There are certain house-keeping tasks that need to be performed at
the very beginning of any Git program, and programs that are not
built-in commands had to do them exactly the same way as "git"
potty does.  It was easy to make mistakes in one-off standalone
programs (like test helpers).  A common "main()" function that
calls cmd_main() of individual program has been introduced to
make it harder to make mistakes.

* jk/common-main:
  mingw: declare main()'s argv as const
  common-main: call git_setup_gettext()
  common-main: call restore_sigpipe_to_default()
  common-main: call sanitize_stdfds()
  common-main: call git_extract_argv0_path()
  add an extra level of indirection to main()

A test that unconditionally used "mktemp" learned that the command
is not necessarily available everywhere.

* ak/lazy-prereq-mktemp:
  t7610: test for mktemp before test execution

"git grep -i" has been taught to fold case in non-ascii locales
correctly.

* nd/icase:
  grep.c: reuse "icase" variable
  diffcore-pickaxe: support case insensitive match on non-ascii
  diffcore-pickaxe: Add regcomp_or_die()
  grep/pcre: support utf-8
  gettext: add is_utf8_locale()
  grep/pcre: prepare locale-dependent tables for icase matching
  grep: rewrite an if/else condition to avoid duplicate expression
  grep/icase: avoid kwsset when -F is specified
  grep/icase: avoid kwsset on literal non-ascii strings
  test-regex: expose full regcomp() to the command line
  test-regex: isolate the bug test code
  grep: break down an "if" stmt in preparation for next changes

Conversion from unsigned char sha1[20] to struct object_id
continues.

* bc/cocci:
  diff: convert prep_temp_blob() to struct object_id
  merge-recursive: convert merge_recursive_generic() to object_id
  merge-recursive: convert leaf functions to use struct object_id
  merge-recursive: convert struct merge_file_info to object_id
  merge-recursive: convert struct stage_data to use object_id
  diff: rename struct diff_filespec's sha1_valid member
  diff: convert struct diff_filespec to struct object_id
  coccinelle: apply object_id Coccinelle transformations
  coccinelle: convert hashcpy() with null_sha1 to hashclr()
  contrib/coccinelle: add basic Coccinelle transforms
  hex: add oid_to_hex_r()

The commands in the "log/diff" family have had an FILE* pointer in the
data structure they pass around for a long time, but some codepaths
used to always write to the standard output.  As a preparatory step
to make "git format-patch" available to the internal callers, these
codepaths have been updated to consistently write into that FILE*
instead.

* js/log-to-diffopt-file:
  mingw: fix the shortlog --output=<file> test
  diff: do not color output when --color=auto and --output=<file> is given
  t4211: ensure that log respects --output=<file>
  shortlog: respect the --output=<file> setting
  format-patch: use stdout directly
  format-patch: avoid freopen()
  format-patch: explicitly switch off color when writing to files
  shortlog: support outputting to streams other than stdout
  graph: respect the diffopt.file setting
  line-log: respect diffopt's configured output file stream
  log-tree: respect diffopt's configured output file stream
  log: prepare log/log-tree to reuse the diffopt.close_file attribute

Fix recently introduced codepaths that are involved in parallel
submodule operations, which gave up on reading too early, and
could have wasted CPU while attempting to write under a corner
case condition.

* sb/submodule-parallel-fetch:
  hoist out handle_nonblock function for xread and xwrite
  xwrite: poll on non-blocking FDs
  xread: retry after poll on EAGAIN/EWOULDBLOCK

Code simplification.

* lf/recv-sideband-cleanup:
  sideband.c: small optimization of strbuf usage
  sideband.c: refactor recv_sideband()

"git blame -M" missed a single line that was moved within the file.

* dk/blame-move-no-reason-for-1-line-context:
  blame: require 0 context lines while finding moved lines with -M

A new configuration variable core.sshCommand has been added to
specify what value for GIT_SSH_COMMAND to use per repository.

* nd/connect-ssh-command-config:
  connect: read $GIT_SSH_COMMAND from config file

As we are not yet moving everything to size_t but still using ulong
internally when talking about the size of object, platforms with
32-bit long will not be able to produce tar archive with 4GB+ file,
and cannot grok 077777777777UL as a constant.  Disable the extended
header feature and do not test it on them.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* maint:
  Git 2.9.2
  t0006: skip "far in the future" test when unsigned long is not long enough

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

Skip tests that are unrunnable on platforms without 64-bit long
to avoid unnecessary test failures.

* jk/tzoffset-fix:
  t0006: skip "far in the future" test when unsigned long is not long enough

Git's source code refers to timestamps as unsigned longs.  On 32-bit
platforms, as well as on Windows, unsigned long is not large enough
to capture dates that are "absurdly far in the future".

While we can fix this issue properly by replacing unsigned long with
a larger type, we want to be a bit more conservative and just skip
those tests on the maint track.
Signed-off-by: NJeff King <peff@peff.net>
Helped-by: NJohannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

A hot-fix to make a test working in mingw again.

* jk/upload-pack-hook:
  mingw: fix regression in t1308-config-set

When we tried to fix in 58461bdf (t1308: do not get fooled by symbolic
links to the source tree, 2016-06-02) an obscure case where the user
cd's into Git's source code via a symbolic link, a regression was
introduced that affects all test runs on Windows.

The original patch introducing the test case in question was careful to
use `$(pwd)` instead of `$PWD`.

This was done to account for the fact that Git's test suite uses shell
scripting even on Windows, where the shell's Unix-y paths are
incompatible with the main Git executable's idea of paths: it only
accepts Windows paths.

It is an awkward but necessary thing, then, to use `$(pwd)` (which gives
us a Windows path) when interacting with the Git executable and `$PWD`
(which gives the shell's idea of the current working directory in Unix-y
form) for shell scripts, including the test suite itself.

Obviously this broke the use case of the Git maintainer when changing
the working directory into Git's source code directory via a symlink,
i.e. when `$(pwd)` does not agree with `$PWD`.

However, we must not fix that use case at the expense of regressing
another use case.

Let's special-case Windows here, even if it is ugly, for lack of a more
elegant solution.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

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

"git archive" learned to handle files that are larger than 8GB and
commits far in the future than expressible by the traditional US-TAR
format.

* jk/big-and-future-archive-tar:
  archive-tar: drop return value
  archive-tar: write extended headers for far-future mtime
  archive-tar: write extended headers for file sizes >= 8GB
  t5000: test tar files that overflow ustar headers
  t9300: factor out portable "head -c" replacement

Git does not know what the contents in the index should be for a
path added with "git add -N" yet, so "git grep --cached" should not
show hits (or show lack of hits, with -L) in such a path, but that
logic does not apply to "git grep", i.e. searching in the working
tree files.  But we did so by mistake, which has been corrected.

* nd/ita-cleanup:
  grep: fix grepping for "intent to add" files
  t7810-grep.sh: fix a whitespace inconsistency
  t7810-grep.sh: fix duplicated test name

"git rebase -i --autostash" did not restore the auto-stashed change
when the operation was aborted.

* ps/rebase-i-auto-unstash-upon-abort:
  rebase -i: restore autostash on abort

Grammofix.

* js/t3404-grammo-fix:
  t3404: fix a grammo (commands are ran -> commands are run)

"git commit --amend --allow-empty-message -S" for a commit without
any message body could have misidentified where the header of the
commit object ends.

* js/sign-empty-commit-fix:
  commit -S: avoid invalid pointer with empty message

More mark-up updates to typeset strings that are expected to
literally typed by the end user in fixed-width font.

* mm/doc-tt:
  doc: typeset HEAD and variants as literal
  CodingGuidelines: formatting HEAD in documentation
  doc: typeset long options with argument as literal
  doc: typeset '--' as literal
  doc: typeset long command-line options as literal
  doc: typeset short command-line options as literal
  Documentation/git-mv.txt: fix whitespace indentation

Add a test to specify the desired behaviour that currently is not
available in "git rebase -Xsubtree=...".

* dg/subtree-rebase-test:
  contrib/subtree: Add a test for subtree rebase that loses commits

Typofix in a doc.

* nd/doc-new-command:
  new-command.txt: correct the command description file

"gc.autoPackLimit" when set to 1 should not trigger a repacking
when there is only one pack, but the code counted poorly and did
so.

* ew/gc-auto-pack-limit-fix:
  gc: fix off-by-one error with gc.autoPackLimit

Grammofix.

* ah/unpack-trees-advice-messages:
  unpack-trees: fix English grammar in do-this-before-that messages

More markings of messages for i18n, with updates to various tests
to pass GETTEXT_POISON tests.

One patch from the original submission dropped due to conflicts
with jk/upload-pack-hook, which is still in flux.

* va/i18n-even-more: (38 commits)
  t5541: become resilient to GETTEXT_POISON
  i18n: branch: mark comment when editing branch description for translation
  i18n: unmark die messages for translation
  i18n: submodule: escape shell variables inside eval_gettext
  i18n: submodule: join strings marked for translation
  i18n: init-db: join message pieces
  i18n: remote: allow translations to reorder message
  i18n: remote: mark URL fallback text for translation
  i18n: standardise messages
  i18n: sequencer: add period to error message
  i18n: merge: change command option help to lowercase
  i18n: merge: mark messages for translation
  i18n: notes: mark options for translation
  i18n: notes: mark strings for translation
  i18n: transport-helper.c: change N_() call to _()
  i18n: bisect: mark strings for translation
  t5523: use test_i18ngrep for negation
  t4153: fix negated test_i18ngrep call
  t9003: become resilient to GETTEXT_POISON
  tests: unpack-trees: update to use test_i18n* functions
  ...

Adjust t4201 to pass on Windows; a couple of test cases need to be
skipped on Windows which leads to a different shortlog than on Linux.

Let's just fix that by limiting the shortlog's commit range to traverse
only one commit: that guarantees that it does not matter how many test
cases were skipped.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* maint:
  Git 2.9.1

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

Portability fix for Windows.

* jc/t2300-setup:
  t2300: "git --exec-path" is not usable in $PATH on Windows as-is

Test clean-up.

* cb/t7810-test-label-fix:
  t7810: fix duplicated test title

Test clean-up.

* sb/t5614-modernize:
  t5614: don't use subshells

The top level documentation "git help git" still pointed at the
documentation set hosted at now-defunct google-code repository.
Update it to point to https://git.github.io/htmldocs/git.html
instead.

* jn/preformatted-doc-url:
  doc: git-htmldocs.googlecode.com is no more

A bug, which caused "git p4" while running under verbose mode to
report paths that are omitted due to branch prefix incorrectly, has
been fixed; the command said "Ignoring file outside of prefix" for
paths that are _inside_.

* ao/p4-has-branch-prefix-fix:
  git-p4: correct hasBranchPrefix verbose output

t/perf needs /usr/bin/time with GNU extension; the invocation of it
is updated to "gtime" on Darwin.

* js/perf-on-apple:
  perf: accommodate for MacOSX

One among four invocations of readlink(1) in our test suite has
been rewritten so that the test can run on systems without the
command (others are in valgrind test framework and t9802).

* ak/t7800-wo-readlink:
  t7800: readlink may not be available

The internal code used to show local timezone offset is not
prepared to handle timestamps beyond year 2100, and gave a
bogus offset value to the caller.  Use a more benign looking
+0000 instead and let "git log" going in such a case, instead
of aborting.

* jk/tzoffset-fix:
  local_tzoffset: detect errors from tm_to_time_t
  t0006: test various date formats
  t0006: rename test-date's "show" to "relative"

Some platform-specific code had non-ANSI strict declarations of C
functions that do not take any parameters, which has been
corrected.

* js/mingw-parameter-less-c-functions:
  mingw: let the build succeed with DEVELOPER=1