git
9 years agoMerge branch 'mr/doc-clean-f-f'
Junio C Hamano [Fri, 6 Mar 2015 23:02:26 +0000 (15:02 -0800)] 
Merge branch 'mr/doc-clean-f-f'

Documentation update.

* mr/doc-clean-f-f:
  Documentation/git-clean.txt: document that -f may need to be given twice

9 years agoMerge branch 'ye/http-accept-language'
Junio C Hamano [Fri, 6 Mar 2015 23:02:24 +0000 (15:02 -0800)] 
Merge branch 'ye/http-accept-language'

Compilation fix for a recent topic in 'master'.

* ye/http-accept-language:
  gettext.c: move get_preferred_languages() from http.c

9 years agoSync with 2.3.2
Junio C Hamano [Fri, 6 Mar 2015 22:59:12 +0000 (14:59 -0800)] 
Sync with 2.3.2

* maint:
  Git 2.3.2

9 years agoGit 2.3.2 v2.3.2
Junio C Hamano [Fri, 6 Mar 2015 22:58:14 +0000 (14:58 -0800)] 
Git 2.3.2

Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoMerge branch 'rj/no-xopen-source-for-cygwin' into maint
Junio C Hamano [Fri, 6 Mar 2015 22:57:58 +0000 (14:57 -0800)] 
Merge branch 'rj/no-xopen-source-for-cygwin' into maint

Code cleanups.

* rj/no-xopen-source-for-cygwin:
  git-compat-util.h: remove redundant code

9 years agoMerge branch 'rs/simple-cleanups' into maint
Junio C Hamano [Fri, 6 Mar 2015 22:57:57 +0000 (14:57 -0800)] 
Merge branch 'rs/simple-cleanups' into maint

Code cleanups.

* rs/simple-cleanups:
  sha1_name: use strlcpy() to copy strings
  pretty: use starts_with() to check for a prefix
  for-each-ref: use skip_prefix() to avoid duplicate string comparison
  connect: use strcmp() for string comparison

9 years agoMerge branch 'mm/am-c-doc' into maint
Junio C Hamano [Fri, 6 Mar 2015 22:57:56 +0000 (14:57 -0800)] 
Merge branch 'mm/am-c-doc' into maint

The configuration variable 'mailinfo.scissors' was hard to
discover in the documentation.

* mm/am-c-doc:
  Documentation/git-am.txt: mention mailinfo.scissors config variable
  Documentation/config.txt: document mailinfo.scissors

9 years agoMerge branch 'ew/svn-maint-fixes' into maint
Junio C Hamano [Fri, 6 Mar 2015 22:57:55 +0000 (14:57 -0800)] 
Merge branch 'ew/svn-maint-fixes' into maint

Correct a breakage to git-svn around v2.2 era that triggers
premature closing of FileHandle.

* ew/svn-maint-fixes:
  Git::SVN::*: avoid premature FileHandle closure
  git-svn: fix localtime=true on non-glibc environments

9 years agoMerge branch 'km/send-email-getopt-long-workarounds' into maint
Junio C Hamano [Fri, 6 Mar 2015 22:57:54 +0000 (14:57 -0800)] 
Merge branch 'km/send-email-getopt-long-workarounds' into maint

Even though we officially haven't dropped Perl 5.8 support, the
Getopt::Long package that came with it does not support "--no-"
prefix to negate a boolean option; manually add support to help
people with older Getopt::Long package.

* km/send-email-getopt-long-workarounds:
  git-send-email.perl: support no- prefix with older GetOptions

9 years agoSync with maint
Junio C Hamano [Thu, 5 Mar 2015 21:16:27 +0000 (13:16 -0800)] 
Sync with maint

* maint:
  Prepare for 2.3.2

9 years agoPrepare for 2.3.2
Junio C Hamano [Thu, 5 Mar 2015 21:15:53 +0000 (13:15 -0800)] 
Prepare for 2.3.2

9 years agoMerge branch 'sb/plug-leak-in-make-cache-entry' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:13 +0000 (13:13 -0800)] 
Merge branch 'sb/plug-leak-in-make-cache-entry' into maint

"update-index --refresh" used to leak when an entry cannot be
refreshed for whatever reason.

* sb/plug-leak-in-make-cache-entry:
  read-cache.c: free cache entry when refreshing fails

9 years agoMerge branch 'jk/fast-import-die-nicely-fix' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:12 +0000 (13:13 -0800)] 
Merge branch 'jk/fast-import-die-nicely-fix' into maint

"git fast-import" used to crash when it could not close and
conclude the resulting packfile cleanly.

* jk/fast-import-die-nicely-fix:
  fast-import: avoid running end_packfile recursively

9 years agoMerge branch 'es/blame-commit-info-fix' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:12 +0000 (13:13 -0800)] 
Merge branch 'es/blame-commit-info-fix' into maint

"git blame" died, trying to free an uninitialized piece of memory.

* es/blame-commit-info-fix:
  builtin/blame: destroy initialized commit_info only

9 years agoMerge branch 'ab/merge-file-prefix' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:10 +0000 (13:13 -0800)] 
Merge branch 'ab/merge-file-prefix' into maint

"git merge-file" did not work correctly in a subdirectory.

* ab/merge-file-prefix:
  merge-file: correctly open files when in a subdir

9 years agoMerge branch 'ps/submodule-sanitize-path-upon-add' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:09 +0000 (13:13 -0800)] 
Merge branch 'ps/submodule-sanitize-path-upon-add' into maint

"git submodule add" failed to squash "path/to/././submodule" to
"path/to/submodule".

* ps/submodule-sanitize-path-upon-add:
  git-submodule.sh: fix '/././' path normalization

9 years agoMerge branch 'jk/prune-mtime' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:08 +0000 (13:13 -0800)] 
Merge branch 'jk/prune-mtime' into maint

In v2.2.0, we broke "git prune" that runs in a repository that
borrows from an alternate object store.

* jk/prune-mtime:
  sha1_file: fix iterating loose alternate objects
  for_each_loose_file_in_objdir: take an optional strbuf path

9 years agoMerge branch 'tc/curl-vernum-output-broken-in-7.11' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:07 +0000 (13:13 -0800)] 
Merge branch 'tc/curl-vernum-output-broken-in-7.11' into maint

Certain older vintages of cURL give irregular output from
"curl-config --vernum", which confused our build system.

* tc/curl-vernum-output-broken-in-7.11:
  Makefile: handle broken curl version number in version check

9 years agoMerge branch 'es/squelch-openssl-warnings-on-macosx' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:06 +0000 (13:13 -0800)] 
Merge branch 'es/squelch-openssl-warnings-on-macosx' into maint

An earlier workaround to squelch unhelpful deprecation warnings
from the complier on Mac OSX unnecessarily set minimum required
version of the OS, which the user might want to raise (or lower)
for other reasons.

* es/squelch-openssl-warnings-on-macosx:
  git-compat-util: do not step on MAC_OS_X_VERSION_MIN_REQUIRED

9 years agoMerge branch 'jc/conf-var-doc' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:05 +0000 (13:13 -0800)] 
Merge branch 'jc/conf-var-doc' into maint

Longstanding configuration variable naming rules has been added to
the documentation.

* jc/conf-var-doc:
  CodingGuidelines: describe naming rules for configuration variables
  config.txt: mark deprecated variables more prominently
  config.txt: clarify that add.ignore-errors is deprecated

9 years agoMerge branch 'av/wincred-with-at-in-username-fix' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:04 +0000 (13:13 -0800)] 
Merge branch 'av/wincred-with-at-in-username-fix' into maint

The credential helper for Windows (in contrib/) used to mishandle
a user name with an at-sign in it.

* av/wincred-with-at-in-username-fix:
  wincred: fix get credential if username has "@"

9 years agoMerge branch 'ch/new-gpg-drops-rfc-1991' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:03 +0000 (13:13 -0800)] 
Merge branch 'ch/new-gpg-drops-rfc-1991' into maint

Older GnuPG implementations may not correctly import the keyring
material we prepare for the tests to use.

* ch/new-gpg-drops-rfc-1991:
  t/lib-gpg: sanity-check that we can actually sign
  t/lib-gpg: include separate public keys in keyring.gpg

9 years agoMerge branch 'jc/remote-set-url-doc' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:03 +0000 (13:13 -0800)] 
Merge branch 'jc/remote-set-url-doc' into maint

Clarify in the documentation that "remote.<nick>.pushURL" and
"remote.<nick>.URL" are there to name the same repository accessed
via different transports, not two separate repositories.

* jc/remote-set-url-doc:
  Documentation/git-remote.txt: stress that set-url is not for triangular

9 years agoMerge branch 'jk/pack-bitmap' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:02 +0000 (13:13 -0800)] 
Merge branch 'jk/pack-bitmap' into maint

The pack bitmap support did not build with older versions of GCC.

* jk/pack-bitmap:
  ewah: fix building with gcc < 3.4.0

9 years agoMerge branch 'jk/config-no-ungetc-eof' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:13:00 +0000 (13:13 -0800)] 
Merge branch 'jk/config-no-ungetc-eof' into maint

Reading configuration from a blob object, when it ends with a lone
CR, use to confuse the configuration parser.

* jk/config-no-ungetc-eof:
  config_buf_ungetc: warn when pushing back a random character
  config: do not ungetc EOF

9 years agoMerge branch 'jk/decimal-width-for-uintmax' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:12:59 +0000 (13:12 -0800)] 
Merge branch 'jk/decimal-width-for-uintmax' into maint

We didn't format an integer that wouldn't fit in "int" but in
"uintmax_t" correctly.

* jk/decimal-width-for-uintmax:
  decimal_width: avoid integer overflow

9 years agoMerge branch 'jc/push-cert' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:12:58 +0000 (13:12 -0800)] 
Merge branch 'jc/push-cert' into maint

"git push --signed" gave an incorrectly worded error message when
the other side did not support the capability.

* jc/push-cert:
  transport-helper: fix typo in error message when --signed is not supported

9 years agoMerge branch 'mh/deref-symref-over-helper-transport' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:12:57 +0000 (13:12 -0800)] 
Merge branch 'mh/deref-symref-over-helper-transport' into maint

"git fetch" over a remote-helper that cannot respond to "list"
command could not fetch from a symbolic reference e.g. HEAD.

* mh/deref-symref-over-helper-transport:
  transport-helper: do not request symbolic refs to remote helpers

9 years agoMerge branch 'ks/rebase-i-abbrev' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:12:56 +0000 (13:12 -0800)] 
Merge branch 'ks/rebase-i-abbrev' into maint

The insn sheet "git rebase -i" creates did not fully honor
core.abbrev settings.

* ks/rebase-i-abbrev:
  rebase -i: use full object name internally throughout the script

9 years agoMerge branch 'dp/remove-duplicated-header-inclusion' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:12:55 +0000 (13:12 -0800)] 
Merge branch 'dp/remove-duplicated-header-inclusion' into maint

Code clean-up.

* dp/remove-duplicated-header-inclusion:
  do not include the same header twice

9 years agoMerge branch 'sb/hex-object-name-is-at-most-41-bytes-long' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:12:54 +0000 (13:12 -0800)] 
Merge branch 'sb/hex-object-name-is-at-most-41-bytes-long' into maint

Code clean-up.

* sb/hex-object-name-is-at-most-41-bytes-long:
  hex.c: reduce memory footprint of sha1_to_hex static buffers

9 years agoMerge branch 'ak/git-pm-typofix' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:12:53 +0000 (13:12 -0800)] 
Merge branch 'ak/git-pm-typofix' into maint

Typofix in comments.

* ak/git-pm-typofix:
  Git.pm: two minor typo fixes

9 years agoMerge branch 'jk/sanity' into maint
Junio C Hamano [Thu, 5 Mar 2015 21:12:52 +0000 (13:12 -0800)] 
Merge branch 'jk/sanity' into maint

The tests that wanted to see that file becomes unreadable after
running "chmod a-r file", and the tests that wanted to make sure it
is not run as root, we used "can we write into the / directory?" as
a cheap substitute, but on some platforms that is not a good
heuristics.  The tests and their prerequisites have been updated to
check what they really require.

* jk/sanity:
  test-lib.sh: set prerequisite SANITY by testing what we really need
  tests: correct misuses of POSIXPERM
  t/lib-httpd: switch SANITY check for NOT_ROOT

9 years agoPost 2.3 cycle (batch #7)
Junio C Hamano [Thu, 5 Mar 2015 20:48:18 +0000 (12:48 -0800)] 
Post 2.3 cycle (batch #7)

Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoMerge branch 'ew/svn-fixes'
Junio C Hamano [Thu, 5 Mar 2015 20:45:45 +0000 (12:45 -0800)] 
Merge branch 'ew/svn-fixes'

* ew/svn-fixes:
  git-svn: lazy load some modules

9 years agoMerge branch 'ew/svn-maint-fixes'
Junio C Hamano [Thu, 5 Mar 2015 20:45:45 +0000 (12:45 -0800)] 
Merge branch 'ew/svn-maint-fixes'

Correct a breakage to git-svn around v2.2 era that triggers
premature closing of FileHandle.

* ew/svn-maint-fixes:
  Git::SVN::*: avoid premature FileHandle closure
  git-svn: fix localtime=true on non-glibc environments

9 years agoMerge branch 'tb/connect-ipv6-parse-fix'
Junio C Hamano [Thu, 5 Mar 2015 20:45:44 +0000 (12:45 -0800)] 
Merge branch 'tb/connect-ipv6-parse-fix'

We did not parse username followed by literal IPv6 address in SSH
transport URLs, e.g. ssh://user@[2001:db8::1]:22/repo.git
correctly.

* tb/connect-ipv6-parse-fix:
  t5500: show user name and host in diag-url
  t5601: add more test cases for IPV6
  connect.c: allow ssh://user@[2001:db8::1]/repo.git

9 years agoMerge branch 'jc/diff-test-updates'
Junio C Hamano [Thu, 5 Mar 2015 20:45:43 +0000 (12:45 -0800)] 
Merge branch 'jc/diff-test-updates'

Test clean-up.

* jc/diff-test-updates:
  test_ln_s_add: refresh stat info of fake symbolic links
  t4008: modernise style
  t/diff-lib: check exact object names in compare_diff_raw
  tests: do not borrow from COPYING and README from the real source
  t4010: correct expected object names
  t9300: correct expected object names
  t4008: correct stale comments

9 years agoMerge branch 'rs/simple-cleanups'
Junio C Hamano [Thu, 5 Mar 2015 20:45:42 +0000 (12:45 -0800)] 
Merge branch 'rs/simple-cleanups'

Code cleanups.

* rs/simple-cleanups:
  sha1_name: use strlcpy() to copy strings
  pretty: use starts_with() to check for a prefix
  for-each-ref: use skip_prefix() to avoid duplicate string comparison
  connect: use strcmp() for string comparison

9 years agoMerge branch 'rj/no-xopen-source-for-cygwin'
Junio C Hamano [Thu, 5 Mar 2015 20:45:41 +0000 (12:45 -0800)] 
Merge branch 'rj/no-xopen-source-for-cygwin'

Code cleanups.

* rj/no-xopen-source-for-cygwin:
  git-compat-util.h: remove redundant code

9 years agoMerge branch 'mm/am-c-doc'
Junio C Hamano [Thu, 5 Mar 2015 20:45:40 +0000 (12:45 -0800)] 
Merge branch 'mm/am-c-doc'

The configuration variable 'mailinfo.scissors' was hard to
discover in the documentation.

* mm/am-c-doc:
  Documentation/git-am.txt: mention mailinfo.scissors config variable
  Documentation/config.txt: document mailinfo.scissors

9 years agoMerge branch 'mh/refs-have-new'
Junio C Hamano [Thu, 5 Mar 2015 20:45:39 +0000 (12:45 -0800)] 
Merge branch 'mh/refs-have-new'

Simplify the ref transaction API around how "the ref should be
pointing at this object" is specified.

* mh/refs-have-new:
  refs.h: remove duplication in function docstrings
  update_ref(): improve documentation
  ref_transaction_verify(): new function to check a reference's value
  ref_transaction_delete(): check that old_sha1 is not null_sha1
  ref_transaction_create(): check that new_sha1 is valid
  commit: avoid race when creating orphan commits
  commit: add tests of commit races
  ref_transaction_delete(): remove "have_old" parameter
  ref_transaction_update(): remove "have_old" parameter
  struct ref_update: move "have_old" into "flags"
  refs.c: change some "flags" to "unsigned int"
  refs: remove the gap in the REF_* constant values
  refs: move REF_DELETING to refs.c

9 years agoPost 2.3 cycle (batch #6)
Junio C Hamano [Tue, 3 Mar 2015 22:39:10 +0000 (14:39 -0800)] 
Post 2.3 cycle (batch #6)

Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoMerge branch 'jk/daemon-interpolate'
Junio C Hamano [Tue, 3 Mar 2015 22:37:05 +0000 (14:37 -0800)] 
Merge branch 'jk/daemon-interpolate'

The "interpolated-path" option of "git daemon" inserted any string
client declared on the "host=" capability request without checking.
Sanitize and limit %H and %CH to a saner and a valid DNS name.

* jk/daemon-interpolate:
  daemon: sanitize incoming virtual hostname
  t5570: test git-daemon's --interpolated-path option
  git_connect: let user override virtual-host we send to daemon

9 years agoMerge branch 'rs/daemon-interpolate'
Junio C Hamano [Tue, 3 Mar 2015 22:37:04 +0000 (14:37 -0800)] 
Merge branch 'rs/daemon-interpolate'

"git daemon" looked up the hostname even when "%CH" and "%IP"
interpolations are not requested, which was unnecessary.

* rs/daemon-interpolate:
  daemon: use callback to build interpolated path
  daemon: look up client-supplied hostname lazily

9 years agoMerge branch 'km/send-email-getopt-long-workarounds'
Junio C Hamano [Tue, 3 Mar 2015 22:37:03 +0000 (14:37 -0800)] 
Merge branch 'km/send-email-getopt-long-workarounds'

Even though we officially haven't dropped Perl 5.8 support, the
Getopt::Long package that came with it does not support "--no-"
prefix to negate a boolean option; manually add support to help
people with older Getopt::Long package.

* km/send-email-getopt-long-workarounds:
  git-send-email.perl: support no- prefix with older GetOptions

9 years agoMerge branch 'jc/apply-ws-fix-expands-report'
Junio C Hamano [Tue, 3 Mar 2015 22:37:02 +0000 (14:37 -0800)] 
Merge branch 'jc/apply-ws-fix-expands-report'

"git apply --whitespace=fix" fixed whitespace errors in the common
context lines but did so without reporting.

* jc/apply-ws-fix-expands-report:
  apply: detect and mark whitespace errors in context lines when fixing

9 years agoMerge branch 'jc/apply-beyond-symlink'
Junio C Hamano [Tue, 3 Mar 2015 22:37:01 +0000 (14:37 -0800)] 
Merge branch 'jc/apply-beyond-symlink'

"git apply" was not very careful about reading from, removing,
updating and creating paths outside the working tree (under
--index/--cached) or the current directory (when used as a
replacement for GNU patch).

* jc/apply-beyond-symlink:
  apply: do not touch a file beyond a symbolic link
  apply: do not read from beyond a symbolic link
  apply: do not read from the filesystem under --index
  apply: reject input that touches outside the working area

9 years agogettext.c: move get_preferred_languages() from http.c
Jeff King [Thu, 26 Feb 2015 03:04:16 +0000 (22:04 -0500)] 
gettext.c: move get_preferred_languages() from http.c

Calling setlocale(LC_MESSAGES, ...) directly from http.c, without
including <locale.h>, was causing compilation warnings.  Move the
helper function to gettext.c that already includes the header and
where locale-related issues are handled.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoMerge branch 'svn-maint-fixes' into svn-fixes
Junio C Hamano [Thu, 26 Feb 2015 22:03:57 +0000 (14:03 -0800)] 
Merge branch 'svn-maint-fixes' into svn-fixes

* svn-maint-fixes:
  Git::SVN::*: avoid premature FileHandle closure
  git-svn: fix localtime=true on non-glibc environments

9 years agoGit::SVN::*: avoid premature FileHandle closure
Kyle J. McKay [Thu, 26 Feb 2015 13:49:34 +0000 (05:49 -0800)] 
Git::SVN::*: avoid premature FileHandle closure

Since b19138b (git-svn: Make it incrementally faster by minimizing temp
files, v1.6.0), git-svn has been using the Git.pm temp_acquire and
temp_release mechanism to avoid unnecessary temp file churn and provide
a speed boost.

However, that change introduced a call to temp_acquire inside the
Git::SVN::Fetcher::close_file function for an 'svn_hash' temp file.
Because an SVN::Pool is active at the time this function is called, if
the Git::temp_acquire function ends up actually creating a new
FileHandle for the temp file (which it will the first time it's called
with the name 'svn_hash') that FileHandle will end up in the SVN::Pool
and should that pool have SVN::Pool::clear called on it that FileHandle
will be closed out from under Git::temp_acquire.

Since the only call site to Git::temp_acquire with the name 'svn_hash'
is inside the close_file function, if an 'svn_hash' temp file is ever
created its FileHandle is guaranteed to be created in the active
SVN::Pool.

This has not been a problem in the past because the SVN::Pool was not
being cleared.  However, since dfa72fdb (git-svn: reload RA every
log-window-size, v2.2.0) the pool has been getting cleared periodically
at which point the FileHandle for the 'svn_hash' temp file gets closed.
Any subsequent calls to Git::temp_acquire for 'svn_hash', however,
succeed without creating/opening a new temporary file since it still has
the now invalid FileHandle in its cache.  Callers that then attempt to
use that FileHandle fail with an error.

We avoid this problem by making sure the 'svn_hash' temp file is created
in the same place the 'svn_delta_...' and 'git_blob_...' temp files are
(and then temp_release'd) so that it can be safely used inside the
close_file function without having its FileHandle end up in an SVN::Pool
that gets cleared.

Additionally the Git.pm cat_blob function creates a bidirectional pipe
FileHandle using the IPC::Open2::open2 function.  If that handle is
created too late, it also gets caught up in the SVN::Pool and incorrectly
closed by the SVN::Pool::clear call.  But this only seems to happen with
more recent versions of Perl and svn.

To avoid this problem we add an explicit call to _open_cat_blob_if_needed
before the first call to SVN::Pool->new_default to make sure the open2
handle does not end up in the SVN::Pool.

Signed-off-by: Kyle J. McKay <mackyle@gmail.com>
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agogit-svn: fix localtime=true on non-glibc environments
Ryuichi Kokubo [Wed, 25 Feb 2015 16:04:41 +0000 (01:04 +0900)] 
git-svn: fix localtime=true on non-glibc environments

git svn uses POSIX::strftime('%s', $sec, $min, ...) to make unix epoch time.
But lowercase %s formatting character is a GNU extention. This causes problem
in git svn fetch --localtime on non-glibc systems, such as msys or cygwin.
Using Time::Local::timelocal($sec, $min, ...) fixes it.

Signed-off-by: Ryuichi Kokubo <ryu1kkb@gmail.com>
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Notes:
    lowercase %s format character in strftime is a GNU extension and not widely supported.
    POSIX::strftime affected by underlying crt's strftime because POSIX::strftime just calls crt's one.
    Time::Local is good function to replace POSIX::strftime because it's a perl core module function.

    Document about Time::Local.
     http://perldoc.perl.org/Time/Local.html

    These are specifications of strftime.

    The GNU C Library Reference Manual.
     http://www.gnu.org/software/libc/manual/html_node/Formatting-Calendar-Time.html

    perl POSIX module's strftime document. It does not have '%s'.
     http://perldoc.perl.org/POSIX.html

    strftime document of Microsort Windows C Run-Time library.
     https://msdn.microsoft.com/en-us/library/fe06s4ak.aspx

    The Open Group's old specification does not have '%s' too.
     http://pubs.opengroup.org/onlinepubs/007908799/xsh/strftime.html

    On my environment, following problems happened.
    - msys   : git svn fetch does not progress at all with perl.exe consuming CPU.
    - cygwin : git svn fetch progresses but time stamp information is dropped.
       Every commits have unix epoch timestamp.

    I would like to thank git developer and contibutors.
    git helps me so much everyday.
    Thank you.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoDocumentation/git-clean.txt: document that -f may need to be given twice
Mikko Rapeli [Thu, 26 Feb 2015 13:16:49 +0000 (15:16 +0200)] 
Documentation/git-clean.txt: document that -f may need to be given twice

This is needed in build automation where the tree really needs to
be reset to known state.

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agogit-svn: lazy load some modules
Eric Wong [Thu, 15 Jan 2015 08:54:22 +0000 (08:54 +0000)] 
git-svn: lazy load some modules

We can delay loading some modules until we need them for uncommon
code paths.  For example, persistent memoization is not often
needed, so we can avoid loading the modules for it until we
encounter svn::mergeinfo during fetch.

This gives a tiny reduction in syscalls (from 15641 to 15305) when
running "git svn info" and counting via "strace -fc".  Further,
more invasive work will be needed to noticeably improve performance.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
9 years agoPost 2.3 cyle (batch #5)
Junio C Hamano [Wed, 25 Feb 2015 23:44:04 +0000 (15:44 -0800)] 
Post 2.3 cyle (batch #5)

Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoMerge branch 'ak/git-pm-typofix'
Junio C Hamano [Wed, 25 Feb 2015 23:40:21 +0000 (15:40 -0800)] 
Merge branch 'ak/git-pm-typofix'

Typofix in comments.

* ak/git-pm-typofix:
  Git.pm: two minor typo fixes

9 years agoMerge branch 'sb/plug-leak-in-make-cache-entry'
Junio C Hamano [Wed, 25 Feb 2015 23:40:21 +0000 (15:40 -0800)] 
Merge branch 'sb/plug-leak-in-make-cache-entry'

"update-index --refresh" used to leak when an entry cannot be
refreshed for whatever reason.

* sb/plug-leak-in-make-cache-entry:
  read-cache.c: free cache entry when refreshing fails

9 years agoMerge branch 'mh/transport-capabilities'
Junio C Hamano [Wed, 25 Feb 2015 23:40:19 +0000 (15:40 -0800)] 
Merge branch 'mh/transport-capabilities'

The transport-helper did not give transport options such as
verbosity, progress, cloning, etc. to import and export based
helpers, like it did for fetch and push based helpers, robbing them
the chance to honor the wish of the end-users better.

* mh/transport-capabilities:
  transport-helper: ask the helper to set the same options for import as for fetch
  transport-helper: ask the helper to set progress and verbosity options after asking for its capabilities

9 years agoMerge branch 'jc/send-email-sensible-encoding'
Junio C Hamano [Wed, 25 Feb 2015 23:40:19 +0000 (15:40 -0800)] 
Merge branch 'jc/send-email-sensible-encoding'

"git send-email" used to accept a mistaken "y" (or "yes") as an
answer to "What encoding do you want to use [UTF-8]? " without
questioning.  Now it asks for confirmation when the answer looks
too short to be a valid encoding name.

* jc/send-email-sensible-encoding:
  send-email: ask confirmation if given encoding name is very short

9 years agoMerge branch 'jk/sanity'
Junio C Hamano [Wed, 25 Feb 2015 23:40:18 +0000 (15:40 -0800)] 
Merge branch 'jk/sanity'

The tests that wanted to see that file becomes unreadable after
running "chmod a-r file", and the tests that wanted to make sure it
is not run as root, we used "can we write into the / directory?" as
a cheap substitute, but on some platforms that is not a good
heuristics.  The tests and their prerequisites have been updated to
check what they really require.

* jk/sanity:
  test-lib.sh: set prerequisite SANITY by testing what we really need
  tests: correct misuses of POSIXPERM
  t/lib-httpd: switch SANITY check for NOT_ROOT

9 years agoMerge branch 'sb/hex-object-name-is-at-most-41-bytes-long'
Junio C Hamano [Wed, 25 Feb 2015 23:40:16 +0000 (15:40 -0800)] 
Merge branch 'sb/hex-object-name-is-at-most-41-bytes-long'

Code clean-up.

* sb/hex-object-name-is-at-most-41-bytes-long:
  hex.c: reduce memory footprint of sha1_to_hex static buffers

9 years agoMerge branch 'jk/fast-import-die-nicely-fix'
Junio C Hamano [Wed, 25 Feb 2015 23:40:15 +0000 (15:40 -0800)] 
Merge branch 'jk/fast-import-die-nicely-fix'

"git fast-import" used to crash when it could not close and
conclude the resulting packfile cleanly.

* jk/fast-import-die-nicely-fix:
  fast-import: avoid running end_packfile recursively

9 years agoMerge branch 'dp/remove-duplicated-header-inclusion'
Junio C Hamano [Wed, 25 Feb 2015 23:40:14 +0000 (15:40 -0800)] 
Merge branch 'dp/remove-duplicated-header-inclusion'

Code clean-up.

* dp/remove-duplicated-header-inclusion:
  do not include the same header twice

9 years agoMerge branch 'jc/max-io-size-and-ssize-max'
Junio C Hamano [Wed, 25 Feb 2015 23:40:13 +0000 (15:40 -0800)] 
Merge branch 'jc/max-io-size-and-ssize-max'

Our default I/O size (8 MiB) for large files was too large for some
platforms with smaller SSIZE_MAX, leading to read(2)/write(2)
failures.

* jc/max-io-size-and-ssize-max:
  xread/xwrite: clip MAX_IO_SIZE to SSIZE_MAX

9 years agoMerge branch 'tc/missing-http-proxyauth'
Junio C Hamano [Wed, 25 Feb 2015 23:40:12 +0000 (15:40 -0800)] 
Merge branch 'tc/missing-http-proxyauth'

We did not check the curl library version before using
CURLOPT_PROXYAUTH feature that may not exist.

* tc/missing-http-proxyauth:
  http: support curl < 7.10.7

9 years agoMerge branch 'jk/strbuf-doc-to-header'
Junio C Hamano [Wed, 25 Feb 2015 23:40:10 +0000 (15:40 -0800)] 
Merge branch 'jk/strbuf-doc-to-header'

The strbuf API was explained between the API documentation and in
the header file.  Move missing bits to strbuf.h so that programmers
can check only one place for all necessary information.

* jk/strbuf-doc-to-header:
  strbuf.h: group documentation for trim functions
  strbuf.h: drop boilerplate descriptions of strbuf_split_*
  strbuf.h: reorganize api function grouping headers
  strbuf.h: format asciidoc code blocks as 4-space indent
  strbuf.h: drop asciidoc list formatting from API docs
  strbuf.h: unify documentation comments beginnings
  strbuf.h: integrate api-strbuf.txt documentation

9 years agoMerge branch 'nd/attr-optim'
Junio C Hamano [Wed, 25 Feb 2015 23:40:10 +0000 (15:40 -0800)] 
Merge branch 'nd/attr-optim'

Optimize attribute look-up, mostly useful in "git grep" on a
project that does not use many attributes, by avoiding it when we
(should) know that the attributes are not defined in the first
place.

* nd/attr-optim:
  attr: avoid heavy work when we know the specified attr is not defined
  attr: do not attempt to expand when we know it's not a macro
  attr.c: rename arg name attr_nr to avoid shadowing the global one

9 years agoMerge branch 'jn/doc-api-errors'
Junio C Hamano [Wed, 25 Feb 2015 23:40:09 +0000 (15:40 -0800)] 
Merge branch 'jn/doc-api-errors'

The error handling functions and conventions are now documented in
the API manual.

* jn/doc-api-errors:
  doc: document error handling functions and conventions

9 years agoSync with 2.3.1
Junio C Hamano [Wed, 25 Feb 2015 06:15:49 +0000 (22:15 -0800)] 
Sync with 2.3.1

9 years agoGit 2.3.1 v2.3.1
Junio C Hamano [Wed, 25 Feb 2015 06:14:42 +0000 (22:14 -0800)] 
Git 2.3.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoMerge branch 'ak/add-i-empty-candidates' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:42 +0000 (22:10 -0800)] 
Merge branch 'ak/add-i-empty-candidates' into maint

The interactive "show a list and let the user choose from it"
interface "add -i" used showed and prompted to the user even when
the candidate list was empty, against which the only "choice" the
user could have made was to choose nothing.

* ak/add-i-empty-candidates:
  add -i: return from list_and_choose if there is no candidate

9 years agoMerge branch 'jc/apply-ws-fix-expands' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:41 +0000 (22:10 -0800)] 
Merge branch 'jc/apply-ws-fix-expands' into maint

"git apply --whitespace=fix" used to under-allocate the memory
when the fix resulted in a longer text than the original patch.

* jc/apply-ws-fix-expands:
  apply: count the size of postimage correctly
  apply: make update_pre_post_images() sanity check the given postlen
  apply.c: typofix

9 years agoMerge branch 'jc/doc-log-rev-list-options' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:39 +0000 (22:10 -0800)] 
Merge branch 'jc/doc-log-rev-list-options' into maint

"git log --help" used to show rev-list options that are irrelevant
to the "log" command.

* jc/doc-log-rev-list-options:
  Documentation: what does "git log --indexed-objects" even mean?

9 years agoMerge branch 'mg/commit-author-no-match-malformed-message' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:38 +0000 (22:10 -0800)] 
Merge branch 'mg/commit-author-no-match-malformed-message' into maint

The error message from "git commit", when a non-existing author
name was given as value to the "--author=" parameter, has been
reworded to avoid misunderstanding.

* mg/commit-author-no-match-malformed-message:
  commit: reword --author error message

9 years agoMerge branch 'jk/dumb-http-idx-fetch-fix' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:37 +0000 (22:10 -0800)] 
Merge branch 'jk/dumb-http-idx-fetch-fix' into maint

A broken pack .idx file in the receiving repository prevented the
dumb http transport from fetching a good copy of it from the other
side.

* jk/dumb-http-idx-fetch-fix:
  dumb-http: do not pass NULL path to parse_pack_index

9 years agoMerge branch 'jc/diff-format-doc' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:36 +0000 (22:10 -0800)] 
Merge branch 'jc/diff-format-doc' into maint

The documentation incorrectly said that C(opy) and R(ename) are the
only ones that can be followed by the score number in the output in
the --raw format.

* jc/diff-format-doc:
  diff-format doc: a score can follow M for rewrite

9 years agoMerge branch 'jk/remote-curl-an-array-in-struct-cannot-be-null' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:35 +0000 (22:10 -0800)] 
Merge branch 'jk/remote-curl-an-array-in-struct-cannot-be-null' into maint

Fix a misspelled conditional that is always true.

* jk/remote-curl-an-array-in-struct-cannot-be-null:
  do not check truth value of flex arrays

9 years agoMerge branch 'jk/status-read-branch-name-fix' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:21 +0000 (22:10 -0800)] 
Merge branch 'jk/status-read-branch-name-fix' into maint

Code to read branch name from various files in .git/ directory
would have misbehaved if the code to write them left an empty file.

* jk/status-read-branch-name-fix:
  read_and_strip_branch: fix typo'd address-of operator

9 years agoMerge branch 'mg/push-repo-option-doc' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:19 +0000 (22:10 -0800)] 
Merge branch 'mg/push-repo-option-doc' into maint

The "git push" documentation made the "--repo=<there>" option
easily misunderstood.

* mg/push-repo-option-doc:
  git-push.txt: document the behavior of --repo

9 years agoMerge branch 'bc/http-fallback-to-password-after-krb-fails' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:17 +0000 (22:10 -0800)] 
Merge branch 'bc/http-fallback-to-password-after-krb-fails' into maint

After attempting and failing a password-less authentication
(e.g. kerberos), libcURL refuses to fall back to password based
Basic authentication without a bit of help/encouragement.

* bc/http-fallback-to-password-after-krb-fails:
  remote-curl: fall back to Basic auth if Negotiate fails

9 years agoMerge branch 'dk/format-patch-ignore-diff-submodule' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:15 +0000 (22:10 -0800)] 
Merge branch 'dk/format-patch-ignore-diff-submodule' into maint

Setting diff.submodule to 'log' made "git format-patch" produce
broken patches.

* dk/format-patch-ignore-diff-submodule:
  format-patch: ignore diff.submodule setting
  t4255: test am submodule with diff.submodule

9 years agoMerge branch 'jn/rerere-fail-on-auto-update-failure' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:10:13 +0000 (22:10 -0800)] 
Merge branch 'jn/rerere-fail-on-auto-update-failure' into maint

"git rerere" (invoked internally from many mergy operations) did
not correctly signal errors when told to update the working tree
files and failed to do so for whatever reason.

* jn/rerere-fail-on-auto-update-failure:
  rerere: error out on autoupdate failure

9 years agoMerge branch 'jk/blame-commit-label' into maint
Junio C Hamano [Wed, 25 Feb 2015 06:09:54 +0000 (22:09 -0800)] 
Merge branch 'jk/blame-commit-label' into maint

"git blame HEAD -- missing" failed to correctly say "HEAD" when it
tried to say "No such path 'missing' in HEAD".

* jk/blame-commit-label:
  blame.c: fix garbled error message
  use xstrdup_or_null to replace ternary conditionals
  builtin/commit.c: use xstrdup_or_null instead of envdup
  builtin/apply.c: use xstrdup_or_null instead of null_strdup
  git-compat-util: add xstrdup_or_null helper

9 years agotest_ln_s_add: refresh stat info of fake symbolic links
Johannes Sixt [Mon, 23 Feb 2015 18:14:47 +0000 (19:14 +0100)] 
test_ln_s_add: refresh stat info of fake symbolic links

We have a helper function test_ln_s_add that inserts a symbolic link
into the index even if the file system does not support symbolic links.
There is a small flaw in the emulation path: the added entry does not
pick up stat information of the fake symbolic link from the file system,
as a consequence, the index is not exactly the same as for the "regular"
path (where symbolic links are available). To fix this, just call
git update-index again.

This flaw was revealed by the earlier change that tightened
compare_diff_raw(), because a test case in t4008 depends on the
correctly updated index.

Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agogit-compat-util.h: remove redundant code
Ramsay Jones [Mon, 23 Feb 2015 00:07:14 +0000 (00:07 +0000)] 
git-compat-util.h: remove redundant code

Since commit 3a0a3a89 ("git-compat-util.h: don't define _XOPEN_SOURCE
on cygwin", 23-11-2014) removed the definition of _XOPEN_SOURCE on
cygwin, the code within a pre-processor conditional further down the
file became redundant. Remove the redundant code.

This effectively reverts commit 41b20017 ("Fix an "implicit function
definition" warning", 03-03-2007).

Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoPost 2.3 cycle (batch #4)
Junio C Hamano [Sun, 22 Feb 2015 20:29:36 +0000 (12:29 -0800)] 
Post 2.3 cycle (batch #4)

Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoMerge branch 'jc/conf-var-doc'
Junio C Hamano [Sun, 22 Feb 2015 20:28:31 +0000 (12:28 -0800)] 
Merge branch 'jc/conf-var-doc'

Longstanding configuration variable naming rules has been added to
the documentation.

* jc/conf-var-doc:
  CodingGuidelines: describe naming rules for configuration variables
  config.txt: mark deprecated variables more prominently
  config.txt: clarify that add.ignore-errors is deprecated

9 years agoMerge branch 'es/squelch-openssl-warnings-on-macosx'
Junio C Hamano [Sun, 22 Feb 2015 20:28:30 +0000 (12:28 -0800)] 
Merge branch 'es/squelch-openssl-warnings-on-macosx'

An earlier workaround to squelch unhelpful deprecation warnings
from the complier on Mac OSX unnecessarily set minimum required
version of the OS, which the user might want to raise (or lower)
for other reasons.

* es/squelch-openssl-warnings-on-macosx:
  git-compat-util: do not step on MAC_OS_X_VERSION_MIN_REQUIRED

9 years agoMerge branch 'tc/curl-vernum-output-broken-in-7.11'
Junio C Hamano [Sun, 22 Feb 2015 20:28:29 +0000 (12:28 -0800)] 
Merge branch 'tc/curl-vernum-output-broken-in-7.11'

Certain older vintages of cURL give irregular output from
"curl-config --vernum", which confused our build system.

* tc/curl-vernum-output-broken-in-7.11:
  Makefile: handle broken curl version number in version check

9 years agoMerge branch 'jk/prune-mtime'
Junio C Hamano [Sun, 22 Feb 2015 20:28:28 +0000 (12:28 -0800)] 
Merge branch 'jk/prune-mtime'

In v2.2.0, we broke "git prune" that runs in a repository that
borrows from an alternate object store.

* jk/prune-mtime:
  sha1_file: fix iterating loose alternate objects
  for_each_loose_file_in_objdir: take an optional strbuf path

9 years agoMerge branch 'jc/diff-files-ita'
Junio C Hamano [Sun, 22 Feb 2015 20:28:27 +0000 (12:28 -0800)] 
Merge branch 'jc/diff-files-ita'

Code cleanup.

* jc/diff-files-ita:
  run_diff_files(): clarify computation of sha1 validity

9 years agoMerge branch 'ps/submodule-sanitize-path-upon-add'
Junio C Hamano [Sun, 22 Feb 2015 20:28:26 +0000 (12:28 -0800)] 
Merge branch 'ps/submodule-sanitize-path-upon-add'

"git submodule add" failed to squash "path/to/././submodule" to
"path/to/submodule".

* ps/submodule-sanitize-path-upon-add:
  git-submodule.sh: fix '/././' path normalization

9 years agoMerge branch 'ab/merge-file-prefix'
Junio C Hamano [Sun, 22 Feb 2015 20:28:25 +0000 (12:28 -0800)] 
Merge branch 'ab/merge-file-prefix'

"git merge-file" did not work correctly in a subdirectory.

* ab/merge-file-prefix:
  merge-file: correctly open files when in a subdir

9 years agoMerge branch 'es/blame-commit-info-fix'
Junio C Hamano [Sun, 22 Feb 2015 20:28:24 +0000 (12:28 -0800)] 
Merge branch 'es/blame-commit-info-fix'

"git blame" died, trying to free an uninitialized piece of memory.

* es/blame-commit-info-fix:
  builtin/blame: destroy initialized commit_info only

9 years agoMerge branch 'ss/check-builtins-on-windows'
Junio C Hamano [Sun, 22 Feb 2015 20:28:23 +0000 (12:28 -0800)] 
Merge branch 'ss/check-builtins-on-windows'

* ss/check-builtins-on-windows:
  check-builtins: strip executable suffix $X when enumerating builtins

9 years agot5500: show user name and host in diag-url
Torsten Bögershausen [Sat, 21 Feb 2015 15:52:55 +0000 (16:52 +0100)] 
t5500: show user name and host in diag-url

The URL for ssh may have include a username before the hostname,
like ssh://user@host/repo.
When literal IPV6 addresses are used together with a username,
the substring "user@[::1]" must be converted into "user@::1".

Make that conversion visible for the user, and write userandhost
in the diagnostics

Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agot5601: add more test cases for IPV6
Torsten Bögershausen [Sat, 21 Feb 2015 15:53:01 +0000 (16:53 +0100)] 
t5601: add more test cases for IPV6

Test the parsing of literall IPV6 addresses more systematically:
- with and without brackets (e.g. ::1 [::1])
- with brackets and port number: (e.g. [::1]:22)
- with username (e.g. user@::1)
- with username and brackets:
  Because user@[::1] was not supported on older Git version,
  [user@::1] had to be used as a workaround.
  Test that user@::1 user@[::1] and [user@::1] all do the same.

Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agoconnect.c: allow ssh://user@[2001:db8::1]/repo.git
Torsten Bögershausen [Sat, 21 Feb 2015 15:52:48 +0000 (16:52 +0100)] 
connect.c: allow ssh://user@[2001:db8::1]/repo.git

The ssh:// syntax was added in 2386d658 (Add first cut at "git
protocol" connect logic., 2005-07-13), it accepted
ssh://user@2001:db8::1/repo.git, which is now legacy.

Over the years the parser was improved to support [] and port numbers,
but the combination of ssh://user@[2001:db8::1]:222/repo.git did
never work.

The only only way to use a user name, a literall IPV6 address and a port
number was ssh://[user@2001:db8::1]:222/repo.git

(Thanks to Christian Taube <lists@hcf.yourweb.de> for reporting this long
standing issue)

New users would use ssh://user@[2001:db8::1]:222/repo.git,
so change the parser to handle it correctly.

Support the old legacy URLs as well, to be backwards compatible,
and avoid regressions for users which upgrade an existing installation
to a later Git version.

Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agosha1_name: use strlcpy() to copy strings
René Scharfe [Sat, 21 Feb 2015 19:55:22 +0000 (20:55 +0100)] 
sha1_name: use strlcpy() to copy strings

Use strlcpy() instead of calling strncpy() and then setting the last
byte of the target buffer to NUL explicitly.  This shortens and
simplifies the code a bit.

Signed-of-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
9 years agopretty: use starts_with() to check for a prefix
René Scharfe [Sat, 21 Feb 2015 19:53:09 +0000 (20:53 +0100)] 
pretty: use starts_with() to check for a prefix

Simplify the code and avoid duplication by using starts_with() instead
of strlen() and strncmp() to check if a line starts with "encoding ".

Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>