The "diagnose" feature to create a zip archive for diagnostic
material has been lifted from "scalar" and made into a feature of
"git bugreport".
* vd/scalar-generalize-diagnose:
scalar: update technical doc roadmap
scalar-diagnose: use 'git diagnose --mode=all'
builtin/bugreport.c: create '--diagnose' option
builtin/diagnose.c: add '--mode' option
builtin/diagnose.c: create 'git diagnose' builtin
diagnose.c: add option to configure archive contents
scalar-diagnose: move functionality to common location
scalar-diagnose: move 'get_disk_info()' to 'compat/'
scalar-diagnose: add directory to archiver more gently
scalar-diagnose: avoid 32-bit overflow of size_t
scalar-diagnose: use "$GIT_UNZIP" in test
The "git remote rm" command's option parsing is fairly primitive: it
insists on a single argument, which it treats as the remote name, and
displays a usage message otherwise.
This is OK, and maybe even convenient, as you could run:
git remote rm --foo
to drop a remote named "--foo". But it's also weirdly unlike most of the
rest of Git, which would complain that there is no option "--foo". The
right way to spell it by our conventions is:
git remote rm -- --foo
but this doesn't currently work.
So let's bring the command in line with the rest of Git (including its
sibling subcommands!) by feeding argv to parse_options(). We already
have an empty options array for the usage helper.
Note that we have to adjust the argc index down by one, as
parse_options() eats the program name from the start of the array.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Several of the git-maintenance subcommands don't take any options, so
they don't bother looking at argv at all. This means they'll silently
accept garbage, like:
$ git maintenance register --foo
[no output]
$ git maintenance stop bar
[no output]
Let's give them the basic boilerplate to detect and handle these cases:
$ git maintenance register --foo
error: unknown option `foo'
usage: git maintenance register
$ git maintenance stop bar
usage: git maintenance stop
We could reduce the number of lines of code here a bit with a shared
helper function. But it's worth building out the boilerplate, as it may
serve as the base for adding options later.
Note one complication: maintenance_start() calls directly into
maintenance_register(), so it now needs to pass a plausible argv (we
don't care, but parse_options() is expecting there to at least be an
argv[0] program name). This is an extra line of code, but it eliminates
the need for an explanatory comment.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Recent commits such as bf0a6b65fc (builtin/multi-pack-index.c: let
parse-options parse subcommands, 2022-08-19) converted a few functions
to match our usual argc/argv/prefix conventions, but the prefix argument
remains unused.
However, there is a good use for it: they should pass it to their own
parse_options() functions, where it may be used to adjust the value of
any filename options. In all but one of these functions, there's no
behavior change, since they don't use OPT_FILENAME. But this is an
actual fix for one option, which you can see by modifying the test suite
like so:
diff --git a/t/t5326-multi-pack-bitmaps.sh b/t/t5326-multi-pack-bitmaps.sh
index 4fe57414c1..d0974d4371 100755
--- a/t/t5326-multi-pack-bitmaps.sh
+++ b/t/t5326-multi-pack-bitmaps.sh
@@ -186,7 +186,11 @@ test_expect_success 'writing a bitmap with --refs-snapshot' '
# Then again, but with a refs snapshot which only sees
# refs/tags/one.
- git multi-pack-index write --bitmap --refs-snapshot=snapshot &&
+ (
+ mkdir subdir &&
+ cd subdir &&
+ git multi-pack-index write --bitmap --refs-snapshot=../snapshot
+ ) &&
test_path_is_file $midx &&
test_path_is_file $midx-$(midx_checksum $objdir).bitmap &&
I'd emphasize that this wasn't broken by bf0a6b65fc; it has been broken
all along, because the sub-function never got to see the prefix. It is
that commit which is actually enabling us to fix it (and which also
brought attention to the problem because it triggers -Wunused-parameter!)
The other functions changed here don't use OPT_FILENAME at all. In their
cases this isn't fixing anything visible, but it's following the usual
pattern and future-proofing them against somebody adding new options and
being surprised.
I didn't include a test for the one visible case above. We don't
generally test routine parse-options behavior for individual options.
The challenge here was finding the problem, and now that this has been
done, it's not likely to regress. Likewise, we could apply the patch
above to cover it "for free" but it makes reading the rest of the test
unnecessarily complicated.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
After our loop through the selected strategies, we compare best_strategy
to wt_strategy. This is fine, but the fact that the code setting
best_strategy sets it to use_strategies[i]->name requires a little bit
of extra checking to determine that at the time of setting, that's the
same as wt_strategy. Just setting best_strategy to wt_strategy makes it
a little easier to verify what the loop is doing, at least for this
reader.
Further, use_strategies[i]->name is used in a number of places, where we
could just use wt_strategy. The latter takes less time for this reader
to parse (one variable name instead of three), so just use wt_strategy
to make the code slightly faster for human readers to parse.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
builtin/merge.c has a loop over the specified strategies, where if they
all fail with conflicts, it picks the one with the least number of
conflicts.
In the codepath that finds a successful merge, if an automatic commit
was wanted, the code breaks out of the above loop, which makes sense.
However, if the user requested there be no automatic commit, the loop
would continue. That seems weird; --no-commit should not affect the
choice of merge strategy, but the code as written makes one think it
does. However, since the loop itself embeds "!merge_was_ok" as a
condition on continuing to loop, it actually would also exit early if
--no-commit was specified, it just exited from a different location.
Restructure the code slightly to make it clear that the loop will
immediately exit whenever we find a merge strategy that is successful.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If a merge failed and we are leaving conflicts in the working directory
for the user to resolve, we should not attempt to apply any autostash.
Further, if we fail to apply the autostash (because either the merge
failed, or the user requested --no-commit), then we should instruct the
user how to apply it later.
Add a testcase verifying we have corrected this behavior.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The --bundle-uri option was added in 5556891961 (clone: add
--bundle-uri option, 2022-08-09), but this also introduced a call to
repo_init() whose return value was ignored. Fix that ignored value by
warning that the bundle URI process could not continue if it failed.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the spirit of 517fe807d6 (assert NOARG/NONEG behavior of
parse-options callbacks, 2018-11-05), this asserts that our callbacks
were invoked using the right flags (since otherwise they'd segfault on
the NULL arg). Both cases are already correct here, so this is mostly
about annotating the functions, and appeasing -Wunused-parameters.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The start_async(), etc, functions need a "proc" callback that conforms
to a particular interface. Not every callback needs every parameter
(e.g., the caller might not even ask to open an input descriptor, in
which case there is no point in the callback looking at it). Let's mark
these for -Wunused-parameter.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We pass a callback to read_tree_recursive(), but not every callback
needs every parameter. Let's mark the unused ones to satisfy
-Wunused-parameter.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Hashmap comparison functions must conform to a particular callback
interface, but many don't use all of their parameters. Especially the
void cmp_data pointer, but some do not use keydata either (because they
can easily form a full struct to pass when doing lookups). Let's mark
these to make -Wunused-parameter happy.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The callback passed to git_config() must conform to a particular
interface. But most callbacks don't actually look at the extra "void
*data" parameter. Let's mark the unused parameters to make
-Wunused-parameter happy.
Note there's one unusual case here in get_remote_default() where we
actually ignore the "value" parameter. That's because it's only checking
whether the option is found at all, and not parsing its value.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Functions used with for_each_reflog_ent() need to conform to a
particular interface, but not every function needs all of the
parameters. Mark the unused ones to make -Wunused-parameter happy.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Functions used with for_each_ref(), etc, need to conform to the
each_ref_fn interface. But most of them don't need every parameter;
let's annotate the unused ones to quiet -Wunused-parameter.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git worktree' parses its subcommands with a long list of if
statements. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
handling missing or unknown subcommands, and listing subcommands for
Bash completion.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git stash' parses its subcommands with a long list of if-else if
statements. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
and listing subcommands for Bash completion.
Note that the push_stash() function implementing the 'push' subcommand
accepts an extra flag parameter to indicate whether push was assumed,
so add a wrapper function with the standard subcommand function
signature.
Note also that this change "hides" the '-h' option in 'git stash push
-h' from the parse_option() call in cmd_stash(), as it comes after the
subcommand. Consequently, from now on it will emit the usage of the
'push' subcommand instead of the usage of 'git stash'. We had a
failing test for this case, which can now be flipped to expect
success.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git sparse-checkout' parses its subcommands with a couple of if
statements. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
handling missing or unknown subcommands, and listing subcommands for
Bash completion.
Note that some of the functions implementing each subcommand only
accept the 'argc' and '**argv' parameters, so add a (unused) '*prefix'
parameter to make them match the type expected by parse-options, and
thus avoid casting function pointers.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git remote' parses its subcommands with a long list of if-else if
statements. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
handling unknown subcommands, and listing subcommands for Bash
completion. Make sure that the default operation mode doesn't accept
any arguments; and while at it remove the capitalization of the error
message and adjust the test checking it accordingly.
Note that 'git remote' has both 'remove' and 'rm' subcommands, and the
former is preferred [1], so hide the latter for completion.
Note also that the functions implementing each subcommand only accept
the 'argc' and '**argv' parameters, so add a (unused) '*prefix'
parameter to make them match the type expected by parse-options, and
thus avoid casting a bunch of function pointers.
[1] e17dba8fe1 (remote: prefer subcommand name 'remove' to 'rm',
2012-09-06)
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git reflog' parses its subcommands with a couple of if-else if
statements. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
and listing subcommands for Bash completion.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git notes' parses its subcommands with a long list of if-else if
statements. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
handling unknown subcommands, and listing subcommands for Bash
completion. Make sure that the default operation mode doesn't accept
any arguments.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git multi-pack-index' parses its subcommands with a couple of if-else
if statements. parse-options has just learned to parse subcommands,
so let's use that facility instead, with the benefits of shorter code,
handling missing or unknown subcommands, and listing subcommands for
Bash completion.
Note that the functions implementing each subcommand only accept the
'argc' and '**argv' parameters, so add a (unused) '*prefix' parameter
to make them match the type expected by parse-options, and thus avoid
casting function pointers.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git hook' parses its currently only subcommand with an if statement.
parse-options has just learned to parse subcommands, so let's use that
facility instead, with the benefits of shorter code, handling missing
or unknown subcommands, and listing subcommands for Bash completion.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git maintenanze' parses its subcommands with a couple of if
statements. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
handling missing or unknown subcommands, and listing subcommands for
Bash completion.
This change makes 'git maintenance' consistent with other commands in
that the help text shown for '-h' goes to standard output, not error,
in the exit code and error message on unknown subcommand, and the
error message on missing subcommand. There is a test checking these,
which is now updated accordingly.
Note that some of the functions implementing each subcommand don't
accept any parameters, so add the (unused) 'argc', '**argv' and
'*prefix' parameters to make them match the type expected by
parse-options, and thus avoid casting function pointers.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git commit-graph' parses its subcommands with an if-else if
statement. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
handling missing or unknown subcommands, and listing subcommands for
Bash completion.
Note that the functions implementing each subcommand only accept the
'argc' and '**argv' parameters, so add a (unused) '*prefix' parameter
to make them match the type expected by parse-options, and thus avoid
casting function pointers.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git bundle' parses its subcommands with a couple of if-else if
statements. parse-options has just learned to parse subcommands, so
let's use that facility instead, with the benefits of shorter code,
handling missing or unknown subcommands, and listing subcommands for
Bash completion.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Several Git commands have subcommands to implement mutually exclusive
"operation modes", and they usually parse their subcommand argument
with a bunch of if-else if statements.
Teach parse-options to handle subcommands as well, which will result
in shorter and simpler code with consistent error handling and error
messages on unknown or missing subcommand, and it will also make
possible for our Bash completion script to handle subcommands
programmatically.
The approach is guided by the following observations:
- Most subcommands [1] are implemented in dedicated functions, and
most of those functions [2] either have a signature matching the
'int cmd_foo(int argc, const char **argc, const char *prefix)'
signature of builtin commands or can be trivially converted to
that signature, because they miss only that last prefix parameter
or have no parameters at all.
- Subcommand arguments only have long form, and they have no double
dash prefix, no negated form, and no description, and they don't
take any arguments, and can't be abbreviated.
- There must be exactly one subcommand among the arguments, or zero
if the command has a default operation mode.
- All arguments following the subcommand are considered to be
arguments of the subcommand, and, conversely, arguments meant for
the subcommand may not preceed the subcommand.
So in the end subcommand declaration and parsing would look something
like this:
parse_opt_subcommand_fn *fn = NULL;
struct option builtin_commit_graph_options[] = {
OPT_STRING(0, "object-dir", &opts.obj_dir, N_("dir"),
N_("the object directory to store the graph")),
OPT_SUBCOMMAND("verify", &fn, graph_verify),
OPT_SUBCOMMAND("write", &fn, graph_write),
OPT_END(),
};
argc = parse_options(argc, argv, prefix, options,
builtin_commit_graph_usage, 0);
return fn(argc, argv, prefix);
Here each OPT_SUBCOMMAND specifies the name of the subcommand and the
function implementing it, and the address of the same 'fn' subcommand
function pointer. parse_options() then processes the arguments until
it finds the first argument matching one of the subcommands, sets 'fn'
to the function associated with that subcommand, and returns, leaving
the rest of the arguments unprocessed. If none of the listed
subcommands is found among the arguments, parse_options() will show
usage and abort.
If a command has a default operation mode, 'fn' should be initialized
to the function implementing that mode, and parse_options() should be
invoked with the PARSE_OPT_SUBCOMMAND_OPTIONAL flag. In this case
parse_options() won't error out when not finding any subcommands, but
will return leaving 'fn' unchanged. Note that if that default
operation mode has any --options, then the PARSE_OPT_KEEP_UNKNOWN_OPT
flag is necessary as well (otherwise parse_options() would error out
upon seeing the unknown option meant to the default operation mode).
Some thoughts about the implementation:
- The same pointer to 'fn' must be specified as 'value' for each
OPT_SUBCOMMAND, because there can be only one set of mutually
exclusive subcommands; parse_options() will BUG() otherwise.
There are other ways to tell parse_options() where to put the
function associated with the subcommand given on the command line,
but I didn't like them:
- Change parse_options()'s signature by adding a pointer to
subcommand function to be set to the function associated with
the given subcommand, affecting all callsites, even those that
don't have subcommands.
- Introduce a specific parse_options_and_subcommand() variant
with that extra funcion parameter.
- I decided against automatically calling the subcommand function
from within parse_options(), because:
- There are commands that have to perform additional actions
after option parsing but before calling the function
implementing the specified subcommand.
- The return code of the subcommand is usually the return code
of the git command, but preserving the return code of the
automatically called subcommand function would have made the
API awkward.
- Also add a OPT_SUBCOMMAND_F() variant to allow specifying an
option flag: we have two subcommands that are purposefully
excluded from completion ('git remote rm' and 'git stash save'),
so they'll have to be specified with the PARSE_OPT_NOCOMPLETE
flag.
- Some of the 'parse_opt_flags' don't make sense with subcommands,
and using them is probably just an oversight or misunderstanding.
Therefore parse_options() will BUG() when invoked with any of the
following flags while the options array contains at least one
OPT_SUBCOMMAND:
- PARSE_OPT_KEEP_DASHDASH: parse_options() stops parsing
arguments when encountering a "--" argument, so it doesn't
make sense to expect and keep one before a subcommand, because
it would prevent the parsing of the subcommand.
However, this flag is allowed in combination with the
PARSE_OPT_SUBCOMMAND_OPTIONAL flag, because the double dash
might be meaningful for the command's default operation mode,
e.g. to disambiguate refs and pathspecs.
- PARSE_OPT_STOP_AT_NON_OPTION: As its name suggests, this flag
tells parse_options() to stop as soon as it encouners a
non-option argument, but subcommands are by definition not
options... so how could they be parsed, then?!
- PARSE_OPT_KEEP_UNKNOWN: This flag can be used to collect any
unknown --options and then pass them to a different command or
subsystem. Surely if a command has subcommands, then this
functionality should rather be delegated to one of those
subcommands, and not performed by the command itself.
However, this flag is allowed in combination with the
PARSE_OPT_SUBCOMMAND_OPTIONAL flag, making possible to pass
--options to the default operation mode.
- If the command with subcommands has a default operation mode, then
all arguments to the command must preceed the arguments of the
subcommand.
AFAICT we don't have any commands where this makes a difference,
because in those commands either only the command accepts any
arguments ('notes' and 'remote'), or only the default subcommand
('reflog' and 'stash'), but never both.
- The 'argv' array passed to subcommand functions currently starts
with the name of the subcommand. Keep this behavior. AFAICT no
subcommand functions depend on the actual content of 'argv[0]',
but the parse_options() call handling their options expects that
the options start at argv[1].
- To support handling subcommands programmatically in our Bash
completion script, 'git cmd --git-completion-helper' will now list
both subcommands and regular --options, if any. This means that
the completion script will have to separate subcommands (i.e.
words without a double dash prefix) from --options on its own, but
that's rather easy to do, and it's not much work either, because
the number of subcommands a command might have is rather low, and
those commands accept only a single --option or none at all. An
alternative would be to introduce a separate option that lists
only subcommands, but then the completion script would need not
one but two git invocations and command substitutions for commands
with subcommands.
Note that this change doesn't affect the behavior of our Bash
completion script, because when completing the --option of a
command with subcommands, e.g. for 'git notes --<TAB>', then all
subcommands will be filtered out anyway, as none of them will
match the word to be completed starting with that double dash
prefix.
[1] Except 'git rerere', because many of its subcommands are
implemented in the bodies of the if-else if statements parsing the
command's subcommand argument.
[2] Except 'credential', 'credential-store' and 'fsmonitor--daemon',
because some of the functions implementing their subcommands take
special parameters.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The description of 'PARSE_OPT_KEEP_UNKNOWN' starts with "Keep unknown
arguments instead of erroring out". This is a bit misleading, as this
flag only applies to unknown --options, while non-option arguments are
kept even without this flag.
Update the description to clarify this, and rename the flag to
PARSE_OPTIONS_KEEP_UNKNOWN_OPT to make this obvious just by looking at
the flag name.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git rev-list --disk-usage" learned to take an optional value
"human" to show the reported value in human-readable format, like
"3.40MiB".
* ll/disk-usage-humanise:
rev-list: support human-readable output for `--disk-usage`
"git rm" has become more aware of the sparse-index feature.
* sy/sparse-rm:
rm: integrate with sparse-index
rm: expand the index only when necessary
pathspec.h: move pathspec_needs_expanded_index() from reset.c to here
t1092: add tests for `git-rm`
Fixes to sparse index compatibility work for "reset" and "checkout"
commands.
* vd/sparse-reset-checkout-fixes:
unpack-trees: unpack new trees as sparse directories
cache.h: create 'index_name_pos_sparse()'
oneway_diff: handle removed sparse directories
checkout: fix nested sparse directory diff in sparse index
Expose a lot of "tech docs" via "git help" interface.
* ab/tech-docs-to-help:
docs: move http-protocol docs to man section 5
docs: move cruft pack docs to gitformat-pack
docs: move pack format docs to man section 5
docs: move signature docs to man section 5
docs: move index format docs to man section 5
docs: move protocol-related docs to man section 5
docs: move commit-graph format docs to man section 5
git docs: add a category for file formats, protocols and interfaces
git docs: add a category for user-facing file, repo and command UX
git help doc: use "<doc>" instead of "<guide>"
help.c: remove common category behavior from drop_prefix() behavior
help.c: refactor drop_prefix() to use a "switch" statement"
Create a '--diagnose' option for 'git bugreport' to collect additional
information about the repository and write it to a zipped archive.
The '--diagnose' option behaves effectively as an alias for simultaneously
running 'git bugreport' and 'git diagnose'. In the documentation, users are
explicitly recommended to attach the diagnostics alongside a bug report to
provide additional context to readers, ideally reducing some back-and-forth
between reporters and those debugging the issue.
Note that '--diagnose' may take an optional string arg (either 'stats' or
'all'). If specified without the arg, the behavior corresponds to running
'git diagnose' without '--mode'. As with 'git diagnose', this default is
intended to help reduce unintentional leaking of sensitive information).
Users can also explicitly specify '--diagnose=(stats|all)' to generate the
respective archive created by 'git diagnose --mode=(stats|all)'.
Suggested-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Create '--mode=<mode>' option in 'git diagnose' to allow users to optionally
select non-default diagnostic information to include in the output archive.
Additionally, document the currently-available modes, emphasizing the
importance of not sharing a '--mode=all' archive publicly due to the
presence of sensitive information.
Note that the option parsing callback - 'option_parse_diagnose()' - is added
to 'diagnose.c' rather than 'builtin/diagnose.c' so that it may be reused in
future callers configuring a diagnostics archive.
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Create a 'git diagnose' builtin to generate a standalone zip archive of
repository diagnostics.
The "diagnose" functionality was originally implemented for Scalar in
aa5c79a331 (scalar: implement `scalar diagnose`, 2022-05-28). However, the
diagnostics gathered are not specific to Scalar-cloned repositories and
can be useful when diagnosing issues in any Git repository.
Helped-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Plug a bit more leaks in the revisions API.
* ab/plug-revisions-leak:
revisions API: don't leak memory on argv elements that need free()-ing
bisect.c: partially fix bisect_rev_setup() memory leak
log: refactor "rev.pending" code in cmd_show()
log: fix a memory leak in "git show <revision>..."
test-fast-rebase helper: use release_revisions() (again)
bisect.c: add missing "goto" for release_revisions()
"git symbolic-ref symref non..sen..se" is now diagnosed as an error.
* lt/symbolic-ref-sanity:
symbolic-ref: refuse to set syntactically invalid target
The '--disk-usage' option for git-rev-list was introduced in 16950f8384
(rev-list: add --disk-usage option for calculating disk usage, 2021-02-09).
This is very useful for people inspect their git repo's objects usage
infomation, but the resulting number is quit hard for a human to read.
Teach git rev-list to output a human readable result when using
'--disk-usage'.
Signed-off-by: Li Linchao <lilinchao@oschina.cn>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git checkout" miscounted the paths it updated, which has been
corrected.
source: <cover.1657799213.git.matheus.bernardino@usp.br>
* mt/checkout-count-fix:
checkout: fix two bugs on the final count of updated entries
checkout: show bug about failed entries being included in final report
checkout: document bug where delayed checkout counts entries twice
The resolve-undo information in the index was not protected against
GC, which has been corrected.
source: <xmqq35f7kzad.fsf@gitster.g>
* jc/resolve-undo:
fsck: do not dereference NULL while checking resolve-undo data
revision: mark blobs needed for resolve-undo as reachable
A previous change added the '--bundle-uri' option, but did not check
if the --depth parameter was included. Since bundles are not compatible
with shallow clones, provide an error message to the user who is
attempting this combination.
I am leaving this as its own change, separate from the one that
implements '--bundle-uri', because this is more of an advisory for the
user. There is nothing wrong with bootstrapping with bundles and then
fetching a shallow clone. However, that is likely going to involve too
much work for the client _and_ the server. The client will download all
of this bundle information containing the full history of the
repository only to ignore most of it. The server will get a shallow
fetch request, but with a list of haves that might cause a more painful
computation of that shallow pack-file.
Reviewed-by: Josh Steadmon <steadmon@google.com>
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Cloning a remote repository is one of the most expensive operations in
Git. The server can spend a lot of CPU time generating a pack-file for
the client's request. The amount of data can clog the network for a long
time, and the Git protocol is not resumable. For users with poor network
connections or are located far away from the origin server, this can be
especially painful.
Add a new '--bundle-uri' option to 'git clone' to bootstrap a clone from
a bundle. If the user is aware of a bundle server, then they can tell
Git to bootstrap the new repository with these bundles before fetching
the remaining objects from the origin server.
Reviewed-by: Josh Steadmon <steadmon@google.com>
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add checking logic for overwriting when moving from in-cone to
out-of-cone. It is the index version of the original overwrite logic.
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add an advice.
When the user use `git mv --sparse <dirty-path> <destination>`, Git
will warn the user to use `git add --sparse <paths>` then use
`git sparse-checkout reapply` to apply the sparsity rules.
Add a few lines to previous "move dirty path" tests so we can test
this new advice is working.
Suggested-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Originally, moving from-in-to-out may leave an empty <source>
directory on-disk (this kind of directory is marked as
WORKING_DIRECTORY).
Cleanup such directories if they are empty (don't have any entries
under them).
Modify two tests that take <source> as WORKING_DIRECTORY to test
this behavior.
Suggested-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Originally, moving an in-cone <source> to an out-of-cone <destination>
was not possible, mainly because such <destination> is a directory that
is not present in the working tree.
Change the behavior so that we can move an in-cone <source> to
out-of-cone <destination> when --sparse is supplied.
Notice that <destination> can also be an out-of-cone file path, rather
than a directory.
Such <source> can be either clean or dirty, and moving it results in
different behaviors:
A clean move should move <source> to <destination> in the index (do
*not* create <destination> in the worktree), then delete <source> from
the worktree.
A dirty move should move the <source> to the <destination>, both in the
working tree and the index, but should *not* remove the resulted path
from the working tree and should *not* turn on its CE_SKIP_WORKTREE bit.
Optional reading
================
We are strict about cone mode when <destination> is a file path.
The reason is that some of the previous tests that use no-cone mode in
t7002 are keep breaking, mainly because the `dst_mode = SPARSE;` line
added in this patch.
Most features developed in both "from-out-to-in" and "from-in-to-out"
only care about cone mode situation, as no-cone mode is becoming
irrelevant. And because assigning `SPARSE` to `dst_mode` when the
repo is in no-cone mode causes miscellaneous bugs, we should just leave
this new functionality to be exclusive cone mode and save some time.
Helped-by: Derrick Stolee <derrickstolee@github.com>
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since BOTH is not used anywhere in the code and its meaning is unclear,
remove it.
Helped-by: Derrick Stolee <derrickstolee@github.com>
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Originally, <destination> is assumed to be in the working tree. If it is
not found as a directory, then it is determined to be either a regular file
path, or error out if used under the second form (move into a directory)
of 'git-mv'. Such behavior is not ideal, mainly because Git does not
look into the index for <destination>, which could potentially be a
SKIP_WORKTREE_DIR, which we need to determine for the later "moving from
in-cone to out-of-cone" patch.
Change the logic so that Git first check if <destination> is a directory
with all its contents sparsified (a SKIP_WORKTREE_DIR).
If <destination> is such a sparse directory, then we should modify the
index the same way as we would if this were a non-sparse directory. We
must be careful to ensure that the <destination> is marked with
SKIP_WORKTREE_DIR.
Also add a `dst_w_slash` to reuse the result from `add_slash()`, which
was everywhere and can be simplified.
Helped-by: Derrick Stolee <derrickstolee@github.com>
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
with_slash may be a malloc'd pointer, and when it is, free it.
Helped-by: Derrick Stolee <derrickstolee@github.com>
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Method check_dir_in_index() introduced in b91a2b6594 (mv: add
check_dir_in_index() and solve general dir check issue, 2022-06-30)
does not describe its intent and behavior well.
Change its name to empty_dir_has_sparse_contents(), which more
precisely describes its purpose.
Reverse the return values, check_dir_in_index() return 0 for success
and 1 for failure; reverse the values so empty_dir_has_sparse_contents()
return 1 for success and 0 for failure. These values are more intuitive
because 1 usually means "has" and 0 means "not found".
Also modify the documentation to better align with the method's
intent and behavior.
Helped-by: Derrick Stolee <derrickstolee@github.com>
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Enable the sparse index within the `git-rm` command.
The `p2000` tests demonstrate a ~92% execution time reduction for
'git rm' using a sparse index.
Test HEAD~1 HEAD
--------------------------------------------------------------------------
2000.74: git rm ... (full-v3) 0.41(0.37+0.05) 0.43(0.36+0.07) +4.9%
2000.75: git rm ... (full-v4) 0.38(0.34+0.05) 0.39(0.35+0.05) +2.6%
2000.76: git rm ... (sparse-v3) 0.57(0.56+0.01) 0.05(0.05+0.00) -91.2%
2000.77: git rm ... (sparse-v4) 0.57(0.55+0.02) 0.03(0.03+0.00) -94.7%
----
Also, normalize a behavioral difference of `git-rm` under sparse-index.
See related discussion [1].
`git-rm` a sparse-directory entry within a sparse-index enabled repo
behaves differently from a sparse directory within a sparse-checkout
enabled repo.
For example, in a sparse-index repo, where 'folder1' is a
sparse-directory entry, `git rm -r --sparse folder1` provides this:
rm 'folder1/'
Whereas in a sparse-checkout repo *without* sparse-index, doing so
provides this:
rm 'folder1/0/0/0'
rm 'folder1/0/1'
rm 'folder1/a'
Because `git rm` a sparse-directory entry does not need to expand the
index, therefore we should accept the current behavior, which is faster
than "expand the sparse-directory entry to match the sparse-checkout
situation".
Modify a previous test so such difference is not considered as an error.
[1] https://github.com/ffyuanda/git/pull/6#discussion_r934861398
Helped-by: Victoria Dye <vdye@github.com>
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove the `ensure_full_index()` method so `git-rm` does not always
expand the index when the expansion is unnecessary, i.e. when
<pathspec> does not have any possibilities to match anything outside
of sparse-checkout definition.
Expand the index when the <pathspec> needs an expanded index, i.e. the
<pathspec> contains wildcard that may need a full-index or the
<pathspec> is simply outside of sparse-checkout definition.
Notice that the test 'rm pathspec expands index when necessary' in
t1092 *is* testing this code change behavior, though it will be marked
as 'test_expect_success' only in the next patch, where we officially
mark `command_requires_full_index = 0`, so the index does not expand
unless we tell it to do so.
Notice that because we also want `ensure_full_index` to record the
stdout and stderr from Git command, a corresponding modification
is also included in this patch. The reason we want the "sparse-index-out"
and "sparse-index-err", is that we need to make sure there is no error
from Git command itself, so we can rely on the `test_region` result
and determine if the index is expanded or not.
Helped-by: Victoria Dye <vdye@github.com>
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Method pathspec_needs_expanded_index() in reset.c from 4d1cfc1351
(reset: make --mixed sparse-aware, 2021-11-29) is reusable when we
need to verify if the index needs to be expanded when the command
is utilizing a pathspec rather than a literal path.
Move it to pathspec.h for reusability.
Add a few items to the function so it can better serve its purpose as
a standalone public function:
* Add a check in front so if the index is not sparse, return early since
no expansion is needed.
* It now takes an arbitrary 'struct index_state' pointer instead of
using `the_index` and `active_cache`.
* Add documentation to the function.
Helped-by: Victoria Dye <vdye@github.com>
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add the 'recursive' diff flag to the local changes reporting done by 'git
checkout' in 'show_local_changes()'. Without the flag enabled, unexpanded
sparse directories will not be recursed into to report the diff of each
file's contents, resulting in the reported local changes including
"modified" sparse directories.
The same issue was found and fixed for 'git status' in 2c521b0e49 (status:
fix nested sparse directory diff in sparse index, 2022-03-01)
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Operating modes like "--batch" of "git cat-file" command learned to
take NUL-terminated input, instead of one-item-per-line.
* tb/cat-file-z:
builtin/cat-file.c: support NUL-delimited input with `-z`
t1006: extract --batch-command inputs to variables
"git clone" from a repository with some ref whose HEAD is unborn
did not set the HEAD in the resulting repository correctly, which
has been corrected.
source: <YsdyLS4UFzj0j/wB@coredump.intra.peff.net>
* jk/clone-unborn-confusion:
clone: move unborn head creation to update_head()
clone: use remote branch if it matches default HEAD
clone: propagate empty remote HEAD even with other branches
clone: drop extra newline from warning message
The "refs/prefetch/" namespace is used by 'git fetch --prefetch' as a
replacement of the destination of the refpsec for a remote. Git also
removes refspecs that include tags.
Instead of using string literals for the 'refs/tags/ and
'refs/prefetch/' namespaces, use the entries in the ref_namespaces
array.
This kind of change could be done in many places around the codebase,
but we are isolating only to this change because of the way the
refs/prefetch/ namespace somewhat motivated the creation of the
ref_namespaces array.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This reverts commit 96eaffebbf (maintenance: set
log.excludeDecoration durin prefetch, 2021-01-19).
The previous change created a default decoration filter that does not
include refs/prefetch/, so this modification of the config is no longer
needed.
One issue that can happen from this point on is that users who ran the
prefetch task on previous versions of Git will still have a
log.excludeDecoration value and that will prevent the new default
decoration filter from being active. Thus, when we add the refs/bundle/
namespace as part of the bundle URI feature, those users will see
refs/bundle/ decorations.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous change introduced the --clear-decorations option for users
who do not want their decorations limited to a narrow set of ref
namespaces.
Add a config option that is equivalent to specifying --clear-decorations
by default.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous changes introduced a new default ref filter for decorations
in the 'git log' command. This can be overridden using
--decorate-refs=HEAD and --decorate-refs=refs/, but that is cumbersome
for users.
Instead, add a --clear-decorations option that resets all previous
filters to a blank filter that accepts all refs.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a user runs 'git log', they expect a certain set of helpful
decorations. This includes:
* The HEAD ref
* Branches (refs/heads/)
* Stashes (refs/stash)
* Tags (refs/tags/)
* Remote branches (refs/remotes/)
* Replace refs (refs/replace/ or $GIT_REPLACE_REF_BASE)
Each of these namespaces was selected due to existing test cases that
verify these namespaces appear in the decorations. In particular,
stashes and replace refs can have custom colors from the
color.decorate.<slot> config option.
While one test checks for a decoration from notes, it only applies to
the tip of refs/notes/commit (or its configured ref name). Notes form
their own kind of decoration instead. Modify the expected output for the
tests in t4013 that expect this note decoration. There are several
tests throughout the codebase that verify that --decorate-refs,
--decorate-refs-exclude, and log.excludeDecoration work as designed and
the tests continue to pass without intervention.
However, there are other refs that are less helpful to show as
decoration:
* Prefetch refs (refs/prefetch/)
* Rebase refs (refs/rebase-merge/ and refs/rebase-apply/)
* Bundle refs (refs/bundle/) [!]
[!] The bundle refs are part of a parallel series that bootstraps a repo
from a bundle file, storing the bundle's refs into the repo's
refs/bundle/ namespace.
In the case of prefetch refs, 96eaffebbf (maintenance: set
log.excludeDecoration durin prefetch, 2021-01-19) added logic to add
refs/prefetch/ to the log.excludeDecoration config option. Additional
feedback pointed out that having such a side-effect can be confusing and
perhaps not helpful to users. Instead, we should hide these ref
namespaces that are being used by Git for internal reasons but are not
helpful for the users to see.
The way to provide a seamless user experience without setting the config
is to modify the default decoration filters to match our expectation of
what refs the user actually wants to see.
In builtin/log.c, after parsing the --decorate-refs and
--decorate-refs-exclude options from the command-line, call
set_default_decoration_filter(). This method populates the exclusions
from log.excludeDecoration, then checks if the list of pattern
modifications are empty. If none are specified, then the default set is
restricted to the set of inclusions mentioned earlier (HEAD, branches,
etc.). A previous change introduced the ref_namespaces array, which
includes all of these currently-used namespaces. The 'decoration' value
is non-zero when that namespace is associated with a special coloring
and fits into the list of "expected" decorations as described above,
which makes the implementation of this filter very simple.
Note that the logic in ref_filter_match() in log-tree.c follows this
matching pattern:
1. If there are exclusion patterns and the ref matches one, then ignore
the decoration.
2. If there are inclusion patterns and the ref matches one, then
definitely include the decoration.
3. If there are config-based exclusions from log.excludeDecoration and
the ref matches one, then ignore the decoration.
With this logic in mind, we need to ensure that we do not populate our
new defaults if any of these filters are manually set. Specifically, if
a user runs
git -c log.excludeDecoration=HEAD log
then we expect the HEAD decoration to not appear. If we left the default
inclusions in the set, then HEAD would match that inclusion before
reaching the config-based exclusions.
A potential alternative would be to check the list of default inclusions
at the end, after the config-based exclusions. This would still create a
behavior change for some uses of --decorate-refs-exclude=<X>, and could
be overwritten somewhat with --decorate-refs=refs/ and
--decorate-refs=HEAD. However, it no longer becomes possible to include
refs outside of the defaults while also excluding some using
log.excludeDecoration.
Another alternative would be to exclude the known namespaces that are
not intended to be shown. This would reduce the visible effect of the
change for expert users who use their own custom ref namespaces. The
implementation change would be very simple to swap due to our use of
ref_namespaces:
int i;
struct string_list *exclude = decoration_filter->exclude_ref_pattern;
/*
* No command-line or config options were given, so
* populate with sensible defaults.
*/
for (i = 0; i < NAMESPACE__COUNT; i++) {
if (ref_namespaces[i].decoration)
continue;
string_list_append(exclude, ref_namespaces[i].ref);
}
The main downside of this approach is that we expect to add new hidden
namespaces in the future, and that means that Git versions will be less
stable in how they behave as those namespaces are added.
It is critical that we provide ways for expert users to disable this
behavior change via command-line options and config keys. These changes
will be implemented in a future change.
Add a test that checks that the defaults are not added when
--decorate-refs is specified. We verify this by showing that HEAD is not
included as it normally would. Also add a test that shows that the
default filter avoids the unwanted decorations from refs/prefetch,
refs/rebase-merge,
and refs/bundle.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The git_replace_ref_base global is used to store the value of the
GIT_REPLACE_REF_BASE environment variable or the default of
"refs/replace/". This is initialized within setup_git_env().
The ref_namespaces array is a new centralized location for information
such as the ref namespace used for replace refs. Instead of having this
namespace stored in two places, use the ref_namespaces array instead.
For simplicity, create a local git_replace_ref_base variable wherever
the global was previously used.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Create a new "File formats, protocols and other developer interfaces"
section in the main "git help git" manual page and start moving the
documentation that now lives in "Documentation/technical/*.git" over
to it. This complements the newly added and adjacent "Repository,
command and file interfaces" section.
This makes the technical documentation more accessible and
discoverable. Before this we wouldn't install it by default, and had
no ability to build man page versions of them. The links to them from
our existing documentation link to the generated HTML version of these
docs.
So let's start moving those over, starting with just the
"bundle-format.txt" documentation added in 7378ec90e1 (doc: describe
Git bundle format, 2020-02-07). We'll now have a new
gitformat-bundle(5) man page. Subsequent commits will move more git
internal format documentation over.
Unfortunately the syntax of the current Documentation/technical/*.txt
is not the same (when it comes to section headings etc.) as our
Documentation/*.txt documentation, so change the relevant bits of
syntax as we're moving this over.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Create a new "Repository, command and file interfaces" section in the
main "git help git" manual page. Move things that belong under this
new criteria from the generic "Guides" section.
The "Guides" section was added in f442f28a81 (git.txt: add list of
guides, 2020-08-05). It makes sense to have e.g. "giteveryday(7)" and
"gitfaq(7)" listed under "Guides".
But placing e.g. "gitignore(5)" in it is stretching the meaning of
what a "guide" is, ideally that section should list things similar to
"giteveryday(7)" and "gitcore-tutorial(7)".
An alternate name that was considered for this new section was "User
formats", for consistency with the nomenclature used for man section 5
in general. My man(1) lists it as "File formats and conventions,
e.g. /etc/passwd".
So calling this "git help --formats" or "git help --user-formats"
would make sense for e.g. gitignore(5), but would be stretching it
somewhat for githooks(5), and would seem really suspect for the likes
of gitcli(7).
Let's instead pick a name that's closer to the generic term "User
interface", which is really what this documentation discusses: General
user-interface documentation that doesn't obviously belong elsewhere.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Replace the use of "<guide>" originally introduced (as "GUIDE") in
a133737b80 (doc: include --guide option description for "git help",
2013-04-02) with the more generic "<doc>". The "<doc>" placeholder is
more generic, and one we'll be able to use as we introduce new
documentation categories.
Let's also add "<doc>" to the "git help -h" output, when it was made
to use parse_option() in in 41eb33bd0c (help: use parseopt,
2008-02-24).
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When "git merge" finds that it cannot perform a merge, it should
restore the working tree to the state before the command was
initiated, but in some corner cases it didn't.
* en/merge-restore-to-pristine:
merge: do not exit restore_state() prematurely
merge: ensure we can actually restore pre-merge state
merge: make restore_state() restore staged state too
merge: fix save_state() to work when there are stat-dirty files
merge: do not abort early if one strategy fails to handle the merge
merge: abort if index does not match HEAD for trivial merges
merge-resolve: abort if index does not match HEAD
merge-ort-wrappers: make printed message match the one from recursive
"git cat-file" learned an option to use the mailmap when showing
commit and tag objects.
* sa/cat-file-mailmap:
cat-file: add mailmap support
ident: rename commit_rewrite_person() to apply_mailmap_to_header()
ident: move commit_rewrite_person() to ident.c
revision: improve commit_rewrite_person()
Add a "free_removed_argv_elements" member to "struct
setup_revision_opt", and use it to fix several memory leaks.
We have various memory leaks in APIs that take and munge "const
char **argv", e.g. parse_options(). Sometimes these APIs are given the
"argv" we get to the "main" function, in which case we don't leak
memory, but other times we're giving it the "v" member of a "struct
strvec" we created.
There's several potential ways to fix those sort of leaks, we could
add a "nodup" mode to "struct strvec", which would work for the cases
where we push constant strings to it. But that wouldn't work as soon
as we used strvec_pushf(), or otherwise needed to duplicate or create
a string for that "struct strvec".
Let's instead make it the responsibility of the revisions API. If it's
going to clobber elements of argv it can also free() them, which it
will now do if instructed to do so via "free_removed_argv_elements".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Refactor the juggling of "rev.pending" and our replacement for it
amended in the preceding commit so that:
* We use an "unsigned int" instead of an "int" for "i", this matches
the types of "struct rev_info" itself.
* We don't need the "count" and "objects" variables introduced in
5d7eeee2ac (git-show: grok blobs, trees and tags, too, 2006-12-14).
They were originally added since we'd clobber rev.pending in the
loop without restoring it. Since the preceding commit we are
restoring it when we handle OBJ_COMMIT, so the main for-loop can
refer to "rev.pending" didrectly.
* We use the "memcpy a &blank" idiom introduced in
5726a6b401 (*.c *_init(): define in terms of corresponding *_INIT
macro, 2021-07-01).
This is more obvious than relying on us enumerating all of the
relevant members of the "struct object_array" that we need to
clear.
* We comment on why we don't need an object_array_clear() here, see
the analysis in [1].
1. https://lore.kernel.org/git/YuQtJ2DxNKX%2Fy70N@coredump.intra.peff.net/
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a memory leak in code added in 5d7eeee2ac (git-show: grok blobs,
trees and tags, too, 2006-12-14). As we iterate over a "<revision>..."
command-line and encounter ad OBJ_COMMIT we want to use our "struct
rev_info", but with a "pending" array of one element: the one commit
we're showing in the loop.
To do this 5d7eeee2ac saved away a pointer to rev.pending.objects and
rev.pending.nr for its iteration. We'd then clobber those (and alloc)
when we needed to show an OBJ_COMMIT.
We'd therefore leak the "rev.pending" we started out with, and only
free the new "rev.pending" in the "OBJ_COMMIT" case arm as
prepare_revision_walk() would draw it down.
Let's fix this memory leak. Now when we encounter an OBJ_COMMIT we
save away the "rev.pending" before clearing it. We then add a single
commit to it, which our indirect invocation of prepare_revision_walk()
will remove. After that we restore the "rev.pending".
Our "rev.pending" will then get free'd by the release_revisions()
added in f6bfea0ad0 (revisions API users: use release_revisions() in
builtin/log.c, 2022-04-13)
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
You can feed absolute garbage to symbolic-ref as a target like:
git symbolic-ref HEAD refs/heads/foo..bar
While this doesn't technically break the repo entirely (our "is it a git
directory" detector looks only for "refs/" at the start), we would never
resolve such a ref, as the ".." is invalid within a refname.
Let's flag these as invalid at creation time to help the caller realize
that what they're asking for is bogus.
A few notes:
- We use REFNAME_ALLOW_ONELEVEL here, which lets:
git update-ref refs/heads/foo FETCH_HEAD
continue to work. It's unclear whether anybody wants to do something
so odd, but it does work now, so this is erring on the conservative
side. There's a test to make sure we didn't accidentally break this,
but don't take that test as an endorsement that it's a good idea, or
something we might not change in the future.
- The test in t4202-log.sh checks how we handle such an invalid ref on
the reading side, so it has to be updated to touch the HEAD file
directly.
- We need to keep our HEAD-specific check for "does it start with
refs/". The ALLOW_ONELEVEL flag means we won't be enforcing that for
other refs, but HEAD is special here because of the checks in
validate_headref().
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git checkout" miscounted the paths it updated, which has been
corrected.
source: <cover.1657799213.git.matheus.bernardino@usp.br>
* mt/checkout-count-fix:
checkout: fix two bugs on the final count of updated entries
checkout: show bug about failed entries being included in final report
checkout: document bug where delayed checkout counts entries twice
"git rebase -i" learns to update branches whose tip appear in the
rebased range with "--update-refs" option.
source: <pull.1247.v5.git.1658255624.gitgitgadget@gmail.com>
* ds/rebase-update-ref:
sequencer: notify user of --update-refs activity
sequencer: ignore HEAD ref under --update-refs
rebase: add rebase.updateRefs config option
sequencer: rewrite update-refs as user edits todo list
rebase: update refs from 'update-ref' commands
rebase: add --update-refs option
sequencer: add update-ref command
sequencer: define array with enum values
rebase-interactive: update 'merge' description
branch: consider refs under 'update-refs'
t2407: test branches currently using apply backend
t2407: test bisect and rebase as black-boxes
"git mktree --missing" lazily fetched objects that are missing from
the local object store, which was totally unnecessary for the purpose
of creating the tree object(s) from its input.
source: <748f39a9-65aa-2110-cf92-7ddf81b5f507@roku.com>
* ro/mktree-allow-missing-fix:
mktree: do not check type of remote objects
An earlier attempt to plug leaks placed a clean-up label to jump to
at a bogus place, which as been corrected.
source: <Ys0c0ePxPOqZ/5ck@coredump.intra.peff.net>
* jk/diff-files-cleanup-fix:
diff-files: move misplaced cleanup label
Add a new option "--format" that outputs index entries
informations in a custom format, taking inspiration
from the option with the same name in the `git ls-tree`
command.
"--format" cannot used with "-s", "-o", "-k", "-t",
" --resolve-undo","--deduplicate" and "--eol".
Signed-off-by: ZheNing Hu <adlternative@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Previously, if the user:
* Had no local changes before starting the merge
* A merge strategy makes changes to the working tree/index but returns
with exit status 2
Then we'd call restore_state() to clean up the changes and either let
the next merge strategy run (if there is one), or exit telling the user
that no merge strategy could handle the merge. Unfortunately,
restore_state() did not clean up the changes as expected; that function
was a no-op if the stash was a null, and the stash would be null if
there were no local changes before starting the merge. So, instead of
"Rewinding the tree to pristine..." as the code claimed, restore_state()
would leave garbage around in the index and working tree (possibly
including conflicts) for either the next merge strategy or for the user
after aborting the merge. And in the case of aborting the merge, the
user would be unable to run "git merge --abort" to get rid of the
unintended leftover conflicts, because the merge control files were not
written as it was presumed that we had restored to a clean state
already.
Fix the main problem by making sure that restore_state() only skips the
stash application if the stash is null rather than skipping the whole
function.
However, there is a secondary problem -- since merge.c forks
subprocesses to do the cleanup, the in-memory index is left out-of-sync.
While there was a refresh_cache(REFRESH_QUIET) call that attempted to
correct that, that function would not handle cases where the previous
merge strategy added conflicted entries. We need to drop the index and
re-read it to handle such cases.
(Alternatively, we could stop forking subprocesses and instead call some
appropriate function to do the work which would update the in-memory
index automatically. For now, just do the simple fix.)
Also, add a testcase checking this, one for which the octopus strategy
fails on the first commit it attempts to merge, and thus which it
cannot handle at all and must completely bail on (as per the "exit 2"
code path of commit 98efc8f3d8 ("octopus: allow manual resolve on the
last round.", 2006-01-13)).
Reported-by: ZheNing Hu <adlternative@gmail.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Merge strategies can:
* succeed with a clean merge
* succeed with a conflicted merge
* fail to handle the given type of merge
If one is thinking in terms of automatic mergeability, they would use
the word "fail" instead of "succeed" for the second bullet, but I am
focusing here on ability of the merge strategy to handle the given
inputs, not on whether the given inputs are mergeable. The third
category is about the merge strategy failing to know how to handle the
given data; examples include:
* Passing more than 2 branches to 'recursive' or 'ort'
* Passing 2 or fewer branches to 'octopus'
* Trying to do more complicated merges with 'resolve' (I believe
directory/file conflicts will cause it to bail.)
* Octopus running into a merge conflict for any branch OTHER than
the final one (see the "exit 2" codepath of commit 98efc8f3d8
("octopus: allow manual resolve on the last round.", 2006-01-13))
That final one is particularly interesting, because it shows that the
merge strategy can muck with the index and working tree, and THEN bail
and say "sorry, this strategy cannot handle this type of merge; use
something else".
Further, we do not currently expect the individual strategies to clean
up after themselves, but instead expect builtin/merge.c to do so. For
it to be able to, it needs to save the state before trying the merge
strategy so it can have something to restore to. Therefore, remove the
shortcut bypassing the save_state() call.
There is another bug on the restore_state() side of things, so no
testcase will be added until the next commit when we have addressed that
issue as well.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are multiple issues at play here:
1) If `git merge` is invoked with staged changes, it should abort
without doing any merging, and the user's working tree and index
should be the same as before merge was invoked.
2) Merge strategies are responsible for enforcing the index == HEAD
requirement. (See 9822175d2b ("Ensure index matches head before
invoking merge machinery, round N", 2019-08-17) for some history
around this.)
3) Merge strategies can bail saying they are not an appropriate
handler for the merge in question (possibly allowing other
strategies to be used instead).
4) Merge strategies can make changes to the index and working tree,
and have no expectation to clean up after themselves, *even* if
they bail out and say they are not an appropriate handler for
the merge in question. (The `octopus` merge strategy does this,
for example.)
5) Because of (3) and (4), builtin/merge.c stashes state before
trying merge strategies and restores it afterward.
Unfortunately, if users had staged changes before calling `git merge`,
builtin/merge.c could do the following:
* stash the changes, in order to clean up after the strategies
* try all the merge strategies in turn, each of which report they
cannot function due to the index not matching HEAD
* restore the changes via "git stash apply"
But that last step would have the net effect of unstaging the user's
changes. Fix this by adding the "--index" option to "git stash apply".
While at it, also squelch the stash apply output; we already report
"Rewinding the tree to pristine..." and don't need a detailed `git
status` report afterwards. Also while at it, switch to using strvec
so folks don't have to count the arguments to ensure we avoided an
off-by-one error, and so it's easier to add additional arguments to
the command.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When there are stat-dirty files, but no files are modified,
`git stash create` exits with unsuccessful status. This causes merge
to fail. Copy some code from sequencer.c's create_autostash to refresh
the index first to avoid this problem.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
builtin/merge is setup to allow multiple strategies to be specified,
and it will find the "best" result and use it. This is defeated if
some of the merge strategies abort early when they cannot handle the
merge. Fix the logic that calls recursive and ort to not do such an
early abort, but instead return "2" or "unhandled" so that the next
strategy can try to handle the merge.
Coming up with a testcase for this is somewhat difficult, since
recursive and ort both handle nearly any two-headed merge (there is
a separate code path that checks for non-two-headed merges and
already returns "2" for them). So use a somewhat synthetic testcase
of having the index not match HEAD before the merge starts, since all
merge strategies will abort for that.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As noted in the last commit and the links therein (especially commit
9822175d2b ("Ensure index matches head before invoking merge machinery,
round N", 2019-08-17), we have had a very long history of problems with
failing to enforce the requirement that index matches HEAD when starting
a merge.
The "trivial merge" logic in builtin/merge.c is yet another such case
we previously missed. Add a check for it to ensure it aborts if the
index does not match HEAD, and add a testcase where this fix is needed.
Note that the fix here would also incidentally be an alternative fix
for the testcase added in the last patch, but the fix in the last patch
is still needed when multiple merge strategies are in use, so tweak the
testcase from the previous commit so that it continues to exercise the
codepath added in the last commit.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When callers are using `cat-file` via one of the stdin-driven `--batch`
modes, all input is newline-delimited. This presents a problem when
callers wish to ask about, e.g. tree-entries that have a newline
character present in their filename.
To support this niche scenario, introduce a new `-z` mode to the
`--batch`, `--batch-check`, and `--batch-command` suite of options that
instructs `cat-file` to treat its input as NUL-delimited, allowing the
individual commands themselves to have newlines present.
The refactoring here is slightly unfortunate, since we turn loops like:
while (strbuf_getline(&buf, stdin) != EOF)
into:
while (1) {
int ret;
if (opt->nul_terminated)
ret = strbuf_getline_nul(&input, stdin);
else
ret = strbuf_getline(&input, stdin);
if (ret == EOF)
break;
}
It's tempting to think that we could use `strbuf_getwholeline()` and
specify either `\n` or `\0` as the terminating character. But for input
on platforms that include a CR character preceeding the LF, this
wouldn't quite be the same, since `strbuf_getline(...)` will trim any
trailing CR, while `strbuf_getwholeline(&buf, stdin, '\n')` will not.
In the future, we could clean this up further by introducing a variant
of `strbuf_getwholeline()` that addresses the aforementioned gap, but
that approach felt too heavy-handed for this pair of uses.
Some tests are added in t1006 to ensure that `cat-file` produces the
same output in `--batch`, `--batch-check`, and `--batch-command` modes
with and without the new `-z` option.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 4a4b4cda (builtin-remote: Make "remote -v" display push urls,
2009-06-13), the string_list that was initialized with 0 in its
strdup_string member is immediately made to strdup its key strings
by flipping the strdup_string member to true. When 183113a5
(string_list: Add STRING_LIST_INIT macro and make use of it.,
2010-07-04) has introduced STRING_LIST_INIT macros, it mechanically
replaced the initialization to STRING_LIST_INIT_NODUP.
Instead, just use the other initialization macro to make it strdup
the key from the beginning.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
An earlier attempt to plug leaks placed a clean-up label to jump to
at a bogus place, which as been corrected.
* jk/diff-files-cleanup-fix:
diff-files: move misplaced cleanup label
"git clone" from a repository with some ref whose HEAD is unborn
did not set the HEAD in the resulting repository correctly, which
has been corrected.
* jk/clone-unborn-confusion:
clone: move unborn head creation to update_head()
clone: use remote branch if it matches default HEAD
clone: propagate empty remote HEAD even with other branches
clone: drop extra newline from warning message
The resolve-undo information in the index was not protected against
GC, which has been corrected.
* jc/resolve-undo:
fsck: do not dereference NULL while checking resolve-undo data
revision: mark blobs needed for resolve-undo as reachable
The previous change added the --update-refs command-line option. For
users who always want this mode, create the rebase.updateRefs config
option which behaves the same way as rebase.autoSquash does with the
--autosquash option.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When working on a large feature, it can be helpful to break that feature
into multiple smaller parts that become reviewed in sequence. During
development or during review, a change to one part of the feature could
affect multiple of these parts. An interactive rebase can help adjust
the multi-part "story" of the branch.
However, if there are branches tracking the different parts of the
feature, then rebasing the entire list of commits can create commits not
reachable from those "sub branches". It can take a manual step to update
those branches.
Add a new --update-refs option to 'git rebase -i' that adds 'update-ref
<ref>' steps to the todo file whenever a commit that is being rebased is
decorated with that <ref>. At the very end, the rebase process updates
all of the listed refs to the values stored during the rebase operation.
Be sure to iterate after any squashing or fixups are placed. Update the
branch only after those squashes and fixups are complete. This allows a
--fixup commit at the tip of the feature to apply correctly to the sub
branch, even if it is fixing up the most-recent commit in that part.
This change update the documentation and builtin to accept the
--update-refs option as well as updating the todo file with the
'update-ref' commands. Tests are added to ensure that these todo
commands are added in the correct locations.
This change does _not_ include the actual behavior of tracking the
updated refs and writing the new ref values at the end of the rebase
process. That is deferred to a later change.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The way "git multi-pack" uses parse-options API has been improved.
* sg/multi-pack-index-parse-options-fix:
multi-pack-index: simplify handling of unknown --options
Add Coccinelle rules to detect the pattern of initializing and then
finalizing a structure without using it in between at all, which
happens after code restructuring and the compilers fail to
recognize as an unused variable.
* ab/cocci-unused:
cocci: generalize "unused" rule to cover more than "strbuf"
cocci: add and apply a rule to find "unused" strbufs
cocci: have "coccicheck{,-pending}" depend on "coccicheck-test"
cocci: add a "coccicheck-test" target and test *.cocci rules
Makefile & .gitignore: ignore & clean "git.res", not "*.res"
Makefile: remove mandatory "spatch" arguments from SPATCH_FLAGS
Apply Coccinelle rule to turn raw memmove() into MOVE_ARRAY() cpp
macro, which would improve maintainability and readability.
* jc/builtin-mv-move-array:
builtin/mv.c: use the MOVE_ARRAY() macro instead of memmove()
git-cat-file is used by tools like GitLab to get commit tag contents
that are then displayed to users. This content which has author,
committer or tagger information, could benefit from passing through the
mailmap mechanism before being sent or displayed.
This patch adds --[no-]use-mailmap command line option to the git
cat-file command. It also adds --[no-]mailmap option as an alias to
--[no-]use-mailmap.
This patch also introduces new test cases to test the mailmap mechanism in
git cat-file command.
Mentored-by: Christian Couder <christian.couder@gmail.com>
Mentored-by: John Cai <johncai86@gmail.com>
Helped-by: Phillip Wood <phillip.wood@dunelm.org.uk>
Helped-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Siddharth Asthana <siddharthasthana31@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Further preparation to turn git-submodule.sh into a builtin.
* ab/submodule-cleanup:
git-sh-setup.sh: remove "say" function, change last users
git-submodule.sh: use "$quiet", not "$GIT_QUIET"
submodule--helper: eliminate internal "--update" option
submodule--helper: understand --checkout, --merge and --rebase synonyms
submodule--helper: report "submodule" as our name in some "-h" output
submodule--helper: rename "absorb-git-dirs" to "absorbgitdirs"
submodule update: remove "-v" option
submodule--helper: have --require-init imply --init
git-submodule.sh: remove unused top-level "--branch" argument
git-submodule.sh: make the "$cached" variable a boolean
git-submodule.sh: remove unused $prefix variable
git-submodule.sh: remove unused sanitize_submodule_env()
"git mv A B" in a sparsely populated working tree can be asked to
move a path between directories that are "in cone" (i.e. expected
to be materialized in the working tree) and "out of cone"
(i.e. expected to be hidden). The handling of such cases has been
improved.
* sy/mv-out-of-cone:
mv: add check_dir_in_index() and solve general dir check issue
mv: use flags mode for update_mode
mv: check if <destination> exists in index to handle overwriting
mv: check if out-of-cone file exists in index with SKIP_WORKTREE bit
mv: decouple if/else-if checks using goto
mv: update sparsity after moving from out-of-cone to in-cone
t1092: mv directory from out-of-cone to in-cone
t7002: add tests for moving out-of-cone file/directory
Allow large objects read from a packstream to be streamed into a
loose object file straight, without having to keep it in-core as a
whole.
* hx/unpack-streaming:
unpack-objects: use stream_loose_object() to unpack large objects
core doc: modernize core.bigFileThreshold documentation
object-file.c: add "stream_loose_object()" to handle large object
object-file.c: factor out deflate part of write_loose_object()
object-file.c: refactor write_loose_object() to several steps
unpack-objects: low memory footprint for get_data() in dry_run mode
"git merge-tree" learned a new mode where it takes two commits and
computes a tree that would result in the merge commit, if the
histories leading to these two commits were to be merged.
* en/merge-tree:
git-merge-tree.txt: add a section on potentional usage mistakes
merge-tree: add a --allow-unrelated-histories flag
merge-tree: allow `ls-files -u` style info to be NUL terminated
merge-ort: optionally produce machine-readable output
merge-ort: store more specific conflict information
merge-ort: make `path_messages` a strmap to a string_list
merge-ort: store messages in a list, not in a single strbuf
merge-tree: provide easy access to `ls-files -u` style info
merge-tree: provide a list of which files have conflicts
merge-ort: remove command-line-centric submodule message from merge-ort
merge-ort: provide a merge_get_conflicted_files() helper function
merge-tree: support including merge messages in output
merge-ort: split out a separate display_update_messages() function
merge-tree: implement real merges
merge-tree: add option parsing and initial shell for real merge function
merge-tree: move logic for existing merge into new function
merge-tree: rename merge_trees() to trivial_merge_trees()
When sorting the output of `git shortlog` by count, a list of authors in
alphabetical order is then sorted by contribution count. Obviously, the
idea is to maintain the alphabetical order for items with identical
contribution count.
At the moment, this job is performed by `qsort()`. As that function is
not guaranteed to implement a stable sort algorithm, this can lead to
inconsistent and/or surprising behavior: items with identical
contribution count could lose their alphabetical sub-order.
The `qsort()` in MS Visual C's runtime does _not_ implement a stable
sort algorithm, and under certain circumstances this even causes a test
failure in t4201.21 "shortlog can match multiple groups", where two
authors both are listed with 2 contributions, and are listed in inverse
alphabetical order.
Let's instead use the stable sort provided by `git_stable_qsort()` to
avoid this inconsistency.
This is a companion to 2049b8dc65 (diffcore_rename(): use a stable sort,
2019-09-30).
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
At the end of `git checkout <pathspec>`, we get a message informing how
many entries were updated in the working tree. However, this number can
be inaccurate for two reasons:
1) Delayed entries currently get counted twice.
2) Failed entries are included in the count.
The first problem happens because the counter is first incremented
before inserting the entry in the delayed checkout queue, and once again
when finish_delayed_checkout() calls checkout_entry(). And the second
happens because the counter is incremented too early in
checkout_entry(), before the entry was in fact checked out. Fix that by
moving the count increment further down in the call stack and removing
the duplicate increment on delayed entries. Note that we have to keep
a per-entry reference for the counter (both on parallel checkout and
delayed checkout) because not all entries are always accumulated at the
same counter. See checkout_worktree(), at builtin/checkout.c for an
example.
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git remote show [-n] frotz" now pays attention to negative
pathspec.
* jk/remote-show-with-negative-refspecs:
remote: handle negative refspecs in git remote show
"git mktree --missing" lazily fetched objects that are missing from
the local object store, which was totally unnecessary for the purpose
of creating the tree object(s) from its input.
* ro/mktree-allow-missing-fix:
mktree: do not check type of remote objects
* ds/branch-checked-out:
branch: drop unused worktrees variable
fetch: stop passing around unused worktrees variable
branch: fix branch_checked_out() leaks
branch: use branch_checked_out() when deleting refs
fetch: use new branch_checked_out() and add tests
branch: check for bisects and rebases
branch: add branch_checked_out() helper
Commit 0139c58ab9 (revisions API users: add "goto cleanup" for
release_revisions(), 2022-04-13) converted an early return in
cmd_diff_files() into a goto. But it put the cleanup label too early: if
read_cache_preload() returns an error, we'll set result to "-1", but
then jump to calling run_diff_files(), overwriting our result.
We should jump past the call to run_diff_files(). Likewise, we should go
past diff_result_code(), which is expecting to see a code from an actual
diff, not a negative error code.
In practice, I suspect this bug cannot actually be triggered, because
read_cache_preload() does not seem to ever return an error. Its return
value (eventually) comes from do_read_index(), which gives the number of
cache entries found, and calls die() on error. Still, it makes sense to
fix the inadvertent change from 0139c58ab9 first, and we can look into
the overall error handling of read_cache() separately (which is present
in many other callsites).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we found an invalid object recorded in the resolve-undo data,
we would have ended up dereferencing NULL while fsck. Reporting the
problem and going on to the next object is the right thing to do
here.
Noticed by SZEDER Gábor.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Introduce a helper to see if a branch is already being worked on
(hence should not be newly checked out in a working tree), which
performs much better than the existing find_shared_symref() to
replace many uses of the latter.
* ds/branch-checked-out:
branch: drop unused worktrees variable
fetch: stop passing around unused worktrees variable
branch: fix branch_checked_out() leaks
branch: use branch_checked_out() when deleting refs
fetch: use new branch_checked_out() and add tests
branch: check for bisects and rebases
branch: add branch_checked_out() helper
Prior to 4f37d45706 (clone: respect remote unborn HEAD, 2021-02-05),
creation of the local HEAD was always done in update_head(). That commit
added code to handle an unborn head in an empty repository, and just did
all symref creation and config setup there.
This makes the code flow a little bit confusing, especially as new
corner cases have been covered (like the previous commit to match our
default branch name to a non-HEAD remote branch).
Let's move the creation of the unborn symref into update_head(). This
matches the other HEAD-creation cases, and now the logic is consistently
separated: the main cmd_clone() function only examines the situation and
sets variables based on what it finds, and update_head() actually
performs the update.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Although parse_options() can handle unknown --options just fine, none
of 'git multi-pack-index's subcommands rely on it, but do it on their
own: they invoke parse_options() with the PARSE_OPT_KEEP_UNKNOWN flag,
then check whether there are any unparsed arguments left, and print
usage and quit if necessary.
Drop that PARSE_OPT_KEEP_UNKNOWN flag to let parse_options() handle
unknown options instead, which has the additional benefit that it
prints not only the usage but an "error: unknown option `foo'" message
as well.
Do leave the unparsed arguments check to catch any unexpected
non-option arguments, though, e.g. 'git multi-pack-index write foo'.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The variables 'source', 'destination', and 'submodule_gitfile' are
all of type "const char **", and an element of such an array is of
"type const char *", but these memmove() calls were written as if
these variables are of type "char **".
Once these memmove() calls are fixed to use the correct type to
compute the number of bytes to be moved, e.g.
- memmove(source + i, source + i + 1, n * sizeof(char *));
+ memmove(source + i, source + i + 1, n * sizeof(const char *));
existing contrib/coccinelle/array.cocci rules can recognize them as
candidates for turning into MOVE_ARRAY().
While at it, use CALLOC_ARRAY() instead of xcalloc() to allocate the
modes[] array that is involved in the change.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Usually clone tries to use the same local HEAD as the remote (unless the
user has given --branch explicitly). Even if the remote HEAD is detached
or unborn, we can detect those situations with modern versions of Git.
If the remote is too old to support the "unborn" extension (or it has
been disabled via config), then we can't know the name of the remote's
unborn HEAD, and we fall back whatever the local default branch name is
configured to be.
But that leads to one weird corner case. It's rare because it needs a
number of factors:
- the remote has an unborn HEAD
- the remote is too old to support "unborn", or has disabled it
- the remote has another branch "foo"
- the local default branch name is "foo"
In that case you end up with a local clone on an unborn "foo" branch,
disconnected completely from the remote's "foo". This is rare in
practice, but the result is quite confusing.
When choosing "foo", we can double check whether the remote has such a
name, and if so, start our local "foo" at the same spot, rather than
making it unborn.
Note that this causes a test failure in t5605, which is cloning from a
bundle that doesn't contain HEAD (so it behaves like a remote that
doesn't support "unborn"), but has a single "main" branch. That test
expects that we end up in the weird "unborn main" case, where we don't
actually check out the remote branch of the same name. Even though we
have to update the test, this seems like an argument in favor of this
patch: checking out main is what I'd expect from such a bundle.
So this patch updates the test for the new behavior and adds an adjacent
one that checks what the original was going for: if there's no HEAD and
the bundle _doesn't_ have a branch that matches our local default name,
then we end up with nothing checked out.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Unless "--branch" was given, clone generally tries to match the local
HEAD to the remote one. For most repositories, this is easy: the remote
tells us which branch HEAD was pointing to, and we call our local
checkout() function on that branch.
When cloning an empty repository, it's a little more tricky: we have
special code that checks the transport's "unborn" extension, or falls back
to our local idea of what the default branch should be. In either case,
we point the new HEAD to that, and set up the branch.* config.
But that leaves one case unhandled: when the remote repository _isn't_
empty, but its HEAD is unborn. The checkout() function is smart enough
to realize we didn't fetch the remote HEAD and it bails with a warning.
But we'll have ignored any information the remote gave us via the unborn
extension. This leads to nonsense outcomes:
- If the remote has its HEAD pointing to an unborn "foo" and contains
another branch "bar", cloning will get branch "bar" but leave the
local HEAD pointing at "master" (or whatever our local default is),
which is useless. The project does not use "master" as a branch.
- Worse, if the other branch "bar" is instead called "master" (but
again, the remote HEAD is not pointing to it), then we end up with a
local unborn branch "master", which is not connected to the remote
"master" (it shares no history, and there's no branch.* config).
Instead, we should try to use the remote's HEAD, even if its unborn, to
be consistent with the other cases.
The reason this case was missed is that cmd_clone() handles empty and
non-empty repositories on two different sides of a conditional:
if (we have any refs) {
fetch refs;
check for --branch;
otherwise, try to point our head at remote head;
otherwise, our head is NULL;
} else {
check for --branch;
otherwise, try to use "unborn" extension;
otherwise, fall back to our default name name;
}
So the smallest change would be to repeat the "unborn" logic at the end
of the first block. But we can note some other overlaps and
inconsistencies:
- both sides have to handle --branch (though note that it's always an
error for the empty repo case, since an empty repo by definition
does not have a matching branch)
- the fall back to the default name is much more explicit in the
empty-repo case. The non-empty case eventually ends up bailing
from checkout() with a warning, which produces a similar result, but
fails to set up the branch config we do in the empty case.
So let's pull the HEAD setup out of this conditional entirely. This
de-duplicates some of the code and the result is easy to follow, because
helper functions like find_ref_by_name() do the right thing even in the
empty-repo case (i.e., by returning NULL).
There are two subtleties:
- for a remote with a detached HEAD, it will advertise an oid for HEAD
(which we store in our "remote_head" variable), but we won't find a
matching refname (so our "remote_head_points_at" is NULL). In this
case we make a local detached HEAD to match. Right now this happens
implicitly by reaching update_head() with a non-NULL remote_head
(since we skip all of the unborn-fallback). We'll now need to
account for it explicitly before doing the fallback.
- for an empty repo, we issue a warning to the user that they've
cloned an empty repo. The text of that warning doesn't make sense
for a non-empty repo with an unborn HEAD, so we'll have to
differentiate the two cases there. We could just use different text,
but instead let's allow the code to continue down to checkout(),
which will issue an appropriate warning, like:
remote HEAD refers to nonexistent ref, unable to checkout
Continuing down to checkout() will make it easier to do more fixes
on top (see below).
Note that this patch fixes the case where the other side reports an
unborn head to us using the protocol extension. It _doesn't_ fix the
case where the other side doesn't tell us, we locally guess "master",
and the other side happens to have a "master" which its HEAD doesn't
point. But it doesn't make anything worse there, and it should actually
make it easier to fix that problem on top.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We don't need to put a "\n" in calls to warning(), since it adds one
itself (and the user sees an extra blank line). Drop it, and while we're
here, drop the full-stop from the message, which goes against our
guidelines.
This bug dates all the way back to 8434c2f1af (Build in clone,
2008-04-27), but presumably nobody noticed because it's hard to trigger:
you have to clone a repository whose HEAD is unborn, but which is not
otherwise empty.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Generalize the newly added "unused.cocci" rule to find more than just
"struct strbuf", let's have it find the same unused patterns for
"struct string_list", as well as other code that uses
similar-looking *_{release,clear,free}() and {release,clear,free}_*()
functions.
We're intentionally loose in accepting e.g. a "strbuf_init(&sb)"
followed by a "string_list_clear(&sb, 0)". It's assumed that the
compiler will catch any such invalid code, i.e. that our
constructors/destructors don't take a "void *".
See [1] for example of code that would be covered by the
"get_worktrees()" part of this rule. We'd still need work that the
series is based on (we were passing "worktrees" to a function), but
could now do the change in [1] automatically.
1. https://lore.kernel.org/git/Yq6eJFUPPTv%2Fzc0o@coredump.intra.peff.net/
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a coccinelle rule to remove "struct strbuf" initialization
followed by calling "strbuf_release()" function, without any uses of
the strbuf in the same function.
See the tests in contrib/coccinelle/tests/unused.{c,res} for what it's
intended to find and replace.
The inclusion of "contrib/scalar/scalar.c" is because "spatch" was
manually run on it (we don't usually run spatch on contrib).
Per the "buggy code" comment we also match a strbuf_init() before the
xmalloc(), but we're not seeking to be so strict as to make checks
that the compiler will catch for us redundant. Saying we'll match
either "init" or "xmalloc" lines makes the rule simpler.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Originally, moving a <source> directory which is not on-disk due
to its existence outside of sparse-checkout cone, "giv mv" command
errors out with "bad source".
Add a helper check_dir_in_index() function to see if a directory
name exists in the index. Also add a SKIP_WORKTREE_DIR bit to mark
such directories.
Change the checking logic, so that such <source> directory makes
"giv mv" command warns with "advise_on_updating_sparse_paths()"
instead of "bad source"; also user now can supply a "--sparse" flag so
this operation can be carried out successfully.
Helped-by: Victoria Dye <vdye@github.com>
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As suggested by Derrick [1], move the in-line definition of
"enum update_mode" to the top of the file and make it use "flags"
mode (each state is a different bit in the word).
Change the flag assignments from '=' (single assignment) to '|='
(additive). Also change flag evaluation from '==' to '&', etc.
[1] https://lore.kernel.org/git/22aadea2-9330-aa9e-7b6a-834585189144@github.com/
Helped-by: Victoria Dye <vdye@github.com>
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Originally, moving a sparse file into cone can result in unwarned
overwrite of existing entry. The expected behavior is that if the
<destination> exists in the entry, user should be prompted to supply
a [-f|--force] to carry out the operation, or the operation should
fail.
Add a check mechanism to do that.
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Originally, moving a <source> file which is not on-disk but exists in
index as a SKIP_WORKTREE enabled cache entry, "giv mv" command errors
out with "bad source".
Change the checking logic, so that such <source>
file makes "giv mv" command warns with "advise_on_updating_sparse_paths()"
instead of "bad source"; also user now can supply a "--sparse" flag so
this operation can be carried out successfully.
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Previous if/else-if chain are highly nested and hard to develop/extend.
Refactor to decouple this if/else-if chain by using goto to jump ahead.
Suggested-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Originally, "git mv" a sparse file from out-of-cone to
in-cone does not update the moved file's sparsity (remove its
SKIP_WORKTREE bit). And the corresponding cache entry is, unexpectedly,
not checked out in the working tree.
Update the behavior so that:
1. Moving from out-of-cone to in-cone removes the SKIP_WORKTREE bit from
corresponding cache entry.
2. The moved cache entry is checked out in the working tree to reflect
the updated sparsity.
Helped-by: Victoria Dye <vdye@github.com>
Signed-off-by: Shaoxuan Yuan <shaoxuan.yuan02@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a memory leak introduced in 44c175c7a4 (pull: error on no merge
candidates, 2015-06-18). As a result we can mark several tests as
passing with SANITIZE=leak using "TEST_PASSES_SANITIZE_LEAK=true".
Removing the "int ret = 0" assignment added here in a6d7eb2c7a (pull:
optionally rebase submodules (remote submodule changes only),
2017-06-23) is not a logic error, it could always have been left
uninitialized (as "int ret"), now that we'll use the "ret" from the
upper scope we can drop the assignment in the "opt_rebase" branch.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a memory leak where "cat-file" will leak the "path" member. See
e5fba602e5 (textconv: support for cat_file, 2010-06-15) for the code
that introduced the offending get_oid_with_context() call (called
get_sha1_with_context() at the time).
As a result we can mark several tests as passing with SANITIZE=leak
using "TEST_PASSES_SANITIZE_LEAK=true".
As noted in dc944b65f1 (get_sha1_with_context: dynamically allocate
oc->path, 2017-05-19) callers must free the "path" member. That same
commit added the relevant free() to this function, but we weren't
catching cases where we'd return early.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a memory leak in code added in 41abfe15d9 (maintenance: add
pack-refs task, 2021-02-09), we need to call strvec_clear() on the
"struct strvec" that we initialized.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 1c41d2805e (unpack_trees_options: free messages when done,
2018-05-21) we started calling clear_unpack_trees_porcelain() on this
codepath, but missed this error path.
We could call clear_unpack_trees_porcelain() just before we error()
and return when unmerged_cache() fails, but the more correct fix is to
not have the unmerged_cache() check happen in the middle of our
"topts" setup.
Before 23cbf11b5c (merge-recursive: porcelain messages for checkout,
2010-08-11) we would not malloc() to setup our "topts", which is when
this started to leak on the error path.
Before that this code wasn't conflating the setup of "topts" and the
unmerged_cache() call in any meaningful way. The initial version in
782c2d65c2 (Build in checkout, 2008-02-07) just does a "memset" of
it, and initializes a single struct member.
Then in 8ccba008ee (unpack-trees: allow Porcelain to give different
error messages, 2008-05-17) we added the initialization of the error
message, which as noted above finally started leaking in 23cbf11b5c.
Let's fix the memory leak, and avoid future issues by initializing the
"topts" with a helper function. There are no functional changes here.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a memory leak in "merge-file", we need to loop over the "mmfs"
array and free() what we've got so far when we error out. As a result
we can mark a test as passing with SANITIZE=leak using
"TEST_PASSES_SANITIZE_LEAK=true".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Refactor the code in builtin/merge-file.c to:
* Use the initializer to zero out "mmfs", and use modern C syntax for
the rest.
* Refactor the the inner loop to use a variable and "if/else if"
pattern followed by "return". This will make a change to change it to
a "goto cleanup" pattern smaller.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a memory leak introduced in 440c705ea6 (cat-file: add
--batch-command mode, 2022-02-18). The free_cmds() function was only
called on "queued_nr" if we had a "flush" command. As the "without
flush for blob info" test added in the same commit shows we can't rely
on that, so let's call free_cmds() again at the end.
Since "nr" follows the usual pattern of being set to 0 if we've
free()'d the memory already it's OK to call it twice, even in cases
where we are doing a "flush".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Call the release_revisions() function added in
1878b5edc0 (revision.[ch]: provide and start using a
release_revisions(), 2022-04-13) in cmd_revert(), as well as freeing
the xmalloc()'d "revs" member itself.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a memory leak added in 0ec4b1650c (clone: fix ref selection in
--single-branch --branch=xxx, 2012-06-22).
Whether we get our "remote_head" from copy_ref() directly, or with a
call to guess_remote_head() it'll be the result of a copy_ref() in
either case, as guess_remote_head() is a wrapper for copy_ref() (or it
returns NULL).
We can't mark any tests passing passing with SANITIZE=leak using
"TEST_PASSES_SANITIZE_LEAK=true" as a result of this change, but
e.g. "t/t1500-rev-parse.sh" now gets closer to passing.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a memory leak in "git check-ref-format" that's been present in the
code in one form or another since 38eedc634b (git check-ref-format
--print, 2009-10-12), the code got substantially refactored in
cfbe22f03f (check-ref-format: handle subcommands in separate
functions, 2010-08-05).
As a result we can mark a test as passing with SANITIZE=leak using
"TEST_PASSES_SANITIZE_LEAK=true".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
All invocations of do_get_submodule_displaypath() pass
get_super_prefix() as the super_prefix arg, which is exactly the same
as get_submodule_displaypath().
Replace all calls to do_get_submodule_displaypath() with
get_submodule_displaypath(), and since it has no more callers, remove
it.
Signed-off-by: Glen Choo <chooglen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Unlike the other subcommands, "git submodule--helper update" uses the
"--recursive-prefix" flag instead of "--super-prefix". The two flags are
otherwise identical (they only serve to compute the 'display path' of a
submodule), except that there is a dedicated helper function to get the
value of "--super-prefix".
This inconsistency exists because "git submodule update" used to pass
"--recursive-prefix" between shell and C (introduced in [1]) before
"--super-prefix" was introduced (in [2]), and for simplicity, we kept
this name when "git submodule--helper update" was created.
Remove "--recursive-prefix" and its associated code from "git
submodule--helper update", replacing it with "--super-prefix".
To use "--super-prefix", module_update is marked with
SUPPORT_SUPER_PREFIX. Note that module_clone must also be marked with
SUPPORT_SUPER_PREFIX, otherwise the "git submodule--helper clone"
subprocess will fail check because "--super-prefix" is propagated via
the environment.
[1] 48308681b0 (git submodule update: have a dedicated helper for
cloning, 2016-02-29)
[2] 74866d7579 (git: make super-prefix option, 2016-10-07)
Signed-off-by: Glen Choo <chooglen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove the SUPPORT_SUPER_PREFIX flag from "add", "init" and
"summary". For the "add" command it hasn't been used since [1],
likewise for "init" and "summary" since [2] and [3], respectively.
As implemented in 74866d7579 (git: make super-prefix option,
2016-10-07) the SUPPORT_SUPER_PREFIX flag in git.c applies for the
entire command, but as implemented in 89c8626557 (submodule helper:
support super prefix, 2016-12-08) we assert here in
cmd_submodule__helper() that we're not getting the flag unexpectedly.
1. 8c8195e9c3 (submodule--helper: introduce add-clone subcommand,
2021-07-10)
2. 6e7c14e65c (submodule update --init: display correct path from
submodule, 2017-01-06)
3. 1cf823d8f0 (submodule: remove unnecessary `prefix` based option
logic, 2021-06-22)
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Glen Choo <chooglen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Replace a chunk of code in update_submodule() with an equivalent
do_get_submodule_displaypath() invocation. This is already tested by
t/t7406-submodule-update.sh:'submodule update --init --recursive from
subdirectory', so no tests are added.
The two are equivalent because:
- Exactly one of recursive_prefix|prefix is non-NULL at a time; prefix
is set at the superproject level, and recursive_prefix is set when
recursing into submodules. There is also a BUG() statement in
get_submodule_displaypath() that asserts that both cannot be non-NULL.
- In get_submodule_displaypath(), get_super_prefix() always returns NULL
because "--super-prefix" is never passed. Thus calling it is
equivalent to calling do_get_submodule_displaypath() with super_prefix
= NULL.
Therefore:
- When recursive_prefix is non-NULL, prefix is NULL, and thus
get_submodule_displaypath() just returns prefixed_path. This is
identical to calling do_get_submodule_displaypath() with super_prefix
= recursive_prefix because the return value is still the concatenation
of recursive_prefix + update_data->sm_path.
- When prefix is non-NULL, prefixed_path = update_data->sm_path. Thus
calling get_submodule_displaypath() with prefixed_path is equivalent
to calling do_get_submodule_displaypath() with update_data->sm_path
Signed-off-by: Glen Choo <chooglen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
update_submodule() uses duplicated code to compute
update_data->displaypath and next.recursive_prefix. The latter is just
the former with "/" appended to it, and since update_data->displaypath
not changed outside of this statement, we can just reuse the already
computed result.
We can go one step further and remove the reference to
next.recursive_prefix altogether. Since it is only used in
update_data_to_args() (to compute the "--recursive-prefix" flag for the
recursive update child process) we can just use the already computed
.displaypath value of there.
Delete the duplicated code, and remove the unnecessary reference to
next.recursive_prefix. As a bonus, this fixes a memory leak where
prefixed_path was never freed (this leak was first reported in [1]).
[1] https://lore.kernel.org/git/877a45867ae368bf9e053caedcb6cf421e02344d.1655336146.git.gitgitgadget@gmail.com
Signed-off-by: Glen Choo <chooglen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are two locations in prepare_to_clone_next_submodule() that
manually calculate the submodule display path, but should just use
do_get_submodule_displaypath() for consistency.
Do this replacement and reorder the code slightly to avoid computing
the display path twice.
Until the preceding commit this code had never been tested, with our
newly added tests we can see that both these sites have been computing
the display path incorrectly ever since they were introduced in
48308681b0 (git submodule update: have a dedicated helper for cloning,
2016-02-29) [1]:
- The first hunk puts a "/" between recursive_prefix and ce->name, but
recursive_prefix already ends with "/".
- The second hunk calls relative_path() on recursive_prefix and
ce->name, but relative_path() only makes sense when both paths share
the same base directory. This is never the case here:
- recursive_prefix is the path from the topmost superproject to the
current submodule
- ce->name is the path from the root of the current submodule to its
submodule.
so, e.g. recursive_prefix="super" and ce->name="submodule" produces
displayname="../super" instead of "super/submodule".
[1] I verified this by applying the tests to 48308681b0.
Signed-off-by: Glen Choo <chooglen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Follow-up on the preceding commit which taught "git submodule--helper
update" to understand "--merge", "--checkout" and "--rebase" and use
those options instead of "--update=(rebase|merge|checkout|none)" when
the command invokes itself.
Unlike the preceding change this isn't strictly necessary to
eventually change "git-submodule.sh" so that it invokes "git
submodule--helper update" directly, but let's remove this
inconsistency in the command-line interface. We shouldn't need to
carry special synonyms for existing options in "git submodule--helper"
when that command can use the primary documented names instead.
But, as seen in the post-image this makes the control flow within
"builtin/submodule--helper.c" simpler, we can now write directly to
the "update_default" member of "struct update_data" when parsing the
options in "module_update()".
Signed-off-by: Glen Choo <chooglen@google.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Understand --checkout, --merge and --rebase synonyms for
--update={checkout,merge,rebase}, as well as the short options that
'git submodule' itself understands.
This removes a difference between the CLI API of "git submodule" and
"git submodule--helper", making it easier to make the latter an alias
for the former. See 48308681b0 (git submodule update: have a
dedicated helper for cloning, 2016-02-29) for the initial addition of
--update.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change the user-facing "git submodule--helper" commands so that
they'll report their name as being "git submodule". To a user these
commands are internal implementation details, and it doesn't make
sense to emit usage about an internal helper when "git submodule" is
invoked with invalid options.
Before this we'd emit e.g.:
$ git submodule absorbgitdirs --blah
error: unknown option `blah'
usage: git submodule--helper absorbgitdirs [<options>] [<path>...]
[...]
And:
$ git submodule set-url -- --
usage: git submodule--helper set-url [--quiet] <path> <newurl>
[...]
Now we'll start with "usage: git submodule [...]" in both of those
cases. This change does not alter the "list", "name", "clone",
"config" and "create-branch" commands, those are internal-only (as an
aside; their usage info should probably invoke BUG(...)). This only
changes the user-facing commands.
The "status", "deinit" and "update" commands are not included in this
change, because their usage information already used "submodule"
rather than "submodule--helper".
I don't think it's currently possible to emit some of this usage
information in practice, as git-submodule.sh will catch unknown
options, and e.g. it doesn't seem to be possible to get "add" to emit
its usage information from "submodule--helper".
Though that change may be superfluous now, it's also harmless, and
will allow us to eventually dispatch further into "git
submodule--helper" from git-submodule.sh, while emitting the correct
usage output.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rename the "absorb-git-dirs" subcommand to "absorbgitdirs", which is
what the "git submodule" command itself has called it since the
subcommand was implemented in f6f8586140 (submodule: add
absorb-git-dir function, 2016-12-12).
Having these two be different will make it more tedious to dispatch to
eventually dispatch "git submodule--helper" directly, as we'd need to
retain this name mapping. So let's get rid of this needless
inconsistency.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Adjust code added in 0060fd1511 (clone --recurse-submodules: prevent
name squatting on Windows, 2019-09-12) to have the internal
--require-init option imply --init, rather than having
"git-submodule.sh" add it implicitly.
This change doesn't make any difference now, but eliminates another
special-case where "git submodule--helper update"'s behavior was
different from "git submodule update". This will make it easier to
eventually replace the cmd_update() function in git-submodule.sh.
We'll still need to keep the distinction between "--init" and
"--require-init" in git-submodule.sh. Once cmd_update() gets
re-implemented in C we'll be able to change variables and other code
related to that, but not yet.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Folks may want to merge histories that have no common ancestry; provide
a flag with the same name as used by `git merge` to allow this.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Much as `git ls-files` has a -z option, let's add one to merge-tree so
that the conflict-info section can be NUL terminated (and avoid quoting
of unusual filenames).
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
With the new `detailed` parameter, a new mode can be triggered when
displaying the merge messages: The `detailed` mode prints NUL-delimited
fields of the following form:
<path-count> NUL <path>... NUL <conflict-type> NUL <message>
The `<path-count>` field determines how many `<path>` fields there are.
The intention of this mode is to support server-side operations, where
worktree-less merges can lead to conflicts and depending on the type
and/or path count, the caller might know how to handle said conflict.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Much like `git merge` updates the index with information of the form
(mode, oid, stage, name)
provide this output for conflicted files for merge-tree as well.
Provide a --name-only option for users to exclude the mode, oid, and
stage and only get the list of conflicted filenames.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Callers of `git merge-tree --write-tree` will often want to know which
files had conflicts. While they could potentially attempt to parse the
CONFLICT notices printed, those messages are not meant to be machine
readable. Provide a simpler mechanism of just printing the files (in
the same format as `git ls-files` with quoting, but restricted to
unmerged files) in the output before the free-form messages.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When running `git merge-tree --write-tree`, we previously would only
return an exit status reflecting the cleanness of a merge, and print out
the toplevel tree of the resulting merge. Merges also have
informational messages, such as:
* "Auto-merging <PATH>"
* "CONFLICT (content): ..."
* "CONFLICT (file/directory)"
* etc.
In fact, when non-content conflicts occur (such as file/directory,
modify/delete, add/add with differing modes, rename/rename (1to2),
etc.), these informational messages may be the only notification the
user gets since these conflicts are not representable in the contents
of the file.
Add a --[no-]messages option so that callers can request these messages
be included at the end of the output. Include such messages by default
when there are conflicts, and omit them by default when the merge is
clean.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This adds the ability to perform real merges rather than just trivial
merges (meaning handling three way content merges, recursive ancestor
consolidation, renames, proper directory/file conflict handling, and so
forth). However, unlike `git merge`, the working tree and index are
left alone and no branch is updated.
The only output is:
- the toplevel resulting tree printed on stdout
- exit status of 0 (clean), 1 (conflicts present), anything else
(merge could not be performed; unknown if clean or conflicted)
This output is meant to be used by some higher level script, perhaps in
a sequence of steps like this:
NEWTREE=$(git merge-tree --write-tree $BRANCH1 $BRANCH2)
test $? -eq 0 || die "There were conflicts..."
NEWCOMMIT=$(git commit-tree $NEWTREE -p $BRANCH1 -p $BRANCH2)
git update-ref $BRANCH1 $NEWCOMMIT
Note that higher level scripts may also want to access the
conflict/warning messages normally output during a merge, or have quick
access to a list of files with conflicts. That is not available in this
preliminary implementation, but subsequent commits will add that
ability (meaning that NEWTREE would be a lot more than a tree in the
case of conflicts).
This also marks the traditional trivial merge of merge-tree as
deprecated. The trivial merge not only had limited applicability, the
output format was also difficult to work with (and its format
undocumented), and will generally be less performant than real merges.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let merge-tree accept a `--write-tree` parameter for choosing real
merges instead of trivial merges, and accept an optional
`--trivial-merge` option to get the traditional behavior. Note that
these accept different numbers of arguments, though, so these names
need not actually be used.
Note that real merges differ from trivial merges in that they handle:
- three way content merges
- recursive ancestor consolidation
- renames
- proper directory/file conflict handling
- etc.
Basically all the stuff you'd expect from `git merge`, just without
updating the index and working tree. The initial shell added here does
nothing more than die with "real merges are not yet implemented", but
that will be fixed in subsequent commits.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In preparation for adding a non-trivial merge capability to merge-tree,
move the existing merge logic for trivial merges into a new function.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
merge-recursive.h defined its own merge_trees() function, different than
the one found in builtin/merge-tree.c. That was okay in the past, but
we want merge-tree to be able to use the merge-ort functions, which will
end up including merge-recursive.h. Rename the function found in
builtin/merge-tree.c to avoid the conflict.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This patch adds a command line option analogous to that of GNU
grep(1)'s -m / --max-count, which users might already be used to.
This makes it possible to limit the amount of matches shown in the
output while keeping the functionality of other options such as -C
(show code context) or -p (show containing function), which would be
difficult to do with a shell pipeline (e.g. head(1)).
Signed-off-by: Carlos López 00xc@protonmail.com
Signed-off-by: Junio C Hamano <gitster@pobox.com>
With 31c8221a (mktree: validate entry type in input, 2009-05-14), we
called the sha1_object_info() API to obtain the type information, but
allowed the call to silently fail when the object was missing locally,
so that we can sanity-check the types opportunistically when the
object did exist.
The implementation is understandable because back then there was no
lazy/on-demand downloading of individual objects from the promisor
remotes that causes a long delay and materializes the object, hence
defeating the point of using "--missing". The design is hurting us
now.
We could bypass the opportunistic type/mode consistency check
altogether when "--missing" is given, but instead, use the
oid_object_info_extended() API and tell it that we are only interested
in objects that locally exist and are immediately available by passing
OBJECT_INFO_SKIP_FETCH_OBJECT bit to it. That way, we will still
retain the cheap and opportunistic sanity check for local objects.
Signed-off-by: Richard Oliver <roliver@roku.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
After b489b9d9aa (branch: use branch_checked_out() when deleting refs,
2022-06-14), we no longer look at our local "worktrees" variable, since
branch_checked_out() handles it under the hood. The compiler didn't
notice the unused variable because we call functions to initialize and
free it (so it's not totally unused, it just doesn't do anything
useful).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 12d47e3b1f (fetch: use new branch_checked_out() and add tests,
2022-06-14), fetch's update_local_ref() function stopped using its
"worktrees" parameter. It doesn't need it, since the
branch_checked_out() function examines the global worktrees under the
hood.
So we can not only drop the unused parameter from that function, but
also from its entire call chain. And as we do so all the way up to
do_fetch(), we can see that nobody uses it at all, and we can drop the
local variable there entirely.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is an option rather than command. Make the message convey this
similar to the other messages in the file.
Signed-off-by: Alexander Shopov <ash@kambanaria.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some config variables are combinations of multiple words, and we
typically write them in camelCase forms in manpage and translatable
strings. It's not easy to find mismatches for these camelCase config
variables during code reviews, but occasionally they are identified
during localization translations.
To check for mismatched config variables, I introduced a new feature
in the helper program for localization[^1]. The following mismatched
config variables have been identified by running the helper program,
such as "git-po-helper check-pot".
Lowercase in manpage should use camelCase:
* Documentation/config/http.txt: http.pinnedpubkey
Lowercase in translable strings should use camelCase:
* builtin/fast-import.c: pack.indexversion
* builtin/gc.c: gc.logexpiry
* builtin/index-pack.c: pack.indexversion
* builtin/pack-objects.c: pack.indexversion
* builtin/repack.c: pack.writebitmaps
* commit.c: i18n.commitencoding
* gpg-interface.c: user.signingkey
* http.c: http.postbuffer
* submodule-config.c: submodule.fetchjobs
Mismatched camelCases, choose the former:
* Documentation/config/transfer.txt: transfer.credentialsInUrl
remote.c: transfer.credentialsInURL
[^1]: https://github.com/git-l10n/git-po-helper
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
By default, the git remote show command will query data from remotes to
show data about what might be done on a future git fetch. This process
currently does not handle negative refspecs. This can be confusing,
because the show command will list refs as if they would be fetched. For
example if the fetch refspec "^refs/heads/pr/*", it still displays the
following:
* remote jdk19
Fetch URL: git@github.com:openjdk/jdk19.git
Push URL: git@github.com:openjdk/jdk19.git
HEAD branch: master
Remote branches:
master tracked
pr/1 new (next fetch will store in remotes/jdk19)
pr/2 new (next fetch will store in remotes/jdk19)
pr/3 new (next fetch will store in remotes/jdk19)
Local ref configured for 'git push':
master pushes to master (fast-forwardable)
Fix this by adding an additional check inside of get_ref_states. If a
ref matches one of the negative refspecs, mark it as skipped instead of
marking it as new or tracked.
With this change, we now report remote branches that are skipped due to
negative refspecs properly:
* remote jdk19
Fetch URL: git@github.com:openjdk/jdk19.git
Push URL: git@github.com:openjdk/jdk19.git
HEAD branch: master
Remote branches:
master tracked
pr/1 skipped
pr/2 skipped
pr/3 skipped
Local ref configured for 'git push':
master pushes to master (fast-forwardable)
By showing the refs as skipped, it helps clarify that these references
won't actually be fetched.
This does not properly handle refs going stale due to a newly added
negative refspec. In addition, git remote prune doesn't handle that
negative refspec case either. Fixing that requires digging into
get_stale_heads and handling the case of a ref which exists on the
remote but is omitted due to a negative refspec locally.
Add a new test case which covers the functionality above, as well as a
new expected failure indicating the poor overlap with stale refs.
Reported-by: Pavel Rappo <pavel.rappo@gmail.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In c51f8f94e5 (submodule--helper: run update procedures from C,
2021-08-24), we added code that first obtains the default remote, and
then adds that to a `strvec`.
However, we never released the default remote's memory.
Reported by Coverity.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Various error messages that talk about the removal of
"--preserve-merges" in "rebase" have been strengthened, and "rebase
--abort" learned to get out of a state that was left by an earlier
use of the option.
* po/rebase-preserve-merges:
rebase: translate a die(preserve-merges) message
rebase: note `preserve` merges may be a pull config option
rebase: help users when dying with `preserve-merges`
rebase.c: state preserve-merges has been removed
"git revert" learns "--reference" option to use more human-readable
reference to the commit it reverts in the message template it
prepares for the user.
* jc/revert-show-parent-info:
revert: --reference should apply only to 'revert', not 'cherry-pick'
revert: optionally refer to commit in the "reference" format
This was found during l10n process by Jiang Xin.
Reported-by: Jiang Xin <worldhello.net@gmail.com>
Signed-off-by: Fangyi Zhou <me@fangyi.io>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is the last current use of find_shared_symref() that can easily be
replaced by branch_checked_out(). The benefit of this switch is that the
code is a bit simpler, but also it is faster on repeated calls.
The remaining uses of find_shared_symref() are non-trivial to remove, so
we probably should not continue in that direction:
* builtin/notes.c uses find_shared_symref() with "NOTES_MERGE_REF"
instead of "HEAD", so it doesn't have an immediate analogue with
branch_checked_out(). Perhaps we should consider extending it to
include that symref in addition to HEAD, BISECT_HEAD, and
REBASE_HEAD.
* receive-pack.c checks to see if a worktree has a checkout for the ref
that is being updated. The tricky part is that it can actually decide
to update the worktree directly instead of just skipping the update.
This all depends on the receive.denyCurrentBranch config option. The
implementation currenty cares about receiving the worktree in the
result, so the current branch_checked_out() prototype is insufficient
currently. This is something to investigate later, though, since a
large number of refs could be updated at the same time and using the
strmap implementation of branch_checked_out() could be beneficial.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When fetching refs from a remote, it is possible that the refspec will
cause use to overwrite a ref that is checked out in a worktree. The
existing logic in builtin/fetch.c uses a possibly-slow mechanism. Update
those sections to use the new, more efficient branch_checked_out()
helper.
These uses were not previously tested, so add a test case that can be
used for these kinds of collisions. There is only one test now, but more
tests will be added as other consumers of branch_checked_out() are
added.
Note that there are two uses in builtin/fetch.c, but only one of the
messages is tested. This is because the tested check is run before
completing the fetch, and the untested check is not reachable without
concurrent updates to the filesystem. Thus, it is beneficial to keep
that extra check for the sake of defense-in-depth. However, we should
not attempt to test the check, as the effort required is too
complicated to be worth the effort. This use in update_local_ref()
also requires a change in the error message because we no longer have
access to the worktree struct, only the path of the worktree. This error
is so rare that making a distinction between the two is not critical.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git show-ref --heads" (and "--tags") still iterated over all the
refs only to discard refs outside the specified area, which has
been corrected.
* tb/show-ref-optim:
builtin/show-ref.c: avoid over-iterating with --heads, --tags
Make use of the stream_loose_object() function introduced in the
preceding commit to unpack large objects. Before this we'd need to
malloc() the size of the blob before unpacking it, which could cause
OOM with very large blobs.
We could use the new streaming interface to unpack all blobs, but
doing so would be much slower, as demonstrated e.g. with this
benchmark using git-hyperfine[0]:
rm -rf /tmp/scalar.git &&
git clone --bare https://github.com/Microsoft/scalar.git /tmp/scalar.git &&
mv /tmp/scalar.git/objects/pack/*.pack /tmp/scalar.git/my.pack &&
git hyperfine \
-r 2 --warmup 1 \
-L rev origin/master,HEAD -L v "10,512,1k,1m" \
-s 'make' \
-p 'git init --bare dest.git' \
-c 'rm -rf dest.git' \
'./git -C dest.git -c core.bigFileThreshold={v} unpack-objects </tmp/scalar.git/my.pack'
Here we'll perform worse with lower core.bigFileThreshold settings
with this change in terms of speed, but we're getting lower memory use
in return:
Summary
'./git -C dest.git -c core.bigFileThreshold=10 unpack-objects </tmp/scalar.git/my.pack' in 'origin/master' ran
1.01 ± 0.01 times faster than './git -C dest.git -c core.bigFileThreshold=1k unpack-objects </tmp/scalar.git/my.pack' in 'origin/master'
1.01 ± 0.01 times faster than './git -C dest.git -c core.bigFileThreshold=1m unpack-objects </tmp/scalar.git/my.pack' in 'origin/master'
1.01 ± 0.02 times faster than './git -C dest.git -c core.bigFileThreshold=1m unpack-objects </tmp/scalar.git/my.pack' in 'HEAD'
1.02 ± 0.00 times faster than './git -C dest.git -c core.bigFileThreshold=512 unpack-objects </tmp/scalar.git/my.pack' in 'origin/master'
1.09 ± 0.01 times faster than './git -C dest.git -c core.bigFileThreshold=1k unpack-objects </tmp/scalar.git/my.pack' in 'HEAD'
1.10 ± 0.00 times faster than './git -C dest.git -c core.bigFileThreshold=512 unpack-objects </tmp/scalar.git/my.pack' in 'HEAD'
1.11 ± 0.00 times faster than './git -C dest.git -c core.bigFileThreshold=10 unpack-objects </tmp/scalar.git/my.pack' in 'HEAD'
A better benchmark to demonstrate the benefits of that this one, which
creates an artificial repo with a 1, 25, 50, 75 and 100MB blob:
rm -rf /tmp/repo &&
git init /tmp/repo &&
(
cd /tmp/repo &&
for i in 1 25 50 75 100
do
dd if=/dev/urandom of=blob.$i count=$(($i*1024)) bs=1024
done &&
git add blob.* &&
git commit -mblobs &&
git gc &&
PACK=$(echo .git/objects/pack/pack-*.pack) &&
cp "$PACK" my.pack
) &&
git hyperfine \
--show-output \
-L rev origin/master,HEAD -L v "512,50m,100m" \
-s 'make' \
-p 'git init --bare dest.git' \
-c 'rm -rf dest.git' \
'/usr/bin/time -v ./git -C dest.git -c core.bigFileThreshold={v} unpack-objects </tmp/repo/my.pack 2>&1 | grep Maximum'
Using this test we'll always use >100MB of memory on
origin/master (around ~105MB), but max out at e.g. ~55MB if we set
core.bigFileThreshold=50m.
The relevant "Maximum resident set size" lines were manually added
below the relevant benchmark:
'/usr/bin/time -v ./git -C dest.git -c core.bigFileThreshold=50m unpack-objects </tmp/repo/my.pack 2>&1 | grep Maximum' in 'origin/master' ran
Maximum resident set size (kbytes): 107080
1.02 ± 0.78 times faster than '/usr/bin/time -v ./git -C dest.git -c core.bigFileThreshold=512 unpack-objects </tmp/repo/my.pack 2>&1 | grep Maximum' in 'origin/master'
Maximum resident set size (kbytes): 106968
1.09 ± 0.79 times faster than '/usr/bin/time -v ./git -C dest.git -c core.bigFileThreshold=100m unpack-objects </tmp/repo/my.pack 2>&1 | grep Maximum' in 'origin/master'
Maximum resident set size (kbytes): 107032
1.42 ± 1.07 times faster than '/usr/bin/time -v ./git -C dest.git -c core.bigFileThreshold=100m unpack-objects </tmp/repo/my.pack 2>&1 | grep Maximum' in 'HEAD'
Maximum resident set size (kbytes): 107072
1.83 ± 1.02 times faster than '/usr/bin/time -v ./git -C dest.git -c core.bigFileThreshold=50m unpack-objects </tmp/repo/my.pack 2>&1 | grep Maximum' in 'HEAD'
Maximum resident set size (kbytes): 55704
2.16 ± 1.19 times faster than '/usr/bin/time -v ./git -C dest.git -c core.bigFileThreshold=512 unpack-objects </tmp/repo/my.pack 2>&1 | grep Maximum' in 'HEAD'
Maximum resident set size (kbytes): 4564
This shows that if you have enough memory this new streaming method is
slower the lower you set the streaming threshold, but the benefit is
more bounded memory use.
An earlier version of this patch introduced a new
"core.bigFileStreamingThreshold" instead of re-using the existing
"core.bigFileThreshold" variable[1]. As noted in a detailed overview
of its users in [2] using it has several different meanings.
Still, we consider it good enough to simply re-use it. While it's
possible that someone might want to e.g. consider objects "small" for
the purposes of diffing but "big" for the purposes of writing them
such use-cases are probably too obscure to worry about. We can always
split up "core.bigFileThreshold" in the future if there's a need for
that.
0. https://github.com/avar/git-hyperfine/
1. https://lore.kernel.org/git/20211210103435.83656-1-chiyutianyi@gmail.com/
2. https://lore.kernel.org/git/20220120112114.47618-5-chiyutianyi@gmail.com/
Helped-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Helped-by: Derrick Stolee <stolee@gmail.com>
Helped-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Han Xin <chiyutianyi@gmail.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As the name implies, "get_data(size)" will allocate and return a given
amount of memory. Allocating memory for a large blob object may cause the
system to run out of memory. Before preparing to replace calling of
"get_data()" to unpack large blob objects in latter commits, refactor
"get_data()" to reduce memory footprint for dry_run mode.
Because in dry_run mode, "get_data()" is only used to check the
integrity of data, and the returned buffer is not used at all, we can
allocate a smaller buffer and use it as zstream output. Make the function
return NULL in the dry-run mode, as no callers use the returned buffer.
The "find [...]objects/?? -type f | wc -l" test idiom being used here
is adapted from the same "find" use added to another test in
d9545c7f46 (fast-import: implement unpack limit, 2016-04-25).
Suggested-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Han Xin <chiyutianyi@gmail.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A new bug() and BUG_if_bug() API is introduced to make it easier to
uniformly log "detect multiple bugs and abort in the end" pattern.
* ab/bug-if-bug:
cache-tree.c: use bug() and BUG_if_bug()
receive-pack: use bug() and BUG_if_bug()
parse-options.c: use optbug() instead of BUG() "opts" check
parse-options.c: use new bug() API for optbug()
usage.c: add a non-fatal bug() function to go with BUG()
common-main.c: move non-trace2 exit() behavior out of trace2.c
More fsmonitor--daemon.
* jh/builtin-fsmonitor-part3: (30 commits)
t7527: improve implicit shutdown testing in fsmonitor--daemon
fsmonitor--daemon: allow --super-prefix argument
t7527: test Unicode NFC/NFD handling on MacOS
t/lib-unicode-nfc-nfd: helper prereqs for testing unicode nfc/nfd
t/helper/hexdump: add helper to print hexdump of stdin
fsmonitor: on macOS also emit NFC spelling for NFD pathname
t7527: test FSMonitor on case insensitive+preserving file system
fsmonitor: never set CE_FSMONITOR_VALID on submodules
t/perf/p7527: add perf test for builtin FSMonitor
t7527: FSMonitor tests for directory moves
fsmonitor: optimize processing of directory events
fsm-listen-darwin: shutdown daemon if worktree root is moved/renamed
fsm-health-win32: force shutdown daemon if worktree root moves
fsm-health-win32: add polling framework to monitor daemon health
fsmonitor--daemon: stub in health thread
fsmonitor--daemon: rename listener thread related variables
fsmonitor--daemon: prepare for adding health thread
fsmonitor--daemon: cd out of worktree root
fsm-listen-darwin: ignore FSEvents caused by xattr changes on macOS
unpack-trees: initialize fsmonitor_has_run_once in o->result
...
Rename .env_array member to .env in the child_process structure.
* ab/env-array:
run-command API users: use "env" not "env_array" in comments & names
run-command API: rename "env_array" to "env"
The resolve-undo extension was added to the index in cfc5789a
(resolve-undo: record resolved conflicts in a new index extension
section, 2009-12-25). This extension records the blob object names
and their modes of conflicted paths when the path gets resolved
(e.g. with "git add"), to allow "undoing" the resolution with
"checkout -m path". These blob objects should be guarded from
garbage-collection while we have the resolve-undo information in the
index (otherwise unresolve operation may try to use a blob object
that has already been pruned away).
But the code called from mark_reachable_objects() for the index
forgets to do so. Teach add_index_objects_to_pending() helper to
also add objects referred to by the resolve-undo extension.
Also make matching changes to "fsck", which has code that is fairly
similar to the reachability stuff, but have parallel implementations
for all these stuff, which may (or may not) someday want to be unified.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git clone --origin X" leaked piece of memory that held value read
from the clone.defaultRemoteName configuration variable, which has
been plugged.
source: <xmqqlevl4ysk.fsf@gitster.g>
* jc/clone-remote-name-leak-fix:
clone: plug a miniscule leak
The path taken by "git multi-pack-index" command from the end user
was compared with path internally prepared by the tool withut first
normalizing, which lead to duplicated paths not being noticed,
which has been corrected.
source: <pull.1221.v2.git.1650911234.gitgitgadget@gmail.com>
* ds/midx-normalize-pathname-before-comparison:
cache: use const char * for get_object_directory()
multi-pack-index: use --object-dir real path
midx: use real paths in lookup_multi_pack_index()
"git rebase --keep-base <upstream> <branch-to-rebase>" computed the
commit to rebase onto incorrectly, which has been corrected.
source: <20220421044233.894255-1-alexhenrie24@gmail.com>
* ah/rebase-keep-base-fix:
rebase: use correct base for --keep-base when a branch is given
The "--current" option of "git show-branch" should have been made
incompatible with the "--reflog" mode, but this was not enforced,
which has been corrected.
source: <xmqqh76mf7s4.fsf_-_@gitster.g>
* jc/show-branch-g-current:
show-branch: -g and --current are incompatible
Plug the memory leaks from the trickiest API of all, the revision
walker.
* ab/plug-leak-in-revisions: (27 commits)
revisions API: add a TODO for diff_free(&revs->diffopt)
revisions API: have release_revisions() release "topo_walk_info"
revisions API: have release_revisions() release "date_mode"
revisions API: call diff_free(&revs->pruning) in revisions_release()
revisions API: release "reflog_info" in release revisions()
revisions API: clear "boundary_commits" in release_revisions()
revisions API: have release_revisions() release "prune_data"
revisions API: have release_revisions() release "grep_filter"
revisions API: have release_revisions() release "filter"
revisions API: have release_revisions() release "cmdline"
revisions API: have release_revisions() release "mailmap"
revisions API: have release_revisions() release "commits"
revisions API users: use release_revisions() for "prune_data" users
revisions API users: use release_revisions() with UNLEAK()
revisions API users: use release_revisions() in builtin/log.c
revisions API users: use release_revisions() in http-push.c
revisions API users: add "goto cleanup" for release_revisions()
stash: always have the owner of "stash_info" free it
revisions API users: use release_revisions() needing REV_INFO_INIT
revision.[ch]: document and move code declared around "init"
...
This is a user facing message for a situation seen in the wild.
Translate it.
Signed-off-by: Philip Oakley <philipoakley@iee.email>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `--preserve-merges` option was removed by v2.34.0. However
users may not be aware that it is also a Pull configuration option,
which is still offered by major IDE vendors such as Visual Studio.
Extend the `--preserve-merges` die message to also direct users to
the possible use of the `preserve` option in the `pull.rebase` config.
This is an additional 'belt and braces' information statement.
Signed-off-by: Philip Oakley <philipoakley@iee.email>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Git would die if a "rebase --preserve-merges" was in progress.
Users could neither --quit, --abort, nor --continue the rebase.
Make the `rebase --abort` option available to allow users to remove
traces of any preserve-merges rebase, even if they had upgraded
during a rebase.
One trigger case was an unexpectedly difficult to resolve conflict, as
reported on the `git-users` group.
(https://groups.google.com/g/git-for-windows/c/3jMWbBlXXHM)
Other potential use-cases include git-experts using the portable
'Git on a stick' to help users with an older git version.
Signed-off-by: Philip Oakley <philipoakley@iee.email>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since feebd2d256 (rebase: hide --preserve-merges option, 2019-10-18)
this option is now removed as stated in the subsequent release notes.
Fix and reflow the option tip.
Signed-off-by: Philip Oakley <philipoakley@iee.email>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When `show-ref` is combined with the `--heads` or `--tags` options, it
can avoid iterating parts of a repository's references that it doesn't
care about.
But it doesn't take advantage of this potential optimization. When this
command was introduced back in 358ddb62cf (Add "git show-ref" builtin
command, 2006-09-15), `for_each_ref_in()` did exist. But since most
repositories don't have many (any?) references that aren't branches or
tags already, this makes little difference in practice.
Though for repositories with a large imbalance of branches and tags (or,
more likely in the case of server operators, many hidden references),
this can make quite a difference. Take, for example, a repository with
500,000 "hidden" references (all of the form "refs/__hidden__/N"), and
a single branch:
git commit --allow-empty -m "base" &&
seq 1 500000 | sed 's,\(.*\),create refs/__hidden__/\1 HEAD,' |
git update-ref --stdin &&
git pack-refs --all
Outputting the existence of that single branch currently takes on the
order of ~50ms on my machine. The vast majority of this time is wasted
iterating through references that we know we're going to discard.
Instead, teach `show-ref` that it can iterate just "refs/heads" and/or
"refs/tags" when given `--heads` and/or `--tags`, respectively. A few
small interesting things to note:
- When given either option, we can avoid the general-purpose
for_each_ref() call altogether, since we know that it won't give us
any references that we wouldn't filter out already.
- We can make two separate calls to `for_each_fullref_in()` (and
avoid, say, the more specialized `for_each_fullref_in_prefixes()`,
since we know that the set of references enumerated by each is
disjoint, so we'll never see the same reference appear in both
calls.
- We have to use the "fullref" variant (instead of just
`for_each_branch_ref()` and `for_each_tag_ref()`), since we expect
fully-qualified reference names to appear in `show-ref`'s output.
When either of `heads_only` or `tags_only` is set, we can eliminate the
strcmp() calls in `builtin/show-ref.c::show_ref()` altogether, since we
know that `show_ref()` will never see a non-branch or tag reference.
Unfortunately, we can't use `for_each_fullref_in_prefixes()` to enhance
`show-ref`'s pattern matching, since `show-ref` patterns match on the
_suffix_ (e.g., the pattern "foo" shows "refs/heads/foo",
"refs/tags/foo", and etc, not "foo/*").
Nonetheless, in our synthetic example above, this provides a significant
speed-up ("git" is roughly v2.36, "git.compile" is this patch):
$ hyperfine -N 'git show-ref --heads' 'git.compile show-ref --heads'
Benchmark 1: git show-ref --heads
Time (mean ± σ): 49.9 ms ± 6.2 ms [User: 45.6 ms, System: 4.1 ms]
Range (min … max): 46.1 ms … 73.6 ms 43 runs
Benchmark 2: git.compile show-ref --heads
Time (mean ± σ): 2.8 ms ± 0.4 ms [User: 1.4 ms, System: 1.2 ms]
Range (min … max): 1.3 ms … 5.6 ms 957 runs
Summary
'git.compile show-ref --heads' ran
18.03 ± 3.38 times faster than 'git show-ref --heads'
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A mechanism to pack unreachable objects into a "cruft pack",
instead of ejecting them into loose form to be reclaimed later, has
been introduced.
* tb/cruft-packs:
sha1-file.c: don't freshen cruft packs
builtin/gc.c: conditionally avoid pruning objects via loose
builtin/repack.c: add cruft packs to MIDX during geometric repack
builtin/repack.c: use named flags for existing_packs
builtin/repack.c: allow configuring cruft pack generation
builtin/repack.c: support generating a cruft pack
builtin/pack-objects.c: --cruft with expiration
reachable: report precise timestamps from objects in cruft packs
reachable: add options to add_unseen_recent_objects_to_traversal
builtin/pack-objects.c: --cruft without expiration
builtin/pack-objects.c: return from create_object_entry()
t/helper: add 'pack-mtimes' test-tool
pack-mtimes: support writing pack .mtimes files
chunk-format.h: extract oid_version()
pack-write: pass 'struct packing_data' to 'stage_tmp_packfiles'
pack-mtimes: support reading .mtimes files
Documentation/technical: add cruft-packs.txt
A workflow change for translators are being proposed.
* jx/l10n-workflow-change:
l10n: Document the new l10n workflow
Makefile: add "po-init" rule to initialize po/XX.po
Makefile: add "po-update" rule to update po/XX.po
po/git.pot: don't check in result of "make pot"
po/git.pot: this is now a generated file
Makefile: remove duplicate and unwanted files in FOUND_SOURCE_FILES
i18n CI: stop allowing non-ASCII source messages in po/git.pot
Makefile: have "make pot" not "reset --hard"
Makefile: generate "po/git.pot" from stable LOCALIZED_C
Makefile: sort source files before feeding to xgettext
Teach "git repack --geometric" work better with "--keep-pack" and
avoid corrupting the repository when packsize limit is used.
* tb/geom-repack-with-keep-and-max:
builtin/repack.c: ensure that `names` is sorted
t7703: demonstrate object corruption with pack.packSizeLimit
repack: respect --keep-pack with geometric repack
"sparse-checkout" learns to work well with the sparse-index
feature.
* ds/sparse-sparse-checkout:
sparse-checkout: integrate with sparse index
p2000: add test for 'git sparse-checkout [add|set]'
sparse-index: complete partial expansion
sparse-index: partially expand directories
sparse-checkout: --no-sparse-index needs a full index
cache-tree: implement cache_tree_find_path()
sparse-index: introduce partially-sparse indexes
sparse-index: create expand_index()
t1092: stress test 'git sparse-checkout set'
t1092: refactor 'sparse-index contents' test
The multi-pack-index code did not protect the packfile it is going
to depend on from getting removed while in use, which has been
corrected.
* tb/midx-race-in-pack-objects:
builtin/pack-objects.c: ensure pack validity from MIDX bitmap objects
builtin/pack-objects.c: ensure included `--stdin-packs` exist
builtin/pack-objects.c: avoid redundant NULL check
pack-bitmap.c: check preferred pack validity when opening MIDX bitmap
Preliminary code refactoring around transport and bundle code.
* ds/bundle-uri:
bundle.h: make "fd" version of read_bundle_header() public
remote: allow relative_url() to return an absolute url
remote: move relative_url()
http: make http_get_file() external
fetch-pack: move --keep=* option filling to a function
fetch-pack: add a deref_without_lazy_fetch_extended()
dir API: add a generalized path_match_flags() function
connect.c: refactor sending of agent & object-format
Introduce a filesystem-dependent mechanism to optimize the way the
bits for many loose object files are ensured to hit the disk
platter.
* ns/batch-fsync:
core.fsyncmethod: performance tests for batch mode
t/perf: add iteration setup mechanism to perf-lib
core.fsyncmethod: tests for batch mode
test-lib-functions: add parsing helpers for ls-files and ls-tree
core.fsync: use batch mode and sync loose objects by default on Windows
unpack-objects: use the bulk-checkin infrastructure
update-index: use the bulk-checkin infrastructure
builtin/add: add ODB transaction around add_files_to_cache
cache-tree: use ODB transaction around writing a tree
core.fsyncmethod: batched disk flushes for loose-objects
bulk-checkin: rebrand plug/unplug APIs as 'odb transactions'
bulk-checkin: rename 'state' variable and separate 'plugged' boolean
Deprecate non-cone mode of the sparse-checkout feature.
* en/sparse-cone-becomes-default:
Documentation: some sparsity wording clarifications
git-sparse-checkout.txt: mark non-cone mode as deprecated
git-sparse-checkout.txt: flesh out pattern set sections a bit
git-sparse-checkout.txt: add a new EXAMPLES section
git-sparse-checkout.txt: shuffle some sections and mark as internal
git-sparse-checkout.txt: update docs for deprecation of 'init'
git-sparse-checkout.txt: wording updates for the cone mode default
sparse-checkout: make --cone the default
tests: stop assuming --no-cone is the default mode for sparse-checkout
Follow-up on a preceding commit which changed all references to the
"env_array" when referring to the "struct child_process" member. These
changes are all unnecessary for the compiler, but help the code's
human readers.
All the comments that referred to "env_array" have now been updated,
as well as function names and variables that had "env_array" in their
name, they now refer to "env".
In addition the "out" name for the submodule.h prototype was
inconsistent with the function definition's use of "env_array" in
submodule.c. Both of them use "env" now.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
