git/git.git
3 years agorepository: introduce the repository object
Brandon Williams [Thu, 22 Jun 2017 18:43:32 +0000 (11:43 -0700)]
repository: introduce the repository object

Introduce the repository object 'struct repository' which can be used to
hold all state pertaining to a git repository.

Some of the benefits of object-ifying a repository are:

  1. Make the code base more readable and easier to reason about.

  2. Allow for working on multiple repositories, specifically
     submodules, within the same process.  Currently the process for
     working on a submodule involves setting up an argv_array of options
     for a particular command and then launching a child process to
     execute the command in the context of the submodule.  This is
     clunky and can require lots of little hacks in order to ensure
     correctness.  Ideally it would be nice to simply pass a repository
     and an options struct to a command.

  3. Eliminating reliance on global state will make it easier to
     enable the use of threading to improve performance.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoenvironment: remove namespace_len variable
Brandon Williams [Tue, 20 Jun 2017 19:19:34 +0000 (12:19 -0700)]
environment: remove namespace_len variable

Use 'skip_prefix' instead of 'starts_with' so that we can drop the need
to keep around 'namespace_len'.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agosetup: add comment indicating a hack
Brandon Williams [Tue, 20 Jun 2017 19:19:33 +0000 (12:19 -0700)]
setup: add comment indicating a hack

'GIT_TOPLEVEL_PREFIX_ENVIRONMENT' was added in (b58a68c1c setup: allow
for prefix to be passed to git commands) to aid in fixing a bug where
'ls-files' and 'grep' were not able to properly recurse when called from
within a subdirectory.  Add a 'NEEDSWORK' comment indicating that this
envvar should be removed once 'ls-files' and 'grep' can recurse
in-process.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agosetup: don't perform lazy initialization of repository state
Brandon Williams [Tue, 20 Jun 2017 19:19:32 +0000 (12:19 -0700)]
setup: don't perform lazy initialization of repository state

Under some circumstances (bogus GIT_DIR value or the discovered gitdir
is '.git') 'setup_git_directory()' won't initialize key repository
state.  This leads to inconsistent state after running the setup code.
To account for this inconsistent state, lazy initialization is done once
a caller asks for the repository's gitdir or some other piece of
repository state.  This is confusing and can be error prone.

Instead let's tighten the expected outcome of 'setup_git_directory()'
and ensure that it initializes repository state in all cases that would
have been handled by lazy initialization.

This also lets us drop the requirement to have 'have_git_dir()' check if
the environment variable GIT_DIR was set as that will be handled by the
end of the setup code.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoMerge branches 'bw/ls-files-sans-the-index' and 'bw/config-h' into bw/repo-object
Junio C Hamano [Wed, 21 Jun 2017 22:20:44 +0000 (15:20 -0700)]
Merge branches 'bw/ls-files-sans-the-index' and 'bw/config-h' into bw/repo-object

* bw/ls-files-sans-the-index:
  ls-files: factor out tag calculation
  ls-files: factor out debug info into a function
  ls-files: convert show_files to take an index
  ls-files: convert show_ce_entry to take an index
  ls-files: convert prune_cache to take an index
  ls-files: convert ce_excluded to take an index
  ls-files: convert show_ru_info to take an index
  ls-files: convert show_other_files to take an index
  ls-files: convert show_killed_files to take an index
  ls-files: convert write_eolinfo to take an index
  ls-files: convert overlay_tree_on_cache to take an index
  tree: convert read_tree to take an index parameter
  convert: convert renormalize_buffer to take an index
  convert: convert convert_to_git to take an index
  convert: convert convert_to_git_filter_fd to take an index
  convert: convert crlf_to_git to take an index
  convert: convert get_cached_convert_stats_ascii to take an index

* bw/config-h:
  config: don't implicitly use gitdir or commondir
  config: respect commondir
  setup: teach discover_git_directory to respect the commondir
  config: don't include config.h by default
  config: remove git_config_iter
  config: create config.h
  alias: use the early config machinery to expand aliases
  t7006: demonstrate a problem with aliases in subdirectories
  t1308: relax the test verifying that empty alias values are disallowed
  help: use early config when autocorrecting aliases
  config: report correct line number upon error
  discover_git_directory(): avoid setting invalid git_dir

3 years agoNinth batch for 2.14
Junio C Hamano [Mon, 19 Jun 2017 19:41:12 +0000 (12:41 -0700)]
Ninth batch for 2.14

Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoMerge branch 'jk/consistent-h'
Junio C Hamano [Mon, 19 Jun 2017 19:38:45 +0000 (12:38 -0700)]
Merge branch 'jk/consistent-h'

"git $cmd -h" for builtin commands calls the implementation of the
command (i.e. cmd_$cmd() function) without doing any repository
set-up, and the commands that expect RUN_SETUP is done by the Git
potty needs to be prepared to show the help text without barfing.

* jk/consistent-h:
  t0012: test "-h" with builtins
  git: add hidden --list-builtins option
  version: convert to parse-options
  diff- and log- family: handle "git cmd -h" early
  submodule--helper: show usage for "-h"
  remote-{ext,fd}: print usage message on invalid arguments
  upload-archive: handle "-h" option early
  credential: handle invalid arguments earlier

3 years agoMerge branch 'ab/perf-remove-index-lock'
Junio C Hamano [Mon, 19 Jun 2017 19:38:44 +0000 (12:38 -0700)]
Merge branch 'ab/perf-remove-index-lock'

When an existing repository is used for t/perf testing, we first
create bit-for-bit copy of it, which may grab a transient state of
the repository and freeze it into the repository used for testing,
which then may cause Git operations to fail.  Single out "the index
being locked" case and forcibly drop the lock from the copy.

* ab/perf-remove-index-lock:
  perf: work around the tested repo having an index.lock

3 years agoMerge branch 'bw/object-id'
Junio C Hamano [Mon, 19 Jun 2017 19:38:44 +0000 (12:38 -0700)]
Merge branch 'bw/object-id'

Conversion from uchar[20] to struct object_id continues.

* bw/object-id: (33 commits)
  diff: rename diff_fill_sha1_info to diff_fill_oid_info
  diffcore-rename: use is_empty_blob_oid
  tree-diff: convert path_appendnew to object_id
  tree-diff: convert diff_tree_paths to struct object_id
  tree-diff: convert try_to_follow_renames to struct object_id
  builtin/diff-tree: cleanup references to sha1
  diff-tree: convert diff_tree_sha1 to struct object_id
  notes-merge: convert write_note_to_worktree to struct object_id
  notes-merge: convert verify_notes_filepair to struct object_id
  notes-merge: convert find_notes_merge_pair_ps to struct object_id
  notes-merge: convert merge_from_diffs to struct object_id
  notes-merge: convert notes_merge* to struct object_id
  tree-diff: convert diff_root_tree_sha1 to struct object_id
  combine-diff: convert find_paths_* to struct object_id
  combine-diff: convert diff_tree_combined to struct object_id
  diff: convert diff_flush_patch_id to struct object_id
  patch-ids: convert to struct object_id
  diff: finish conversion for prepare_temp_file to struct object_id
  diff: convert reuse_worktree_file to struct object_id
  diff: convert fill_filespec to struct object_id
  ...

3 years agoMerge branch 'sb/submodule-rm-absorb'
Junio C Hamano [Mon, 19 Jun 2017 19:38:44 +0000 (12:38 -0700)]
Merge branch 'sb/submodule-rm-absorb'

Doc update to a recently graduated topic.

* sb/submodule-rm-absorb:
  Documentation/git-rm: correct submodule description

3 years agoMerge branch 'ab/pcre-v2'
Junio C Hamano [Mon, 19 Jun 2017 19:38:43 +0000 (12:38 -0700)]
Merge branch 'ab/pcre-v2'

Update "perl-compatible regular expression" support to enable JIT
and also allow linking with the newer PCRE v2 library.

* ab/pcre-v2:
  grep: add support for PCRE v2
  grep: un-break building with PCRE >= 8.32 without --enable-jit
  grep: un-break building with PCRE < 8.20
  grep: un-break building with PCRE < 8.32
  grep: add support for the PCRE v1 JIT API
  log: add -P as a synonym for --perl-regexp
  grep: skip pthreads overhead when using one thread
  grep: don't redundantly compile throwaway patterns under threading

3 years agoMerge branch 'jk/pathspec-magic-disambiguation'
Junio C Hamano [Mon, 19 Jun 2017 19:38:42 +0000 (12:38 -0700)]
Merge branch 'jk/pathspec-magic-disambiguation'

The convention for a command line is to follow "git cmdname
--options" with revisions followed by an optional "--"
disambiguator and then finally pathspecs.  When "--" is not there,
we make sure early ones are all interpretable as revs (and do not
look like paths) and later ones are the other way around.  A
pathspec with "magic" (e.g. ":/p/a/t/h" that matches p/a/t/h from
the top-level of the working tree, no matter what subdirectory you
are working from) are conservatively judged as "not a path", which
required disambiguation more often.  The command line parser
learned to say "it's a pathspec" a bit more often when the syntax
looks like so.

* jk/pathspec-magic-disambiguation:
  verify_filename(): flip order of checks
  verify_filename(): treat ":(magic)" as a pathspec
  check_filename(): handle ":^" path magic
  check_filename(): use skip_prefix
  check_filename(): refactor ":/" handling
  t4208: add check for ":/" without matching file

3 years agosub-process: correct path to API docs in a comment
Christian Couder [Wed, 14 Jun 2017 15:12:25 +0000 (17:12 +0200)]
sub-process: correct path to API docs in a comment

Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Acked-by: Ben Peart <peartben@gmail.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoMerge branch 'svn-doc' of git://bogomips.org/git-svn
Junio C Hamano [Thu, 15 Jun 2017 21:15:03 +0000 (14:15 -0700)]
Merge branch 'svn-doc' of git://bogomips.org/git-svn

* 'svn-doc' of git://bogomips.org/git-svn:
  git-svn: document special options for commit-diff

3 years agoconfig: don't implicitly use gitdir or commondir
Brandon Williams [Wed, 14 Jun 2017 18:07:39 +0000 (11:07 -0700)]
config: don't implicitly use gitdir or commondir

'git_config_with_options()' takes a 'config_options' struct which
contains feilds for 'git_dir' and 'commondir'.  If those feilds happen
to be NULL the config machinery falls back to querying global repository
state.  Let's change this and instead use these fields in the
'config_options' struct explicilty all the time.  Since the API is
slightly changing to require these two fields to be set if callers want
the config machinery to load the repository's config, let's change the
name to 'config_with_optison()'.  This allows the config machinery to
not implicitly rely on any global repository state.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconfig: respect commondir
Brandon Williams [Wed, 14 Jun 2017 18:07:38 +0000 (11:07 -0700)]
config: respect commondir

Worktrees present an interesting problem when it comes to the config.
Historically we could assume that the per-repository config lives at
'gitdir/config', but since worktrees were introduced this isn't the case
anymore.  There is currently no way to specify per-worktree
configuration, and as such the repository config is shared with all
worktrees and is located at 'commondir/config'.

Many users of the config machinery correctly set
'config_options.git_dir' with the repository's commondir, allowing the
config to be properly loaded when operating in a worktree.  But other's,
like 'read_early_config()', set 'config_options.git_dir' with the
repository's gitdir which can be incorrect when using worktrees.

To fix this issue, and to make things less ambiguous, lets add a
'commondir' field to the 'config_options' struct and have all callers
properly set both the 'git_dir' and 'commondir' fields so that the
config machinery is able to properly find the repository's config.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agosetup: teach discover_git_directory to respect the commondir
Brandon Williams [Wed, 14 Jun 2017 18:07:37 +0000 (11:07 -0700)]
setup: teach discover_git_directory to respect the commondir

Currently 'discover_git_directory' only looks at the gitdir to determine
if a git directory was discovered.  This causes a problem in the event
that the gitdir which was discovered was in fact a per-worktree git
directory and not the common git directory.  This is because the
repository config, which is checked to verify the repository's format,
is stored in the commondir and not in the per-worktree gitdir.  Correct
this behavior by checking the config stored in the commondir.

It will also be of use for callers to have access to the commondir, so
lets also return that upon successfully discovering a git directory.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconfig: don't include config.h by default
Brandon Williams [Wed, 14 Jun 2017 18:07:36 +0000 (11:07 -0700)]
config: don't include config.h by default

Stop including config.h by default in cache.h.  Instead only include
config.h in those files which require use of the config system.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconfig: remove git_config_iter
Brandon Williams [Wed, 14 Jun 2017 18:07:35 +0000 (11:07 -0700)]
config: remove git_config_iter

Since there is no implementation of the function 'git_config_iter',
let's stop exporting it and remove the prototype from config.h.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconfig: create config.h
Brandon Williams [Wed, 14 Jun 2017 18:07:34 +0000 (11:07 -0700)]
config: create config.h

Move all config related declarations from cache.h to a new config.h
header file.  This makes cache.h smaller and allows for the opportunity
in a following patch to only include config.h when needed.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoalias: use the early config machinery to expand aliases
Johannes Schindelin [Wed, 14 Jun 2017 11:36:00 +0000 (13:36 +0200)]
alias: use the early config machinery to expand aliases

Instead of discovering the .git/ directory, reading the config and then
trying to painstakingly reset all the global state if we did not find a
matching alias, let's use the early config machinery instead.

It may look like unnecessary work to discover the .git/ directory in the
early config machinery and then call setup_git_directory_gently() in the
case of a shell alias, repeating the very same discovery *again*.
However, we have to do this as the early config machinery takes pains
*not* to touch any global state, while shell aliases expect a possibly
changed working directory and at least the GIT_PREFIX and GIT_DIR
variables to be set.

This change also fixes a known issue where Git tried to read the pager
config from an incorrect path in a subdirectory of a Git worktree if an
alias expanded to a shell command.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agot7006: demonstrate a problem with aliases in subdirectories
Johannes Schindelin [Wed, 14 Jun 2017 11:35:56 +0000 (13:35 +0200)]
t7006: demonstrate a problem with aliases in subdirectories

When expanding aliases, the git_dir is set during the alias expansion
(by virtue of running setup_git_directory_gently()).

This git_dir may be relative to the current working directory, and
indeed often is simply ".git/".

When the alias expands to a shell command, we restore the original
working directory, though, yet we do not reset git_dir.

As a consequence, subsequent read_early_config() runs will mistake the
git_dir to be populated properly and not find the correct config.

Demonstrate this problem by adding a test case.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agot1308: relax the test verifying that empty alias values are disallowed
Johannes Schindelin [Wed, 14 Jun 2017 11:35:53 +0000 (13:35 +0200)]
t1308: relax the test verifying that empty alias values are disallowed

We are about to change the way aliases are expanded, to use the early
config machinery.

This machinery reports errors in a slightly different manner than the
cached config machinery.

Let's not get hung up by the precise wording of the message mentioning
the line number. It is really sufficient to verify that all the relevant
information is given to the user.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agohelp: use early config when autocorrecting aliases
Johannes Schindelin [Wed, 14 Jun 2017 11:35:50 +0000 (13:35 +0200)]
help: use early config when autocorrecting aliases

Git has this feature which suggests similar commands (including aliases)
in case the user specified an unknown command.

This feature currently relies on a side effect of the way we expand
aliases right now: when a command is not a builtin, we use the regular
config machinery (meaning: discovering the .git/ directory and
initializing global state such as the config cache) to see whether the
command refers to an alias.

However, we will change the way aliases are expanded in the next
commits, to use the early config instead. That means that the
autocorrect feature can no longer discover the available aliases by
looking at the config cache (because it has not yet been initialized).

So let's just use the early config machinery instead.

This is slightly less performant than the previous way, as the early
config is used *twice*: once to see whether the command refers to an
alias, and then to see what aliases are most similar. However, this is
hardly a performance-critical code path, so performance is less important
here.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconfig: report correct line number upon error
Johannes Schindelin [Wed, 14 Jun 2017 11:35:46 +0000 (13:35 +0200)]
config: report correct line number upon error

When get_value() parses a key/value pair, it is possible that the line
number is decreased (because the \n has been consumed already) before the
key/value pair is passed to the callback function, to allow for the
correct line to be attributed in case of an error.

However, when git_parse_source() asks get_value() to parse the key/value
pair, the error reporting is performed *after* get_value() returns.

Which means that we have to be careful not to increase the line number
in get_value() after the callback function returned an error.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agodiscover_git_directory(): avoid setting invalid git_dir
Johannes Schindelin [Wed, 14 Jun 2017 11:35:26 +0000 (13:35 +0200)]
discover_git_directory(): avoid setting invalid git_dir

When discovering a .git/ directory, we take pains to ensure that its
repository format version matches Git's expectations, and we return NULL
otherwise.

However, we still appended the invalid path to the strbuf passed as
argument.

Let's just reset the strbuf to the state before we appended the .git/
directory that was eventually rejected.

There is another early return path in that function, when
setup_git_directory_gently_1() returns GIT_DIR_NONE or an error. In that
case, the gitdir parameter has not been touched, therefore there is no
need for an equivalent change in that code path.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agogit-svn: document special options for commit-diff
Andreas Heiduk [Wed, 14 Jun 2017 09:31:38 +0000 (11:31 +0200)]
git-svn: document special options for commit-diff

Some options specific for `git svn commit-diff` where not documented
so far.

Signed-off-by: Andreas Heiduk <asheiduk@gmail.com>
Signed-off-by: Eric Wong <e@80x24.org>
3 years agoSync with maint
Junio C Hamano [Tue, 13 Jun 2017 20:52:53 +0000 (13:52 -0700)]
Sync with maint

3 years agoEighth batch for 2.14
Junio C Hamano [Tue, 13 Jun 2017 20:52:29 +0000 (13:52 -0700)]
Eighth batch for 2.14

Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoMerge branch 'jk/pack-idx-corruption-safety'
Junio C Hamano [Tue, 13 Jun 2017 20:47:10 +0000 (13:47 -0700)]
Merge branch 'jk/pack-idx-corruption-safety'

A flaky test has been corrected.

* jk/pack-idx-corruption-safety:
  t5313: make extended-table test more deterministic

3 years agoMerge branch 'nd/fopen-errors'
Junio C Hamano [Tue, 13 Jun 2017 20:47:09 +0000 (13:47 -0700)]
Merge branch 'nd/fopen-errors'

We often try to open a file for reading whose existence is
optional, and silently ignore errors from open/fopen; report such
errors if they are not due to missing files.

* nd/fopen-errors:
  mingw_fopen: report ENOENT for invalid file names
  mingw: verify that paths are not mistaken for remote nicknames
  log: fix memory leak in open_next_file()
  rerere.c: move error_errno() closer to the source system call
  print errno when reporting a system call error
  wrapper.c: make warn_on_inaccessible() static
  wrapper.c: add and use fopen_or_warn()
  wrapper.c: add and use warn_on_fopen_errors()
  config.mak.uname: set FREAD_READS_DIRECTORIES for Darwin, too
  config.mak.uname: set FREAD_READS_DIRECTORIES for Linux and FreeBSD
  clone: use xfopen() instead of fopen()
  use xfopen() in more places
  git_fopen: fix a sparse 'not declared' warning

3 years agoMerge branch 'rf/completion'
Junio C Hamano [Tue, 13 Jun 2017 20:47:09 +0000 (13:47 -0700)]
Merge branch 'rf/completion'

Completion updates.

* rf/completion:
  completion: add git config credentialCache.ignoreSIGHUP
  completion: add git config credential completions
  completion: add git config advice completions
  completion: add git config am.threeWay completion
  completion: add git config core completions
  completion: add git config gc completions

3 years agoMerge branch 'jc/diff-tree-stale-comment'
Junio C Hamano [Tue, 13 Jun 2017 20:47:08 +0000 (13:47 -0700)]
Merge branch 'jc/diff-tree-stale-comment'

Comment fix.

* jc/diff-tree-stale-comment:
  diff-tree: update stale in-code comments

3 years agoMerge branch 'sb/submodule-blanket-recursive'
Junio C Hamano [Tue, 13 Jun 2017 20:47:07 +0000 (13:47 -0700)]
Merge branch 'sb/submodule-blanket-recursive'

Many commands learned to pay attention to submodule.recurse
configuration.

* sb/submodule-blanket-recursive:
  builtin/fetch.c: respect 'submodule.recurse' option
  builtin/push.c: respect 'submodule.recurse' option
  builtin/grep.c: respect 'submodule.recurse' option
  Introduce 'submodule.recurse' option for worktree manipulators
  submodule loading: separate code path for .gitmodules and config overlay
  reset/checkout/read-tree: unify config callback for submodule recursion
  submodule test invocation: only pass additional arguments
  submodule recursing: do not write a config variable twice

3 years agoMerge branch 'jc/noent-notdir'
Junio C Hamano [Tue, 13 Jun 2017 20:47:06 +0000 (13:47 -0700)]
Merge branch 'jc/noent-notdir'

Our code often opens a path to an optional file, to work on its
contents when we can successfully open it.  We can ignore a failure
to open if such an optional file does not exist, but we do want to
report a failure in opening for other reasons (e.g. we got an I/O
error, or the file is there, but we lack the permission to open).

The exact errors we need to ignore are ENOENT (obviously) and
ENOTDIR (less obvious).  Instead of repeating comparison of errno
with these two constants, introduce a helper function to do so.

* jc/noent-notdir:
  treewide: use is_missing_file_error() where ENOENT and ENOTDIR are checked
  compat-util: is_missing_file_error()

3 years agoPrepare for 2.13.2
Junio C Hamano [Tue, 13 Jun 2017 20:30:16 +0000 (13:30 -0700)]
Prepare for 2.13.2

Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoMerge branch 'ad/pull-remote-doc' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:06 +0000 (13:27 -0700)]
Merge branch 'ad/pull-remote-doc' into maint

Docfix.

* ad/pull-remote-doc:
  docs: fix formatting and grammar

3 years agoMerge branch 'jk/drop-free-refspecs' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:06 +0000 (13:27 -0700)]
Merge branch 'jk/drop-free-refspecs' into maint

Code clean-up.

* jk/drop-free-refspecs:
  remote: drop free_refspecs() function

3 years agoMerge branch 'jk/connect-symref-info-leak-fix' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:05 +0000 (13:27 -0700)]
Merge branch 'jk/connect-symref-info-leak-fix' into maint

Leakfix.

* jk/connect-symref-info-leak-fix:
  connect.c: fix leak in parse_one_symref_info()

3 years agoMerge branch 'rf/completion-config-commit' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:05 +0000 (13:27 -0700)]
Merge branch 'rf/completion-config-commit' into maint

Completion update.

* rf/completion-config-commit:
  completion: add completions for git config commit

3 years agoMerge branch 'ab/t3070-test-dedup' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:04 +0000 (13:27 -0700)]
Merge branch 'ab/t3070-test-dedup' into maint

Test cleanup.

* ab/t3070-test-dedup:
  wildmatch test: remove redundant duplicate test

3 years agoMerge branch 'jh/memihash-opt' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:03 +0000 (13:27 -0700)]
Merge branch 'jh/memihash-opt' into maint

perf-test update.

* jh/memihash-opt:
  p0004: don't error out if test repo is too small
  p0004: don't abort if multi-threaded is too slow
  p0004: use test_perf
  p0004: avoid using pipes
  p0004: simplify calls of test-lazy-init-name-hash

3 years agoMerge branch 'tb/pull-ff-rebase-autostash' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:03 +0000 (13:27 -0700)]
Merge branch 'tb/pull-ff-rebase-autostash' into maint

"git pull --rebase --autostash" didn't auto-stash when the local history
fast-forwards to the upstream.

* tb/pull-ff-rebase-autostash:
  pull: ff --rebase --autostash works in dirty repo

3 years agoMerge branch 'jh/close-index-before-stat' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:02 +0000 (13:27 -0700)]
Merge branch 'jh/close-index-before-stat' into maint

The timestamp of the index file is now taken after the file is
closed, to help Windows, on which a stale timestamp is reported by
fstat() on a file that is opened for writing and data was written
but not yet closed.

* jh/close-index-before-stat:
  read-cache: close index.lock in do_write_index

3 years agoMerge branch 'sl/clean-d-ignored-fix' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:01 +0000 (13:27 -0700)]
Merge branch 'sl/clean-d-ignored-fix' into maint

"git clean -d" used to clean directories that has ignored files,
even though the command should not lose ignored ones without "-x".
"git status --ignored"  did not list ignored and untracked files
without "-uall".  These have been corrected.

* sl/clean-d-ignored-fix:
  clean: teach clean -d to preserve ignored paths
  dir: expose cmp_name() and check_contains()
  dir: hide untracked contents of untracked dirs
  dir: recurse into untracked dirs for ignored files
  t7061: status --ignored should search untracked dirs
  t7300: clean -d should skip dirs with ignored files

3 years agoMerge branch 'dk/send-email-avoid-net-smtp-ssl-when-able' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:01 +0000 (13:27 -0700)]
Merge branch 'dk/send-email-avoid-net-smtp-ssl-when-able' into maint

A hotfix to a topic in 'master'.

* dk/send-email-avoid-net-smtp-ssl-when-able:
  send-email: Net::SMTP::starttls was introduced in v2.34
  send-email: Net::SMTP::SSL is obsolete, use only when necessary

3 years agoMerge branch 'jc/skip-test-in-the-middle' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:27:00 +0000 (13:27 -0700)]
Merge branch 'jc/skip-test-in-the-middle' into maint

A recent update to t5545-push-options.sh started skipping all the
tests in the script when a web server testing is disabled or
unavailable, not just the ones that require a web server.  Non HTTP
tests have been salvaged to always run in this script.

* jc/skip-test-in-the-middle:
  t5545: enhance test coverage when no http server is installed
  test: allow skipping the remainder

3 years agoMerge branch 'bw/forking-and-threading' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:26:59 +0000 (13:26 -0700)]
Merge branch 'bw/forking-and-threading' into maint

The "run-command" API implementation has been made more robust
against dead-locking in a threaded environment.

* bw/forking-and-threading:
  usage.c: drop set_error_handle()
  run-command: restrict PATH search to executable files
  run-command: expose is_executable function
  run-command: block signals between fork and execve
  run-command: add note about forking and threading
  run-command: handle dup2 and close errors in child
  run-command: eliminate calls to error handling functions in child
  run-command: don't die in child when duping /dev/null
  run-command: prepare child environment before forking
  string-list: add string_list_remove function
  run-command: use the async-signal-safe execv instead of execvp
  run-command: prepare command before forking
  t0061: run_command executes scripts without a #! line
  t5550: use write_script to generate post-update hook

3 years agoMerge branch 'jk/bug-to-abort' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:26:59 +0000 (13:26 -0700)]
Merge branch 'jk/bug-to-abort' into maint

Introduce the BUG() macro to improve die("BUG: ...").

* jk/bug-to-abort:
  usage: add NORETURN to BUG() function definitions
  config: complain about --local outside of a git repo
  setup_git_env: convert die("BUG") to BUG()
  usage.c: add BUG() function

3 years agoMerge branch 'sb/checkout-recurse-submodules' into maint
Junio C Hamano [Tue, 13 Jun 2017 20:26:59 +0000 (13:26 -0700)]
Merge branch 'sb/checkout-recurse-submodules' into maint

"git checkout --recurse-submodules" did not quite work with a
submodule that itself has submodules.

* sb/checkout-recurse-submodules:
  submodule: properly recurse for read-tree and checkout
  submodule: avoid auto-discovery in new working tree manipulator code
  submodule_move_head: reuse child_process structure for futher commands

3 years agols-files: factor out tag calculation
Brandon Williams [Mon, 12 Jun 2017 22:14:08 +0000 (15:14 -0700)]
ls-files: factor out tag calculation

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: factor out debug info into a function
Brandon Williams [Mon, 12 Jun 2017 22:14:07 +0000 (15:14 -0700)]
ls-files: factor out debug info into a function

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert show_files to take an index
Brandon Williams [Mon, 12 Jun 2017 22:14:06 +0000 (15:14 -0700)]
ls-files: convert show_files to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert show_ce_entry to take an index
Brandon Williams [Mon, 12 Jun 2017 22:14:05 +0000 (15:14 -0700)]
ls-files: convert show_ce_entry to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert prune_cache to take an index
Brandon Williams [Mon, 12 Jun 2017 22:14:04 +0000 (15:14 -0700)]
ls-files: convert prune_cache to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert ce_excluded to take an index
Brandon Williams [Mon, 12 Jun 2017 22:14:03 +0000 (15:14 -0700)]
ls-files: convert ce_excluded to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert show_ru_info to take an index
Brandon Williams [Mon, 12 Jun 2017 22:14:02 +0000 (15:14 -0700)]
ls-files: convert show_ru_info to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert show_other_files to take an index
Brandon Williams [Mon, 12 Jun 2017 22:14:01 +0000 (15:14 -0700)]
ls-files: convert show_other_files to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert show_killed_files to take an index
Brandon Williams [Mon, 12 Jun 2017 22:14:00 +0000 (15:14 -0700)]
ls-files: convert show_killed_files to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert write_eolinfo to take an index
Brandon Williams [Mon, 12 Jun 2017 22:13:59 +0000 (15:13 -0700)]
ls-files: convert write_eolinfo to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agols-files: convert overlay_tree_on_cache to take an index
Brandon Williams [Mon, 12 Jun 2017 22:13:58 +0000 (15:13 -0700)]
ls-files: convert overlay_tree_on_cache to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agotree: convert read_tree to take an index parameter
Brandon Williams [Mon, 12 Jun 2017 22:13:57 +0000 (15:13 -0700)]
tree: convert read_tree to take an index parameter

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconvert: convert renormalize_buffer to take an index
Brandon Williams [Mon, 12 Jun 2017 22:13:56 +0000 (15:13 -0700)]
convert: convert renormalize_buffer to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconvert: convert convert_to_git to take an index
Brandon Williams [Mon, 12 Jun 2017 22:13:55 +0000 (15:13 -0700)]
convert: convert convert_to_git to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconvert: convert convert_to_git_filter_fd to take an index
Brandon Williams [Mon, 12 Jun 2017 22:13:54 +0000 (15:13 -0700)]
convert: convert convert_to_git_filter_fd to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconvert: convert crlf_to_git to take an index
Brandon Williams [Mon, 12 Jun 2017 22:13:53 +0000 (15:13 -0700)]
convert: convert crlf_to_git to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoconvert: convert get_cached_convert_stats_ascii to take an index
Brandon Williams [Mon, 12 Jun 2017 22:13:52 +0000 (15:13 -0700)]
convert: convert get_cached_convert_stats_ascii to take an index

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoMerge branch 'master' of git://bogomips.org/git-svn
Junio C Hamano [Sat, 10 Jun 2017 05:29:26 +0000 (14:29 +0900)]
Merge branch 'master' of git://bogomips.org/git-svn

* 'master' of git://bogomips.org/git-svn:
  doc: describe git svn init --ignore-refs

3 years agodoc: describe git svn init --ignore-refs
Andreas Heiduk [Wed, 7 Jun 2017 18:32:14 +0000 (20:32 +0200)]
doc: describe git svn init --ignore-refs

Add the missing documentation for `git svn init --ignore-refs`.

Signed-off-by: Andreas Heiduk <asheiduk@gmail.com>
Signed-off-by: Eric Wong <e@80x24.org>
3 years agoSync with maint
Junio C Hamano [Wed, 7 Jun 2017 00:32:04 +0000 (09:32 +0900)]
Sync with maint

* maint:
  sha1dc: update from upstream
  sha1dc: ignore indent-with-non-tab whitespace violations

3 years agoMerge branch 'ab/sha1dc-maint' into maint
Junio C Hamano [Wed, 7 Jun 2017 00:31:53 +0000 (09:31 +0900)]
Merge branch 'ab/sha1dc-maint' into maint

* ab/sha1dc-maint:
  sha1dc: update from upstream
  sha1dc: ignore indent-with-non-tab whitespace violations

3 years agosha1dc: update from upstream
Ævar Arnfjörð Bjarmason [Tue, 6 Jun 2017 15:12:29 +0000 (15:12 +0000)]
sha1dc: update from upstream

Update sha1dc from the latest version by the upstream
maintainer[1].

See commit a0103914c2 ("sha1dc: update from upstream", 2017-05-20) for
the latest update. That update was done sans some whitespace changes
by upstream, which is why the diff here isn't the same as the upstream
cc46554..e139984.

It also brings in a change[2] upstream made which should hopefully
address the breakage in 2.13.1 on Cygwin, see [3]. Cygwin defines both
_BIG_ENDIAN and _LITTLE_ENDIAN.

Adam Dinwoodie reports on the mailing list that that upstream commit
fixes the issue on Cygwin[4].

1. https://github.com/cr-marcstevens/sha1collisiondetection/commit/e1399840b501a68ac6c8d7ed9a5cb1455480200e
2. https://github.com/cr-marcstevens/sha1collisiondetection/commit/a24eef58c0684078405f8c7a89f9b78271432005
3. <20170606100355.GC25777@dinwoodie.org> (https://public-inbox.org/git/20170606100355.GC25777@dinwoodie.org/)
4. <20170606124323.GD25777@dinwoodie.org> (https://public-inbox.org/git/20170606124323.GD25777@dinwoodie.org/)

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agosha1dc: ignore indent-with-non-tab whitespace violations
Jeff King [Tue, 6 Jun 2017 19:01:11 +0000 (15:01 -0400)]
sha1dc: ignore indent-with-non-tab whitespace violations

The upstream sha1dc code indents some lines with spaces.
While this doesn't match Git's coding guidelines, it's better
to leave this imported code untouched than to try to make it
match our style. However, we can use .gitattributes to tell
"diff --check" and "git am" not to bother us about it.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agot5313: make extended-table test more deterministic
Jeff King [Mon, 5 Jun 2017 19:15:25 +0000 (15:15 -0400)]
t5313: make extended-table test more deterministic

Commit a1283866b (t5313: test bounds-checks of
corrupted/malicious pack/idx files, 2016-02-25) added a test
that requires our corrupted pack index to have two objects.
The entry for the first one remains untouched, but we
corrupt the entry for second one. Since the index stores the
entries in sha1-sorted order, this means that the test must
make sure that the sha1 of the object we expect to be
corrupted ("$object") sorts after the other placeholder
object.

That commit used the HEAD commit as the placeholder, but the
script never calls test_tick. That means that the commit
object (and thus its sha1) depends on the timestamp when the
test script is run. This usually works in practice, because
the sha1 of $object starts with "fff". The commit object
will sort after that only 1 in 4096 times, but when it does
the test will fail.

One obvious solution is to add the test_tick call to get a
deterministic commit sha1. But since we're relying on the
sort order for the test to function, let's make that very
explicit by just generating a second blob with a known sha1.

Reported-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agot0012: test "-h" with builtins
Jeff King [Tue, 30 May 2017 05:19:30 +0000 (01:19 -0400)]
t0012: test "-h" with builtins

Since commit 99caeed05 (Let 'git <command> -h' show usage
without a git dir, 2009-11-09), the git wrapper handles "-h"
specially, skipping any repository setup but still calling
the builtin's cmd_foo() function. This means that every
cmd_foo() must be ready to handle this case, but we don't
have any systematic tests. This led to "git am -h" being
broken for some time without anybody noticing.

This patch just tests that "git foo -h" works for every
builtin, where we see a 129 exit code (the normal code for
our usage() helper), and that the word "usage" appears in
the output.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agogit: add hidden --list-builtins option
Jeff King [Tue, 30 May 2017 05:18:43 +0000 (01:18 -0400)]
git: add hidden --list-builtins option

It can be useful in the test suite to be able to iterate
over the list of builtins. We could do this with some
Makefile magic. But since the authoritative list is in the
commands array inside git.c, and since this could also be
handy for debugging, let's add a hidden command-line option
to dump that list.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoversion: convert to parse-options
Jeff King [Tue, 30 May 2017 05:17:42 +0000 (01:17 -0400)]
version: convert to parse-options

The "git version" command didn't traditionally accept any
options, and in fact ignores any you give it. When we added
simple option parsing for "--build-options" in 6b9c38e14, we
didn't improve this; we just loop over the arguments and
pick out the one we recognize.

Instead, let's move to a real parsing loop, complain about
nonsense options, and recognize conventions like "-h".

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agodiff- and log- family: handle "git cmd -h" early
Junio C Hamano [Thu, 1 Jun 2017 04:38:16 +0000 (13:38 +0900)]
diff- and log- family: handle "git cmd -h" early

"git $builtin -h" bypasses the usual repository setup and calls the
cmd_$builtin() function, expecting it to show the help text.

Unfortunately the commands in the log- and the diff- family want to
call into the revisions machinery, which by definition needs to have
a repository already discovered.  Strictly speaking, they may not
need a repository only for parsing "-h", but it is a good discipline
to future-proof codepath to ensure that setup_revisions() is called
after we know that a repository is there.

Handle the "git $builtin -h" special case very early in these
commands to work around potential issues.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agodiff: rename diff_fill_sha1_info to diff_fill_oid_info
Brandon Williams [Tue, 30 May 2017 17:31:09 +0000 (10:31 -0700)]
diff: rename diff_fill_sha1_info to diff_fill_oid_info

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agodiffcore-rename: use is_empty_blob_oid
Brandon Williams [Tue, 30 May 2017 17:31:08 +0000 (10:31 -0700)]
diffcore-rename: use is_empty_blob_oid

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agotree-diff: convert path_appendnew to object_id
Brandon Williams [Tue, 30 May 2017 17:31:07 +0000 (10:31 -0700)]
tree-diff: convert path_appendnew to object_id

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agotree-diff: convert diff_tree_paths to struct object_id
Brandon Williams [Tue, 30 May 2017 17:31:06 +0000 (10:31 -0700)]
tree-diff: convert diff_tree_paths to struct object_id

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agotree-diff: convert try_to_follow_renames to struct object_id
Brandon Williams [Tue, 30 May 2017 17:31:05 +0000 (10:31 -0700)]
tree-diff: convert try_to_follow_renames to struct object_id

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agobuiltin/diff-tree: cleanup references to sha1
Brandon Williams [Tue, 30 May 2017 17:31:04 +0000 (10:31 -0700)]
builtin/diff-tree: cleanup references to sha1

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agodiff-tree: convert diff_tree_sha1 to struct object_id
Brandon Williams [Tue, 30 May 2017 17:31:03 +0000 (10:31 -0700)]
diff-tree: convert diff_tree_sha1 to struct object_id

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agonotes-merge: convert write_note_to_worktree to struct object_id
Brandon Williams [Tue, 30 May 2017 17:31:02 +0000 (10:31 -0700)]
notes-merge: convert write_note_to_worktree to struct object_id

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agonotes-merge: convert verify_notes_filepair to struct object_id
Brandon Williams [Tue, 30 May 2017 17:31:01 +0000 (10:31 -0700)]
notes-merge: convert verify_notes_filepair to struct object_id

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agonotes-merge: convert find_notes_merge_pair_ps to struct object_id
Brandon Williams [Tue, 30 May 2017 17:31:00 +0000 (10:31 -0700)]
notes-merge: convert find_notes_merge_pair_ps to struct object_id

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agonotes-merge: convert merge_from_diffs to struct object_id
Brandon Williams [Tue, 30 May 2017 17:30:59 +0000 (10:30 -0700)]
notes-merge: convert merge_from_diffs to struct object_id

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agonotes-merge: convert notes_merge* to struct object_id
Brandon Williams [Tue, 30 May 2017 17:30:58 +0000 (10:30 -0700)]
notes-merge: convert notes_merge* to struct object_id

Convert notes_merge and notes_merge_commit to use struct object_id.

Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoperf: work around the tested repo having an index.lock
Ævar Arnfjörð Bjarmason [Fri, 2 Jun 2017 10:33:30 +0000 (10:33 +0000)]
perf: work around the tested repo having an index.lock

When the tested repo has an index.lock file it should be removed. This
file may be present if e.g. git-status previously crashed in that
repo, and it will make a lot of git commands fail. Let's try harder
and remove the lock.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoSync with v2.13.1
Junio C Hamano [Mon, 5 Jun 2017 00:33:16 +0000 (09:33 +0900)]
Sync with v2.13.1

3 years agoSeventh batch for 2.14
Junio C Hamano [Mon, 5 Jun 2017 00:32:25 +0000 (09:32 +0900)]
Seventh batch for 2.14

Signed-off-by: Junio C Hamano <gitster@pobox.com>
3 years agoMerge branch 'ad/pull-remote-doc'
Junio C Hamano [Mon, 5 Jun 2017 00:18:13 +0000 (09:18 +0900)]
Merge branch 'ad/pull-remote-doc'

Docfix.

* ad/pull-remote-doc:
  docs: fix formatting and grammar

3 years agoMerge branch 'tb/pull-ff-rebase-autostash'
Junio C Hamano [Mon, 5 Jun 2017 00:18:13 +0000 (09:18 +0900)]
Merge branch 'tb/pull-ff-rebase-autostash'

"git pull --rebase --autostash" didn't auto-stash when the local history
fast-forwards to the upstream.

* tb/pull-ff-rebase-autostash:
  pull: ff --rebase --autostash works in dirty repo

3 years agoMerge branch 'jk/drop-free-refspecs'
Junio C Hamano [Mon, 5 Jun 2017 00:18:13 +0000 (09:18 +0900)]
Merge branch 'jk/drop-free-refspecs'

Code clean-up.

* jk/drop-free-refspecs:
  remote: drop free_refspecs() function

3 years agoMerge branch 'jk/connect-symref-info-leak-fix'
Junio C Hamano [Mon, 5 Jun 2017 00:18:12 +0000 (09:18 +0900)]
Merge branch 'jk/connect-symref-info-leak-fix'

Leakfix.

* jk/connect-symref-info-leak-fix:
  connect.c: fix leak in parse_one_symref_info()

3 years agoMerge branch 'js/blame-lib'
Junio C Hamano [Mon, 5 Jun 2017 00:18:11 +0000 (09:18 +0900)]
Merge branch 'js/blame-lib'

The internal logic used in "git blame" has been libified to make it
easier to use by cgit.

* js/blame-lib: (29 commits)
  blame: move entry prepend to libgit
  blame: move scoreboard setup to libgit
  blame: move scoreboard-related methods to libgit
  blame: move fake-commit-related methods to libgit
  blame: move origin-related methods to libgit
  blame: move core structures to header
  blame: create entry prepend function
  blame: create scoreboard setup function
  blame: create scoreboard init function
  blame: rework methods that determine 'final' commit
  blame: wrap blame_sort and compare_blame_final
  blame: move progress updates to a scoreboard callback
  blame: make sanity_check use a callback in scoreboard
  blame: move no_whole_file_rename flag to scoreboard
  blame: move xdl_opts flags to scoreboard
  blame: move show_root flag to scoreboard
  blame: move reverse flag to scoreboard
  blame: move contents_from to scoreboard
  blame: move copy/move thresholds to scoreboard
  blame: move stat counters to scoreboard
  ...

3 years agoMerge branch 'mh/packed-ref-store-prep'
Junio C Hamano [Mon, 5 Jun 2017 00:18:11 +0000 (09:18 +0900)]
Merge branch 'mh/packed-ref-store-prep'

The implementation of "ref" API around the "packed refs" have been
cleaned up, in preparation for further changes.

* mh/packed-ref-store-prep: (25 commits)
  cache_ref_iterator_begin(): avoid priming unneeded directories
  ref-filter: limit traversal to prefix
  create_ref_entry(): remove `check_name` option
  refs_ref_iterator_begin(): handle `GIT_REF_PARANOIA`
  read_packed_refs(): report unexpected fopen() failures
  read_packed_refs(): do more of the work of reading packed refs
  get_packed_ref_cache(): assume "packed-refs" won't change while locked
  should_pack_ref(): new function, extracted from `files_pack_refs()`
  ref_update_reject_duplicates(): add a sanity check
  ref_update_reject_duplicates(): use `size_t` rather than `int`
  ref_update_reject_duplicates(): expose function to whole refs module
  ref_transaction_prepare(): new optional step for reference updates
  ref_transaction_commit(): check for valid `transaction->state`
  files_transaction_cleanup(): new helper function
  files_ref_store: put the packed files lock directly in this struct
  files-backend: move `lock` member to `files_ref_store`
  lockfile: add a new method, is_lock_file_locked()
  ref_store: take a `msg` parameter when deleting references
  refs: use `size_t` indexes when iterating over ref transaction updates
  refs_ref_iterator_begin(): don't check prefixes redundantly
  ...

3 years agoMerge branch 'mb/diff-default-to-indent-heuristics'
Junio C Hamano [Mon, 5 Jun 2017 00:18:10 +0000 (09:18 +0900)]
Merge branch 'mb/diff-default-to-indent-heuristics'

Make the "indent" heuristics the default in "diff" and diff.indentHeuristics
configuration variable an escape hatch for those who do no want it.

* mb/diff-default-to-indent-heuristics:
  add--interactive: drop diff.indentHeuristic handling
  diff: enable indent heuristic by default
  diff: have the diff-* builtins configure diff before initializing revisions
  diff: make the indent heuristic part of diff's basic configuration