path: worktree_git_path() should not use file relocation
git_path is a convenience function that usually produces a string
$GIT_DIR/<path>. Since v2.5.0-rc0~143^2~35 (git_path(): be aware of
file relocation in $GIT_DIR, 2014-11-30), as a side benefit callers
get support for path relocation variables like $GIT_OBJECT_DIRECTORY:
- git_path("index") is $GIT_INDEX_FILE when set
- git_path("info/grafts") is $GIT_GRAFTS_FILE when set
- git_path("objects/<foo>") is $GIT_OBJECT_DIRECTORY/<foo> when set
- git_path("hooks/<foo>") is <foo> under core.hookspath when set
- git_path("refs/<foo>") etc (see path.c::common_list) is relative
to $GIT_COMMON_DIR instead of $GIT_DIR
worktree_git_path, by comparison, is designed to resolve files in a
specific worktree's git dir. Unfortunately, it shares code with
git_path and performs the same relocation. The result is that paths
that are meant to be relative to the specified worktree's git dir end
up replaced by paths from environment variables within the current git
dir.
Luckily, no current callers pass such arguments. The relocation was
noticed when testing the result of merging two patches under review,
one of which introduces a caller:
* The first patch made git prune check the index file in each
worktree's git dir (using worktree_git_path(wt, "index")) for
objects not to prune. This would trigger the unwanted relocation
when GIT_INDEX_FILE is set, causing objects reachable from the
index to be pruned.
* The second patch simplified the relocation logic for index,
info/grafts, objects, and hooks to happen unconditionally instead of
based on whether environment or configuration variables are set.
This caused the relocation to trigger even when GIT_INDEX_FILE is
not set.
[jn: rewrote commit message; skipping all relocation instead of just
GIT_INDEX_FILE]
Signed-off-by: Brandon Williams <bmwill@google.com> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
path: always pass in commondir to update_common_dir
Instead of passing in 'NULL' and having 'update_common_dir()' query for
the commondir, have the callers of 'update_common_dir()' be responsible
for providing the commondir.
Signed-off-by: Brandon Williams <bmwill@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Move all path related declarations from cache.h to a new path.h header
file. This makes cache.h smaller and makes it easier to add new path
related functions.
Signed-off-by: Brandon Williams <bmwill@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
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>
'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>
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>
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
Newly added tests to t3420 in this series prepare expected
human-readable output from "git rebase -i" and then compare the
actual output with it. As the output from the command is designed
to go through i18n/l10n, we need to use test_i18ncmp to tell
GETTEXT_POISON build that it is OK the output does not match.
As there is no portable way to pass timezone information to
strftime, some output format from "git log" and friends are
impossible to produce. Teach our own strbuf_addftime to replace %z
and %Z with caller-supplied values to help working around this.
* rs/strbuf-addftime-zZ:
date: use localtime() for "-local" time formats
t0006: check --date=format zone offsets
strbuf: let strbuf_addftime handle %z and %Z itself
* sg/revision-parser-skip-prefix:
revision.c: use skip_prefix() in handle_revision_pseudo_opt()
revision.c: use skip_prefix() in handle_revision_opt()
revision.c: stricter parsing of '--early-output'
revision.c: stricter parsing of '--no-{min,max}-parents'
revision.h: turn rev_info.early_output back into an unsigned int
The "add" section for 'git-submodule' is redundant in its
description and the short synopsis line. Fix it.
Remove the redundant mentioning of the 'repository' argument
being mandatory.
The text is hard to read because of back-references, so remove
those.
Replace the word "humanish" by "canonical" as that conveys better
what we do to guess the path.
While at it, quote all occurrences of '.gitmodules' as that is an
important file in the submodule context, also link to it on its
first mention.
Helped-by: Stefan Beller <sbeller@google.com> Signed-off-by: Kaartic Sivaraam <kaarticsivaraam91196@gmail.com> Reviewed-by: Stefan Beller <sbeller@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
sha1_name: cache readdir(3) results in find_short_object_filename()
Read each loose object subdirectory at most once when looking for unique
abbreviated hashes. This speeds up commands like "git log --pretty=%h"
considerably, which previously caused one readdir(3) call for each
candidate, even for subdirectories that were visited before.
The new cache is kept until the program ends and never invalidated. The
same is already true for pack indexes. The inherent racy nature of
finding unique short hashes makes it still fit for this purpose -- a
conflicting new object may be added at any time. Tasks with higher
consistency requirements should not use it, though.
The cached object names are stored in an oid_array, which is quite
compact. The bitmap for remembering which subdir was already read is
stored as a char array, with one char per directory -- that's not quite
as compact, but really simple and incurs only an overhead equivalent to
11 hashes after all.
Suggested-by: Jeff King <peff@peff.net> Helped-by: Jeff King <peff@peff.net> Signed-off-by: Rene Scharfe <l.s.r@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
status: contextually notify user about an initial commit
The existing message, "Initial commit", makes sense for the commit template
notifying users that it's their initial commit, but is confusing when
merely checking the status of a fresh repository (or orphan branch)
without having any commits yet.
Change the output of "status" to say "No commits yet" when "git
status" is run on a fresh repo (or orphan branch), while retaining the
current "Initial commit" message displayed in the template that's
displayed in the editor when the initial commit is being authored.
Correspondingly change the output of "short status" to "No commits yet
on " when "git status -sb" is run on a fresh repo (or orphan branch).
A few alternatives considered were,
* Waiting for initial commit
* Your current branch does not have any commits
* Current branch waiting for initial commit
The most succint one among the alternatives was chosen.
[with help on tests from Ævar]
Helped-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Kaartic Sivaraam <kaarticsivaraam91196@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
read_object() and sha1_object_info_extended() both implement mechanisms
such as object replacement, retrying the packed store after failing to
find the object in the packed store then the loose store, and being able
to mark a packed object as bad and then retrying the whole process.
Consolidating these mechanisms would be a great help to maintainability.
Therefore, consolidate them by extending sha1_object_info_extended() to
support the functionality needed, and then modifying read_object() to
use sha1_object_info_extended().
Signed-off-by: Jonathan Tan <jonathantanmy@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The LOOKUP_REPLACE_OBJECT flag controls whether the
lookup_replace_object() function is invoked by
sha1_object_info_extended(), read_sha1_file_extended(), and
lookup_replace_object_extended(), but it is not immediately clear which
functions accept that flag.
Therefore restrict this flag to only sha1_object_info_extended(),
renaming it appropriately to OBJECT_INFO_LOOKUP_REPLACE and adding some
documentation. Update read_sha1_file_extended() to have a boolean
parameter instead, and delete lookup_replace_object_extended().
parse_sha1_header() also passes this flag to
parse_sha1_header_extended() since commit 46f0344 ("sha1_file: support
reading from a loose object of unknown type", 2015-05-03), but that has
had no effect since that commit. Therefore this patch also removes this
flag from that invocation.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The LOOKUP_UNKNOWN_OBJECT flag was introduced in commit 46f0344
("sha1_file: support reading from a loose object of unknown type",
2015-05-03) in order to support a feature in cat-file subsequently
introduced in commit 39e4ae3 ("cat-file: teach cat-file a
'--allow-unknown-type' option", 2015-05-03). Despite its name and
location in cache.h, this flag is used neither in
read_sha1_file_extended() nor in any of the lookup functions, but used
only in sha1_object_info_extended().
Therefore rename this flag to OBJECT_INFO_ALLOW_UNKNOWN_TYPE, taking the
name of the cat-file flag that invokes this feature, and move it closer
to the declaration of sha1_object_info_extended(). Also add
documentation for this flag.
OBJECT_INFO_ALLOW_UNKNOWN_TYPE is defined to 2, not 1, to avoid
conflicting with LOOKUP_REPLACE_OBJECT. Avoidance of this conflict is
necessary because sha1_object_info_extended() supports both flags.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
die(): stop hiding errors due to overzealous recursion guard
Change the recursion limit for the default die routine from a *very*
low 1 to 1024. This ensures that infinite recursions are broken, but
doesn't lose the meaningful error messages under threaded execution
where threads concurrently start to die.
The intent of the existing code, as explained in commit cd163d4b4e ("usage.c: detect recursion in die routines and bail out
immediately", 2012-11-14), is to break infinite recursion in cases
where the die routine itself calls die(), and would thus infinitely
recurse.
However, doing that very aggressively by immediately printing out
"recursion detected in die handler" if we've already called die() once
means that threaded invocations of git can end up only printing out
the "recursion detected" error, while hiding the meaningful error.
An example of this is running a threaded grep which dies on execution
against pretty much any repo, git.git will do:
With the current version of git this will print some combination of
multiple PCRE failures that caused the abort and multiple "recursion
detected", some invocations will print out multiple "recursion
detected" errors with no PCRE error at all!
Before this change, running the above grep command 1000 times against
git.git[1] and taking the top 20 results will on my system yield the
following distribution of actual errors ("E") and recursion
errors ("R"):
322 E R
306 E
116 E R R
65 R R
54 R E
49 E E
44 R
15 E R R R
9 R R R
7 R E R
5 R R E
3 E R R R R
2 E E R
1 R R R R
1 R R R E
1 R E R R
The exact results are obviously random and system-dependent, but this
shows the race condition in this code. Some small part of the time
we're about to print out the actual error ("E") but another thread's
recursion error beats us to it, and sometimes we print out nothing but
the recursion error.
With this change we get, now with "W" to mean the new warning being
emitted indicating that we've called die() many times:
502 E
160 E W E
120 E E
53 E W
35 E W E E
34 W E E
29 W E E E
16 E E W
16 E E E
11 W E E E E
7 E E W E
4 W E
3 W W E E
2 E W E E E
1 W W E
1 W E W E
1 E W W E E E
1 E W W E E
1 E W W E
1 E W E E W
Which still sucks a bit, due to a still present race-condition in this
code we're sometimes going to print out several errors still, or
several warnings, or two duplicate errors without the warning.
But we will never have a case where we completely hide the actual
error as we do now.
Now, git-grep could make use of the pluggable error facility added in
commit c19a490e37 ("usage: allow pluggable die-recursion checks",
2013-04-16). There's other threaded code that calls set_die_routine()
or set_die_is_recursing_routine().
But this is about fixing the general die() behavior with threading
when we don't have such a custom routine yet. Right now the common
case is not an infinite recursion in the handler, but us losing error
messages by default because we're overly paranoid about our recursion
check.
So let's just set the recursion limit to a number higher than the
number of threads we're ever likely to spawn. Now we won't lose
errors, and if we have a recursing die handler we'll still die within
microseconds.
There are race conditions in this code itself, in particular the
"dying" variable is not thread mutexed, so we e.g. won't be dying at
exactly 1024, or for that matter even be able to accurately test
"dying == 2", see the cases where we print out more than one "W"
above.
But that doesn't really matter, for the recursion guard we just need
to die "soon", not at exactly 1024 calls, and for printing the correct
error and only one warning most of the time in the face of threaded
death this is good enough and a net improvement on the current code.
1. for i in {1..1000}; do git grep -P --threads=8 '(*LIMIT_MATCH=1)-?-?-?---$' 2>&1|perl -pe 's/^fatal: r.*/R/; s/^fatal: p.*/E/; s/^warning.*/W/' | tr '\n' ' '; echo; done | sort | uniq -c | sort -nr | head -n 20
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since c9d961647 (i18n: add--interactive: mark
edit_hunk_manually message for translation, 2016-12-14),
when the user asks to edit a hunk manually, we respect
core.commentChar in generating the edit instructions.
However, when we then strip out comment lines, we use a
simple regex like:
/^$commentChar/
If your chosen comment character is a regex metacharacter,
then that will behave in a confusing manner ("$", for
instance, would only eliminate blank lines, not actual
comment lines).
We can fix that by telling perl not to respect
metacharacters.
Reported-by: Christian Rösch <christian@croesch.de> Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The prompt_yesno function loops indefinitely waiting for a
"y" or "n" response. But it doesn't handle EOF, meaning
that we can end up in an infinite loop of reading EOF from
stdin. One way to simulate that is with:
echo e | GIT_EDITOR='echo corrupt >' git add -p
Let's break out of the loop and propagate the undef to the
caller. Without modifying the callers that effectively turns
it into a "no" response. This is reasonable for both of the
current callers, and it leaves room for any future caller to
check for undef explicitly.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
When help.autoCorrect is enabled, an invalid git command prints a
warning and a continuation message, which differs depending on
whether or not the value of help.autoCorrect is positive or
negative.
With help.autoCorrect = 15:
WARNING: You called a Git command named 'lgo', which does not exist.
Continuing under the assumption that you meant 'log'
in 1.5 seconds automatically...
With help.autoCorrect < 0:
WARNING: You called a Git command named 'lgo', which does not exist.
Continuing under the assumption that you meant 'log'
The continuation message's phrasing is awkward. This commit cleans it up.
As a bonus, we now use full-sentence strings which make translation easier.
With help.autoCorrect = 15:
WARNING: You called a Git command named 'lgo', which does not exist.
Continuing in 1.5 seconds, assuming that you meant 'log'.
With help.autoCorrect < 0:
WARNING: You called a Git command named 'lgo', which does not exist.
Continuing under the assumption that you meant 'log'.
Signed-off-by: Marc Branchaud <marcnarc@xiplink.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change the indentation from "\t " to "\t". This indenting issue was
introduced when the test was added in commit 1d2f393ac9
("status/commit: show staged submodules regardless of ignore
config", 2014-04-05).
Helped-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Kaartic Sivaraam <kaarticsivaraam91196@gmail.com> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
grep: fix erroneously copy/pasted variable in check/assert pattern
Fix an erroneously copy/pasted check for the pcre2_jit_stack variable
to check pcre2_match_context instead. The former was already checked
in the preceding "if" statement.
This is a trivial and obvious error introduced in my commit 94da9193a6 ("grep: add support for PCRE v2", 2017-06-01).
In practice if pcre2_match_context_create() returned NULL we were
likely in a situation where malloc() was returning NULL, and were thus
screwed anyway, but if only the pcre2_match_context_create() call
returned NULL (through some transitory bug) PCRE v2 would just
allocate and supply its own context object when matching, and we'd run
normally at the trivial expense of not getting a slight speedup by
sharing the context object between successive matches.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
"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
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
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
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
rebase: add more regression tests for console output
Check the console output when using --autostash and the stash does not
apply is what we expect. The test is quite strict but should catch any
changes to the console output from the various rebase flavors.
Thanks-to: Johannes Schindelin <Johannes.Schindelin@gmx.de> Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Check the console output when using --autostash and the stash applies
cleanly is what we expect. The test is quite strict but should catch
any changes to the console output from the various rebase flavors.
Thanks-to: Johannes Schindelin <Johannes.Schindelin@gmx.de> Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The rebase messages are printed to stderr traditionally. However due
to a bug introduced in 587947750bd (rebase: implement --[no-]autostash
and rebase.autostash, 2013-05-12) which was faithfully copied when
reimplementing parts of the interactive rebase in the sequencer the
autostash messages are printed to stdout instead.
It is time to fix that: let's print the autostash messages to stderr
instead of stdout.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Signed-off-by: Junio C Hamano <gitster@pobox.com>
`for_each_bisect_ref()` is called by `for_each_bad_bisect_ref()` with
a term "bad". This used to make it call `for_each_ref_in_submodule()`
with a prefix "refs/bisect/bad". But the latter is the name of the
reference that is being sought, so the empty string was being passed
to the callback as the trimmed refname. Moreover, this questionable
practice was turned into an error by
b9c8e7f2fb prefix_ref_iterator: don't trim too much, 2017-05-22
It makes more sense (and agrees better with the documentation of
`--bisect`) for the callers to receive the full reference names. So
* Add a new function, `for_each_fullref_in_submodule()`, to the refs
API. This plugs a gap in the existing functionality, analogous to
`for_each_fullref_in()` but accepting a `submodule` argument.
* Change `for_each_bad_bisect_ref()` to call the new function rather
than `for_each_ref_in_submodule()`.
* Add a test.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
mergetools/meld: improve compatibiilty with Meld on macOS X
The macOS X fork of Meld[1] requires a "=" in the "--output"
argument, as it uses a wrapper[2] script that munges the
"--output" argument before calling into the common "meld"
script.
The macOS X wrapper script[2] accepts "--output=<filename>"
only, despite the fact that the underlying meld code accepts
both "--output <filename" and "--output=<filename>"[3].
All versions of meld which accept "--output" accept it in
the "--output=<filename>" form, so use "--output=<file>" for
maximum compatibility.
Reported-by: Matthew Groth <mgroth49@gmail.com> Helped-by: Samuel Lijin <sxlijin@gmail.com> Signed-off-by: David Aguilar <davvid@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
coccinelle: make use of the "expression" FREE_AND_NULL() rule
A follow-up to the existing "expression" rule added in an earlier
change. This manually excludes a few occurrences, mostly things that
resulted in many FREE_AND_NULL() on one line, that'll be manually
fixed in a subsequent change.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
coccinelle: make use of the "type" FREE_AND_NULL() rule
Apply the result of the just-added coccinelle rule. This manually
excludes a few occurrences, mostly things that resulted in many
FREE_AND_NULL() on one line, that'll be manually fixed in a subsequent
change.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
wildmatch test: cover a blind spot in "/" matching
A negated character class that does not include '/', e.g. [^a-z]:
- Should match '/' when doing "wildmatch"
- Should not match '/' when doing "pathmatch"
Add two tests to cover these cases.
Helped-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
git-compat-util: add a FREE_AND_NULL() wrapper around free(ptr); ptr = NULL
Add a FREE_AND_NULL() wrapper marco for the common pattern of freeing
a pointer and assigning NULL to it right afterwards.
The implementation is similar to the (currently unused) XDL_PTRFREE
macro in xdiff/xmacros.h added in commit 3443546f6e ("Use a *real*
built-in diff generator", 2006-03-24). The only difference is that
free() is called unconditionally, see [1].
See [2] for a suggested alternative which does this via a function
instead of a macro. As covered in replies to that message, while it's
a viable approach, it would introduce caveats which this approach
doesn't have, so that potential change is left to a future follow-up
change.
This merely allows us to translate exactly what we're doing now to a
less verbose & idiomatic form using a macro, while guaranteeing that
we don't introduce any functional changes.
When we convert seconds-since-epochs timestamps into a
broken-down "struct tm", we do so by adjusting the timestamp
according to the known offset and then using gmtime() to
break down the result. This means that the resulting struct
"knows" that it's in GMT, even though the time it represents
is adjusted for a different zone. The fields where it stores
this data are not portably accessible, so we have no way to
override them to tell them the real zone info.
For the most part, this works. Our date-formatting routines
don't pay attention to these inaccessible fields, and use
the same tz info we provided for adjustment. The one
exception is when we call strftime(), whose %Z format
reveals this hidden timezone data.
We solved that by always showing the empty string for %Z.
This is allowed by POSIX, but not very helpful to the user.
We can't make this work in the general case, as there's no
portable function for setting an arbitrary timezone (and
anyway, we don't have the zone name for the author zones,
only their offsets).
But for the special case of the "-local" formats, we can
just skip the adjustment and use localtime() instead of
gmtime(). This makes --date=format-local:%Z work correctly,
showing the local timezone instead of an empty string.
The new test checks the result for "UTC", our default
test-lib value for $TZ. Using something like EST5 might be
more interesting, but the actual zone string is
system-dependent (for instance, on my system it expands to
just EST). Hopefully "UTC" is vanilla enough that every
system treats it the same.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We already test that "%z" and "%Z" show the right thing, but
we don't actually check that the time we display is the
correct one. Let's add two new tests:
1. Test that "format:" shows the time in the author's
timezone, just like the other time formats.
2. Test that "format-local:" shows time in the local
timezone. We don't want to use our normal UTC for this,
because its offset is zero (so the result would be
"correct" even if the code forgot to apply the offset
or applied it in the wrong direction).
We'll use the EST5 zone, which is already used
elsewhere in the script (and so is assumed to be
available everywhere).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
strbuf: let strbuf_addftime handle %z and %Z itself
There is no portable way to pass timezone information to strftime. Add
parameters for timezone offset and name to strbuf_addftime and let it
handle the timezone-related format specifiers %z and %Z internally.
Callers can opt out for %Z by passing NULL as timezone name. %z is
always handled internally -- this helps on Windows, where strftime would
expand it to a timezone name (same as %Z), in violation of POSIX.
Modifiers are not handled, e.g. %Ez is still passed to strftime.
Use an empty string as timezone name in show_date (the only current
caller) for now because we only have the timezone offset in non-local
mode. POSIX allows %Z to resolve to an empty string in case of missing
information.
Helped-by: Ulrich Mueller <ulm@gentoo.org> Helped-by: Jeff King <peff@peff.net> Signed-off-by: Rene Scharfe <l.s.r@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
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>
configure.ac: loosen FREAD_READS_DIRECTORIES test program
We added an FREAD_READS_DIRECTORIES Makefile knob long ago
in cba22528f (Add compat/fopen.c which returns NULL on
attempt to open directory, 2008-02-08) to handle systems
where reading from a directory returned garbage. This works
by catching the problem at the fopen() stage and returning
NULL.
More recently, we found that there is a class of systems
(including Linux) where fopen() succeeds but fread() fails.
Since the solution is the same (having fopen return NULL),
they use the same Makefile knob as of e2d90fd1c
(config.mak.uname: set FREAD_READS_DIRECTORIES for Linux and
FreeBSD, 2017-05-03).
This works fine except for one thing: the autoconf test in
configure.ac to set FREAD_READS_DIRECTORIES actually checks
whether fread succeeds. Which means that on Linux systems,
the knob isn't set (and we even override the config.mak.uname
default). t1308 catches the failure.
We can fix this by tweaking the autoconf test to cover both
cases. In theory we might care about the distinction between
the traditional "fread reads directories" case and the new
"fopen opens directories". But since our solution catches
the problem at the fopen stage either way, we don't actually
need to know the difference. The "fopen" case is a superset.
This does mean the FREAD_READS_DIRECTORIES name is slightly
misleading. Probably FOPEN_OPENS_DIRECTORIES would be more
accurate. But it would be disruptive to simply change the
name (people's existing build configs would fail), and it's
not worth the complexity of handling both. Let's just add a
comment in the knob description.
Reported-by: Øyvind A. Holm <sunny@sunbase.org> Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
'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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
The diff-so-fancy project is also written in perl, and most
of its users pipe diffs through both diff-highlight and
diff-so-fancy. It would be nice if this could be done in a
single script. So let's pull most of diff-highlight's code
into its own module which can be used by diff-so-fancy.
In addition, we'll abstract a few basic items like reading
from stdio so that a script using the module can do more
processing before or after diff-highlight handles the lines.
See the README update for more details.
One small downside is that the diff-highlight script must
now be built using the Makefile. There are ways around this,
but it quickly gets into perl arcana. Let's go with the
simple solution. As a bonus, our Makefile now respects the
PERL_PATH variable if it is set.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
b9c6232138 (--format=pretty: avoid calculating expensive expansions
twice) optimized adding short hashes multiple times by using the
fact that the output strbuf was only ever simply appended to and
copying the added string from the previous run. That prerequisite
is no longer given; we now have modfiers like %< and %+ that can
cause the cache to lose track of the correct offsets. Remove it.
Reported-by: Michael Giuffrida <michaelpg@chromium.org> Signed-off-by: Rene Scharfe <l.s.r@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
doc: do not use `rm .git/index` when normalizing line endings
When illustrating how to normalize the line endings, the
documentation in gitattributes tells the user to `rm .git/index`.
This is incorrect for two reasons:
- Users shouldn't be instructed to mess around with the internal
implementation of Git using raw file system tools like `rm`.
- Within a submodule or an additional working tree `.git` is just a
file containing a `gitdir: <path>` pointer into the real `.git`
directory. Therefore `rm .git/index` does not work.
The purpose of the `rm .git/index` instruction is to remove all entries
from the index without touching the working tree. The way to do this
with Git is to use `read-tree --empty`.
Signed-off-by: Andreas Heiduk <asheiduk@gmail.com> Helped-by: Junio C Hamano <gitster@pobox.com> Helped-by: Torsten Bögershausen <tboegi@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
sha1_file: teach packed_object_info about typename
In commit 46f0344 ("sha1_file: support reading from a loose object of
unknown type", 2015-05-06), "struct object_info" gained a "typename"
field that could represent a type name from a loose object file, whether
valid or invalid, as opposed to the existing "typep" which could only
represent valid types. Some relatively complex manipulations were added
to avoid breaking packed_object_info() without modifying it, but it is
much easier to just teach packed_object_info() about the new field.
Therefore, teach packed_object_info() as described above.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some submodule tests do some setup outside of a test_expect
block. This is bad because we won't actually check the
outcome of those commands. But it's doubly so because "git
add submodule" now produces a warning to stderr, which is
not suppressed by the test scripts in non-verbose mode.
This patch does the minimal to fix the annoying warnings.
All three of these scripts could use more cleanup of related
setup.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
It's an easy mistake to add a repository inside another
repository, like:
git clone $url
git add .
The resulting entry is a gitlink, but there's no matching
.gitmodules entry. Trying to use "submodule init" (or clone
with --recursive) doesn't do anything useful. Prior to
v2.13, such an entry caused git-submodule to barf entirely.
In v2.13, the entry is considered "inactive" and quietly
ignored. Either way, no clone of your repository can do
anything useful with the gitlink without the user manually
adding the submodule config.
In most cases, the user probably meant to either add a real
submodule, or they forgot to put the embedded repository in
their .gitignore file.
Let's issue a warning when we see this case. There are a few
things to note:
- the warning will go in the git-add porcelain; anybody
wanting to do low-level manipulation of the index is
welcome to create whatever funny states they want.
- we detect the case by looking for a newly added gitlink;
updates via "git add submodule" are perfectly reasonable,
and this avoids us having to investigate .gitmodules
entirely
- there's a command-line option to suppress the warning.
This is needed for git-submodule itself (which adds the
entry before adding any submodule config), but also
provides a mechanism for other scripts doing
submodule-like things.
We could make this a hard error instead of a warning.
However, we do add lots of sub-repos in our test suite. It's
not _wrong_ to do so. It just creates a state where users
may be surprised. Pointing them in the right direction with
a gentle hint is probably the best option.
There is a config knob that can disable the (long) hint. But
I intentionally omitted a config knob to disable the warning
entirely. Whether the warning is sensible or not is
generally about context, not about the user's preferences.
If there's a tool or workflow that adds gitlinks without
matching .gitmodules, it should probably be taught about the
new command-line option, rather than blanket-disabling the
warning.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
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
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()