The built-in `git add -i` machinery obviously has its `the_repository`
structure initialized at the point where `cmd_commit()` calls it, and
therefore does not look at the environment variable `GIT_INDEX_FILE`.
But when being called from `commit --interactive`, it has to, because
the index was already locked in that case, and we want to ask the
interactive add machinery to work on the `index.lock` file instead of
the `index` file.
Technically, we could teach `run_add_i()`, or for that matter
`run_add_p()`, to look specifically at that environment variable, but
the entire idea of passing in a parameter of type `struct repository *`
is to allow working on multiple repositories (and their index files)
independently.
So let's instead override the `index_file` field of that structure
temporarily.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is a straight-forward port of 2f0896ec3a (restore: support
--patch, 2019-04-25) which added support for `git restore -p`.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This patch teaches the built-in `git add -p` machinery all the tricks it
needs to know in order to act as the work horse for `git checkout -p`.
Apart from the minor changes (slightly reworded messages, different
`diff` and `apply --check` invocations), it requires a new function to
actually apply the changes, as `git checkout -p` is a bit special in
that respect: when the desired changes do not apply to the index, but
apply to the work tree, Git does not fail straight away, but asks the
user whether to apply the changes to the worktree at least.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The scripted version of `git stash` called directly into the Perl script
`git-add--interactive.perl`, and this was faithfully converted to C.
However, we have a much better way to do this now: call the internal API
directly, which will now incidentally also respect the
`add.interactive.useBuiltin` setting. Let's just do this.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As `git add` traditionally did not expose the `--patch=<mode>` modes via
command-line options, the scripted version of `git stash` had to call
`git add--interactive` directly.
But this prevents the built-in `add -p` from kicking in, as
`add--interactive` is the scripted version (which does not have a
"fall-back" to the built-in version).
So let's introduce support for internal switch for `git add` that the
scripted `git stash` can use to call the appropriate backend (scripted
or built-in, depending on `add.interactive.useBuiltin`).
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `git stash` and `git reset` commands support a `--patch` option, and
both simply hand off to `git add -p` to perform that work. Let's teach
the built-in version of that command to be able to perform that work, too.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The Perl script backing `git add -p` is used not only for that command,
but also for `git stash -p`, `git reset -p` and `git checkout -p`.
In preparation for teaching the C version of `git add -p` to support
also the latter commands, let's abstract away what is "stage" specific
into a dedicated data structure describing the differences between the
patch modes.
Finally, please note that the Perl version tries to make sure that the
diffs are only generated for the modified files. This is not actually
necessary, as the calls to Git's diff machinery already perform that
work, and perform it well. This makes it unnecessary to port the
`FILTER` field of the `%patch_modes` struct, as well as the
`get_diff_reference()` function.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the previous steps, we re-implemented the main loop of `git add -i`
in C, and most of the commands.
Notably, we left out the actual functionality of `patch`, as the
relevant code makes up more than half of `git-add--interactive.perl`,
and is actually pretty independent of the rest of the commands.
With this commit, we start to tackle that `patch` part. For better
separation of concerns, we keep the code in a separate file,
`add-patch.c`. The new code is still guarded behind the
`add.interactive.useBuiltin` config setting, and for the moment,
it can only be called via `git add -p`.
The actual functionality follows the original implementation of
5cde71d64a (git-add --interactive, 2006-12-10), but not too closely
(for example, we use string offsets rather than copying strings around,
and after seeing whether the `k` and `j` commands are applicable, in the
C version we remember which previous/next hunk was undecided, and use it
rather than looking again when the user asked to jump).
As a further deviation from that commit, We also use a comma instead of
a slash to separate the available commands in the prompt, as the current
version of the Perl script does this, and we also add a line about the
question mark ("print help") to the help text.
While it is tempting to use this conversion of `git add -p` as an excuse
to work on `apply_all_patches()` so that it does _not_ want to read a
file from `stdin` or from a file, but accepts, say, an `strbuf` instead,
we will refrain from this particular rabbit hole at this stage.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Unlike previous conversions to C, where we started with a built-in
helper, we start this conversion by adding an interception in the
`run_add_interactive()` function when the new opt-in
`add.interactive.useBuiltin` config knob is turned on (or the
corresponding environment variable `GIT_TEST_ADD_I_USE_BUILTIN`), and
calling the new internal API function `run_add_i()` that is implemented
directly in libgit.a.
At this point, the built-in version of `git add -i` only states that it
cannot do anything yet. In subsequent patches/patch series, the
`run_add_i()` function will gain more and more functionality, until it
is feature complete. The whole arc of the conversion can be found in the
PRs #170-175 at https://github.com/gitgitgadget/git.
The "--helper approach" can unfortunately not be used here: on Windows
we face the very specific problem that a `system()` call in
Perl seems to close `stdin` in the parent process when the spawned
process consumes even one character from `stdin`. Which prevents us from
implementing the main loop in C and still trying to hand off to the Perl
script.
The very real downside of the approach we have to take here is that the
test suite won't pass with `GIT_TEST_ADD_I_USE_BUILTIN=true` until the
conversion is complete (the `--helper` approach would have let it pass,
even at each of the incremental conversion steps).
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Suppose, from a repository that has ".gitmodules", we clone with
--filter=blob:none:
git clone --filter=blob:none --no-checkout \
https://kernel.googlesource.com/pub/scm/git/git
Then we fetch:
git -C git fetch
This will cause a "unable to load config blob object", because the
fetch_config_from_gitmodules() invocation in cmd_fetch() will attempt to
load ".gitmodules" (which Git knows to exist because the client has the
tree of HEAD) while fetch_if_missing is set to 0.
fetch_if_missing is set to 0 too early - ".gitmodules" here should be
lazily fetched. Git must set fetch_if_missing to 0 before the fetch
because as part of the fetch, packfile negotiation happens (and we do
not want to fetch any missing objects when checking existence of
objects), but we do not need to set it so early. Move the setting of
fetch_if_missing to the earliest possible point in cmd_fetch(), right
before any fetching happens.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
94da9193a6 (grep: add support for PCRE v2, 2017-06-01) didn't include
a way to override the system allocator, and so it is incompatible with
custom allocators (e.g. nedmalloc). This problem became obvious when we
tried to plug a memory leak by `free()`ing a data structure allocated by
PCRE2, triggering a segfault in Windows (where we use nedmalloc by
default).
PCRE2 requires the use of a general context to override the allocator
and therefore, there is a lot more code needed than in PCRE1, including
a couple of wrapper functions.
Extend the grep API with a "destructor" that could be called to cleanup
any objects that were created and used globally.
Update `builtin/grep.c` to use that new API, but any other future users
should make sure to have matching `grep_init()`/`grep_destroy()` calls
if they are using the pattern matching functionality.
Move some of the logic that was before done per thread (in the workers)
into an earlier phase to avoid degrading performance, but as the use of
PCRE2 with custom allocators is better understood it is expected more of
its functions will be instructed to use the custom allocator as well as
was done in the original code[1] this work was based on.
[1] https://public-inbox.org/git/3397e6797f872aedd18c6d795f4976e1c579514b.1565005867.git.gitgitgadget@gmail.com/
Reported-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
git stash push does not recursively stash submodules, but if
submodule.recurse is set, it may recursively reset --hard them. Having
only the destructive action recurse is likely to be surprising
behaviour, and unlikely to be desirable, so the easiest fix should be to
ensure that the call to git reset --hard never recurses into submodules.
This matches the behavior of check_changes_tracked_files, which ignores
submodules.
Signed-off-by: Jakob Jarmar <jakob@jarmar.se>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git format-patch -o <outdir>' did an equivalent of 'mkdir <outdir>'
not 'mkdir -p <outdir>', which is being corrected.
Avoid the usage of 'adjust_shared_perm' on the leading directories which
may have security implications. Achieved by temporarily disabling of
'config.sharedRepository' like 'git init' does.
Signed-off-by: Bert Wesarg <bert.wesarg@googlemail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Comments stating that "struct hashmap_entry" must be the first
member in a struct are no longer valid.
Suggested-by: Phillip Wood <phillip.wood123@gmail.com>
Signed-off-by: Eric Wong <e@80x24.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since these macros already take a `keyvar' pointer of a known type,
we can rely on OFFSETOF_VAR to get the correct offset without
relying on non-portable `__typeof__' and `offsetof'.
Argument order is also rearranged, so `keyvar' and `member' are
sequential as they are used as: `keyvar->member'
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While we cannot rely on a `__typeof__' operator being portable
to use with `offsetof'; we can calculate the pointer offset
using an existing pointer and the address of a member using
pointer arithmetic for compilers without `__typeof__'.
This allows us to simplify usage of hashmap iterator macros
by not having to specify a type when a pointer of that type
is already given.
In the future, list iterator macros (e.g. list_for_each_entry)
may also be implemented using OFFSETOF_VAR to save hackers the
trouble of using container_of/list_entry macros and without
relying on non-portable `__typeof__'.
v3: use `__typeof__' to avoid clang warnings
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
`hashmap_free_entries' behaves like `container_of' and passes
the offset of the hashmap_entry struct to the internal
`hashmap_free_' function, allowing the function to free any
struct pointer regardless of where the hashmap_entry field
is located.
`hashmap_free' no longer takes any arguments aside from
the hashmap itself.
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Inspired by list_for_each_entry in the Linux kernel.
Once again, these are somewhat compromised usability-wise
by compilers lacking __typeof__ support.
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Another step in eliminating the requirement of hashmap_entry
being the first member of a struct.
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Update callers to use hashmap_get_entry, hashmap_get_entry_from_hash
or container_of as appropriate.
This is another step towards eliminating the requirement of
hashmap_entry being the first field in a struct.
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is less error-prone than "void *" as the compiler now
detects invalid types being passed.
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is less error-prone than "const void *" as the compiler
now detects invalid types being passed.
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is less error-prone than "void *" as the compiler now
detects invalid types being passed.
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
C compilers do type checking to make life easier for us. So
rely on that and update all hashmap_entry_init callers to take
"struct hashmap_entry *" to avoid future bugs while improving
safety and readability.
Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When Git wants to spawn a child Git process inside a worktree's
subdirectory while `GIT_DIR` is set, we need to take care of specifying
the work tree's top-level directory explicitly because it cannot be
discovered: the current directory is _not_ the top-level directory of
the work tree, and neither is it inside the parent directory of
`GIT_DIR`.
This fixes the problem where `git stash apply` would report pretty much
everything deleted or untracked when run inside a worktree's
subdirectory.
To make sure that we do not introduce the "reverse problem", i.e. when
`GIT_WORK_TREE` is defined but `GIT_DIR` is not, we simply make sure
that both are set.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
So far, `--jobs=<n>` only parallelizes submodule fetches/clones, not
`--multiple` fetches, which is unintuitive, given that the option's name
does not say anything about submodules in particular.
Let's change that. With this patch, also fetches from multiple remotes
are parallelized.
For backwards-compatibility (and to prepare for a use case where
submodule and multiple-remote fetches may need different parallelization
limits), the config setting `submodule.fetchJobs` still only controls
the submodule part of `git fetch`, while the newly-introduced setting
`fetch.parallel` controls both (but can be overridden for submodules
with `submodule.fetchJobs`).
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Found with "git grep '^#include ' '*.c' | sort | uniq -d".
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This function is marked as `NORETURN`, and it indeed does not want to
return anything. So let's not declare it with the return type `int`.
This fixes the following warning when building with MSVC:
C4646: function declared with 'noreturn' has non-void return type
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a new option, --mark-tags, which will output mark identifiers with
each tag object. This improves the incremental export story with
--export-marks since it will allow us to record that annotated tags have
been exported, and it is also needed as a step towards supporting nested
tags.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
fast-import has support for both an --import-marks flag and an
--import-marks-if-exists flag; the latter of which will not die() if the
file does not exist. fast-export only had support for an --import-marks
flag; add an --import-marks-if-exists flag for consistency.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add trace2 regions in transport.c and builtin/push.c to better track
time spent in various phases of pushing:
* Listing refs
* Checking submodules
* Pushing submodules
* Pushing refs
Signed-off-by: Josh Steadmon <steadmon@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add trace2 regions to fetch-pack.c and builtins/fetch.c to better track
time spent in the various phases of a fetch:
* listing refs
* negotiation for protocol versions v0-v2
* fetching refs
* consuming refs
Signed-off-by: Josh Steadmon <steadmon@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Emit trace2_cmd_mode() messages for each commit-graph
sub-command.
The commit graph commands were in flux when trace2 was
making it's way to git. Now that we have enough sub-commands
in commit-graph, we can label the various modes within them.
Distinguishing between read, write and verify is a great
start.
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Acked-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
fast-export allows specifying revision ranges, which can be used to
export a tag without exporting the commit it tags. fast-export handled
this rather poorly: it would emit a "from :0" directive. Since marks
start at 1 and increase, this means it refers to an unknown commit and
fast-import will choke on the input.
When we are unable to look up a mark for the object being tagged, use a
"from $HASH" directive instead to fix this problem.
Note that this is quite similar to the behavior fast-export exhibits
with commits and parents when --reference-excluded-parents is passed
along with an excluded commit range. For tags of excluded commits we do
not require the --reference-excluded-parents flag because we always have
to tag something. By contrast, when dealing with commits, pruning a
parent is always a viable option, so we need the flag to specify that
parent pruning is not wanted. (It is slightly weird that
--reference-excluded-parents isn't the default with a separate
--prune-excluded-parents flag, but backward compatibility concerns
resulted in the current defaults.)
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When 'git name-rev' is invoked with commit-ish parameters, it tries to
save some work, and doesn't visit commits older than the committer
date of the oldest given commit minus a one day worth of slop. Since
our 'timestamp_t' is an unsigned type, this leads to a timestamp
underflow when the committer date of the oldest given commit is within
a day of the UNIX epoch. As a result the cutoff timestamp ends up
far-far in the future, and 'git name-rev' doesn't visit any commits,
and names each given commit as 'undefined'.
Check whether subtracting the slop from the oldest committer date
would lead to an underflow, and use no cutoff in that case. We don't
have a TIME_MIN constant, dddbad728c (timestamp_t: a new data type for
timestamps, 2017-04-26) didn't add one, so do it now.
Note that the type of the cutoff timestamp variable used to be signed
before 5589e87fd8 (name-rev: change a "long" variable to timestamp_t,
2017-05-20). The behavior was still the same even back then, but the
underflow didn't happen when substracting the slop from the oldest
committer date, but when comparing the signed cutoff timestamp with
unsigned committer dates in name_rev(). IOW, this underflow bug is as
old as 'git name-rev' itself.
Helped-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When converting stash into C, calls to 'git update-index --refresh'
were replaced with the 'refresh_cache()' function. That is fine as
long as the index is only needed in-core, and not re-read from disk.
However in many cases we do actually need the refreshed index to be
written to disk, for example 'merge_recursive_generic()' discards the
in-core index before re-reading it from disk, and in the case of 'apply
--quiet', the 'refresh_cache()' we currently have is pointless without
writing the index to disk.
Always write the index after refreshing it to ensure there are no
regressions in this compared to the scripted stash. In the future we
can consider avoiding the write where possible after making sure none
of the subsequent calls actually need the refreshed cache, and it is
not expected to be refreshed after stash exits or it is written
somewhere else already.
Reported-by: Jeff King <peff@peff.net>
Signed-off-by: Thomas Gummerer <t.gummerer@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use the 'refresh_and_write_cache()' convenience function introduced in
the last commit, instead of refreshing and writing the index manually
in merge.c
Signed-off-by: Thomas Gummerer <t.gummerer@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Getting the lock for the index, refreshing it and then writing it is a
pattern that happens more than once throughout the codebase, and isn't
trivial to get right. Factor out the refresh_and_write_cache function
from builtin/am.c to read-cache.c, so it can be re-used in other
places in a subsequent commit.
Note that we return different error codes for failing to refresh the
cache, and failing to write the index. The current caller only cares
about failing to write the index. However for other callers we're
going to convert in subsequent patches we will need this distinction.
Helped-by: Martin Ågren <martin.agren@gmail.com>
Helped-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Thomas Gummerer <t.gummerer@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add --[no-]progress to git commit-graph write and verify.
The progress feature was introduced in 7b0f229
("commit-graph write: add progress output", 2018-09-17) but
the ability to opt-out was overlooked.
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
cmd_clean() had the following code structure:
struct strbuf abs_path = STRBUF_INIT;
for_each_string_list_item(item, &del_list) {
strbuf_addstr(&abs_path, prefix);
strbuf_addstr(&abs_path, item->string);
PROCESS(&abs_path);
strbuf_reset(&abs_path);
}
where I've elided a bunch of unnecessary details and PROCESS(&abs_path)
represents a big chunk of code rather than an actual function call. One
piece of PROCESS was:
if (lstat(abs_path.buf, &st))
continue;
which would cause the strbuf_reset() to be missed -- meaning that the
next path to be handled would have two paths concatenated. This path
used to use die_errno() instead of continue prior to commit 396049e5fb
("git-clean: refactor git-clean into two phases", 2013-06-25), but my
understanding of how correct_untracked_entries() works is that it will
prevent both dir/ and dir/file from being in the list to clean so this
should be dead code and the die_errno() should be safe. But I hesitate
to remove it since I am not certain.
However, we can fix both this bug and possible similar future bugs by
simply moving the strbuf_reset(&abs_path) to the beginning of the loop.
It'll result in N calls to strbuf_reset() instead of N-1, but that's a
small price to pay to avoid sneaky bugs like this.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Users expect files in a nested git repository to be left alone unless
sufficiently forced (with two -f's). Unfortunately, in certain
circumstances, git would delete both tracked (and possibly dirty) files
and untracked files within a nested repository. To explain how this
happens, let's contrast a couple cases. First, take the following
example setup (which assumes we are already within a git repo):
git init nested
cd nested
>tracked
git add tracked
git commit -m init
>untracked
cd ..
In this setup, everything works as expected; running 'git clean -fd'
will result in fill_directory() returning the following paths:
nested/
nested/tracked
nested/untracked
and then correct_untracked_entries() would notice this can be compressed
to
nested/
and then since "nested/" is a directory, we would call
remove_dirs("nested/", ...), which would
check is_nonbare_repository_dir() and then decide to skip it.
However, if someone also creates an ignored file:
>nested/ignored
then running 'git clean -fd' would result in fill_directory() returning
the same paths:
nested/
nested/tracked
nested/untracked
but correct_untracked_entries() will notice that we had ignored entries
under nested/ and thus simplify this list to
nested/tracked
nested/untracked
Since these are not directories, we do not call remove_dirs() which was
the only place that had the is_nonbare_repository_dir() safety check --
resulting in us deleting both the untracked file and the tracked (and
possibly dirty) file.
One possible fix for this issue would be walking the parent directories
of each path and checking if they represent nonbare repositories, but
that would be wasteful. Even if we added caching of some sort, it's
still a waste because we should have been able to check that "nested/"
represented a nonbare repository before even descending into it in the
first place. Add a DIR_SKIP_NESTED_GIT flag to dir_struct.flags and use
it to prevent fill_directory() and friends from descending into nested
git repos.
With this change, we also modify two regression tests added in commit
91479b9c72 ("t7300: add tests to document behavior of clean and nested
git", 2015-06-15). That commit, nor its series, nor the six previous
iterations of that series on the mailing list discussed why those tests
coded the expectation they did. In fact, it appears their purpose was
simply to test _existing_ behavior to make sure that the performance
changes didn't change the behavior. However, these two tests directly
contradicted the manpage's claims that two -f's were required to delete
files/directories under a nested git repository. While one could argue
that the user gave an explicit path which matched files/directories that
were within a nested repository, there's a slippery slope that becomes
very difficult for users to understand once you go down that route (e.g.
what if they specified "git clean -f -d '*.c'"?) It would also be hard
to explain what the exact behavior was; avoid such problems by making it
really simple.
Also, clean up some grammar errors describing this functionality in the
git-clean manpage.
Finally, there are still a couple bugs with -ffd not cleaning out enough
(e.g. missing the nested .git) and with -ffdX possibly cleaning out the
wrong files (paying attention to outer .gitignore instead of inner).
This patch does not address these cases at all (and does not change the
behavior relative to those flags), it only fixes the handling when given
a single -f. See
https://public-inbox.org/git/20190905212043.GC32087@szeder.dev/ for more
discussion of the -ffd[X?] bugs.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The -d flag pre-dated git-clean's ability to have paths specified. As
such, the default for git-clean was to only remove untracked files in
the current directory, and -d existed to allow it to recurse into
subdirectories.
The interaction of paths and the -d option appears to not have been
carefully considered, as evidenced by numerous bugs and a dearth of
tests covering such pairings in the testsuite. The definition turns out
to be important, so let's look at some of the various ways one could
interpret the -d option:
A) Without -d, only look in subdirectories which contain tracked
files under them; with -d, also look in subdirectories which
are untracked for files to clean.
B) Without specified paths from the user for us to delete, we need to
have some kind of default, so...without -d, only look in
subdirectories which contain tracked files under them; with -d,
also look in subdirectories which are untracked for files to clean.
The important distinction here is that choice B says that the presence
or absence of '-d' is irrelevant if paths are specified. The logic
behind option B is that if a user explicitly asked us to clean a
specified pathspec, then we should clean anything that matches that
pathspec. Some examples may clarify. Should
git clean -f untracked_dir/file
remove untracked_dir/file or not? It seems crazy not to, but a strict
reading of option A says it shouldn't be removed. How about
git clean -f untracked_dir/file1 tracked_dir/file2
or
git clean -f untracked_dir_1/file1 untracked_dir_2/file2
? Should it remove either or both of these files? Should it require
multiple runs to remove both the files listed? (If this sounds like a
crazy question to even ask, see the commit message of "t7300: Add some
testcases showing failure to clean specified pathspecs" added earlier in
this patch series.) What if -ffd were used instead of -f -- should that
allow these to be removed? Should it take multiple invocations with
-ffd? What if a glob (such as '*tracked*') were used instead of
spelling out the directory names? What if the filenames involved globs,
such as
git clean -f '*.o'
or
git clean -f '*/*.o'
?
The current documentation actually suggests a definition that is
slightly different than choice A, and the implementation prior to this
series provided something radically different than either choices A or
B. (The implementation, though, was clearly just buggy). There may be
other choices as well. However, for almost any given choice of
definition for -d that I can think of, some of the examples above will
appear buggy to the user. The only case that doesn't have negative
surprises is choice B: treat a user-specified path as a request to clean
all untracked files which match that path specification, including
recursing into any untracked directories.
Change the documentation and basic implementation to use this
definition.
There were two regression tests that indirectly depended on the current
implementation, but neither was about subdirectory handling. These two
tests were introduced in commit 5b7570cfb4 ("git-clean: add tests for
relative path", 2008-03-07) which was solely created to add coverage for
the changes in commit fb328947c8e ("git-clean: correct printing relative
path", 2008-03-07). Both tests specified a directory that happened to
have an untracked subdirectory, but both were only checking that the
resulting printout of a file that was removed was shown with a relative
path. Update these tests appropriately.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
During git-fetch, the client checks if the advertised tags' OIDs are
already in the fetch request's want OID set. This check is done in a
linear scan. For a repository that has a lot of refs, repeating this
scan takes 15+ minutes. In order to speed this up, create a oid_set for
other refs' OIDs.
Signed-off-by: Masaya Suzuki <masayasuzuki@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The list-objects-filter code has two steps to its initialization:
1. parse_list_objects_filter() makes sure the spec is a filter we know
about and is syntactically correct. This step is done by "rev-list"
or "upload-pack" that is going to apply a filter, but also by "git
clone" or "git fetch" before they send the spec across the wire.
2. list_objects_filter__init() runs the type-specific initialization
(using function pointers established in step 1). This happens at
the start of traverse_commit_list_filtered(), when we're about to
actually use the filter.
It's a good idea to parse as much as we can in step 1, in order to catch
problems early (e.g., a blob size limit that isn't a number). But one
thing we _shouldn't_ do is resolve any oids at that step (e.g., for
sparse-file contents specified by oid). In the case of a fetch, the oid
has to be resolved on the remote side.
The current code does resolve the oid during the parse phase, but
ignores any error (which we must do, because we might just be sending
the spec across the wire). This leads to two bugs:
- if we're not in a repository (e.g., because it's git-clone parsing
the spec), then we trigger a BUG() trying to resolve the name
- if we did hit the error case, we still have to notice that later and
bail. The code path in rev-list handles this, but the one in
upload-pack does not, leading to a segfault.
We can fix both by moving the oid resolution into the sparse-oid init
function. At that point we know we have a repository (because we're
about to traverse), and handling the error there fixes the segfault.
As a bonus, we can drop the NULL sparse_oid_value check in rev-list,
since this is now handled in the sparse-oid-filter init function.
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The commit-graph tool may read a lot of commits, but it only cares about
parsing their metadata (parents, trees, etc) and doesn't ever show the
messages to the user. And so it should not need save_commit_buffer,
which is meant for holding onto the object data of parsed commits so
that we can show them later. In fact, it's quite harmful to do so.
According to massif, the max heap of "git commit-graph write
--reachable" in linux.git before/after this patch (removing the commit
graph file in between) goes from ~1.1GB to ~270MB.
Which isn't surprising, since the difference is about the sum of the
uncompressed sizes of all commits in the repository, and this was
equivalent to leaking them.
This obviously helps if you're under memory pressure, but even without
it, things go faster. My before/after times for that command (without
massif) went from 12.521s to 11.874s, a speedup of ~5%.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When doing 'git rebase --autostash <upstream> <master>' with a dirty worktree
a 'HEAD is now at ...' message is emitted, which is pointless as it refers to
the old active branch which isn't actually moved.
This commit removes the 'HEAD is now at...' message.
Signed-off-by: Ben Wijen <ben@wijen.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Consider the following scenario:
git checkout not-the-master
work work work
git rebase --autostash upstream master
Here 'rebase --autostash <upstream> <branch>' incorrectly moves the
active branch (not-the-master) to master (before the rebase).
The expected behavior: (58794775:/git-rebase.sh:526)
AUTOSTASH=$(git stash create autostash)
git reset --hard
git checkout master
git rebase upstream
git stash apply $AUTOSTASH
The actual behavior: (6defce2b:/builtin/rebase.c:1062)
AUTOSTASH=$(git stash create autostash)
git reset --hard master
git checkout master
git rebase upstream
git stash apply $AUTOSTASH
This commit reinstates the 'legacy script' behavior as introduced with
58794775: rebase: implement --[no-]autostash and rebase.autostash
Signed-off-by: Ben Wijen <ben@wijen.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Johannes Schindelin