"git am" that was recently reimplemented in C had a performance
regression in "git am --abort" that goes back to the version before
an attempted (and failed) patch application.
* pt/am-builtin-abort-fix:
am --skip/--abort: merge HEAD/ORIG_HEAD tree into index
Merge branch 'jk/guess-repo-name-regression-fix' into maint
"git clone $URL" in recent releases of Git contains a regression in
the code that invents a new repository name incorrectly based on
the $URL. This has been corrected.
* jk/guess-repo-name-regression-fix:
clone: use computed length in guess_dir_name
clone: add tests for output directory
The upload_pack shell variable is initialized to an empty string, so
conditional expansion with ${upload_pack+"$upload_pack"} would not
work very well. You need a colon there.
When trying to see that an object does not exist, a state errno
leaked from our "first try to open a packfile with O_NOATIME and
then if it fails retry without it" logic on a system that refuses
O_NOATIME. This confused us and caused us to die, saying that the
packfile is unreadable, when we should have just reported that the
object does not exist in that packfile to the caller.
* cb/open-noatime-clear-errno:
git_open_noatime: return with errno=0 on success
The "lockfile" API has been rebuilt on top of a new "tempfile" API.
* mh/tempfile:
credential-cache--daemon: use tempfile module
credential-cache--daemon: delete socket from main()
gc: use tempfile module to handle gc.pid file
lock_repo_for_gc(): compute the path to "gc.pid" only once
diff: use tempfile module
setup_temporary_shallow(): use tempfile module
write_shared_index(): use tempfile module
register_tempfile(): new function to handle an existing temporary file
tempfile: add several functions for creating temporary files
prepare_tempfile_object(): new function, extracted from create_tempfile()
tempfile: a new module for handling temporary files
commit_lock_file(): use get_locked_file_path()
lockfile: add accessor get_lock_file_path()
lockfile: add accessors get_lock_file_fd() and get_lock_file_fp()
create_bundle(): duplicate file descriptor to avoid closing it twice
lockfile: move documentation to lockfile.h and lockfile.c
To prepare for allowing a different "ref" backend to be plugged in
to the system, update_ref()/delete_ref() have been taught about
ref-like things like MERGE_HEAD that are per-worktree (they will
always be written to the filesystem inside $GIT_DIR).
* dt/refs-pseudo:
pseudoref: check return values from read_ref()
sequencer: replace write_cherry_pick_head with update_ref
bisect: use update_ref
pseudorefs: create and use pseudoref update and delete functions
refs: add ref_type function
refs: introduce pseudoref and per-worktree ref concepts
write_file(): drop caller-supplied LF from calls to create a one-liner file
All of the callsites covered by this change call write_file() or
write_file_gently() to create a one-liner file. Drop the caller
supplied LF and let these callees to append it as necessary.
write_file_v(): do not leave incomplete line at the end
All existing callers to this function use it to produce a text file
or an empty file, and a new callsite that mimick them must end their
payload with a LF. If they forget to do so, the resulting file will
end with an incomplete line.
Teach write_file_v() to complete the incomplete line, if exists, so
that the callers do not have to.
With this, the caller-side fix in builtin/am.c becomes unnecessary.
527ec39 (generate-cmdlist: parse common group commands, 2015-05-21)
replaced generate-cmdlist.sh with a more functional Perl version,
generate-cmdlist.perl. The Perl version gleans named tags from a new
"common groups" section in command-list.txt and recognizes those
tags in "command list" section entries in place of the old 'common'
tag. This allows git-help to, not only recognize, but also group
common commands.
Although the tests require Perl, 527ec39 creates an unconditional
dependence upon Perl in the build system itself, which can not be
overridden with NO_PERL. Such a dependency may be undesirable; for
instance, the 'git-lite' package in the FreeBSD ports tree is
intended as a minimal Git installation (which may, for example, be
useful on servers needing only local clone and update capability),
which, historically, has not depended upon Perl[1].
Therefore, revive generate-cmdlist.sh and extend it to recognize
"common groups" and its named tags. Retire generate-cmdlist.perl.
setup: update the right file in multiple checkouts
This code is introduced in 23af91d (prune: strategies for linked
checkouts - 2014-11-30), and it's supposed to implement this rule from
that commit's message:
- linked checkouts are supposed to keep its location in $R/gitdir up
to date. The use case is auto fixup after a manual checkout move.
Note the name, "$R/gitdir", not "$R/gitfile". Correct the path to be
updated accordingly.
While at there, make sure I/O errors are not silently dropped.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Unlike other 'git describe' options, the '--contains' code path is
implemented by calling 'name-rev' with a bunch of options plus all the
commit-ishes that were passed to 'git describe'. If no commit-ish was
present, then 'name-rev' got invoked with none, which then leads to the
behavior illustrated above.
Porcelain commands usually default to HEAD when no commit-ish is given,
and 'git describe' already does so in all other cases, so it should do
so with '--contains' as well.
Pass HEAD to 'name-rev' when no commit-ish is given on the command line
to make '--contains' behave consistently with other 'git describe'
options. While at it, use argv_array_pushv() instead of the loop to
pass commit-ishes to 'git name-rev'.
'git describe's short help already indicates that the commit-ish is
optional, but the synopsis in the man page doesn't, so update it
accordingly as well.
Signed-off-by: SZEDER Gábor <szeder@ira.uka.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
All callers except three passed 1 for the "fatal" parameter to ask
this function to die upon error, but to a casual reader of the code,
it was not all obvious what that 1 meant. Instead, split the
function into two based on a common write_file_v() that takes the
flag, introduce write_file_gently() as a new way to attempt creating
a file without dying on error, and make three callers to call it.
We forgot to terminate the payload given to write_file() with LF,
resulting in files that end with an incomplete line. Teach the
wrappers builtin/am uses to make sure it adds LF at the end as
necessary.
There are many calls to write_file() that repeat the same pattern in
the implementation of the builtin version of "am". They all share
the same traits, i.e they
- produce a text file with a single string in it;
- have enough information to produce the entire contents of that
file;
- generate the pathname of the file by making a call to am_path(); and
- they ask write_file() to die() upon failure.
The slight differences among the call sites throw them into roughly
three categories:
- many write either "t" or "f" based on a boolean value to a file;
- some write the integer value in decimal text;
- some others write more general string, e.g. an object name in
hex, an empty string (i.e. the presense of the file itself serves
as a flag), etc.
Introduce three helpers, write_state_bool(), write_state_count() and
write_state_text(), to reduce direct calls to write_file().
This is a preparatory step for the next step to ensure that no
"state" file this command leaves in $GIT_DIR is with an incomplete
line at the end.
Suggested-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
rev-list: make it obvious that we do not support notes
The rev-list command does not have the internal
infrastructure to display notes. Running:
git rev-list --notes HEAD
will silently ignore the "--notes" option. Running:
git rev-list --notes --grep=. HEAD
will crash on an assert. Running:
git rev-list --format=%N HEAD
will place a literal "%N" in the output (it does not even
expand to an empty string).
Let's have rev-list tell the user that it cannot fill the
user's request, rather than silently producing wrong data.
Likewise, let's remove mention of the notes options from the
rev-list documentation.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
config: silence warnings for command names with invalid keys
When we are running the git command "foo", we may have to
look up the config keys "pager.foo" and "alias.foo". These
config schemes are mis-designed, as the command names can be
anything, but the config syntax has some restrictions. For
example:
$ git foo_bar
error: invalid key: pager.foo_bar
error: invalid key: alias.foo_bar
git: 'foo_bar' is not a git command. See 'git --help'.
You cannot name an alias with an underscore. And if you have
an external command with one, you cannot configure its
pager.
In the long run, we may develop a different config scheme
for these features. But in the near term (and because we'll
need to support the existing scheme indefinitely), we should
at least squelch the error messages shown above.
These errors come from git_config_parse_key. Ideally we
would pass a "quiet" flag to the config machinery, but there
are many layers between the pager code and the key parsing.
Passing a flag through all of those would be an invasive
change.
Instead, let's provide a config function to report on
whether a key is syntactically valid, and have the pager and
alias code skip lookup for bogus keys. We can build this
easily around the existing git_config_parse_key, with two
minor modifications:
1. We now handle a NULL store_key, to validate but not
write out the normalized key.
2. We accept a "quiet" flag to avoid writing to stderr.
This doesn't need to be a full-blown public "flags"
field, because we can make the existing implementation
a static helper function, keeping the mess contained
inside config.c.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove "复刻" from translation option list of "fork", keeping "派生"
as the only one.
In Git context, by talking about the term "fork", one usually mean the
procedure that divides a branch into two or more history lines. That
is quite like the "fork" concept in programming, which split a process
into two or more independent processes.
On the other hand, "复刻" by itself means "copy" or "mirror". It is
considered more suitable for describing the procedure like
"fork a repository", which is common on public repository hosting
services, such as GitHub. However, this is beyond the scope of
core-Git. As a l10n work for core-Git, "复刻" should be removed here.
There used to be another option - "分岔". Its a good idea to
translate "fork point" as "分岔点". However, "派生点" is as good,
too. So this option is finally discarded.
Also fix a relevant translation issue which was introduced in
`160fb2b`.
When looking for the start of the trailers in the message
we are passed, we should ignore the first line of the message.
The reason is that if we are passed a patch or commit message
then the first line should be the patch title.
If we are passed only trailers we can expect that they start
with an empty line that can be ignored too.
This way we can properly process commit messages that have
only one line with something that looks like a trailer, for
example like "area of code: change we made".
Signed-off-by: Christian Couder <chriscool@tuxfamily.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We defaulted to ignoring merge diffs because long long ago, in a
galaxy far away, we didn't have a great way to show the diffs. The
whole "--cc" option goes back to January '06 and commit d8f4790e6fe7
("diff-tree --cc: denser combined diff output for a merge commit").
And before that option - so for about 8 months - we had no good way
to show the diffs of merges in a good dense way. So the whole
"don't show diffs for merges by default" actually made a lot of
sense originally, because our merge diffs were not very useful.
And this was carried forward to this day. "git log --cc" still
ignores merge commits, and you need to say "git log -m --cc" to view
a sensible rendition of merge and non-merge commits, even with the
previous change to make "--cc" imply "-p".
Teach "git log" that "--cc" means the user wants to see interesting
changes in merge commits by turning "-m" on.
log: when --cc is given, default to -p unless told otherwise
The "--cc" option to "git log" is clearly a request to show some
sort of combined diff (be it --patch or --raw), but traditionally
we required the command line to explicitly ask for "git log -p --cc".
Teach the command line parser to treat a lone "--cc" as if the user
specified "-p --cc". Formats that do ask for other forms of diff
output, e.g. "log --raw --cc", are not overriden.
The revision walking API allows the callers to tweak its
configuration at the last minute, immediately after all the revision
and pathspec parameters are parsed from the command line but before
the default actions are decided based on them, by defining a "tweak"
callback function when calling setup_revisions(). Traditionally,
this facility was used by "git show" to turn on the patch output
"-p" by default when no diff output option (e.g. "--raw" or "-s" to
squelch the output altogether) is given on the command line, and
further give dense combined diffs "--cc" for merge commits when no
option to countermand it (e.g. "-m" to show pairwise patches).
Recently, "git log" started using the same facility, but we named
the callback function "default_follow_tweak()", as if the only kind
of tweaking we would want for "git log" will forever be limited to
turning "--follow" on by default when told by a configuration
variable. That was myopic.
Rename it to more generic name "log_setup_revisions_tweak()", and
match the one used by show "show_setup_revisions_tweak()".
We create a strbuf only to insert a single string, pass the
resulting buffer to a function (which does not modify the
string), and then free it. We can just pass the original
string instead.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
When formatting a config value into a strbuf, we may end
up stringifying it into a fixed-size buffer using sprintf,
and then copying that buffer into the strbuf. We can
eliminate the middle-man (and drop some calls to sprintf!)
by writing directly to the strbuf.
The reason it was written this way in the first place is
that we need to know before writing the value whether to
insert a delimiter. Instead of delaying the write of the
value, we speculatively write the delimiter, and roll it
back in the single case that cares.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
It's unusual for a function which writes to a passed-in
strbuf to call strbuf_init; that will throw away anything
already there, leaking memory. In this case, there are
exactly two callers; one relies on this initialization and
the other passes in an already-initialized buffer.
There's no leak, as the initialized buffer doesn't have
anything in it. But let's bump the strbuf_init out to the
one caller who needs it, making format_config more
idiomatic.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
config: restructure format_config() for better control flow
Commit 578625fa91 (config: add '--name-only' option to list only
variable names, 2015-08-10) modified format_config() such that it
returned from the middle of the function when showing only keys,
resulting in ugly code structure.
Reorganize the if statements and dealing with the key-value delimiter to
make the function easier to read.
Signed-off-by: SZEDER Gábor <szeder@ira.uka.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
ps_matched: xcalloc() takes nmemb and then element size
Even though multiplication is commutative, the order of arguments
should be xcalloc(nmemb, size). ps_matched is an array of 1-byte
element whose size is the same as the number of pathspec elements.
git_path() and mkpath() are handy helper functions but it is easy
to misuse, as the callers need to be careful to keep the number of
active results below 4. Their uses have been reduced.
* jk/git-path:
memoize common git-path "constant" files
get_repo_path: refactor path-allocation
find_hook: keep our own static buffer
refs.c: remove_empty_directories can take a strbuf
refs.c: avoid git_path assignment in lock_ref_sha1_basic
refs.c: avoid repeated git_path calls in rename_tmp_log
refs.c: simplify strbufs in reflog setup and writing
path.c: drop git_path_submodule
refs.c: remove extra git_path calls from read_loose_refs
remote.c: drop extraneous local variable from migrate_file
prefer mkpathdup to mkpath in assignments
prefer git_pathdup to git_path in some possibly-dangerous cases
add_to_alternates_file: don't add duplicate entries
t5700: modernize style
cache.h: complete set of git_path_submodule helpers
cache.h: clarify documentation for git_path, et al
"git clone $URL", when cloning from a site whose sole purpose is to
host a single repository (hence, no path after <scheme>://<site>/),
tried to use the site name as the new repository name, but did not
remove username or password when <site> part was of the form
<user>@<pass>:<host>. The code is taught to redact these.
* ps/guess-repo-name-at-root:
clone: abort if no dir name could be guessed
clone: do not use port number as dir name
clone: do not include authentication data in guessed dir
"git clone $URL" in recent releases of Git contains a regression in
the code that invents a new repository name incorrectly based on
the $URL. This has been corrected.
* jk/guess-repo-name-regression-fix:
clone: use computed length in guess_dir_name
clone: add tests for output directory
Merge branch 'es/doc-clean-outdated-tools' into maint
* es/doc-clean-outdated-tools:
Documentation/git-tools: retire manually-maintained list
Documentation/git-tools: drop references to defunct tools
Documentation/git-tools: fix item text formatting
Documentation/git-tools: improve discoverability of Git wiki
Documentation/git: drop outdated Cogito reference
Merge branch 'mh/fast-import-optimize-current-from' into maint
Often a fast-import stream builds a new commit on top of the
previous commit it built, and it often unconditionally emits a
"from" command to specify the first parent, which can be omitted in
such a case. This caused fast-import to forget the tree of the
previous commit and then re-read it from scratch, which was
inefficient. Optimize for this common case.
* mh/fast-import-optimize-current-from:
fast-import: do less work when given "from" matches current branch head
Merge branch 'ib/scripted-parse-opt-better-hint-string' into maint
The "rev-parse --parseopt" mode parsed the option specification
and the argument hint in a strange way to allow '=' and other
special characters in the option name while forbidding them from
the argument hint. This made it impossible to define an option
like "--pair <key>=<value>" with "pair=key=value" specification,
which instead would have defined a "--pair=key <value>" option.
* ib/scripted-parse-opt-better-hint-string:
rev-parse --parseopt: allow [*=?!] in argument hints
Merge branch 'se/doc-checkout-ours-theirs' into maint
A "rebase" replays changes of the local branch on top of something
else, as such they are placed in stage #3 and referred to as
"theirs", while the changes in the new base, typically a foreign
work, are placed in stage #2 and referred to as "ours". Clarify
the "checkout --ours/--theirs".
* se/doc-checkout-ours-theirs:
checkout: document subtlety around --ours/--theirs
Merge branch 'jx/do-not-crash-receive-pack-wo-head' into maint
An attempt to delete a ref by pushing into a repositorywhose HEAD
symbolic reference points at an unborn branch that cannot be
created due to ref D/F conflict (e.g. refs/heads/a/b exists, HEAD
points at refs/heads/a) failed.
* jx/do-not-crash-receive-pack-wo-head:
receive-pack: crash when checking with non-exist HEAD
push: support signing pushes iff the server supports it
Add a new flag --sign=true (or --sign=false), which means the same
thing as the original --signed (or --no-signed). Give it a third
value --sign=if-asked to tell push and send-pack to send a push
certificate if and only if the server advertised a push cert nonce.
If not, warn the user that their push may not be as secure as they
thought.
Signed-off-by: Dave Borowitz <dborowitz@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The old option parsing code in this plumbing command predates this
API, so option parsing was done more manually. Using the new API
brings send-pack more in line with push, and accepts new variants
like --no-* for negating options.
Signed-off-by: Dave Borowitz <dborowitz@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
config.c: rename git_config_maybe_bool_text and export it as git_parse_maybe_bool
This helper function does not complain about the config variable
but just silently reports failure to the caller. It is useful for
callers that need to parse any string that could be boolean or other
string (e.g. tristate yes/no/auto).
Signed-off-by: Dave Borowitz <dborowitz@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
This field was set in transport_set_option, but never read in the push
code. The push code basically ignores the smart_options field
entirely, and derives its options from the flags arguments to the
push* callbacks. Note that in git_transport_push there are already
several args set from flags that have no corresponding field in
git_transport_options; after this change, push_cert is just like
those.
Signed-off-by: Dave Borowitz <dborowitz@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
submodule: allow erroneous values for the fetchRecurseSubmodules option
We should not die when reading the submodule config cache since the
user might not be able to get out of that situation when the
configuration is part of the history.
We should handle this condition later when the value is about to be
used.
Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net> Signed-off-by: Stefan Beller <sbeller@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
submodule: use new config API for worktree configurations
We remove the extracted functions and directly parse into and read out
of the cache. This allows us to have one unified way of accessing
submodule configuration values specific to single submodules. Regardless
whether we need to access a configuration from history or from the
worktree.
Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net> Signed-off-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Stefan Beller <sbeller@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
submodule: implement a config API for lookup of .gitmodules values
In a superproject some commands need to interact with submodules. They
need to query values from the .gitmodules file either from the worktree
of from certain revisions. At the moment this is quite hard since a
caller would need to read the .gitmodules file from the history and then
parse the values. We want to provide an API for this so we have one
place to get values from .gitmodules from any revision (including the
worktree).
The API is realized as a cache which allows us to lazily read
.gitmodules configurations by commit into a runtime cache which can then
be used to easily lookup values from it. Currently only the values for
path or name are stored but it can be extended for any value needed.
It is expected that .gitmodules files do not change often between
commits. Thats why we lookup the .gitmodules sha1 from a commit and then
either lookup an already parsed configuration or parse and cache an
unknown one for each sha1. The cache is lazily build on demand for each
requested commit.
This cache can be used for all purposes which need knowledge about
submodule configurations. Example use cases are:
* Recursive submodule checkout needs to lookup a submodule name from
its path when a submodule first appears. This needs be done before
this configuration exists in the worktree.
* The implementation of submodule support for 'git archive' needs to
lookup the submodule name to generate the archive when given a
revision that is not checked out.
* 'git fetch' when given the --recurse-submodules=on-demand option (or
configuration) needs to lookup submodule names by path from the
database rather than reading from the worktree. For new submodule it
needs to lookup the name from its path to allow cloning new
submodules into the .git folder so they can be checked out without
any network interaction when the user does a checkout of that
revision.
Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net> Signed-off-by: Stefan Beller <sbeller@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>