git
7 years agoMerge branch 'sb/submodule-parallel-update'
Junio C Hamano [Wed, 23 Aug 2017 21:13:14 +0000 (14:13 -0700)] 
Merge branch 'sb/submodule-parallel-update'

Code clean-up.

* sb/submodule-parallel-update:
  submodule.sh: remove unused variable

7 years agoMerge branch 'jc/diff-sane-truncate-no-more'
Junio C Hamano [Wed, 23 Aug 2017 21:13:13 +0000 (14:13 -0700)] 
Merge branch 'jc/diff-sane-truncate-no-more'

Code clean-up.

* jc/diff-sane-truncate-no-more:
  diff: retire sane_truncate_fn

7 years agoMerge branch 'hv/t5526-andand-chain-fix'
Junio C Hamano [Wed, 23 Aug 2017 21:13:13 +0000 (14:13 -0700)] 
Merge branch 'hv/t5526-andand-chain-fix'

Test fix.

* hv/t5526-andand-chain-fix:
  t5526: fix some broken && chains

7 years agoMerge branch 'as/grep-quiet-no-match-exit-code-fix'
Junio C Hamano [Wed, 23 Aug 2017 21:13:12 +0000 (14:13 -0700)] 
Merge branch 'as/grep-quiet-no-match-exit-code-fix'

"git grep -L" and "git grep --quiet -L" reported different exit
codes; this has been corrected.

* as/grep-quiet-no-match-exit-code-fix:
  git-grep: correct exit code with --quiet and -L

7 years agoMerge branch 'kw/commit-keep-index-when-pre-commit-is-not-run'
Junio C Hamano [Wed, 23 Aug 2017 21:13:11 +0000 (14:13 -0700)] 
Merge branch 'kw/commit-keep-index-when-pre-commit-is-not-run'

"git commit" used to discard the index and re-read from the filesystem
just in case the pre-commit hook has updated it in the middle; this
has been optimized out when we know we do not run the pre-commit hook.

* kw/commit-keep-index-when-pre-commit-is-not-run:
  commit: skip discarding the index if there is no pre-commit hook

7 years agoMerge branch 'sb/sha1-file-cleanup'
Junio C Hamano [Wed, 23 Aug 2017 21:13:10 +0000 (14:13 -0700)] 
Merge branch 'sb/sha1-file-cleanup'

Code clean-up.

* sb/sha1-file-cleanup:
  sha1_file: make read_info_alternates static

7 years agoMerge branch 'rs/t1002-do-not-use-sum'
Junio C Hamano [Wed, 23 Aug 2017 21:13:09 +0000 (14:13 -0700)] 
Merge branch 'rs/t1002-do-not-use-sum'

Test simplification.

* rs/t1002-do-not-use-sum:
  t1002: stop using sum(1)

7 years agoMerge branch 'kd/stash-with-bash-4.4'
Junio C Hamano [Wed, 23 Aug 2017 21:13:08 +0000 (14:13 -0700)] 
Merge branch 'kd/stash-with-bash-4.4'

bash 4.4 or newer gave a warning on NUL byte in command
substitution done in "git stash"; this has been squelched.

* kd/stash-with-bash-4.4:
  stash: prevent warning about null bytes in input

7 years agoMerge branch 'ah/doc-empty-string-is-false'
Junio C Hamano [Wed, 23 Aug 2017 21:13:08 +0000 (14:13 -0700)] 
Merge branch 'ah/doc-empty-string-is-false'

Doc update.

* ah/doc-empty-string-is-false:
  doc: clarify "config --bool" behaviour with empty string

7 years agoMerge branch 'kw/rebase-progress'
Junio C Hamano [Wed, 23 Aug 2017 21:13:07 +0000 (14:13 -0700)] 
Merge branch 'kw/rebase-progress'

"git rebase", especially when it is run by mistake and ends up
trying to replay many changes, spent long time in silence.  The
command has been taught to show progress report when it spends
long time preparing these many changes to replay (which would give
the user a chance to abort with ^C).

* kw/rebase-progress:
  rebase: turn on progress option by default for format-patch
  format-patch: have progress option while generating patches

7 years agoMerge branch 'ks/prepare-commit-msg-sample-fix'
Junio C Hamano [Wed, 23 Aug 2017 21:13:07 +0000 (14:13 -0700)] 
Merge branch 'ks/prepare-commit-msg-sample-fix'

An "oops" fix to a topic that is already in 'master'.

* ks/prepare-commit-msg-sample-fix:
  hook: use correct logical variable

7 years agoMerge branch 'nm/stash-untracked'
Junio C Hamano [Wed, 23 Aug 2017 21:13:07 +0000 (14:13 -0700)] 
Merge branch 'nm/stash-untracked'

"git stash -u" used the contents of the committed version of the
".gitignore" file to decide which paths are ignored, even when the
file has local changes.  The command has been taught to instead use
the locally modified contents.

* nm/stash-untracked:
  stash: clean untracked files before reset

7 years agoMerge branch 'jt/sha1-file-cleanup'
Junio C Hamano [Wed, 23 Aug 2017 21:13:07 +0000 (14:13 -0700)] 
Merge branch 'jt/sha1-file-cleanup'

Preparatory code clean-up.

* jt/sha1-file-cleanup:
  sha1_file: remove read_packed_sha1()
  sha1_file: set whence in storage-specific info fn

7 years agoThe second batch post 2.14
Junio C Hamano [Tue, 22 Aug 2017 17:33:58 +0000 (10:33 -0700)] 
The second batch post 2.14

Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agoMerge branch 'mh/packed-ref-store'
Junio C Hamano [Tue, 22 Aug 2017 17:29:16 +0000 (10:29 -0700)] 
Merge branch 'mh/packed-ref-store'

The "ref-store" code reorganization continues.

* mh/packed-ref-store: (32 commits)
  files-backend: cheapen refname_available check when locking refs
  packed_ref_store: handle a packed-refs file that is a symlink
  read_packed_refs(): die if `packed-refs` contains bogus data
  t3210: add some tests of bogus packed-refs file contents
  repack_without_refs(): don't lock or unlock the packed refs
  commit_packed_refs(): remove call to `packed_refs_unlock()`
  clear_packed_ref_cache(): don't protest if the lock is held
  packed_refs_unlock(), packed_refs_is_locked(): new functions
  packed_refs_lock(): report errors via a `struct strbuf *err`
  packed_refs_lock(): function renamed from lock_packed_refs()
  commit_packed_refs(): use a staging file separate from the lockfile
  commit_packed_refs(): report errors rather than dying
  packed_ref_store: make class into a subclass of `ref_store`
  packed-backend: new module for handling packed references
  packed_read_raw_ref(): new function, replacing `resolve_packed_ref()`
  packed_ref_store: support iteration
  packed_peel_ref(): new function, extracted from `files_peel_ref()`
  repack_without_refs(): take a `packed_ref_store *` parameter
  get_packed_ref(): take a `packed_ref_store *` parameter
  rollback_packed_refs(): take a `packed_ref_store *` parameter
  ...

7 years agoMerge branch 'sb/retire-t1200'
Junio C Hamano [Tue, 22 Aug 2017 17:29:16 +0000 (10:29 -0700)] 
Merge branch 'sb/retire-t1200'

A test script that outlived its usefulness has been removed.

* sb/retire-t1200:
  t1200: remove t1200-tutorial.sh

7 years agoMerge branch 'rs/win32-syslog-leakfix'
Junio C Hamano [Tue, 22 Aug 2017 17:29:16 +0000 (10:29 -0700)] 
Merge branch 'rs/win32-syslog-leakfix'

Memory leak in an error codepath has been plugged.

* rs/win32-syslog-leakfix:
  win32: plug memory leak on realloc() failure in syslog()

7 years agoMerge branch 'rs/unpack-entry-leakfix'
Junio C Hamano [Tue, 22 Aug 2017 17:29:15 +0000 (10:29 -0700)] 
Merge branch 'rs/unpack-entry-leakfix'

Memory leak in an error codepath has been plugged.

* rs/unpack-entry-leakfix:
  sha1_file: release delta_stack on error in unpack_entry()

7 years agoMerge branch 'rs/strbuf-getwholeline-fix'
Junio C Hamano [Tue, 22 Aug 2017 17:29:15 +0000 (10:29 -0700)] 
Merge branch 'rs/strbuf-getwholeline-fix'

A helper function to read a single whole line into strbuf
mistakenly triggered OOM error at EOF under certain conditions,
which has been fixed.

* rs/strbuf-getwholeline-fix:
  strbuf: clear errno before calling getdelim(3)

7 years agoMerge branch 'rs/merge-microcleanup'
Junio C Hamano [Tue, 22 Aug 2017 17:29:14 +0000 (10:29 -0700)] 
Merge branch 'rs/merge-microcleanup'

Code clean-up.

* rs/merge-microcleanup:
  merge: use skip_prefix()

7 years agoMerge branch 'rs/fsck-obj-leakfix'
Junio C Hamano [Tue, 22 Aug 2017 17:29:13 +0000 (10:29 -0700)] 
Merge branch 'rs/fsck-obj-leakfix'

Memory leak in an error codepath has been plugged.

* rs/fsck-obj-leakfix:
  fsck: free buffers on error in fsck_obj()

7 years agoMerge branch 'rs/t4062-obsd'
Junio C Hamano [Tue, 22 Aug 2017 17:29:13 +0000 (10:29 -0700)] 
Merge branch 'rs/t4062-obsd'

Test portability fix.

* rs/t4062-obsd:
  t4062: use less than 256 repetitions in regex

7 years agoMerge branch 'rs/find-pack-entry-bisection'
Junio C Hamano [Tue, 22 Aug 2017 17:29:11 +0000 (10:29 -0700)] 
Merge branch 'rs/find-pack-entry-bisection'

Code clean-up.

* rs/find-pack-entry-bisection:
  sha1_file: avoid comparison if no packed hash matches the first byte

7 years agoMerge branch 'rs/apply-lose-prefix-length'
Junio C Hamano [Tue, 22 Aug 2017 17:29:10 +0000 (10:29 -0700)] 
Merge branch 'rs/apply-lose-prefix-length'

Code clean-up.

* rs/apply-lose-prefix-length:
  apply: remove prefix_length member from apply_state

7 years agoMerge branch 'rj/add-chmod-error-message'
Junio C Hamano [Tue, 22 Aug 2017 17:29:10 +0000 (10:29 -0700)] 
Merge branch 'rj/add-chmod-error-message'

Message fix.

* rj/add-chmod-error-message:
  builtin/add: add detail to a 'cannot chmod' error message

7 years agoMerge branch 'jk/hashcmp-memcmp'
Junio C Hamano [Tue, 22 Aug 2017 17:29:09 +0000 (10:29 -0700)] 
Merge branch 'jk/hashcmp-memcmp'

Code clean-up.

* jk/hashcmp-memcmp:
  hashcmp: use memcmp instead of open-coded loop

7 years agoMerge branch 'jk/drop-sha1-entry-pos'
Junio C Hamano [Tue, 22 Aug 2017 17:29:08 +0000 (10:29 -0700)] 
Merge branch 'jk/drop-sha1-entry-pos'

Code clean-up.

* jk/drop-sha1-entry-pos:
  sha1_file: drop experimental GIT_USE_LOOKUP search

7 years agoMerge branch 'ur/svn-local-zone'
Junio C Hamano [Tue, 22 Aug 2017 17:29:07 +0000 (10:29 -0700)] 
Merge branch 'ur/svn-local-zone'

"git svn" used with "--localtime" option did not compute the tz
offset for the timestamp in question and instead always used the
current time, which has been corrected.

* ur/svn-local-zone:
  git svn fetch: Create correct commit timestamp when using --localtime

7 years agoMerge branch 'pw/am-signoff'
Junio C Hamano [Tue, 22 Aug 2017 17:29:07 +0000 (10:29 -0700)] 
Merge branch 'pw/am-signoff'

"git am -s" has been taught that some input may end with a trailer
block that is not Signed-off-by: and it should refrain from adding
an extra blank line before adding a new sign-off in such a case.

* pw/am-signoff:
  am: fix signoff when other trailers are present

7 years agoMerge branch 'rs/t3700-clean-leftover'
Junio C Hamano [Tue, 22 Aug 2017 17:29:07 +0000 (10:29 -0700)] 
Merge branch 'rs/t3700-clean-leftover'

A test fix.

* rs/t3700-clean-leftover:
  t3700: fix broken test under !POSIXPERM

7 years agoMerge branch 'jc/perl-git-comment-typofix'
Junio C Hamano [Tue, 22 Aug 2017 17:29:06 +0000 (10:29 -0700)] 
Merge branch 'jc/perl-git-comment-typofix'

A comment fix.

* jc/perl-git-comment-typofix:
  perl/Git.pm: typofix in a comment

7 years agoMerge branch 'rs/in-obsd-basename-dirname-take-const'
Junio C Hamano [Tue, 22 Aug 2017 17:29:05 +0000 (10:29 -0700)] 
Merge branch 'rs/in-obsd-basename-dirname-take-const'

Portability fix.

* rs/in-obsd-basename-dirname-take-const:
  test-path-utils: handle const parameter of basename and dirname

7 years agoMerge branch 'rs/obsd-getcwd-workaround'
Junio C Hamano [Tue, 22 Aug 2017 17:29:04 +0000 (10:29 -0700)] 
Merge branch 'rs/obsd-getcwd-workaround'

Test portability fix for BSDs.

* rs/obsd-getcwd-workaround:
  t0001: skip test with restrictive permissions if getpwd(3) respects them

7 years agoMerge branch 'mf/no-dashed-subcommands'
Junio C Hamano [Tue, 22 Aug 2017 17:29:03 +0000 (10:29 -0700)] 
Merge branch 'mf/no-dashed-subcommands'

Code clean-up.

* mf/no-dashed-subcommands:
  scripts: use "git foo" not "git-foo"

7 years agoMerge branch 'ma/parse-maybe-bool'
Junio C Hamano [Tue, 22 Aug 2017 17:29:03 +0000 (10:29 -0700)] 
Merge branch 'ma/parse-maybe-bool'

Code clean-up.

* ma/parse-maybe-bool:
  parse_decoration_style: drop unused argument `var`
  treewide: deprecate git_config_maybe_bool, use git_parse_maybe_bool
  config: make git_{config,parse}_maybe_bool equivalent
  config: introduce git_parse_maybe_bool_text
  t5334: document that git push --signed=1 does not work
  Doc/git-{push,send-pack}: correct --sign= to --signed=

7 years agoMerge branch 'ab/ref-filter-no-contains'
Junio C Hamano [Tue, 22 Aug 2017 17:29:02 +0000 (10:29 -0700)] 
Merge branch 'ab/ref-filter-no-contains'

A test fix.

* ab/ref-filter-no-contains:
  tests: don't give unportable ">" to "test" built-in, use -gt

7 years agoMerge branch 'bw/clone-recursive-quiet'
Junio C Hamano [Tue, 22 Aug 2017 17:29:01 +0000 (10:29 -0700)] 
Merge branch 'bw/clone-recursive-quiet'

"git clone --recurse-submodules --quiet" did not pass the quiet
option down to submodules.

* bw/clone-recursive-quiet:
  clone: teach recursive clones to respect -q

7 years agoMerge branch 'bw/grep-recurse-submodules'
Junio C Hamano [Tue, 22 Aug 2017 17:29:01 +0000 (10:29 -0700)] 
Merge branch 'bw/grep-recurse-submodules'

"git grep --recurse-submodules" has been reworked to give a more
consistent output across submodule boundary (and do its thing
without having to fork a separate process).

* bw/grep-recurse-submodules:
  grep: recurse in-process using 'struct repository'
  submodule: merge repo_read_gitmodules and gitmodules_config
  submodule: check for unmerged .gitmodules outside of config parsing
  submodule: check for unstaged .gitmodules outside of config parsing
  submodule: remove fetch.recursesubmodules from submodule-config parsing
  submodule: remove submodule.fetchjobs from submodule-config parsing
  config: add config_from_gitmodules
  cache.h: add GITMODULES_FILE macro
  repository: have the_repository use the_index
  repo_read_index: don't discard the index

7 years agoMerge branch 'pw/sequence-rerere-autoupdate'
Junio C Hamano [Tue, 22 Aug 2017 17:29:00 +0000 (10:29 -0700)] 
Merge branch 'pw/sequence-rerere-autoupdate'

Commands like "git rebase" accepted the --rerere-autoupdate option
from the command line, but did not always use it.  This has been
fixed.

* pw/sequence-rerere-autoupdate:
  cherry-pick/revert: reject --rerere-autoupdate when continuing
  cherry-pick/revert: remember --rerere-autoupdate
  t3504: use test_commit
  rebase -i: honor --rerere-autoupdate
  rebase: honor --rerere-autoupdate
  am: remember --rerere-autoupdate setting

7 years agoMerge branch 'bw/push-options-recursively-to-submodules'
Junio C Hamano [Tue, 22 Aug 2017 17:29:00 +0000 (10:29 -0700)] 
Merge branch 'bw/push-options-recursively-to-submodules'

"git push --recurse-submodules $there HEAD:$target" was not
propagated down to the submodules, but now it is.

* bw/push-options-recursively-to-submodules:
  submodule--helper: teach push-check to handle HEAD

7 years agogit-grep: correct exit code with --quiet and -L
Anthony Sottile [Fri, 18 Aug 2017 01:38:51 +0000 (18:38 -0700)] 
git-grep: correct exit code with --quiet and -L

The handling of `status_only` no longer interferes with the handling of
`unmatch_name_only`.  `--quiet` no longer affects the exit code when using
`-L`/`--files-without-match`.

Signed-off-by: Anthony Sottile <asottile@umich.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agodiff: retire sane_truncate_fn
Junio C Hamano [Thu, 17 Aug 2017 17:27:28 +0000 (10:27 -0700)] 
diff: retire sane_truncate_fn

Long time ago, 23707811 ("diff: do not chomp hunk-header in the
middle of a character", 2008-01-02) introduced sane_truncate_line()
helper function to trim the "function header" line that is shown at
the end of the hunk header line, in order to avoid chomping it in
the middle of a single UTF-8 character.  It also added a facility to
define a custom callback function to make it possible to extend it
to non UTF-8 encodings.

During the following 8 1/2 years, nobody found need for this custom
callback facility.

A custom callback function is a wrong design to use here anyway---if
your contents need support for non UTF-8 encoding, you shouldn't
have to write a custom function and recompile Git to plumb it in.  A
better approach would be to extend sane_truncate_line() function and
have a new member in emit_callback to conditionally trigger it.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agofiles-backend: cheapen refname_available check when locking refs
Michael Haggerty [Thu, 17 Aug 2017 15:12:50 +0000 (17:12 +0200)] 
files-backend: cheapen refname_available check when locking refs

When locking references in preparation for updating them, we need to
check that none of the newly added references D/F conflict with
existing references (e.g., we don't allow `refs/foo` to be added if
`refs/foo/bar` already exists, or vice versa).

Prior to 524a9fdb51 (refs_verify_refname_available(): use function in
more places, 2017-04-16), conflicts with existing loose references
were checked by looking directly in the filesystem, and then conflicts
with existing packed references were checked by running
`verify_refname_available_dir()` against the packed-refs cache.

But that commit changed the final check to call
`refs_verify_refname_available()` against the *whole* files ref-store,
including both loose and packed references, with the following
comment:

> This means that those callsites now check for conflicts with all
> references rather than just packed refs, but the performance cost
> shouldn't be significant (and will be regained later).

That comment turned out to be too sanguine. User s@kazlauskas.me
reported that fetches involving a very large number of references in
neighboring directories were slowed down by that change.

The problem is that when fetching, each reference is updated
individually, within its own reference transaction. This is done
because some reference updates might succeed even though others fail.
But every time a reference update transaction is finished,
`clear_loose_ref_cache()` is called. So when it is time to update the
next reference, part of the loose ref cache has to be repopulated for
the `refs_verify_refname_available()` call. If the references are all
in neighboring directories, then the cost of repopulating the
reference cache increases with the number of references, resulting in
O(N²) effort.

The comment above also claims that the performance cost "will be
regained later". The idea was that once the packed-refs were finished
being split out into a separate ref-store, we could limit the
`refs_verify_refname_available()` call to the packed references again.
That is what we do now.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agot5526: fix some broken && chains
Heiko Voigt [Thu, 17 Aug 2017 10:36:13 +0000 (12:36 +0200)] 
t5526: fix some broken && chains

Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agosubmodule.sh: remove unused variable
Stefan Beller [Wed, 16 Aug 2017 22:50:42 +0000 (15:50 -0700)] 
submodule.sh: remove unused variable

This could have been part of 48308681b0 (git submodule update: have a
dedicated helper for cloning, 2016-02-29).

Signed-off-by: Stefan Beller <sbeller@google.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agocommit: skip discarding the index if there is no pre-commit hook
Kevin Willford [Mon, 14 Aug 2017 21:54:25 +0000 (15:54 -0600)] 
commit: skip discarding the index if there is no pre-commit hook

If there is not a pre-commit hook, there is no reason to discard
the index and reread it.

This change checks to presence of a pre-commit hook and then only
discards the index if there was one.

Signed-off-by: Kevin Willford <kewillf@microsoft.com>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agosha1_file: make read_info_alternates static
Stefan Beller [Tue, 15 Aug 2017 20:13:19 +0000 (13:13 -0700)] 
sha1_file: make read_info_alternates static

read_info_alternates is not used from outside, so let's make it static.

We have to declare the function before link_alt_odb_entry instead of
moving the code around, link_alt_odb_entry calls read_info_alternates,
which in turn calls link_alt_odb_entry.

Signed-off-by: Stefan Beller <sbeller@google.com>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agot1002: stop using sum(1)
René Scharfe [Mon, 14 Aug 2017 20:16:34 +0000 (22:16 +0200)] 
t1002: stop using sum(1)

sum(1) is a command for calculating checksums of the contents of files.
It was part of early editions of Unix ("Research Unix", 1972/1973, [1]).
cksum(1) appeared in 4.4BSD (1993) as a replacement [2], and became part
of POSIX.1-2008 [3].  OpenBSD 5.6 (2014) removed sum(1).

We only use sum(1) in t1002 to check for changes in three files.  On
MinGW we use md5sum(1) instead.  We could switch to the standard command
cksum(1) for all platforms; MinGW comes with GNU coreutils now, which
provides sum(1), cksum(1) and md5sum(1).  Use our standard method for
checking for file changes instead: test_cmp.

It's more convenient because it shows differences nicely, it's faster on
MinGW because we have a special implementation there based only on
shell-internal commands, it's simpler as it allows us to avoid stripping
out unnecessary entries from the checksum file using grep(1), and it's
more consistent with the rest of the test suite.

We already compare changed files with their expected new contents using
diff(1), so we don't need to check with "test_must_fail test_cmp" if
they differ from their original state.  A later patch could convert the
direct diff(1) calls to test_cmp as well.

With all sum(1) calls gone, remove the MinGW-specific implementation
from test-lib.sh as well.

[1] http://minnie.tuhs.org/cgi-bin/utree.pl?file=V3/man/man1/sum.1
[2] http://minnie.tuhs.org/cgi-bin/utree.pl?file=4.4BSD/usr/share/man/cat1/cksum.0
[3] http://pubs.opengroup.org/onlinepubs/9699919799/utilities/cksum.html

Signed-off-by: René Scharfe <l.s.r@web.de>
Reviewed-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agodoc: clarify "config --bool" behaviour with empty string
Andreas Heiduk [Mon, 14 Aug 2017 22:12:18 +0000 (00:12 +0200)] 
doc: clarify "config --bool" behaviour with empty string

`git config --bool xxx.yyy` returns `true` for `[xxx]yyy` but
`false` for `[xxx]yyy=` or `[xxx]yyy=""`.  This is tested in
t1300-repo-config.sh since 09bc098c2.

Signed-off-by: Andreas Heiduk <asheiduk@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agostash: prevent warning about null bytes in input
Kevin Daudt [Mon, 14 Aug 2017 21:43:33 +0000 (23:43 +0200)] 
stash: prevent warning about null bytes in input

The `no_changes` function calls the `untracked_files` function through
command substitution. `untracked_files` will return null bytes because it
runs ls-files with the '-z' option.

Bash since version 4.4 warns about these null bytes. As they are not
required for the test that is being done, make sure `untracked_files`
does not output null bytes when not required.

This is achieved by adding a parameter to the `untracked_files` function to
specify wither `-z` should be passed to ls-files or not.

This warning is triggered when running git stash save -u resulting in
two warnings:

    git-stash: line 43: warning: command substitution: ignored null byte
    in input

Signed-off-by: Kevin Daudt <me@ikke.info>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agorebase: turn on progress option by default for format-patch
Kevin Willford [Thu, 10 Aug 2017 18:32:56 +0000 (14:32 -0400)] 
rebase: turn on progress option by default for format-patch

Pass the "--progress" option to format-patch when the standard error
stream is connected to the terminal and "--quiet" is not given.

Signed-off-by: Kevin Willford <kewillf@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agoformat-patch: have progress option while generating patches
Kevin Willford [Thu, 10 Aug 2017 18:32:55 +0000 (14:32 -0400)] 
format-patch: have progress option while generating patches

When generating patches for the rebase command, if the user does
not realize the branch they are rebasing onto is thousands of
commits different, there is no progress indication after initial
rewinding message.

The progress meter as presented in this patch assumes the thousands of
patches to have a fine granularity as well as assuming to require all
the same amount of work/time for each, such that a steady progress bar
is achieved.

We do not want to estimate the time for each patch based e.g.
on their size or number of touched files (or parents) as that is too
expensive for just a progress meter.

This patch allows a progress option to be passed to format-patch
so that the user can be informed the progress of generating the
patch.  This option is then used by the rebase command when
calling format-patch.

Signed-off-by: Kevin Willford <kewillf@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agohook: use correct logical variable
Kaartic Sivaraam [Mon, 14 Aug 2017 08:46:46 +0000 (14:16 +0530)] 
hook: use correct logical variable

Sign-off added should be that of the "committer", not that of the
"commit's author"; that is how the rest of Git adds sign-off using
sequencer.c::append_signoff().

Use the correct logical variable that identifies the committer.

Signed-off-by: Kaartic Sivaraam <kaarticsivaraam91196@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agostash: clean untracked files before reset
Nicolas Morey-Chaisemartin [Fri, 11 Aug 2017 17:14:43 +0000 (19:14 +0200)] 
stash: clean untracked files before reset

If calling git stash -u on a repo that contains a file that is not
ignored any more due to a current modification of the gitignore file,
this file is stashed but not remove from the working tree.
This is due to git-stash first doing a reset --hard which clears the
.gitignore file modification and the call git clean, leaving the file
untouched.
This causes git stash pop to fail due to the file existing.

This patch simply switches the order between cleaning and resetting
and adds a test for this usecase.

Reported-by: Sam Partington <sam@whiteoctober.co.uk>
Signed-off-by: Nicolas Morey-Chaisemartin <nicolas@morey-chaisemartin.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agosha1_file: remove read_packed_sha1()
Jonathan Tan [Fri, 11 Aug 2017 20:36:15 +0000 (13:36 -0700)] 
sha1_file: remove read_packed_sha1()

Use read_object() in its place instead. This avoids duplication of code.

This makes force_object_loose() slightly slower (because of a redundant
check of loose object storage), but only in the error case.

Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agosha1_file: set whence in storage-specific info fn
Jonathan Tan [Fri, 11 Aug 2017 20:36:14 +0000 (13:36 -0700)] 
sha1_file: set whence in storage-specific info fn

Move the setting of oi->whence to sha1_loose_object_info() and
packed_object_info().

This allows sha1_object_info_extended() to not need to know about the
delta base cache. This will be useful during a future refactoring in
which packfile-related functions, including the handling of the delta
base cache, will be moved to a separate file.

Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agoThe first batch of topics after the 2.14 cycle master
Junio C Hamano [Fri, 11 Aug 2017 20:34:31 +0000 (13:34 -0700)] 
The first batch of topics after the 2.14 cycle

Notably, let's declare that we aim to make "git add ''" illegal in
the cycle after this one.

The topic to do so, ex/deprecate-empty-pathspec-as-match-all, has
been cooking in 'next' too long, and will stay there during this
cycle, but not after.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agoMerge branch 'ma/pager-per-subcommand-action'
Junio C Hamano [Fri, 11 Aug 2017 20:27:07 +0000 (13:27 -0700)] 
Merge branch 'ma/pager-per-subcommand-action'

The "tag.pager" configuration variable was useless for those who
actually create tag objects, as it interfered with the use of an
editor.  A new mechanism has been introduced for commands to enable
pager depending on what operation is being carried out to fix this,
and then "git tag -l" is made to run pager by default.

* ma/pager-per-subcommand-action:
  git.c: ignore pager.* when launching builtin as dashed external
  tag: change default of `pager.tag` to "on"
  tag: respect `pager.tag` in list-mode only
  t7006: add tests for how git tag paginates
  git.c: provide setup_auto_pager()
  git.c: let builtins opt for handling `pager.foo` themselves
  builtin.h: take over documentation from api-builtin.txt

7 years agoMerge branch 'jk/rev-list-empty-input'
Junio C Hamano [Fri, 11 Aug 2017 20:27:07 +0000 (13:27 -0700)] 
Merge branch 'jk/rev-list-empty-input'

"git log --tag=no-such-tag" showed log starting from HEAD, which
has been fixed---it now shows nothing.

* jk/rev-list-empty-input:
  revision: do not fallback to default when rev_input_given is set
  rev-list: don't show usage when we see empty ref patterns
  revision: add rev_input_given flag
  t6018: flesh out empty input/output rev-list tests

7 years agoMerge branch 'jt/t1450-fsck-corrupt-packfile'
Junio C Hamano [Fri, 11 Aug 2017 20:27:07 +0000 (13:27 -0700)] 
Merge branch 'jt/t1450-fsck-corrupt-packfile'

A test update.

* jt/t1450-fsck-corrupt-packfile:
  tests: ensure fsck fails on corrupt packfiles

7 years agoMerge branch 'js/git-gui-msgfmt-on-windows'
Junio C Hamano [Fri, 11 Aug 2017 20:27:06 +0000 (13:27 -0700)] 
Merge branch 'js/git-gui-msgfmt-on-windows'

Because recent Git for Windows do come with a real msgfmt, the
build procedure for git-gui has been updated to use it instead of a
hand-rolled substitute.

* js/git-gui-msgfmt-on-windows:
  git-gui (MinGW): make use of MSys2's msgfmt
  git gui: allow for a long recentrepo list
  git gui: de-dup selected repo from recentrepo history
  git gui: cope with duplicates in _get_recentrepo
  git-gui: remove duplicate entries from .gitconfig's gui.recentrepo

7 years agoMerge branch 'eb/contacts-reported-by'
Junio C Hamano [Fri, 11 Aug 2017 20:27:06 +0000 (13:27 -0700)] 
Merge branch 'eb/contacts-reported-by'

"git contacts" (in contrib/) now lists the address on the
"Reported-by:" trailer to its output, in addition to those on
S-o-b: and other trailers, to make it easier to notify (and thank)
the original bug reporter.

* eb/contacts-reported-by:
  git-contacts: also recognise "Reported-by:"

7 years agoMerge branch 'dl/credential-cache-socket-in-xdg-cache'
Junio C Hamano [Fri, 11 Aug 2017 20:27:06 +0000 (13:27 -0700)] 
Merge branch 'dl/credential-cache-socket-in-xdg-cache'

A recently added test for the "credential-cache" helper revealed
that EOF detection done around the time the connection to the cache
daemon is torn down were flaky.  This was fixed by reacting to
ECONNRESET and behaving as if we got an EOF.

* dl/credential-cache-socket-in-xdg-cache:
  credential-cache: interpret an ECONNRESET as an EOF

7 years agoMerge branch 'rg/rerere-train-overwrite'
Junio C Hamano [Fri, 11 Aug 2017 20:27:05 +0000 (13:27 -0700)] 
Merge branch 'rg/rerere-train-overwrite'

The "rerere-train" script (in contrib/) learned the "--overwrite"
option to allow overwriting existing recorded resolutions.

* rg/rerere-train-overwrite:
  contrib/rerere-train: optionally overwrite existing resolutions

7 years agoMerge branch 'jb/t8008-cleanup'
Junio C Hamano [Fri, 11 Aug 2017 20:27:05 +0000 (13:27 -0700)] 
Merge branch 'jb/t8008-cleanup'

Code clean-up.

* jb/t8008-cleanup:
  t8008: rely on rev-parse'd HEAD instead of sha1 value

7 years agoMerge branch 'jt/subprocess-handshake'
Junio C Hamano [Fri, 11 Aug 2017 20:27:05 +0000 (13:27 -0700)] 
Merge branch 'jt/subprocess-handshake'

Code cleanup.

* jt/subprocess-handshake:
  sub-process: refactor handshake to common function
  Documentation: migrate sub-process docs to header

7 years agoMerge branch 'dc/fmt-merge-msg-microcleanup'
Junio C Hamano [Fri, 11 Aug 2017 20:27:05 +0000 (13:27 -0700)] 
Merge branch 'dc/fmt-merge-msg-microcleanup'

Code cleanup.

* dc/fmt-merge-msg-microcleanup:
  fmt-merge-msg: fix coding style

7 years agoMerge branch 'ah/doc-wserrorhighlight'
Junio C Hamano [Fri, 11 Aug 2017 20:27:04 +0000 (13:27 -0700)] 
Merge branch 'ah/doc-wserrorhighlight'

Doc update.

* ah/doc-wserrorhighlight:
  doc: add missing values "none" and "default" for diff.wsErrorHighlight

7 years agoMerge branch 'cc/ref-is-hidden-microcleanup'
Junio C Hamano [Fri, 11 Aug 2017 20:27:03 +0000 (13:27 -0700)] 
Merge branch 'cc/ref-is-hidden-microcleanup'

Code cleanup.

* cc/ref-is-hidden-microcleanup:
  refs: use skip_prefix() in ref_is_hidden()

7 years agoMerge branch 'js/run-process-parallel-api-fix'
Junio C Hamano [Fri, 11 Aug 2017 20:27:02 +0000 (13:27 -0700)] 
Merge branch 'js/run-process-parallel-api-fix'

API fix.

* js/run-process-parallel-api-fix:
  run_processes_parallel: change confusing task_cb convention

7 years agoMerge branch 'hb/gitweb-project-list'
Junio C Hamano [Fri, 11 Aug 2017 20:27:01 +0000 (13:27 -0700)] 
Merge branch 'hb/gitweb-project-list'

When a directory is not readable, "gitweb" fails to build the
project list.  Work this around by skipping such a directory.

* hb/gitweb-project-list:
  gitweb: skip unreadable subdirectories

7 years agoMerge branch 'ks/commit-abort-on-empty-message-fix'
Junio C Hamano [Fri, 11 Aug 2017 20:27:01 +0000 (13:27 -0700)] 
Merge branch 'ks/commit-abort-on-empty-message-fix'

"git commit" when seeing an totally empty message said "you did not
edit the message", which is clearly wrong.  The message has been
corrected.

* ks/commit-abort-on-empty-message-fix:
  commit: check for empty message before the check for untouched template

7 years agoMerge branch 'sb/hashmap-cleanup'
Junio C Hamano [Fri, 11 Aug 2017 20:27:01 +0000 (13:27 -0700)] 
Merge branch 'sb/hashmap-cleanup'

Many uses of comparision callback function the hashmap API uses
cast the callback function type when registering it to
hashmap_init(), which defeats the compile time type checking when
the callback interface changes (e.g. gaining more parameters).
The callback implementations have been updated to take "void *"
pointers and cast them to the type they expect instead.

* sb/hashmap-cleanup:
  t/helper/test-hashmap: use custom data instead of duplicate cmp functions
  name-hash.c: drop hashmap_cmp_fn cast
  submodule-config.c: drop hashmap_cmp_fn cast
  remote.c: drop hashmap_cmp_fn cast
  patch-ids.c: drop hashmap_cmp_fn cast
  convert/sub-process: drop cast to hashmap_cmp_fn
  config.c: drop hashmap_cmp_fn cast
  builtin/describe: drop hashmap_cmp_fn cast
  builtin/difftool.c: drop hashmap_cmp_fn cast
  attr.c: drop hashmap_cmp_fn cast

7 years agoMerge branch 'jk/reflog-walk'
Junio C Hamano [Fri, 11 Aug 2017 20:27:00 +0000 (13:27 -0700)] 
Merge branch 'jk/reflog-walk'

Numerous bugs in walking of reflogs via "log -g" and friends have
been fixed.

* jk/reflog-walk:
  reflog-walk: apply --since/--until to reflog dates
  reflog-walk: stop using fake parents
  rev-list: check reflog_info before showing usage
  get_revision_1(): replace do-while with an early return
  log: do not free parents when walking reflog
  log: clarify comment about reflog cycles
  revision: disallow reflog walking with revs->limited
  t1414: document some reflog-walk oddities

7 years agoMerge branch 'ls/filter-process-delayed'
Junio C Hamano [Fri, 11 Aug 2017 20:27:00 +0000 (13:27 -0700)] 
Merge branch 'ls/filter-process-delayed'

The filter-process interface learned to allow a process with long
latency give a "delayed" response.

* ls/filter-process-delayed:
  convert: add "status=delayed" to filter process protocol
  convert: refactor capabilities negotiation
  convert: move multiple file filter error handling to separate function
  convert: put the flags field before the flag itself for consistent style
  t0021: write "OUT <size>" only on success
  t0021: make debug log file name configurable
  t0021: keep filter log files on comparison

7 years agoMerge branch 'st/lib-gpg-kill-stray-agent'
Junio C Hamano [Fri, 11 Aug 2017 20:27:00 +0000 (13:27 -0700)] 
Merge branch 'st/lib-gpg-kill-stray-agent'

Some versions of GnuPG fails to kill gpg-agent it auto-spawned
and such a left-over agent can interfere with a test.  Work it
around by attempting to kill one before starting a new test.

* st/lib-gpg-kill-stray-agent:
  t: lib-gpg: flush gpg agent on startup

7 years agoMerge branch 'rs/pack-objects-pbase-cleanup'
Junio C Hamano [Fri, 11 Aug 2017 20:27:00 +0000 (13:27 -0700)] 
Merge branch 'rs/pack-objects-pbase-cleanup'

Code clean-up.

* rs/pack-objects-pbase-cleanup:
  pack-objects: remove unnecessary NULL check

7 years agoMerge branch 'jt/fsck-code-cleanup'
Junio C Hamano [Fri, 11 Aug 2017 20:26:59 +0000 (13:26 -0700)] 
Merge branch 'jt/fsck-code-cleanup'

Code clean-up.

* jt/fsck-code-cleanup:
  fsck: cleanup unused variable
  object: remove "used" field from struct object
  fsck: remove redundant parse_tree() invocation

7 years agoMerge branch 'jc/http-sslkey-and-ssl-cert-are-paths'
Junio C Hamano [Fri, 11 Aug 2017 20:26:59 +0000 (13:26 -0700)] 
Merge branch 'jc/http-sslkey-and-ssl-cert-are-paths'

The http.{sslkey,sslCert} configuration variables are to be
interpreted as a pathname that honors "~[username]/" prefix, but
weren't, which has been fixed.

* jc/http-sslkey-and-ssl-cert-are-paths:
  http.c: http.sslcert and http.sslkey are both pathnames

7 years agoMerge branch 'jk/c99'
Junio C Hamano [Fri, 11 Aug 2017 20:26:58 +0000 (13:26 -0700)] 
Merge branch 'jk/c99'

Start using selected c99 constructs in small, stable and
essentialpart of the system to catch people who care about
older compilers that do not grok them.

* jk/c99:
  clean.c: use designated initializer
  strbuf: use designated initializers in STRBUF_INIT

7 years agoMerge branch 'jk/ref-filter-colors'
Junio C Hamano [Fri, 11 Aug 2017 20:26:58 +0000 (13:26 -0700)] 
Merge branch 'jk/ref-filter-colors'

"%C(color name)" in the pretty print format always produced ANSI
color escape codes, which was an early design mistake.  They now
honor the configuration (e.g. "color.ui = never") and also tty-ness
of the output medium.

* jk/ref-filter-colors:
  ref-filter: consult want_color() before emitting colors
  pretty: respect color settings for %C placeholders
  rev-list: pass diffopt->use_colors through to pretty-print
  for-each-ref: load config earlier
  color: check color.ui in git_default_config()
  ref-filter: pass ref_format struct to atom parsers
  ref-filter: factor out the parsing of sorting atoms
  ref-filter: make parse_ref_filter_atom a private function
  ref-filter: provide a function for parsing sort options
  ref-filter: move need_color_reset_at_eol into ref_format
  ref-filter: abstract ref format into its own struct
  ref-filter: simplify automatic color reset
  t: use test_decode_color rather than literal ANSI codes
  docs/for-each-ref: update pointer to color syntax
  check return value of verify_ref_format()

7 years agoMerge branch 'wd/rebase-conflict-guide'
Junio C Hamano [Fri, 11 Aug 2017 20:26:58 +0000 (13:26 -0700)] 
Merge branch 'wd/rebase-conflict-guide'

The advice message given when "git rebase" stops for conflicting
changes has been improved.

* wd/rebase-conflict-guide:
  rebase: make resolve message clearer for inexperienced users

7 years agoMerge branch 'rs/stat-data-unaligned-reads-fix'
Junio C Hamano [Fri, 11 Aug 2017 20:26:58 +0000 (13:26 -0700)] 
Merge branch 'rs/stat-data-unaligned-reads-fix'

Code clean-up.

* rs/stat-data-unaligned-reads-fix:
  dir: support platforms that require aligned reads

7 years agoMerge branch 'rs/move-array'
Junio C Hamano [Fri, 11 Aug 2017 20:26:57 +0000 (13:26 -0700)] 
Merge branch 'rs/move-array'

Code clean-up.

* rs/move-array:
  ls-files: don't try to prune an empty index
  apply: use COPY_ARRAY and MOVE_ARRAY in update_image()
  use MOVE_ARRAY
  add MOVE_ARRAY

7 years agoMerge branch 'rs/bswap-ubsan-fix'
Junio C Hamano [Fri, 11 Aug 2017 20:26:57 +0000 (13:26 -0700)] 
Merge branch 'rs/bswap-ubsan-fix'

Code clean-up.

* rs/bswap-ubsan-fix:
  bswap: convert get_be16, get_be32 and put_be32 to inline functions
  bswap: convert to unsigned before shifting in get_be32

7 years agoMerge branch 'ks/prepare-commit-msg-sample'
Junio C Hamano [Fri, 11 Aug 2017 20:26:56 +0000 (13:26 -0700)] 
Merge branch 'ks/prepare-commit-msg-sample'

Remove an example that is now obsolete from a sample hook,
and improve an old example in it that added a sign-off manually
to use the interpret-trailers command.

* ks/prepare-commit-msg-sample:
  hook: add a simple first example
  hook: add sign-off using "interpret-trailers"
  hook: name the positional variables
  hook: cleanup script

7 years agoMerge branch 'bw/object-id'
Junio C Hamano [Fri, 11 Aug 2017 20:26:56 +0000 (13:26 -0700)] 
Merge branch 'bw/object-id'

Conversion from uchar[20] to struct object_id continues.

* bw/object-id:
  receive-pack: don't access hash of NULL object_id pointer
  notes: don't access hash of NULL object_id pointer
  tree-diff: don't access hash of NULL object_id pointer

7 years agoMerge branch 'bc/object-id'
Junio C Hamano [Fri, 11 Aug 2017 20:26:55 +0000 (13:26 -0700)] 
Merge branch 'bc/object-id'

Conversion from uchar[20] to struct object_id continues.

* bc/object-id:
  sha1_name: convert uses of 40 to GIT_SHA1_HEXSZ
  sha1_name: convert GET_SHA1* flags to GET_OID*
  sha1_name: convert get_sha1* to get_oid*
  Convert remaining callers of get_sha1 to get_oid.
  builtin/unpack-file: convert to struct object_id
  bisect: convert bisect_checkout to struct object_id
  builtin/update_ref: convert to struct object_id
  sequencer: convert to struct object_id
  remote: convert struct push_cas to struct object_id
  submodule: convert submodule config lookup to use object_id
  builtin/merge-tree: convert remaining caller of get_sha1 to object_id
  builtin/fsck: convert remaining caller of get_sha1 to object_id

7 years agoMerge branch 'sb/object-id'
Junio C Hamano [Fri, 11 Aug 2017 20:26:55 +0000 (13:26 -0700)] 
Merge branch 'sb/object-id'

Conversion from uchar[20] to struct object_id continues.

* sb/object-id:
  tag: convert gpg_verify_tag to use struct object_id
  commit: convert lookup_commit_graft to struct object_id

7 years agosha1_file: release delta_stack on error in unpack_entry()
René Scharfe [Thu, 10 Aug 2017 09:42:21 +0000 (11:42 +0200)] 
sha1_file: release delta_stack on error in unpack_entry()

When unpack_entry() encounters a broken packed object, it returns early.
It adjusts the reference count of the pack window, but leaks the buffer
for a big delta stack in case the small automatic one was not enough.
Jump to the cleanup code at end instead, which takes care of that.

Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agofsck: free buffers on error in fsck_obj()
René Scharfe [Thu, 10 Aug 2017 09:42:10 +0000 (11:42 +0200)] 
fsck: free buffers on error in fsck_obj()

Move the code for releasing tree buffers and commit buffers in
fsck_obj() to the end of the function and make sure it's executed no
matter of an error is encountered or not.

Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agostrbuf: clear errno before calling getdelim(3)
René Scharfe [Thu, 10 Aug 2017 20:56:40 +0000 (22:56 +0200)] 
strbuf: clear errno before calling getdelim(3)

getdelim(3) returns -1 at the end of the file and if it encounters an
error, but sets errno only in the latter case.  Set errno to zero before
calling it to avoid misdiagnosing an out-of-memory condition due to a
left-over value from some other function call.

Reported-by: Yaroslav Halchenko <yoh@onerussian.com>
Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agowin32: plug memory leak on realloc() failure in syslog()
René Scharfe [Thu, 10 Aug 2017 10:23:45 +0000 (12:23 +0200)] 
win32: plug memory leak on realloc() failure in syslog()

If realloc() fails then the original buffer is still valid.  Free it
before exiting the function.

Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agomerge: use skip_prefix()
René Scharfe [Thu, 10 Aug 2017 16:47:55 +0000 (18:47 +0200)] 
merge: use skip_prefix()

Get rid of a magic string length constant by using skip_prefix() instead
of starts_with().

Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agot1200: remove t1200-tutorial.sh
Stefan Beller [Wed, 9 Aug 2017 21:22:53 +0000 (14:22 -0700)] 
t1200: remove t1200-tutorial.sh

v1.2.0~121 (New tutorial, 2006-01-22) rewrote the tutorial such that the
original intent of 2ae6c70674 (Adapt tutorial to cygwin and add test case,
2005-10-13) to test the examples from the tutorial doesn't hold any more.

There are dedicated tests for the commands used, even "git whatchanged",
such that removing these tests doesn't seem like a reduction in test
coverage.

Signed-off-by: Stefan Beller <sbeller@google.com>
Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agosha1_file: drop experimental GIT_USE_LOOKUP search
Jeff King [Wed, 9 Aug 2017 10:14:32 +0000 (06:14 -0400)] 
sha1_file: drop experimental GIT_USE_LOOKUP search

Long ago in 628522ec14 (sha1-lookup: more memory efficient
search in sorted list of SHA-1, 2007-12-29) we added
sha1_entry_pos(), a binary search that uses the uniform
distribution of sha1s to scale the selection of mid-points.
As this was a performance experiment, we tied it to the
GIT_USE_LOOKUP environment variable and never enabled it by
default.

This code was successful in reducing the number of steps in
each search. But the overhead of the scaling ends up making
it slower when the cache is warm. Here are best-of-five
timings for running rev-list on linux.git, which will have
to look up every object:

  $ time git rev-list --objects --all >/dev/null
  real 0m35.357s
  user 0m35.016s
  sys 0m0.340s

  $ time GIT_USE_LOOKUP=1 git rev-list --objects --all >/dev/null
  real 0m37.364s
  user 0m37.045s
  sys 0m0.316s

The USE_LOOKUP version might have more benefit on a cold
cache, as the time to fault in each page would dominate. But
that would be for a single lookup. In practice, most
operations tend to look up many objects, and the whole pack
.idx will end up warm.

It's possible that the code could be better optimized to
compete with a naive binary search for the warm-cache case,
and we could have the best of both worlds. But over the
years nobody has done so, and this is largely dead code that
is rarely run outside of the test suite. Let's drop it in
the name of simplicity.

This lets us remove sha1_entry_pos() entirely, as the .idx
lookup code was the only caller.  Note that sha1-lookup.c
still contains sha1_pos(), which differs from
sha1_entry_pos() in two ways:

  - it has a different interface; it uses a function pointer
    to access sha1 entries rather than a size/offset pair
    describing the table's memory layout

  - it only scales the initial selection of "mi", rather
    than each iteration of the search

We can't get rid of this function, as it's called from
several places. It may be that we could replace it with a
simple binary search, but that's out of scope for this patch
(and would need benchmarking).

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agohashcmp: use memcmp instead of open-coded loop
Jeff King [Wed, 9 Aug 2017 10:16:45 +0000 (06:16 -0400)] 
hashcmp: use memcmp instead of open-coded loop

In 1a812f3a70 (hashcmp(): inline memcmp() by hand to
optimize, 2011-04-28), it was reported that an open-coded
loop outperformed memcmp() for comparing sha1s.

Discussion[1] a few years later in 2013 showed that this
depends on your libc's version of memcmp(). In particular,
glibc 2.13 optimized their memcmp around 2011. Here are
current timings with glibc 2.24 (best-of-five, on
linux.git):

  [before this patch, open-coded]
  $ time git rev-list --objects --all
  real 0m35.357s
  user 0m35.016s
  sys 0m0.340s

  [after this patch, memcmp]
  real 0m32.930s
  user 0m32.630s
  sys 0m0.300s

Now that we've had 6 years for that version of glibc to
make its way onto people's machines, it's worth revisiting
our benchmarks and switching to memcmp().

It may be that there are other non-glibc systems where
memcmp() isn't as well optimized. But since our single data
point in favor of open-coding was on a now-ancient glibc, we
should probably assume the system memcmp is good unless
proven otherwise. We may end up with a SLOW_MEMCMP Makefile
knob, but we can hold off on that until we actually find
such a system in practice.

[1] https://public-inbox.org/git/20130318073229.GA5551@sigill.intra.peff.net/

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agoapply: remove prefix_length member from apply_state
René Scharfe [Wed, 9 Aug 2017 15:54:46 +0000 (17:54 +0200)] 
apply: remove prefix_length member from apply_state

Use a NULL-and-NUL check to see if we have a prefix and consistently use
C string functions on it instead of storing its length in a member of
struct apply_state.  This avoids strlen() calls and simplifies the code.

Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agobuiltin/add: add detail to a 'cannot chmod' error message
Ramsay Jones [Wed, 9 Aug 2017 00:51:23 +0000 (01:51 +0100)] 
builtin/add: add detail to a 'cannot chmod' error message

In addition to adding the missing newline, add the x-ecutable bit
'mode change' character to the error message. This message now has
the same form as similar messages output by 'update-index'.

Signed-off-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
7 years agosha1_file: avoid comparison if no packed hash matches the first byte
René Scharfe [Tue, 8 Aug 2017 22:07:30 +0000 (00:07 +0200)] 
sha1_file: avoid comparison if no packed hash matches the first byte

find_pack_entry_one() uses the fan-out table of pack indexes to find out
which entries match the first byte of the searched hash and does a
binary search on this subset of the main index table.

If there are no matching entries then lo and hi will have the same
value.  The binary search still starts and compares the hash of the
following entry (which has a non-matching first byte, so won't cause any
trouble), or whatever comes after the sorted list of entries.

The probability of that stray comparison matching by mistake is low, but
let's not take any chances and check when entering the binary search
loop if we're actually done already.

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