* kk/maint-prefix-in-config-mak:
  config.mak.in: allow "configure --sysconfdir=/else/where"

* jk/rebase-head-reflog:
  rebase: write a reflog entry when finishing
  rebase: create HEAD reflog entry when aborting

* jk/maint-docs:
  docs: fix some antique example output
  docs: make sure literal "->" isn't converted to arrow
  docs: update status --porcelain format
  docs: minor grammar fixes to git-status

* jk/read-in-full-stops-on-error:
  read_in_full: always report errors

* jk/maint-remote-mirror-safer:
  remote: allow "-t" with fetch mirrors

* jl/read-tree-m-dry-run:
  Teach read-tree the -n|--dry-run option
  unpack-trees: add the dry_run flag to unpack_trees_options

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

* tr/add-i-no-escape:
  add -i: ignore terminal escape sequences

* vh/config-interactive-singlekey-doc:
  git-reset.txt: better docs for '--patch'
  git-checkout.txt: better docs for '--patch'
  git-stash.txt: better docs for '--patch'
  git-add.txt: document 'interactive.singlekey'
  config.txt: 'interactive.singlekey; is used by...

* ml/test-readme:
  t/README: unify documentation of test function args

* ab/i18n-fixup: (24 commits)
  i18n: use test_i18n{cmp,grep} in t7600, t7607, t7611 and t7811
  i18n: use test_i18n{grep,cmp} in t7508
  i18n: use test_i18ngrep in t7506
  i18n: use test_i18ngrep and test_i18ncmp in t7502
  i18n: use test_i18ngrep in t7501
  i18n: use test_i18ncmp in t7500
  i18n: use test_i18ngrep in t7201
  i18n: use test_i18ncmp and test_i18ngrep in t7102 and t7110
  i18n: use test_i18ncmp and test_i18ngrep in t5541, t6040, t6120, t7004, t7012 and t7060
  i18n: use test_i18ncmp and test_i18ngrep in t3700, t4001 and t4014
  i18n: use test_i18ncmp and test_i18ngrep in t3203, t3501 and t3507
  i18n: use test_i18ngrep in t2020, t2204, t3030, and t3200
  i18n: use test_i18ngrep in lib-httpd and t2019
  i18n: do not overuse C_LOCALE_OUTPUT (grep)
  i18n: use test_i18ncmp in t1200 and t2200
  i18n: .git file is not a human readable message (t5601)
  i18n: do not overuse C_LOCALE_OUTPUT
  i18n: mark init-db messages for translation
  i18n: mark checkout plural warning for translation
  i18n: mark checkout --detach messages for translation
  ...

* jc/rename-degrade-cc-to-c:
  diffcore-rename: fall back to -C when -C -C busts the rename limit
  diffcore-rename: record filepair for rename src
  diffcore-rename: refactor "too many candidates" logic
  builtin/diff.c: remove duplicated call to diff_result_code()

* rr/doc-content-type:
  Documentation: Allow custom diff tools to be specified in 'diff.tool'
  Documentation: Add diff.<driver>.* to config
  Documentation: Move diff.<driver>.* from config.txt to diff-config.txt
  Documentation: Add filter.<driver>.* to config

Commit 8f323c00 (drop support for GIT_CONFIG_NOGLOBAL, 15-03-2011)
removed the git_config_global() function, among other things, since
it is no longer required. Unfortunately, this function has since
been unintentionally restored by a faulty conflict resolution.

Remove it.
Signed-off-by: NRamsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* jc/fmt-req-fix:
  userformat_find_requirements(): find requirement for the correct format

* jk/maint-config-alias-fix:
  handle_options(): do not miscount how many arguments were used
  config: always parse GIT_CONFIG_PARAMETERS during git_config
  git_config: don't peek at global config_parameters
  config: make environment parsing routines static

Conflicts:
	config.c

* maint:
  git-submodule.sh: separate parens by a space to avoid confusing some shells
  Documentation/technical/api-diff.txt: correct name of diff_unmerge()
  read_gitfile_gently: use ssize_t to hold read result
  remove tests of always-false condition
  rerere.c: diagnose a corrupt MERGE_RR when hitting EOF between TAB and '\0'

* jm/maint-misc-fix:
  read_gitfile_gently: use ssize_t to hold read result
  remove tests of always-false condition
  rerere.c: diagnose a corrupt MERGE_RR when hitting EOF between TAB and '\0'

* bc/maint-submodule-fix-parked:
  git-submodule.sh: separate parens by a space to avoid confusing some shells

* bc/maint-api-doc-parked:
  Documentation/technical/api-diff.txt: correct name of diff_unmerge()

* mk/grep-pcre:
  git-grep: Fix problems with recently added tests
  git-grep: Update tests (mainly for -P)
  Makefile: Pass USE_LIBPCRE down in GIT-BUILD-OPTIONS
  git-grep: update tests now regexp type is "last one wins"
  git-grep: do not die upon -F/-P when grep.extendedRegexp is set.
  git-grep: Bail out when -P is used with -F or -E
  grep: Add basic tests
  configure: Check for libpcre
  git-grep: Learn PCRE
  grep: Extract compile_regexp_failed() from compile_regexp()
  grep: Fix a typo in a comment
  grep: Put calls to fixmatch() and regmatch() into patmatch()
  contrib/completion: --line-number to git grep
  Documentation: Add --line-number to git-grep synopsis

Brian Gernhardt reported that test 'git grep -E -F -G a\\+b' fails on
OS X 10.6.7. This is because I assumed \+ is part of BRE, which isn't
true on all platforms.

The easiest way to make this test pass is to just update expected
output, but that would make the test pointless. Its real purpose is to
check whether 'git grep -E -F -G' is different from 'git grep -E -G -F'.
To check that, let's change pattern to "a+b*c". This should return
different match for -G, -F and -E.

I also made two small tweaks to the tests. First, I added path "ab" to
all calls to future-proof tests. Second, I updated last two tests to
better show that 'git grep -P -E' is different from 'git grep -E -P'.
Signed-off-by: NMichał Kiedrowicz <michal.kiedrowicz@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* jc/notes-batch-removal:
  show: --ignore-missing
  notes remove: --stdin reads from the standard input
  notes remove: --ignore-missing
  notes remove: allow removing more than one

* jk/haves-from-alternate-odb:
  receive-pack: eliminate duplicate .have refs
  bisect: refactor sha1_array into a generic sha1 list
  refactor refs_from_alternate_cb to allow passing extra data

* jn/run-command-error-failure:
  run-command: handle short writes and EINTR in die_child
  tests: check error message from run_command

When we finish a rebase, our detached HEAD is at the final
result. We update the original branch ref with this result,
and then point the HEAD symbolic ref at the updated branch.
We write a reflog for the branch update, but not for the
update of HEAD.

Because we're already at the final result on the detached
HEAD, moving to the branch actually doesn't change our
commit sha1 at all. So in that sense, a reflog entry would
be pointless.

However, humans do read reflogs, and an entry saying "rebase
finished: returning to refs/heads/master" can be helpful in
understanding what is going on.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When we abort a rebase, we return to the original value of
HEAD. Failing to write a reflog entry means we create a
gap in the reflog (which can cause "git show
HEAD@{5.minutes.ago}" to issue a warning). Plus having the
extra entry makes the reflog easier to follow for a human.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

We do allow vanilla Makefile users to say make sysconfdir=/else/where
and config.mak can also be tweaked manually for the same effect. Give
the same configurablity to ./configure users as well.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

These diff-index and diff-tree sample outputs date back to
the first month of git's existence. The output format has
changed slightly since then, so let's have it match the
current output.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Recent versions of asciidoc will treat "->" as a
single-glyph arrow symbol, unless it is inside a literal
code block. This is a problem if we are discussing literal
output and want to show the ASCII characters.

Our usage falls into three categories:

  1. Inside a code block. These can be left as-is.

  2. Discussing literal output or code, but inside a
     paragraph. This patch escapes these as "\->".

  3. Using the arrow as a symbolic element, such as "use the
     Edit->Account Settings menu". In this case, the
     arrow symbol is preferable, so we leave it as-is.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The --porcelain format was originally identical to the
--short format, but designed to be stable as the short
format changed. Since this was written, the short format
picked up a few incompatible niceties, but this description
was never changed.

Let's mention the differences. While we're at it, let's add
some sub-section headings to make the "output" section a
little easier to navigate.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Some shells interpret '(( ))' according to the rules for arithmetic
expansion.  This may not follow POSIX, but is prevalent in commonly used
shells.  Bash does not have a problem with this particular instance of
'((', likely because it is not followed by a '))', but the public domain
ksh does, and so does ksh on IRIX 6.5.

So, add a space between the parenthesis to avoid confusing these shells.
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The read_in_full function repeatedly calls read() to fill a
buffer. If the first read() returns an error, we notify the
caller by returning the error. However, if we read some data
and then get an error on a subsequent read, we simply return
the amount of data that we did read, and the caller is
unaware of the error.

This makes the tradeoff that seeing the partial data is more
important than the fact that an error occurred. In practice,
this is generally not the case; we care more if an error
occurred, and should throw away any partial data.

I audited the current callers. In most cases, this will make
no difference at all, as they do:

  if (read_in_full(fd, buf, size) != size)
	  error("short read");

However, it will help in a few cases:

  1. In sha1_file.c:index_stream, we would fail to notice
     errors in the incoming stream.

  2. When reading symbolic refs in resolve_ref, we would
     fail to notice errors and potentially use a truncated
     ref name.

  3. In various places, we will get much better error
     messages. For example, callers of safe_read would
     erroneously print "the remote end hung up unexpectedly"
     instead of showing the read error.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Commit 13fc2c18 (remote: disallow some nonsensical option
combinations, 2011-03-30) made it impossible to use "remote
add -t foo --mirror". The argument was that specifying
specific branches is useless because:

  1. Push mirrors do not want a refspec at all.

  2. The point of fetch mirroring is to use a broad refspec
     like "refs/*", but using "-t" overrides that.

Point (1) is valid; "-t" with push mirrors is useless. But
point (2) ignored another side effect of using --mirror: it
fetches the refs directly into the refs/ namespace as they
are found upstream, instead of placing them in a
separate-remote layout.

So 13fc2c18 was overly constrictive, and disallowed
reasonable specific-branch mirroring, like:

  git remote add -t heads/foo -t heads/bar --mirror=fetch

which makes the local "foo" and "bar" branches direct
mirrors of the remote, but does not fetch anything else.

This patch restores the original behavior, but only for
fetch mirrors.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Otherwise, a negative error return becomes a very large read
value. We catch this in practice because we compare the
expected and actual numbers of bytes (and you are not likely
to be reading (size_t)-1 bytes), but this makes the
correctness a little more obvious.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* fsck.c (fsck_error_function): Don't test obj->sha1 == 0.
It can never be true, since that sha1 member is an array.
* transport.c (set_upstreams): Likewise for ref->new_sha1.
Signed-off-by: NJim Meyering <meyering@redhat.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJim Meyering <jim@meyering.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>