Code

git.git
12 years agot5550: repack everything into one file
Clemens Buchacher [Wed, 4 Jan 2012 15:55:34 +0000 (16:55 +0100)]
t5550: repack everything into one file

Subsequently we assume that there is only one pack. Currently this is
true only by accident. Pass '-a -d' to repack in order to guarantee that
assumption to hold true.

The prune-packed command is now redundant since repack -d already calls
it.

Signed-off-by: Clemens Buchacher <drizzd@aon.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'nd/maint-parse-depth' into maint
Junio C Hamano [Wed, 4 Jan 2012 17:43:26 +0000 (09:43 -0800)]
Merge branch 'nd/maint-parse-depth' into maint

* nd/maint-parse-depth:
  Catch invalid --depth option passed to clone or fetch

12 years agoCatch invalid --depth option passed to clone or fetch
Nguyễn Thái Ngọc Duy [Wed, 4 Jan 2012 10:01:55 +0000 (17:01 +0700)]
Catch invalid --depth option passed to clone or fetch

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'maint-1.7.7' into maint
Junio C Hamano [Tue, 3 Jan 2012 21:47:46 +0000 (13:47 -0800)]
Merge branch 'maint-1.7.7' into maint

* maint-1.7.7:
  docs: describe behavior of relative submodule URLs
  Documentation: read-tree --prefix works with existing subtrees
  Add MYMETA.json to perl/.gitignore

12 years agoMerge branch 'maint-1.7.6' into maint-1.7.7
Junio C Hamano [Tue, 3 Jan 2012 21:47:15 +0000 (13:47 -0800)]
Merge branch 'maint-1.7.6' into maint-1.7.7

* maint-1.7.6:
  Documentation: read-tree --prefix works with existing subtrees
  Add MYMETA.json to perl/.gitignore

12 years agodocs: describe behavior of relative submodule URLs
Jens Lehmann [Sun, 1 Jan 2012 15:13:16 +0000 (16:13 +0100)]
docs: describe behavior of relative submodule URLs

Since the relative submodule URLs have been introduced in f31a522a2d, they
do not conform to the rules for resolving relative URIs but rather to
those of relative directories.

Document that behavior.

Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agofix hang in git fetch if pointed at a 0 length bundle
Brian Harring [Tue, 3 Jan 2012 13:46:03 +0000 (05:46 -0800)]
fix hang in git fetch if pointed at a 0 length bundle

git-repo if interupted at the exact wrong time will generate zero
length bundles- literal empty files.  git-repo is wrong here, but
git fetch shouldn't effectively spin loop if pointed at a zero
length bundle.

Signed-off-by: Brian Harring <ferringb@chromium.org>
Helped-by: Johannes Sixt
Helped-by: Nguyen Thai Ngoc Duy
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoDocumentation: read-tree --prefix works with existing subtrees
Clemens Buchacher [Sat, 31 Dec 2011 11:50:56 +0000 (12:50 +0100)]
Documentation: read-tree --prefix works with existing subtrees

Since 34110cd4 (Make 'unpack_trees()' have a separate source and
destination index) it is no longer true that a subdirectory with
the same prefix must not exist.

Signed-off-by: Clemens Buchacher <drizzd@aon.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoAdd MYMETA.json to perl/.gitignore
Jack Nagel [Thu, 29 Dec 2011 04:42:05 +0000 (22:42 -0600)]
Add MYMETA.json to perl/.gitignore

ExtUtils::MakeMaker generates MYMETA.json in addition to MYMETA.yml
since version 6.57_07. As it suggests, it is just meta information about
the build and is cleaned up with 'make clean', so it should be ignored.

Signed-off-by: Jack Nagel <jacknagel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoGit 1.7.8.2 v1.7.8.2
Junio C Hamano [Wed, 28 Dec 2011 19:49:09 +0000 (11:49 -0800)]
Git 1.7.8.2

Contains accumulated fixes since 1.7.8 that have been merged to the
'master' branch in preparation for the 1.7.9 release.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'jv/maint-config-set' into maint
Junio C Hamano [Wed, 28 Dec 2011 20:03:19 +0000 (12:03 -0800)]
Merge branch 'jv/maint-config-set' into maint

* jv/maint-config-set:
  Fix an incorrect reference to --set-all.

12 years agoMerge branch 'jk/follow-rename-score' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:49:37 +0000 (11:49 -0800)]
Merge branch 'jk/follow-rename-score' into maint

* jk/follow-rename-score:
  use custom rename score during --follow

12 years agoMerge branch 'jc/checkout-m-twoway' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:44:54 +0000 (11:44 -0800)]
Merge branch 'jc/checkout-m-twoway' into maint

* jc/checkout-m-twoway:
  t/t2023-checkout-m.sh: fix use of test_must_fail
  checkout_merged(): squelch false warning from some gcc
  Test 'checkout -m -- path'
  checkout -m: no need to insist on having all 3 stages

12 years agoMerge branch 'tr/doc-sh-setup' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:42:51 +0000 (11:42 -0800)]
Merge branch 'tr/doc-sh-setup' into maint

* tr/doc-sh-setup:
  git-sh-setup: make require_clean_work_tree part of the interface

12 years agoMerge branch 'jk/maint-strbuf-missing-init' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:42:46 +0000 (11:42 -0800)]
Merge branch 'jk/maint-strbuf-missing-init' into maint

* jk/maint-strbuf-missing-init:
  commit, merge: initialize static strbuf

12 years agoMerge branch 'jk/maint-push-v-is-verbose' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:42:42 +0000 (11:42 -0800)]
Merge branch 'jk/maint-push-v-is-verbose' into maint

* jk/maint-push-v-is-verbose:
  make "git push -v" actually verbose

12 years agoMerge branch 'jk/http-push-to-empty' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:42:37 +0000 (11:42 -0800)]
Merge branch 'jk/http-push-to-empty' into maint

* jk/http-push-to-empty:
  remote-curl: don't pass back fake refs

Conflicts:
remote-curl.c

12 years agoMerge branch 'jk/doc-fsck' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:42:33 +0000 (11:42 -0800)]
Merge branch 'jk/doc-fsck' into maint

* jk/doc-fsck:
  docs: brush up obsolete bits of git-fsck manpage

12 years agoMerge branch 'jc/maint-lf-to-crlf-keep-crlf' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:42:27 +0000 (11:42 -0800)]
Merge branch 'jc/maint-lf-to-crlf-keep-crlf' into maint

* jc/maint-lf-to-crlf-keep-crlf:
  lf_to_crlf_filter(): resurrect CRLF->CRLF hack

12 years agoMerge branch 'ef/setenv-putenv' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:42:24 +0000 (11:42 -0800)]
Merge branch 'ef/setenv-putenv' into maint

* ef/setenv-putenv:
  compat/setenv.c: error if name contains '='
  compat/setenv.c: update errno when erroring out

12 years agoMerge branch 'jc/advice-doc' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:39 +0000 (11:32 -0800)]
Merge branch 'jc/advice-doc' into maint

* jc/advice-doc:
  advice: Document that they all default to true

12 years agoMerge branch 'jn/maint-sequencer-fixes' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:39 +0000 (11:32 -0800)]
Merge branch 'jn/maint-sequencer-fixes' into maint

* jn/maint-sequencer-fixes:
  revert: stop creating and removing sequencer-old directory
  Revert "reset: Make reset remove the sequencer state"
  revert: do not remove state until sequence is finished
  revert: allow single-pick in the middle of cherry-pick sequence
  revert: pass around rev-list args in already-parsed form
  revert: allow cherry-pick --continue to commit before resuming
  revert: give --continue handling its own function

12 years agoMerge branch 'jk/maint-snprintf-va-copy' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:38 +0000 (11:32 -0800)]
Merge branch 'jk/maint-snprintf-va-copy' into maint

* jk/maint-snprintf-va-copy:
  compat/snprintf: don't look at va_list twice

12 years agoMerge branch 'jk/maint-push-over-dav' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:37 +0000 (11:32 -0800)]
Merge branch 'jk/maint-push-over-dav' into maint

* jk/maint-push-over-dav:
  http-push: enable "proactive auth"
  t5540: test DAV push with authentication

12 years agoMerge branch 'jk/maint-mv' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:36 +0000 (11:32 -0800)]
Merge branch 'jk/maint-mv' into maint

* jk/maint-mv:
  mv: be quiet about overwriting
  mv: improve overwrite warning
  mv: make non-directory destination error more clear
  mv: honor --verbose flag
  docs: mention "-k" for both forms of "git mv"

12 years agoMerge branch 'jk/fetch-no-tail-match-refs' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:36 +0000 (11:32 -0800)]
Merge branch 'jk/fetch-no-tail-match-refs' into maint

* jk/fetch-no-tail-match-refs:
  connect.c: drop path_match function
  fetch-pack: match refs exactly
  t5500: give fully-qualified refs to fetch-pack
  drop "match" parameter from get_remote_heads

12 years agoMerge branch 'ew/keepalive' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:36 +0000 (11:32 -0800)]
Merge branch 'ew/keepalive' into maint

* ew/keepalive:
  enable SO_KEEPALIVE for connected TCP sockets

12 years agoMerge branch 'ci/stripspace-docs' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:35 +0000 (11:32 -0800)]
Merge branch 'ci/stripspace-docs' into maint

* ci/stripspace-docs:
  Update documentation for stripspace

12 years agoMerge branch 'jh/fast-import-notes' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:35 +0000 (11:32 -0800)]
Merge branch 'jh/fast-import-notes' into maint

* jh/fast-import-notes:
  fast-import: Fix incorrect fanout level when modifying existing notes refs
  t9301: Add 2nd testcase exposing bugs in fast-import's notes fanout handling
  t9301: Fix testcase covering up a bug in fast-import's notes fanout handling

12 years agoMerge branch 'aw/rebase-i-stop-on-failure-to-amend' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:34 +0000 (11:32 -0800)]
Merge branch 'aw/rebase-i-stop-on-failure-to-amend' into maint

* aw/rebase-i-stop-on-failure-to-amend:
  rebase -i: interrupt rebase when "commit --amend" failed during "reword"

12 years agoMerge branch 'tj/maint-imap-send-remove-unused' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:34 +0000 (11:32 -0800)]
Merge branch 'tj/maint-imap-send-remove-unused' into maint

* tj/maint-imap-send-remove-unused:
  imap-send: Remove unused 'use_namespace' variable

12 years agoMerge branch 'jn/branch-move-to-self' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:33 +0000 (11:32 -0800)]
Merge branch 'jn/branch-move-to-self' into maint

* jn/branch-move-to-self:
  Allow checkout -B <current-branch> to update the current branch
  branch: allow a no-op "branch -M <current-branch> HEAD"

12 years agoMerge branch 'na/strtoimax' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:33 +0000 (11:32 -0800)]
Merge branch 'na/strtoimax' into maint

* na/strtoimax:
  Support sizes >=2G in various config options accepting 'g' sizes.
  Compatibility: declare strtoimax() under NO_STRTOUMAX
  Add strtoimax() compatibility function.

12 years agoMerge branch 'jk/refresh-porcelain-output' into maint
Junio C Hamano [Wed, 28 Dec 2011 19:32:32 +0000 (11:32 -0800)]
Merge branch 'jk/refresh-porcelain-output' into maint

* jk/refresh-porcelain-output:
  refresh_index: make porcelain output more specific
  refresh_index: rename format variables
  read-cache: let refresh_cache_ent pass up changed flags

12 years agoFix an incorrect reference to --set-all.
Jelmer Vernooij [Tue, 27 Dec 2011 02:03:45 +0000 (03:03 +0100)]
Fix an incorrect reference to --set-all.

Signed-off-by: Jelmer Vernooij <jelmer@samba.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoGit 1.7.8.1 v1.7.8.1
Junio C Hamano [Wed, 21 Dec 2011 20:02:13 +0000 (12:02 -0800)]
Git 1.7.8.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'jl/submodule-status-failure-report' into maint
Junio C Hamano [Wed, 21 Dec 2011 19:42:45 +0000 (11:42 -0800)]
Merge branch 'jl/submodule-status-failure-report' into maint

* jl/submodule-status-failure-report:
  diff/status: print submodule path when looking for changes fails

12 years agoMerge branch 'tr/userdiff-c-returns-pointer' into maint
Junio C Hamano [Wed, 21 Dec 2011 19:42:45 +0000 (11:42 -0800)]
Merge branch 'tr/userdiff-c-returns-pointer' into maint

* tr/userdiff-c-returns-pointer:
  userdiff: allow * between cpp funcname words

12 years agoMerge branch 'bc/maint-apply-check-no-patch' into maint
Junio C Hamano [Wed, 21 Dec 2011 19:42:45 +0000 (11:42 -0800)]
Merge branch 'bc/maint-apply-check-no-patch' into maint

* bc/maint-apply-check-no-patch:
  builtin/apply.c: report error on failure to recognize input
  t/t4131-apply-fake-ancestor.sh: fix broken test

12 years agoMerge branch 'cn/maint-lf-to-crlf-filter' into maint
Junio C Hamano [Wed, 21 Dec 2011 19:42:44 +0000 (11:42 -0800)]
Merge branch 'cn/maint-lf-to-crlf-filter' into maint

* cn/maint-lf-to-crlf-filter:
  lf_to_crlf_filter(): tell the caller we added "\n" when draining
  convert: track state in LF-to-CRLF filter

12 years agoMerge branch 'jk/maint-upload-archive' into maint
Junio C Hamano [Wed, 21 Dec 2011 19:42:44 +0000 (11:42 -0800)]
Merge branch 'jk/maint-upload-archive' into maint

* jk/maint-upload-archive:
  archive: don't let remote clients get unreachable commits

12 years agoclone: the -o option has nothing to do with <branch>
Carlos Martín Nieto [Wed, 21 Dec 2011 18:14:09 +0000 (18:14 +0000)]
clone: the -o option has nothing to do with <branch>

It is to give an alternate <name> instead of "origin" to the remote
we are cloning from.

Signed-off-by: Carlos Martín Nieto <cmn@elego.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agobuiltin/log: remove redundant initialization
Michael Schubert [Wed, 21 Dec 2011 12:05:27 +0000 (13:05 +0100)]
builtin/log: remove redundant initialization

"abbrev" and "commit_format" in struct rev_info get initialized in
init_revisions - no need to reinit in cmd_log_init_defaults.

Signed-off-by: Michael Schubert <mschub@elegosoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'ms/commit-cc-option-helpstring' into maint
Junio C Hamano [Wed, 21 Dec 2011 18:50:20 +0000 (10:50 -0800)]
Merge branch 'ms/commit-cc-option-helpstring' into maint

* ms/commit-cc-option-helpstring:
  builtin/commit: add missing '/' in help message

12 years agobuiltin/commit: add missing '/' in help message
Michael Schubert [Wed, 21 Dec 2011 14:56:00 +0000 (15:56 +0100)]
builtin/commit: add missing '/' in help message

Signed-off-by: Michael Schubert <mschub@elegosoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agobuiltin/init-db.c: eliminate -Wformat warning on Solaris
Ævar Arnfjörð Bjarmason [Tue, 20 Dec 2011 23:27:41 +0000 (23:27 +0000)]
builtin/init-db.c: eliminate -Wformat warning on Solaris

On Solaris systems we'd warn about an implicit cast of mode_t when we
printed things out with the %d format. We'd get this warning under GCC
4.6.0 with Solaris headers:

    builtin/init-db.c: In function ‘separate_git_dir’:
    builtin/init-db.c:354:4: warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has type ‘mode_t’ [-Wformat]

We've been doing this ever since v1.7.4.1-296-gb57fb80. Just work
around this by adding an explicit cast.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agogit-sh-setup: make require_clean_work_tree part of the interface
Thomas Rast [Tue, 20 Dec 2011 21:42:39 +0000 (22:42 +0100)]
git-sh-setup: make require_clean_work_tree part of the interface

92c62a3 (Porcelain scripts: Rewrite cryptic "needs update" error
message, 2010-10-19) refactored git's own checking to a function in
git-sh-setup.  This is a very useful thing for script writers, so
document it.

Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agot/t2023-checkout-m.sh: fix use of test_must_fail
Ævar Arnfjörð Bjarmason [Tue, 20 Dec 2011 20:37:45 +0000 (20:37 +0000)]
t/t2023-checkout-m.sh: fix use of test_must_fail

Change an invocation of test_must_fail() to be inside a
test_expect_success() as is our usual pattern. Having it outside
caused our tests to fail under prove(1) since we wouldn't print a
newline before TAP output:

    CONFLICT (content): Merge conflict in both.txt
    # GETTEXT POISON #ok 2 - -m restores 2-way conflicted+resolved file

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoremote-curl: don't pass back fake refs
Jeff King [Sat, 17 Dec 2011 10:45:39 +0000 (05:45 -0500)]
remote-curl: don't pass back fake refs

When receive-pack advertises its list of refs, it generally hides the
capabilities information after a NUL at the end of the first ref.

However, when we have an empty repository, there are no refs, and
therefore receive-pack writes a fake ref "capabilities^{}" with the
capabilities afterwards.

On the client side, git reads the result with get_remote_heads(). We pick
the capabilities from the end of the line, and then call check_ref() to
make sure the ref name is valid. We see that it isn't, and don't bother
adding it to our list of refs.

However, the call to check_ref() is enabled by passing the REF_NORMAL flag
to get_remote_heads. For the regular git transport, we pass REF_NORMAL in
get_refs_via_connect() if we are doing a push (since only receive-pack
uses this fake ref).  But in remote-curl, we never use this flag, and we
accept the fake ref as a real one, passing it back from the helper to the
parent git-push.

Most of the time this bug goes unnoticed, as the fake ref won't match our
refspecs. However, if "--mirror" is used, then we see it as remote cruft
to be pruned, and try to pass along a deletion refspec for it. Of course
this refspec has bogus syntax (because of the ^{}), and the helper
complains, aborting the push.

Let's have remote-curl mirror what the builtin get_refs_via_connect() does
(at least for the case of using git protocol; we can leave the dumb
info/refs reader as it is).

This also fixes pushing with --mirror to a smart-http remote that uses
alternates. The fake ".have" refs the server gives to avoid unnecessary
network transfer has a similar bad interactions with the machinery.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoadvice: Document that they all default to true
Junio C Hamano [Mon, 19 Dec 2011 05:35:01 +0000 (21:35 -0800)]
advice: Document that they all default to true

By definition, the default value of "advice.*" variables must be true and
they all control various additional help messages that are designed to aid
new users. Setting one to false is to tell Git that the user understands
the nature of the error and does not need the additional verbose help
message.

Also fix the asciidoc markup for linkgit:git-checkout[1] in the
description of the detachedHead advice by removing an excess colon.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agolf_to_crlf_filter(): resurrect CRLF->CRLF hack
Junio C Hamano [Fri, 16 Dec 2011 23:44:18 +0000 (15:44 -0800)]
lf_to_crlf_filter(): resurrect CRLF->CRLF hack

The non-streaming version of the filter counts CRLF and LF in the whole
buffer, and returns without doing anything when they match (i.e. what is
recorded in the object store already uses CRLF). This was done to help
people who added files from the DOS world before realizing they want to go
cross platform and adding .gitattributes to tell Git that they only want
CRLF in their working tree.

The streaming version of the filter does not want to read the whole thing
before starting to work, as that defeats the whole point of streaming. So
we instead check what byte follows CR whenever we see one, and add CR
before LF only when the LF does not immediately follow CR already to keep
CRLF as is.

Reported-and-tested-by: Ralf Thielow
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoUpdate jk/maint-strbuf-missing-init to builtin/ rename
Junio C Hamano [Sun, 18 Dec 2011 08:27:59 +0000 (00:27 -0800)]
Update jk/maint-strbuf-missing-init to builtin/ rename

12 years agocommit, merge: initialize static strbuf
Jeff King [Sun, 18 Dec 2011 05:03:22 +0000 (00:03 -0500)]
commit, merge: initialize static strbuf

Strbufs cannot rely on static all-zero initialization; instead, they must
use STRBUF_INIT to point to the "slopbuf".

Without this patch, "git commit --no-message" segfaults reliably. Fix the
same issue in builtin/merge.c as well.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agomake "git push -v" actually verbose
Jeff King [Sat, 17 Dec 2011 09:37:15 +0000 (04:37 -0500)]
make "git push -v" actually verbose

Providing a single "-v" to "git push" currently does
nothing. Giving two flags ("git push -v -v") turns on the
first level of verbosity.

This is caused by a regression introduced in 8afd8dc (push:
support multiple levels of verbosity, 2010-02-24). Before
the series containing 8afd8dc, the verbosity handling for
fetching and pushing was completely separate. Commit bde873c
refactored the verbosity handling out of the fetch side, and
then 8afd8dc converted push to use the refactored code.

However, the fetch and push sides numbered and passed along
their verbosity levels differently. For both, a verbosity
level of "-1" meant "quiet", and "0" meant "default output".
But from there they differed.

For fetch, a verbosity level of "1" indicated to the "fetch"
program that it should make the status table slightly more
verbose, showing up-to-date entries. A verbosity level of
"2" meant that we should pass a verbose flag to the
transport; in the case of fetch-pack, this displays protocol
debugging information.

As a result, the refactored code in bde873c checks for
"verbosity >= 2", and only then passes it on to the
transport. From the transport code's perspective, a
verbosity of 0 or 1 both meant "0".

Push, on the other hand, does not show its own status table;
that is always handled by the transport layer or below
(originally send-pack itself, but these days it is done by
the transport code). So a verbosity level of 1 meant that we
should pass the verbose flag to send-pack, so that it knows
we want a verbose status table. However, once 8afd8dc
switched it to the refactored fetch code, a verbosity level
of 1 was now being ignored.  Thus, you needed to
artificially bump the verbosity to 2 (via "-v -v") to have
any effect.

We can fix this by letting the transport code know about the
true verbosity level (i.e., let it distinguish level 0 or
1).

We then have to also make an adjustment to any transport
methods that assumed "verbose > 0" meant they could spew
lots of debugging information. Before, they could only get
"0" or "2", but now they will also receive "1". They need to
adjust their condition for turning on such spew from
"verbose > 0" to "verbose > 1".

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agodocs: brush up obsolete bits of git-fsck manpage
Jeff King [Fri, 16 Dec 2011 11:33:10 +0000 (06:33 -0500)]
docs: brush up obsolete bits of git-fsck manpage

After the description and options, the fsck manpage contains
some discussion about what it does. Over time, this
discussion has become somewhat obsolete, both in content and
formatting. In particular:

  1. There are many options now, so starting the discussion
     with "It tests..." makes it unclear whether we are
     talking about the last option, or about the tool in
     general. Let's start a new "discussion" section and
     make our antecedent more clear.

  2. It gave an example for --unreachable using for-each-ref
     to mention all of the heads, saying that it will do "a
     _lot_ of verification". This is hopelessly out-of-date,
     as giving no arguments will check much more (reflogs,
     the index, non-head refs).

  3. It goes on to mention tests "to be added" (like tree
     object sorting). We now have these tests.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agolf_to_crlf_filter(): tell the caller we added "\n" when draining
Junio C Hamano [Fri, 16 Dec 2011 22:39:37 +0000 (14:39 -0800)]
lf_to_crlf_filter(): tell the caller we added "\n" when draining

This can only happen when the input size is multiple of the
buffer size of the cascade filter (16k) and ends with an LF,
but in such a case, the code forgot to tell the caller that
it added the "\n" it could not add during the last round.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agouse custom rename score during --follow
Jeff King [Fri, 16 Dec 2011 11:27:50 +0000 (06:27 -0500)]
use custom rename score during --follow

If you provide a custom rename score on the command line,
like:

  git log -M50 --follow foo.c

it is completely ignored, and there is no way to --follow
with a looser rename score. Instead, let's use the same
rename score that will be used for generating diffs. This is
convenient, and mirrors what we do with the break-score.

You can see an example of it being useful in git.git:

  $ git log --oneline --summary --follow \
    Documentation/technical/api-string-list.txt
  86d4b52 string-list: Add API to remove an item from an unsorted list
  1d2f80f string_list: Fix argument order for string_list_append
  e242148 string-list: add unsorted_string_list_lookup()
  0dda1d1 Fix two leftovers from path_list->string_list
  c455c87 Rename path_list to string_list
   create mode 100644 Documentation/technical/api-string-list.txt

  $ git log --oneline --summary -M40 --follow \
  Documentation/technical/api-string-list.txt
  86d4b52 string-list: Add API to remove an item from an unsorted list
  1d2f80f string_list: Fix argument order for string_list_append
  e242148 string-list: add unsorted_string_list_lookup()
  0dda1d1 Fix two leftovers from path_list->string_list
  c455c87 Rename path_list to string_list
   rename Documentation/technical/{api-path-list.txt => api-string-list.txt} (47%)
  328a475 path-list documentation: document all functions and data structures
  530e741 Start preparing the API documents.
   create mode 100644 Documentation/technical/api-path-list.txt

You could have two separate rename scores, one for following
and one for diff. But almost nobody is going to want that,
and it would just be unnecessarily confusing. Besides which,
we re-use the diff results from try_to_follow_renames for
the actual diff output, which means having them as separate
scores is actively wrong. E.g., with the current code, you
get:

  $ git log --oneline --diff-filter=R --name-status \
            -M90 --follow git.spec.in
  27dedf0 GIT 0.99.9j aka 1.0rc3
  R084    git-core.spec.in        git.spec.in
  f85639c Rename the RPM from "git" to "git-core"
  R098    git.spec.in     git-core.spec.in

The first one should not be considered a rename by the -M
score we gave, but we print it anyway, since we blindly
re-use the diff information from the follow (which uses the
default score). So this could also be considered simply a
bug-fix, as with the current code "-M" is completely ignored
when using "--follow".

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agocheckout_merged(): squelch false warning from some gcc
Junio C Hamano [Thu, 15 Dec 2011 18:10:11 +0000 (10:10 -0800)]
checkout_merged(): squelch false warning from some gcc

gcc 4.6.2 (there may be others) does not realize that the variable "mode"
can never be used uninitialized in this function and issues a false warning
under -Wuninitialized option.

Squelch it with an unnecessary initialization; it is not like a single
assignment matters to the performance in this codepath that writes out
to the filesystem with checkout_entry() anyway.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agocompat/setenv.c: error if name contains '='
Erik Faye-Lund [Wed, 14 Dec 2011 14:07:09 +0000 (15:07 +0100)]
compat/setenv.c: error if name contains '='

According to POSIX, setenv should error out with EINVAL if it's
asked to set an environment variable whose name contains an equals
sign. Implement this detail in our compatibility-fallback.

Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agocompat/setenv.c: update errno when erroring out
Erik Faye-Lund [Wed, 14 Dec 2011 14:07:08 +0000 (15:07 +0100)]
compat/setenv.c: update errno when erroring out

Previously, gitsetenv didn't update errno as it should when
erroring out. Fix this.

Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoUpdate draft release notes for 1.7.8.1
Junio C Hamano [Wed, 14 Dec 2011 06:08:52 +0000 (22:08 -0800)]
Update draft release notes for 1.7.8.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'jc/maint-pack-object-cycle' into maint
Junio C Hamano [Wed, 14 Dec 2011 06:04:50 +0000 (22:04 -0800)]
Merge branch 'jc/maint-pack-object-cycle' into maint

* jc/maint-pack-object-cycle:
  pack-object: tolerate broken packs that have duplicated objects

Conflicts:
builtin/pack-objects.c

12 years agoMerge branch 'jc/index-pack-reject-dups' into maint
Junio C Hamano [Wed, 14 Dec 2011 06:03:36 +0000 (22:03 -0800)]
Merge branch 'jc/index-pack-reject-dups' into maint

* jc/index-pack-reject-dups:
  receive-pack, fetch-pack: reject bogus pack that records objects twice

12 years agoMerge branch 'mf/curl-select-fdset' into maint
Junio C Hamano [Wed, 14 Dec 2011 06:03:17 +0000 (22:03 -0800)]
Merge branch 'mf/curl-select-fdset' into maint

* mf/curl-select-fdset:
  http: drop "local" member from request struct
  http.c: Rely on select instead of tracking whether data was received
  http.c: Use timeout suggested by curl instead of fixed 50ms timeout
  http.c: Use curl_multi_fdset to select on curl fds instead of just sleeping

12 years agoMerge branch 'nd/misc-cleanups' into maint
Junio C Hamano [Wed, 14 Dec 2011 06:02:51 +0000 (22:02 -0800)]
Merge branch 'nd/misc-cleanups' into maint

* nd/misc-cleanups:
  unpack_object_header_buffer(): clear the size field upon error
  tree_entry_interesting: make use of local pointer "item"
  tree_entry_interesting(): give meaningful names to return values
  read_directory_recursive: reduce one indentation level
  get_tree_entry(): do not call find_tree_entry() on an empty tree
  tree-walk.c: do not leak internal structure in tree_entry_len()

12 years agoMerge branch 'maint-1.7.7' into maint
Junio C Hamano [Wed, 14 Dec 2011 05:58:51 +0000 (21:58 -0800)]
Merge branch 'maint-1.7.7' into maint

* maint-1.7.7:
  Git 1.7.7.5
  Git 1.7.6.5
  blame: don't overflow time buffer
  fetch: create status table using strbuf
  checkout,merge: loosen overwriting untracked file check based on info/exclude
  cast variable in call to free() in builtin/diff.c and submodule.c
  apply: get rid of useless x < 0 comparison on a size_t type

Conflicts:
Documentation/git.txt
GIT-VERSION-GEN
RelNotes
builtin/fetch.c

12 years agoGit 1.7.7.5 v1.7.7.5
Junio C Hamano [Wed, 14 Dec 2011 05:55:31 +0000 (21:55 -0800)]
Git 1.7.7.5

Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'ab/clang-lints' into maint-1.7.7
Junio C Hamano [Wed, 14 Dec 2011 05:47:51 +0000 (21:47 -0800)]
Merge branch 'ab/clang-lints' into maint-1.7.7

* ab/clang-lints:
  cast variable in call to free() in builtin/diff.c and submodule.c
  apply: get rid of useless x < 0 comparison on a size_t type

12 years agoMerge branch 'nd/maint-ignore-exclude' into maint-1.7.7
Junio C Hamano [Wed, 14 Dec 2011 05:47:08 +0000 (21:47 -0800)]
Merge branch 'nd/maint-ignore-exclude' into maint-1.7.7

* nd/maint-ignore-exclude:
  checkout,merge: loosen overwriting untracked file check based on info/exclude

12 years agoMerge branch 'maint-1.7.6' into maint-1.7.7
Junio C Hamano [Wed, 14 Dec 2011 05:44:56 +0000 (21:44 -0800)]
Merge branch 'maint-1.7.6' into maint-1.7.7

* maint-1.7.6:
  Git 1.7.6.5
  blame: don't overflow time buffer
  fetch: create status table using strbuf

Conflicts:
Documentation/git.txt
GIT-VERSION-GEN
RelNotes

12 years agoGit 1.7.6.5 v1.7.6.5
Junio C Hamano [Wed, 14 Dec 2011 05:30:40 +0000 (21:30 -0800)]
Git 1.7.6.5

Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'jk/maint-fetch-status-table' into maint-1.7.6
Junio C Hamano [Wed, 14 Dec 2011 05:21:30 +0000 (21:21 -0800)]
Merge branch 'jk/maint-fetch-status-table' into maint-1.7.6

* jk/maint-fetch-status-table:
  fetch: create status table using strbuf

12 years agoMerge branch 'jc/maint-name-rev-all' into maint-1.7.6
Junio C Hamano [Wed, 14 Dec 2011 05:12:34 +0000 (21:12 -0800)]
Merge branch 'jc/maint-name-rev-all' into maint-1.7.6

* jc/maint-name-rev-all:
  name-rev --all: do not even attempt to describe non-commit object

12 years agoMerge branch 'ml/mailmap' into maint-1.7.6
Junio C Hamano [Wed, 14 Dec 2011 05:12:14 +0000 (21:12 -0800)]
Merge branch 'ml/mailmap' into maint-1.7.6

* ml/mailmap:
  mailmap: xcalloc mailmap_info

Conflicts:
mailmap.c

12 years agoblame: don't overflow time buffer
Jeff King [Thu, 8 Dec 2011 10:25:54 +0000 (05:25 -0500)]
blame: don't overflow time buffer

When showing the raw timestamp, we format the numeric
seconds-since-epoch into a buffer, followed by the timezone
string. This string has come straight from the commit
object. A well-formed object should have a timezone string
of only a few bytes, but we could be operating on data
pushed by a malicious user.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agohttp-push: enable "proactive auth"
Jeff King [Wed, 14 Dec 2011 00:11:56 +0000 (19:11 -0500)]
http-push: enable "proactive auth"

Before commit 986bbc08, git was proactive about asking for
http passwords. It assumed that if you had a username in
your URL, you would also want a password, and asked for it
before making any http requests.

However, this could interfere with the use of .netrc (see
986bbc08 for details). And it was also unnecessary, since
the http fetching code had learned to recognize an HTTP 401
and prompt the user then. Furthermore, the proactive prompt
could interfere with the usage of .netrc (see 986bbc08 for
details).

Unfortunately, the http push-over-DAV code never learned to
recognize HTTP 401, and so was broken by this change. This
patch does a quick fix of re-enabling the "proactive auth"
strategy only for http-push, leaving the dumb http fetch and
smart-http as-is.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agot5540: test DAV push with authentication
Jeff King [Tue, 13 Dec 2011 20:17:04 +0000 (15:17 -0500)]
t5540: test DAV push with authentication

We don't currently test this case at all, and instead just
test the DAV mechanism over an unauthenticated push. That
isn't very realistic, as most people will want to
authenticate pushes.

Two of the tests expect_failure as they reveal bugs:

  1. Pushing without a username in the URL fails to ask for
     credentials when we get an HTTP 401. This has always
     been the case, but it would be nice if it worked like
     smart-http.

  2. Pushing with a username fails to ask for the password
     since 986bbc0 (http: don't always prompt for password,
     2011-11-04). This is a severe regression in v1.7.8, as
     authenticated push-over-DAV is now totally unusable
     unless you have credentials in your .netrc.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoconnect.c: drop path_match function
Jeff King [Tue, 13 Dec 2011 00:49:59 +0000 (19:49 -0500)]
connect.c: drop path_match function

This function was used for comparing local and remote ref
names during fetch (which makes it a candidate for "most
confusingly named function of the year").

It no longer has any callers, so let's get rid of it.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agofetch-pack: match refs exactly
Jeff King [Tue, 13 Dec 2011 00:48:08 +0000 (19:48 -0500)]
fetch-pack: match refs exactly

When we are determining the list of refs to fetch via
fetch-pack, we have two sets of refs to compare: those on
the remote side, and a "match" list of things we want to
fetch. We iterate through the remote refs alphabetically,
seeing if each one is wanted by the "match" list.

Since def88e9 (Commit first cut at "git-fetch-pack",
2005-07-04), we have used the "path_match" function to do a
suffix match, where a remote ref is considered wanted if
any of the "match" elements is a suffix of the remote
refname.

This enables callers of fetch-pack to specify unqualified
refs and have them matched up with remote refs (e.g., ask
for "A" and get remote's "refs/heads/A"). However, if you
provide a fully qualified ref, then there are corner cases
where we provide the wrong answer. For example, given a
remote with two refs:

   refs/foo/refs/heads/master
   refs/heads/master

asking for "refs/heads/master" will first match
"refs/foo/refs/heads/master" by the suffix rule, and we will
erroneously fetch it instead of refs/heads/master.

As it turns out, all callers of fetch_pack do provide
fully-qualified refs for the match list. There are two ways
fetch_pack can get match lists:

  1. Through the transport code (i.e., via git-fetch)

  2. On the command-line of git-fetch-pack

In the first case, we will always be providing the names of
fully-qualified refs from "struct ref" objects. We will have
pre-matched those ref objects already (since we have to
handle more advanced matching, like wildcard refspecs), and
are just providing a list of the refs whose objects we need.

In the second case, users could in theory be providing
non-qualified refs on the command-line. However, the
fetch-pack documentation claims that refs should be fully
qualified (and has always done so since it was written in
2005).

Let's change this path_match call to simply check for string
equality, matching what the callers of fetch_pack are
expecting.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agot5500: give fully-qualified refs to fetch-pack
Jeff King [Tue, 13 Dec 2011 00:44:40 +0000 (19:44 -0500)]
t5500: give fully-qualified refs to fetch-pack

The fetch-pack documentation is very clear that refs given
on the command line are to be full refs:

  <refs>...::
          The remote heads to update from. This is relative to
          $GIT_DIR (e.g. "HEAD", "refs/heads/master").  When
          unspecified, update from all heads the remote side has.

and this has been the case since fetch-pack was originally documented in
8b3d9dc ([PATCH] Documentation: clone/fetch/upload., 2005-07-14).

Let's follow our own documentation to set a good example,
and to avoid breaking when this restriction is enforced in
the next patch.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agodrop "match" parameter from get_remote_heads
Jeff King [Tue, 13 Dec 2011 00:41:37 +0000 (19:41 -0500)]
drop "match" parameter from get_remote_heads

The get_remote_heads function reads the list of remote refs
during git protocol session. It dates all the way back to
def88e9 (Commit first cut at "git-fetch-pack", 2005-07-04).
At that time, the idea was to come up with a list of refs we
were interested in, and then filter the list as we got it
from the remote side.

Later, 1baaae5 (Make maximal use of the remote refs,
2005-10-28) stopped filtering at the get_remote_heads layer,
letting us use the non-matching refs to find common history.

As a result, all callers now simply pass an empty match
list (and any future callers will want to do the same). So
let's drop these now-useless parameters.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoUpdate documentation for stripspace
Conrad Irwin [Mon, 12 Dec 2011 23:52:51 +0000 (15:52 -0800)]
Update documentation for stripspace

Tell the user what this command is intended for, and expand the
description of what it does.

Signed-off-by: Conrad Irwin <conrad.irwin@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agomv: be quiet about overwriting
Jeff King [Mon, 12 Dec 2011 21:54:42 +0000 (16:54 -0500)]
mv: be quiet about overwriting

When a user asks us to force a mv and overwrite the
destination, we print a warning. However, since a typical
use would be:

  $ git mv one two
  fatal: destination exists, source=one, destination=two
  $ git mv -f one two
  warning: overwriting 'two'

this warning is just noise. We already know we're
overwriting; that's why we gave -f!

This patch silences the warning unless "--verbose" is given.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agomv: improve overwrite warning
Jeff King [Mon, 12 Dec 2011 21:54:17 +0000 (16:54 -0500)]
mv: improve overwrite warning

When we try to "git mv" over an existing file, the error
message is fairly informative:

  $ git mv one two
  fatal: destination exists, source=one, destination=two

When the user forces the overwrite, we give a warning:

  $ git mv -f one two
  warning: destination exists; will overwrite!

This is less informative, but still sufficient in the simple
rename case, as there is only one rename happening.

But when moving files from one directory to another, it
becomes useless:

  $ mkdir three
  $ touch one two three/one
  $ git add .
  $ git mv one two three
  fatal: destination exists, source=one, destination=three/one
  $ git mv -f one two three
  warning: destination exists; will overwrite!

The first message is helpful, but the second one gives us no
clue about what was overwritten. Let's mention the name of
the destination file:

  $ git mv -f one two three
  warning: overwriting 'three/one'

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agorevert: stop creating and removing sequencer-old directory
Jonathan Nieder [Sat, 10 Dec 2011 13:06:12 +0000 (07:06 -0600)]
revert: stop creating and removing sequencer-old directory

Now that "git reset" no longer implicitly removes .git/sequencer that
the operator may or may not have wanted to keep, the logic to write a
backup copy of .git/sequencer and remove it when stale is not needed
any more.  Simplify the sequencer API and repository layout by
dropping it.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoRevert "reset: Make reset remove the sequencer state"
Jonathan Nieder [Sat, 10 Dec 2011 13:03:48 +0000 (07:03 -0600)]
Revert "reset: Make reset remove the sequencer state"

This reverts commit 95eb88d8ee588d89b4f06d2753ed4d16ab13b39f, which
was a UI experiment that did not reflect how "git reset" actually gets
used.  The reversion also fixes a test, indicated in the patch.

Encouraged-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agorevert: do not remove state until sequence is finished
Jonathan Nieder [Sat, 10 Dec 2011 13:02:12 +0000 (07:02 -0600)]
revert: do not remove state until sequence is finished

As v1.7.8-rc0~141^2~4 (2011-08-04) explains, git cherry-pick removes
the sequencer state just before applying the final patch.  In the
single-pick case, that was a good thing, since --abort and --continue
work fine without access to such state and removing it provides a
signal that git should not complain about the need to clobber it ("a
cherry-pick or revert is already in progress") in sequences like the
following:

git cherry-pick foo
git read-tree -m -u HEAD; # forget that; let's try a different one
git cherry-pick bar

After the recent patch "allow single-pick in the middle of cherry-pick
sequence" we don't need that hack any more.  In the new regime, a
traditional "git cherry-pick <commit>" command never looks at
.git/sequencer, so we do not need to cripple "git cherry-pick
<commit>..<commit>" for it any more.

So now you can run "git cherry-pick --abort" near the end of a
multi-pick sequence and it will abort the entire sequence, instead of
misbehaving and aborting just the final commit.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agorevert: allow single-pick in the middle of cherry-pick sequence
Jonathan Nieder [Sat, 10 Dec 2011 12:59:48 +0000 (06:59 -0600)]
revert: allow single-pick in the middle of cherry-pick sequence

After messing up a difficult conflict resolution in the middle of a
cherry-pick sequence, it can be useful to be able to

git checkout HEAD . && git cherry-pick that-one-commit

to restart the conflict resolution. The current code however errors out
saying that another cherry-pick is already in progress.

Suggested-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agorevert: pass around rev-list args in already-parsed form
Jonathan Nieder [Sat, 13 Aug 2011 17:06:23 +0000 (12:06 -0500)]
revert: pass around rev-list args in already-parsed form

Since 7e2bfd3f (revert: allow cherry-picking more than one commit,
2010-07-02), the pick/revert machinery has kept track of the set of
commits to be cherry-picked or reverted using commit_argc and
commit_argv variables, storing the corresponding command-line
parameters.

Future callers as other commands are built in (am, rebase, sequencer)
may find it easier to pass rev-list options to this machinery in
already-parsed form.  Teach cmd_cherry_pick and cmd_revert to parse
the rev-list arguments in advance and pass the commit set to
pick_revisions() as a rev_info structure.

Original patch by Jonathan, tweaks and test from Ram.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Improved-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agorevert: allow cherry-pick --continue to commit before resuming
Jonathan Nieder [Sat, 10 Dec 2011 12:49:25 +0000 (06:49 -0600)]
revert: allow cherry-pick --continue to commit before resuming

When "git cherry-pick ..bar" encounters conflicts, permit the operator
to use cherry-pick --continue after resolving them as a shortcut for
"git commit && git cherry-pick --continue" to record the resolution
and carry on with the rest of the sequence.

This improves the analogy with "git rebase" (in olden days --continue
was the way to preserve authorship when a rebase encountered
conflicts) and fits well with a general UI goal of making "git cmd
--continue" save humans the trouble of deciding what to do next.

Example: after encountering a conflict from running "git cherry-pick
foo bar baz":

CONFLICT (content): Merge conflict in main.c
error: could not apply f78a8d98c... bar!
hint: after resolving the conflicts, mark the corrected paths
hint: with 'git add <paths>' or 'git rm <paths>'
hint: and commit the result with 'git commit'

We edit main.c to resolve the conflict, mark it acceptable with "git
add main.c", and can run "cherry-pick --continue" to resume the
sequence.

$ git cherry-pick --continue
[editor opens to confirm commit message]
[master 78c8a8c98] bar!
 1 files changed, 1 insertions(+), 1 deletions(-)
[master 87ca8798c] baz!
 1 files changed, 1 insertions(+), 1 deletions(-)

This is done for both codepaths to pick multiple commits and a single
commit.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agorevert: give --continue handling its own function
Jonathan Nieder [Sat, 10 Dec 2011 12:47:36 +0000 (06:47 -0600)]
revert: give --continue handling its own function

This makes pick_revisions() a little shorter and easier to read
straight through.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agomv: make non-directory destination error more clear
Jeff King [Mon, 12 Dec 2011 07:51:36 +0000 (02:51 -0500)]
mv: make non-directory destination error more clear

If you try to "git mv" multiple files onto another
non-directory file, you confusingly get the "usage" message:

  $ touch one two three
  $ git add .
  $ git mv one two three
  usage: git mv [options] <source>... <destination>
  [...]

From the user's perspective, that makes no sense. They just
gave parameters that exactly match that usage!

This behavior dates back to the original C version of "git
mv", which had a usage message like:

  usage: git mv (<source> <destination> | <source>...  <destination>)

This was slightly less confusing, because it at least
mentions that there are two ways to invoke (but it still
isn't clear why what the user provided doesn't work).

Instead, let's show an error message like:

  $ git mv one two three
  fatal: destination 'three' is not a directory

We could leave the usage message in place, too, but it
doesn't actually help here. It contains no hints that there
are two forms, nor that multi-file form requires that the
endpoint be a directory. So it just becomes useless noise
that distracts from the real error.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agomv: honor --verbose flag
Jeff King [Mon, 12 Dec 2011 07:51:24 +0000 (02:51 -0500)]
mv: honor --verbose flag

The code for a verbose flag has been here since "git mv" was
converted to C many years ago, but actually getting the "-v"
flag from the command line was accidentally lost in the
transition.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agodocs: mention "-k" for both forms of "git mv"
Jeff King [Mon, 12 Dec 2011 07:50:31 +0000 (02:50 -0500)]
docs: mention "-k" for both forms of "git mv"

The "git mv" synopsis shows two forms: renaming a file, and
moving files into a directory. They can both make use of the
"-k" flag to ignore errors, so mention it in both places.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agocompat/snprintf: don't look at va_list twice
Jeff King [Mon, 12 Dec 2011 14:25:51 +0000 (09:25 -0500)]
compat/snprintf: don't look at va_list twice

If you define SNPRINTF_RETURNS_BOGUS, we use a special
git_vsnprintf wrapper assumes that vsnprintf returns "-1"
instead of the number of characters that you would need to
store the result.

To do this, it invokes vsnprintf multiple times, growing a
heap buffer until we have enough space to hold the result.
However, this means we evaluate the va_list parameter
multiple times, which is generally a bad thing (it may be
modified by calls to vsnprintf, yielding undefined
behavior).

Instead, we must va_copy it and hand the copy to vsnprintf,
so we always have a pristine va_list.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoTest 'checkout -m -- path'
Pete Harlan [Wed, 7 Dec 2011 07:01:28 +0000 (23:01 -0800)]
Test 'checkout -m -- path'

Signed-off-by: Pete Harlan <pgit@pcharlan.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agofetch: create status table using strbuf
Jeff King [Thu, 8 Dec 2011 08:43:19 +0000 (03:43 -0500)]
fetch: create status table using strbuf

When we fetch from a remote, we print a status table like:

  From url
   * [new branch]   foo -> origin/foo

We create this table in a static buffer using sprintf. If
the remote refnames are long, they can overflow this buffer
and smash the stack.

Instead, let's use a strbuf to build the string.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agoMerge branch 'maint-1.7.7' into maint
Junio C Hamano [Fri, 9 Dec 2011 21:33:39 +0000 (13:33 -0800)]
Merge branch 'maint-1.7.7' into maint

* maint-1.7.7:
  am: don't persist keepcr flag
  mingw: give waitpid the correct signature
  git symbolic-ref: documentation fix

12 years agoam: don't persist keepcr flag
Martin von Zweigbergk [Fri, 9 Dec 2011 07:30:45 +0000 (23:30 -0800)]
am: don't persist keepcr flag

The keepcr flag is only used in the split_patches function, which is
only called before a patch application has to stopped for user input,
not after resuming. It is therefore unnecessary to persist the
flag. This seems to have been the case since it was introduced in
ad2c928 (git-am: Add command line parameter `--keep-cr` passing it to
git-mailsplit, 2010-02-27).

Signed-off-by: Martin von Zweigbergk <martin.von.zweigbergk@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
12 years agomingw: give waitpid the correct signature
Erik Faye-Lund [Thu, 8 Dec 2011 19:39:57 +0000 (20:39 +0100)]
mingw: give waitpid the correct signature

POSIX says that last parameter to waitpid should be 'int',
so let's make it so.

Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>