 +0000 (22:00 +1000)]
bisect: move argument parsing before state modification.

Currently 'git bisect start' modifies some state prior to checking
that its arguments are valid.

This change moves argument validation before state modification
with the effect that state modification does not occur
unless argument validations succeeds.

An existing test is changed to check that new bisect state
is not created if arguments are invalid.

A new test is added to check that existing bisect state
is not modified if arguments are invalid.

Signed-off-by: Jon Seymour <jon.seymour@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agogitweb: pass string after encoding in utf-8 to syntax highlighter
张忠山 [Thu, 4 Aug 2011 15:52:55 +0000 (23:52 +0800)]
gitweb: pass string after encoding in utf-8 to syntax highlighter

Otherwise the highlight filter would work on a corrupt byte sequence.

Signed-off-by: 张忠山 <zzs213@126.com>
Acked-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoAdd option hooks.diffopts to customize change summary in post-receive-email
Jon Jensen [Thu, 4 Aug 2011 03:36:08 +0000 (21:36 -0600)]
Add option hooks.diffopts to customize change summary in post-receive-email

This makes it easy to customize the git diff-tree options, for example
to include -p to include inline diffs.

It defaults to the current options "--stat --summary --find-copies-harder"
and thus is backward-compatible.

Signed-off-by: Jon Jensen <jon@endpoint.com>
Improved-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agodiff -c/--cc: do not mistake "resolved as deletion" as "use working tree"
Junio C Hamano [Thu, 4 Aug 2011 18:39:10 +0000 (11:39 -0700)]
diff -c/--cc: do not mistake "resolved as deletion" as "use working tree"

The combined diff machinery can be used to compare:

 - a merge commit with its parent commits;
 - a working-tree file with multiple stages in an unmerged index; or
 - a working-tree file with the HEAD and the index.

The internal function combine-diff.c:show_patch_diff() checked if it needs
to read the "result" from the working tree by looking at the object name
of the result --- if it is null_sha1, it read from the working tree.

This mistook a merge that records a deletion as the conflict resolution
as if it is a cue to read from the working tree. Pass this information
explicitly from the caller instead.

Noticed and reported by Johan Herland.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoMerge branch 'maint'
Junio C Hamano [Wed, 3 Aug 2011 21:16:17 +0000 (14:16 -0700)]
Merge branch 'maint'

* maint:
  add gitignore entry to description about how to write a builtin
  gitattributes: Reword "attribute macro" to "macro attribute"
  gitattributes: Clarify discussion of attribute macros

13 years agoadd gitignore entry to description about how to write a builtin
Heiko Voigt [Wed, 3 Aug 2011 18:06:17 +0000 (20:06 +0200)]
add gitignore entry to description about how to write a builtin

If the author forgets the gitignore entry the built result will show up
as new file in the git working directory.

Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
13 years agoMakefile: some changes for http-related flag documentation
Tay Ray Chuan [Wed, 3 Aug 2011 12:07:57 +0000 (20:07 +0800)]
Makefile: some changes for http-related flag documentation

Rename git-http-pull to git-http-fetch. This was passed over in 215a7ad
(Big tool rename, Wed Sep 7 17:26:23 2005 -0700).

Also, distinguish between dumb and smart in flag docs, as the "warnings"
in NO_CURL and NO_EXPACT are no longer accurate given the introduction
of smart http(s).

Signed-off-by: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>