grep does not follow the conventions used by other Git commands when
printing paths that contain unusual characters (as double-quotes or
newlines). Commands such as ls-files, commit, status and diff will:
- Quote and escape unusual pathnames, by default.
- Print names verbatim and unquoted when "-z" is used.
But grep *never* quotes/escapes absolute paths with unusual chars and
*always* quotes/escapes relative ones, even with "-z". Besides being
inconsistent in its own output, the deviation from other Git commands
can be confusing. So let's make it follow the two rules above and add
some tests for this new behavior. Note that, making grep quote/escape
all unusual paths by default, also make it fully compliant with the
core.quotePath configuration, which is currently ignored for absolute
paths.
Reported-by: Greg Hurrell <greg@hurrell.net>
Helped-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The changed-path Bloom filters help reduce the amount of tree
parsing required during history queries. Before calculating a
diff, we can ask the filter if a path changed between a commit
and its first parent. If the filter says "no" then we can move
on without parsing trees. If the filter says "maybe" then we
parse trees to discover if the answer is actually "yes" or "no".
When computing a blame, there is a section in find_origin() that
computes a diff between a commit and one of its parents. When this
is the first parent, we can check the Bloom filters before calling
diff_tree_oid().
In order to make this work with the blame machinery, we need to
initialize a struct bloom_key with the initial path. But also, we
need to add more keys to a list if a rename is detected. We then
check to see if _any_ of these keys answer "maybe" in the diff.
During development, I purposefully left out this "add a new key
when a rename is detected" to see if the test suite would catch
my error. That is how I discovered the issues with
GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS from the previous change.
With that change, we can feel some confidence in the coverage of
this change.
If a user requests copy detection using "git blame -C", then there
are more places where the set of "important" files can expand. I
do not know enough about how this happens in the blame machinery.
Thus, the Bloom filter integration is explicitly disabled in this
mode. A later change could expand the bloom_key data with an
appropriate call (or calls) to add_bloom_key().
If we did not disable this mode, then the following tests would
fail:
t8003-blame-corner-cases.sh
t8011-blame-split-file.sh
Generally, this is a performance enhancement and should not
change the behavior of 'git blame' in any way. If a repo has a
commit-graph file with computed changed-path Bloom filters, then
they should notice improved performance for their 'git blame'
commands.
Here are some example timings that I found by blaming some paths
in the Linux kernel repository:
git blame arch/x86/kernel/topology.c >/dev/null
Before: 0.83s
After: 0.24s
git blame kernel/time/time.c >/dev/null
Before: 0.72s
After: 0.24s
git blame tools/perf/ui/stdio/hist.c >/dev/null
Before: 0.27s
After: 0.11s
I specifically looked for "deep" paths that were also edited many
times. As a counterpoint, the MAINTAINERS file was edited many
times but is located in the root tree. This means that the cost of
computing a diff relative to the pathspec is very small. Here are
the timings for that command:
git blame MAINTAINERS >/dev/null
Before: 20.1s
After: 18.0s
These timings are the best of five. The worst-case runs were on the
order of 2.5 minutes for both cases. Note that the MAINTAINERS file
has 18,740 lines across 17,000+ commits. This happens to be one of
the cases where this change provides the least improvement.
The lack of improvement for the MAINTAINERS file and the relatively
modest improvement for the other examples can be easily explained.
The blame machinery needs to compute line-level diffs to determine
which lines were changed by each commit. That makes up a large
proportion of the computation time, and this change does not
attempt to improve on that section of the algorithm. The
MAINTAINERS file is large and changed often, so it takes time to
determine which lines were updated by which commit. In contrast,
the code files are much smaller, and it takes longer to comute
the line-by-line diff for a single patch on the Linux mailing
lists.
Outside of the "-C" integration, I believe there is little more to
gain from the changed-path Bloom filters for 'git blame' after this
patch.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The GIT_TEST_COMMIT_GRAPH environment variable updates the commit-
graph file whenever "git commit" is run, ensuring that we always
have an updated commit-graph throughout the test suite. The
GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS environment variable was
introduced to write the changed-path Bloom filters whenever "git
commit-graph write" is run. However, the GIT_TEST_COMMIT_GRAPH
trick doesn't launch a separate process and instead writes it
directly.
To expand the number of tests that have commits in the commit-graph
file, add a helper method that computes the commit-graph and place
that helper inside "git commit" and "git merge".
In the helper method, check GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS
to ensure we are writing changed-path Bloom filters whenever
possible.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Starting in 3ac68a93fd, help.o began to depend on builtin/branch.o,
builtin/clean.o, and builtin/config.o. This meant that help.o was
unusable outside of the context of the main Git executable.
To make help.o usable by other commands again, move list_config_help()
into builtin/help.c (where it makes sense to assume other builtin libraries
are present).
When command-list.h is included but a member is not used, we start to
hear a compiler warning. Since the config list is generated in a fairly
different way than the command list, and since commands and config
options are semantically different, move the config list into its own
header and move the generator into its own script and build rule.
For reasons explained in 976aaedc (msvc: add a Makefile target to
pre-generate the Visual Studio solution, 2019-07-29), some build
artifacts we consider non-source files cannot be generated in the
Visual Studio environment, and we already have some Makefile tweaks
to help Visual Studio to use generated command-list.h header file.
Do the same to a new generated file, config-list.h, introduced by
this change.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
In 'git log', the --decorate-refs-exclude option appends a pattern
to a string_list. This list is used to prevent showing some refs
in the decoration output, or even by --simplify-by-decoration.
Users may want to use their refs space to store utility refs that
should not appear in the decoration output. For example, Scalar [1]
runs a background fetch but places the "new" refs inside the
refs/scalar/hidden/<remote>/* refspace instead of refs/<remote>/*
to avoid updating remote refs when the user is not looking. However,
these "hidden" refs appear during regular 'git log' queries.
A similar idea to use "hidden" refs is under consideration for core
Git [2].
Add the 'log.excludeDecoration' config option so users can exclude
some refs from decorations by default instead of needing to use
--decorate-refs-exclude manually. The config value is multi-valued
much like the command-line option. The documentation is careful to
point out that the config value can be overridden by the
--decorate-refs option, even though --decorate-refs-exclude would
always "win" over --decorate-refs.
Since the 'log.excludeDecoration' takes lower precedence to
--decorate-refs, and --decorate-refs-exclude takes higher
precedence, the struct decoration_filter needed another field.
This led also to new logic in load_ref_decorations() and
ref_filter_match().
There are several tests in t4202-log.sh that test the
--decorate-refs-(include|exclude) options, so these are extended.
Since the expected output is already stored as a file, most tests
could simply replace a "--decorate-refs-exclude" option with an
in-line config setting. Other tests involve the precedence of
the config option compared to command-line options and needed more
modification.
[1] https://github.com/microsoft/scalar
[2] https://lore.kernel.org/git/77b1da5d3063a2404cd750adfe3bb8be9b6c497d.1585946894.git.gitgitgadget@gmail.com/
Helped-by: Junio C Hamano <gister@pobox.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When operating on a stream of commit OIDs on stdin, 'git commit-graph
write' checks that each OID refers to an object that is indeed a commit.
This is convenient to make sure that the given input is well-formed, but
can sometimes be undesirable.
For example, server operators may wish to feed the refnames that were
updated during a push to 'git commit-graph write --input=stdin-commits',
and silently discard refs that don't point at commits. This can be done
by combing the output of 'git for-each-ref' with '--format
%(*objecttype)', but this requires opening up a potentially large number
of objects. Instead, it is more convenient to feed the updated refs to
the commit-graph machinery, and let it throw out refs that don't point
to commits.
Introduce '--[no-]check-oids' to make such a behavior possible. With
'--check-oids' (the default behavior to retain backwards compatibility),
'git commit-graph write' will barf on a non-commit line in its input.
With 'no-check-oids', such lines will be silently ignored, making the
above possible by specifying this option.
No matter which is supplied, 'git commit-graph write' retains the
behavior from the previous commit of rejecting non-OID inputs like
"HEAD" and "refs/heads/foo" as before.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'write_commit_graph()' function takes in either a string list of
pack indices, or a string list of hexadecimal commit OIDs. These
correspond to the '--stdin-packs' and '--stdin-commits' mode(s) from
'git commit-graph write'.
Using a string_list of hexadecimal commit IDs is not the most efficient
use of memory, since we can instead use the 'struct oidset', which is
more well-suited for this case.
This has another benefit which will become apparent in the following
commit. This is that we are about to disambiguate the kinds of errors we
produce with '--stdin-commits' into "non-hex input" and "hex-input, but
referring to a non-commit object". By having 'write_commit_graph' take
in a 'struct oidset *' of commits, we place the burden on the caller (in
this case, the builtin) to handle the first case, and the commit-graph
machinery can handle the second case.
Suggested-by: Jeff King <peff@peff.net>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When using split commit-graphs, it is sometimes useful to completely
replace the commit-graph chain with a new base.
For example, consider a scenario in which a repository builds a new
commit-graph incremental for each push. Occasionally (say, after some
fixed number of pushes), they may wish to rebuild the commit-graph chain
with all reachable commits.
They can do so with
$ git commit-graph write --reachable
but this removes the chain entirely and replaces it with a single
commit-graph in 'objects/info/commit-graph'. Unfortunately, this means
that the next push will have to move this commit-graph into the first
layer of a new chain, and then write its new commits on top.
Avoid such copying entirely by allowing the caller to specify that they
wish to replace the entirety of their commit-graph chain, while also
specifying that the new commit-graph should become the basis of a fresh,
length-one chain.
This addresses the above situation by making it possible for the caller
to instead write:
$ git commit-graph write --reachable --split=replace
which writes a new length-one chain to 'objects/info/commit-graphs',
making the commit-graph incremental generated by the subsequent push
relatively cheap by avoiding the aforementioned copy.
In order to do this, remove an assumption in 'write_commit_graph_file'
that chains are always at least two incrementals long.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the previous commit, we laid the groundwork for supporting different
splitting strategies. In this commit, we introduce the first splitting
strategy: 'no-merge'.
Passing '--split=no-merge' is useful for callers which wish to write a
new incremental commit-graph, but do not want to spend effort condensing
the incremental chain [1]. Previously, this was possible by passing
'--size-multiple=0', but this no longer the case following 63020f175f
(commit-graph: prefer default size_mult when given zero, 2020-01-02).
When '--split=no-merge' is given, the commit-graph machinery will never
condense an existing chain, and it will always write a new incremental.
[1]: This might occur when, for example, a server administrator running
some program after each push may want to ensure that each job runs
proportional in time to the size of the push, and does not "jump" when
the commit-graph machinery decides to trigger a merge.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
With '--split', the commit-graph machinery writes new commits in another
incremental commit-graph which is part of the existing chain, and
optionally decides to condense the chain into a single commit-graph.
This is done to ensure that the asymptotic behavior of looking up a
commit in an incremental chain is not dominated by the number of
incrementals in that chain. It can be controlled by the '--max-commits'
and '--size-multiple' options.
In the next two commits, we will introduce additional splitting
strategies that can exert additional control over:
- when a split commit-graph is and isn't written, and
- when the existing commit-graph chain is discarded completely and
replaced with another graph
To prepare for this, make '--split' take an optional strategy (as in
'--split[=<strategy>]'), and add a new enum to describe which strategy
is being used. For now, no strategies are given, and the only enumerated
value is 'COMMIT_GRAPH_SPLIT_UNSPECIFIED', indicating the absence of a
strategy.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of using `starts_with()`, the magic number 7, `strlen()` and a
fair number of additions to verify the three parts of the config key
"branch.<branch>.mergeoptions", use `skip_prefix()` to jump through them
more explicitly.
We need to introduce a new variable for this (we certainly can't modify
`k` just because we see "branch."!). With `skip_prefix()` we often use
quite bland names like `p` or `str`. Let's do the same. If and when this
function needs to do more prefix-skipping, we'll have a generic variable
ready for this.
Signed-off-by: Martin Ågren <martin.agren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When rebasing against an upstream that has had many commits since the
original branch was created:
O -- O -- ... -- O -- O (upstream)
\
-- O (my-dev-branch)
it must read the contents of every novel upstream commit, in addition to
the tip of the upstream and the merge base, because "git rebase"
attempts to exclude commits that are duplicates of upstream ones. This
can be a significant performance hit, especially in a partial clone,
wherein a read of an object may end up being a fetch.
Add a flag to "git rebase" to allow suppression of this feature. This
flag only works when using the "merge" backend.
This flag changes the behavior of sequencer_make_script(), called from
do_interactive_rebase() <- run_rebase_interactive() <-
run_specific_rebase() <- cmd_rebase(). With this flag, limit_list()
(indirectly called from sequencer_make_script() through
prepare_revision_walk()) will no longer call cherry_pick_list(), and
thus PATCHSAME is no longer set. Refraining from setting PATCHSAME both
means that the intermediate commits in upstream are no longer read (as
shown by the test) and means that no PATCHSAME-caused skipping of
commits is done by sequencer_make_script(), either directly or through
make_script_with_merges().
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the user specifies the apply backend with options that only work
with the merge backend, such as
git rebase --apply --exec /bin/true HEAD~3
the error message has always been
fatal: --exec requires an interactive rebase
This error message is misleading and was one of the reasons we renamed
the interactive backend to the merge backend. Update the error message
to state that these options merely require use of the merge backend.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit d48e5e21da ("rebase (interactive-backend): make --keep-empty the
default", 2020-02-15) turned --keep-empty (for keeping commits which
start empty) into the default. The logic underpinning that commit was:
1) 'git commit' errors out on the creation of empty commits without an
override flag
2) Once someone determines that the override is worthwhile, it's
annoying and/or harmful to required them to take extra steps in
order to keep such commits around (and to repeat such steps with
every rebase).
While the logic on which the decision was made is sound, the result was
a bit of an overcorrection. Instead of jumping to having --keep-empty
being the default, it jumped to making --keep-empty the only available
behavior. There was a simple workaround, though, which was thought to
be good enough at the time. People could still drop commits which
started empty the same way the could drop any commits: by firing up an
interactive rebase and picking out the commits they didn't want from the
list. However, there are cases where external tools might create enough
empty commits that picking all of them out is painful. As such, having
a flag to automatically remove start-empty commits may be beneficial.
Provide users a way to drop commits which start empty using a flag that
existed for years: --no-keep-empty. Interpret --keep-empty as
countermanding any previous --no-keep-empty, but otherwise leaving
--keep-empty as the default.
This might lead to some slight weirdness since commands like
git rebase --empty=drop --keep-empty
git rebase --empty=keep --no-keep-empty
look really weird despite making perfect sense (the first will drop
commits which become empty, but keep commits that started empty; the
second will keep commits which become empty, but drop commits which
started empty). However, --no-keep-empty was named years ago and we are
predominantly keeping it for backward compatibility; also we suspect it
will only be used rarely since folks already have a simple way to drop
commits they don't want with an interactive rebase.
Reported-by: Bryan Turner <bturner@atlassian.com>
Reported-by: Sami Boukortt <sami@boukortt.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We return the length to a subset of a string using an "int *"
out-parameter. This is fine most of the time, as we'd expect config keys
to be relatively short, but it could behave oddly if we had a gigantic
config key. A more appropriate type is size_t.
Let's switch over, which lets our callers use size_t as appropriate
(they are bound by our type because they must pass the out-parameter as
a pointer). This is mostly just a cleanup to make it clear this code
handles long strings correctly. In practice, our config parser already
chokes on long key names (because of a similar int/size_t mixup!).
When doing an int/size_t conversion, we have to be careful that nobody
was trying to assign a negative value to the variable. I manually
confirmed that for each case here. They tend to just feed the result to
xmemdupz() or similar; in a few cases I adjusted the parameter types for
helper functions to make sure the size_t is preserved.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are quite a few code locations (e.g. `git clean --interactive`)
where Git asks the user for an answer. In preparation for fixing a bug
shared by all of them, and also to DRY up the code, let's refactor it.
Please note that most of these callers trimmed white-space both at the
beginning and at the end of the answer, instead of trimming only the
end (as the caller in `add-patch.c` does).
Therefore, technically speaking, we change behavior in this patch. At
the same time, it can be argued that this is actually a bug fix.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before, `--autostash` only worked with `git pull --rebase`. However, in
the last patch, merge learned `--autostash` as well so there's no reason
why we should have this restriction anymore. Teach pull to pass
`--autostash` to merge, just like it did for rebase.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In rebase, one can pass the `--autostash` option to cause the worktree
to be automatically stashed before continuing with the rebase. This
option is missing in merge, however.
Implement the `--autostash` option and corresponding `merge.autoStash`
option in merge which stashes before merging and then pops after.
This option is useful when a developer has some local changes on a topic
branch but they realize that their work depends on another branch.
Previously, they had to run something like
git fetch ...
git stash push
git merge FETCH_HEAD
git stash pop
but now, that is reduced to
git fetch ...
git merge --autostash FETCH_HEAD
When an autostash is generated, it is automatically reapplied to the
worktree only in three explicit situations:
1. An incomplete merge is commit using `git commit`.
2. A merge completes successfully.
3. A merge is aborted using `git merge --abort`.
In all other situations where the merge state is removed using
remove_merge_branch_state() such as aborting a merge via
`git reset --hard`, the autostash is saved into the stash reflog
instead keeping the worktree clean.
Helped-by: Phillip Wood <phillip.wood@dunelm.org.uk>
Suggested-by: Alban Gruin <alban.gruin@gmail.com>
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Lib-ify the autostash code by extracting perform_autostash() from rebase
into sequencer. In a future commit, this will be used to implement
`--autostash` in other builtins.
This patch is best viewed with `--color-moved`.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the future, we plan on lib-ifying create_autostash() so we need it to
be more generic. Make it more generic by making it accept a
`struct repository` argument instead of implicitly using the non-repo
functions and `the_repository`. Also, make it accept a `path` argument
so that we no longer rely have to rely on `struct rebase_options`.
Finally, make it accept a `default_reflog_action` argument so we no
longer have to rely on `DEFAULT_REFLOG_ACTION`.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In a future commit, we will lib-ify this code. In preparation for
this, extract the code into the create_autostash() function so that it
can be cleaned up before it is finally lib-ified.
This patch is best viewed with `--color-moved` and
`--color-moved-ws=allow-indentation-change`.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Continue the process of lib-ifying the autostash code. In a future
commit, this will be used to implement `--autostash` in other builtins.
This patch is best viewed with `--color-moved`.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the future, we plan on lib-ifying reset_head() so we need it to
be more generic. Make it more generic by making it accept a
`struct repository` argument instead of implicitly using the non-repo
functions. Also, make it accept a `const char *default_reflog_action`
argument so that the default action of "rebase" isn't hardcoded in.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The apply_autostash() function in builtin/rebase.c is similar enough to
the apply_autostash() function in sequencer.c that they are almost
interchangeable, except for the type of arg they accept. Make the
sequencer.c version extern and use it in rebase.
The rebase version was introduced in 6defce2b02 (builtin rebase: support
`--autostash` option, 2018-09-04) as part of the shell to C conversion.
It opted to duplicate the function because, at the time, there was
another in-progress project converting interactive rebase from shell to
C as well and they did not want to clash with them by refactoring
sequencer.c version of apply_autostash(). Since both efforts are long
done, we can freely combine them together now.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since in sequencer.c, read_one() basically duplicates the functionality
of read_oneliner(), reduce code duplication by replacing read_one() with
read_oneliner().
This was done with the following Coccinelle script
@@
expression a, b;
@@
- read_one(a, b)
+ !read_oneliner(b, a, READ_ONELINER_WARN_MISSING)
and long lines were manually broken up.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we're comparing a push cert nonce, we currently do so using strcmp.
Most implementations of strcmp short-circuit and exit as soon as they
know whether two values are equal. This, however, is a problem when
we're comparing the output of HMAC, as it leaks information in the time
taken about how much of the two values match if they do indeed differ.
In our case, the nonce is used to prevent replay attacks against our
server via the embedded timestamp and replay attacks using requests from
a different server via the HMAC. Push certs, which contain the nonces,
are signed, so an attacker cannot tamper with the nonces without
breaking validation of the signature. They can, of course, create their
own signatures with invalid nonces, but they can also create their own
signatures with valid nonces, so there's nothing to be gained. Thus,
there is no security problem.
Even though it doesn't appear that there are any negative consequences
from the current technique, for safety and to encourage good practices,
let's use a constant time comparison function for nonce verification.
POSIX does not provide one, but they are easy to write.
The technique we use here is also used in NaCl and the Go standard
library and relies on the fact that bitwise or and xor are constant time
on all known architectures.
We need not be concerned about exiting early if the actual and expected
lengths differ, since the standard cryptographic assumption is that
everyone, including an attacker, knows the format of and algorithm used
in our nonces (and in any event, they have the source code and can
determine it easily). As a result, we assume everyone knows how long
our nonces should be. This philosophy is also taken by the Go standard
library and other cryptographic libraries when performing constant time
comparisons on HMAC values.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When trying to stash part of the worktree changes by splitting a hunk
and then only partially accepting the split bits and pieces, the user
is presented with a rather cryptic error:
error: patch failed: <file>:<line>
error: test: patch does not apply
Cannot remove worktree changes
and the command would fail to stash the desired parts of the worktree
changes (even if the `stash` ref was actually updated correctly).
We even have a test case demonstrating that failure, carrying it for
four years already.
The explanation: when splitting a hunk, the changed lines are no longer
separated by more than 3 lines (which is the amount of context lines
Git's diffs use by default), but less than that. So when staging only
part of the diff hunk for stashing, the resulting diff that we want to
apply to the worktree in reverse will contain those changes to be
dropped surrounded by three context lines, but since the diff is
relative to HEAD rather than to the worktree, these context lines will
not match.
Example time. Let's assume that the file README contains these lines:
We
the
people
and the worktree added some lines so that it contains these lines
instead:
We
are
the
kind
people
and the user tries to stash the line containing "are", then the command
will internally stage this line to a temporary index file and try to
revert the diff between HEAD and that index file. The diff hunk that
`git stash` tries to revert will look somewhat like this:
@@ -1776,3 +1776,4
We
+are
the
people
It is obvious, now, that the trailing context lines overlap with the
part of the original diff hunk that the user did *not* want to stash.
Keeping in mind that context lines in diffs serve the primary purpose of
finding the exact location when the diff does not apply precisely (but
when the exact line number in the file to be patched differs from the
line number indicated in the diff), we work around this by reducing the
amount of context lines: the diff was just generated.
Note: this is not a *full* fix for the issue. Just as demonstrated in
t3701's 'add -p works with pathological context lines' test case, there
are ambiguities in the diff format. It is very rare in practice, of
course, to encounter such repeated lines.
The full solution for such cases would be to replace the approach of
generating a diff from the stash and then applying it in reverse by
emulating `git revert` (i.e. doing a 3-way merge). However, in `git
stash -p` it would not apply to `HEAD` but instead to the worktree,
which makes this non-trivial to implement as long as we also maintain a
scripted version of `add -i`.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When commit subjects or authors have non-ASCII characters, git
format-patch Q-encodes them so they can be safely sent over email.
However, if the patch transfer method is something other than email (web
review tools, sneakernet), this only serves to make the patch metadata
harder to read without first applying it (unless you can decode RFC 2047
in your head). git am as well as some email software supports
non-Q-encoded mail as described in RFC 6531.
Add --[no-]encode-email-headers and format.encodeEmailHeaders to let the
user control this behavior.
Signed-off-by: Emma Brooks <me@pluvano.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS test flag to the test setup suite
in order to toggle writing Bloom filters when running any of the git tests.
If set to true, we will compute and write Bloom filters every time a test
calls `git commit-graph write`, as if the `--changed-paths` option was
passed in.
The test suite passes when GIT_TEST_COMMIT_GRAPH and
GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS are enabled.
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add --changed-paths option to git commit-graph write. This option will
allow users to compute information about the paths that have changed
between a commit and its first parent, and write it into the commit graph
file. If the option is passed to the write subcommand we set the
COMMIT_GRAPH_WRITE_BLOOM_FILTERS flag and pass it down to the
commit-graph logic.
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are 3 callers to promisor_remote_get_direct() that first check if
the number of objects to be fetched is equal to 0. Fold that check into
promisor_remote_get_direct(), and in doing so, be explicit as to what
promisor_remote_get_direct() does if oid_nr is 0 (it returns 0, success,
immediately).
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The git-update-ref(1) command can only handle queueing transactions
right now via its "--stdin" parameter, but there is no way for users to
handle the transaction itself in a more explicit way. E.g. in a
replicated scenario, one may imagine a coordinator that spawns
git-update-ref(1) for multiple repositories and only if all agree that
an update is possible will the coordinator send a commit. Such a
transactional session could look like
> start
< start: ok
> update refs/heads/master $OLD $NEW
> prepare
< prepare: ok
# All nodes have returned "ok"
> commit
< commit: ok
or
> start
< start: ok
> create refs/heads/master $OLD $NEW
> prepare
< fatal: cannot lock ref 'refs/heads/master': reference already exists
# On all other nodes:
> abort
< abort: ok
In order to allow for such transactional sessions, this commit
introduces four new commands for git-update-ref(1), which matches those
we have internally already with the exception of "start":
- start: start a new transaction
- prepare: prepare the transaction, that is try to lock all
references and verify their current value matches the
expected one
- commit: explicitly commit a session, that is update references to
match their new expected state
- abort: abort a session and roll back all changes
By design, git-update-ref(1) will commit as soon as standard input is
being closed. While fine in a non-transactional world, it is definitely
unexpected in a transactional world. Because of this, as soon as any of
the new transactional commands is used, the default will change to
aborting without an explicit "commit". To avoid a race between queueing
updates and the first "prepare" that starts a transaction, the "start"
command has been added to start an explicit transaction.
Add some tests to exercise this new functionality.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The git-update-ref(1) supports a `--stdin` mode that allows it to read
all reference updates from standard input. This is mainly used to allow
for atomic reference updates that are all or nothing, so that either all
references will get updated or none.
Currently, git-update-ref(1) reads all commands as a single block of up
to 1000 characters and only starts processing after stdin gets closed.
This is less flexible than one might wish for, as it doesn't really
allow for longer-lived transactions and doesn't allow any verification
without committing everything. E.g. one may imagine the following
exchange:
> start
< start: ok
> update refs/heads/master $NEWOID1 $OLDOID1
> update refs/heads/branch $NEWOID2 $OLDOID2
> prepare
< prepare: ok
> commit
< commit: ok
When reading all input as a whole block, the above interactive protocol
is obviously impossible to achieve. But by converting the command to
read commands linewise, we can make it more interactive than before.
Obviously, the linewise interface is only a first step in making
git-update-ref(1) work in a more transaction-oriented way. Missing is
most importantly support for transactional commands that manage the
current transaction.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While the actual logic to update the transaction is handled in
`update_refs_stdin()`, the transaction itself is started and committed
in `cmd_update_ref()` itself. This makes it hard to handle transaction
abortion and commits as part of `update_refs_stdin()` itself, which is
required in order to introduce transaction handling features to `git
update-refs --stdin`.
Refactor the code to move all transaction handling into
`update_refs_stdin()` to prepare for transaction handling features.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We currently pass both an `strbuf` containing the current command line
as well as the `next` pointer pointing to the first argument to
commands. This is both confusing and makes code more intertwined.
Convert this to use a simple pointer as well as a pointer pointing to
the end of the input as a preparatory step to line-wise reading of
stdin.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `parse_refname` function accepts a `struct strbuf *input` argument
that isn't used at all. As we're about to convert commands to not use a
strbuf anymore but instead an end pointer, let's drop this argument now
to make the converting commit easier to review.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We currently manually wire up all commands known to `git-update-ref
--stdin`, making it harder than necessary to preprocess arguments after
the command is determined. To make this more extensible, let's refactor
the code to use an array of known commands instead. While this doesn't
add a lot of value now, it is a preparatory step to implement line-wise
reading of commands.
As we're going to introduce commands without trailing spaces, this
commit also moves whitespace parsing into the respective commands.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The commit-graph builtin has an --expire-time option that takes a
datetime using OPT_EXPIRY_DATE(). However, the implementation inside
expire_commit_graphs() was treating a non-zero value as a number of
seconds to subtract from "now".
Update t5323-split-commit-graph.sh to demonstrate the correct value
of the --expire-time option by actually creating a crud .graph file
with mtime earlier than the expire time. Instead of using a super-
early time (1980) we use an explicit, and recent, time. Using
test-tool chmtime to create two files on either end of an exact
second, we create a test that catches this failure no matter the
current time. Using a fixed date is more portable than trying to
format a relative date string into the --expiry-date input.
I noticed this when inspecting some Scalar repos that had an excess
number of commit-graph files. In Scalar, we were using this second
interpretation by using "--expire-time=3600" to mean "delete graphs
older than one hour ago" to avoid deleting a commit-graph that a
foreground process may be trying to load.
Also I noticed that the help text was copied from the --max-commits
option. Fix that help text.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Traditionally, the expected calling convention for the dir.c API was:
fill_directory(&dir, ..., pathspec)
foreach entry in dir->entries:
if (dir_path_match(entry, pathspec))
process_or_display(entry)
This may have made sense once upon a time, because the fill_directory() call
could use cheap checks to avoid doing full pathspec matching, and an external
caller may have wanted to do other post-processing of the results anyway.
However:
* this structure makes it easy for users of the API to get it wrong
* this structure actually makes it harder to understand
fill_directory() and the functions it uses internally. It has
tripped me up several times while trying to fix bugs and
restructure things.
* relying on post-filtering was already found to produce wrong
results; pathspec matching had to be added internally for multiple
cases in order to get the right results (see commits 404ebceda0
(dir: also check directories for matching pathspecs, 2019-09-17)
and 89a1f4aaf7 (dir: if our pathspec might match files under a
dir, recurse into it, 2019-09-17))
* it's bad for performance: fill_directory() already has to do lots
of checks and knows the subset of cases where it still needs to do
more checks. Forcing external callers to do full pathspec
matching means they must re-check _every_ path.
So, add the pathspec matching within the fill_directory() internals, and
remove it from external callers.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When cloning with --single-branch, we implement git-fetch's usual
tag-following behavior, grabbing any tag objects that point to objects
we have locally.
When we're a partial clone, though, our has_object_file() check will
actually lazy-fetch each tag. That not only defeats the purpose of
--single-branch, but it does it incredibly slowly, potentially kicking
off a new fetch for each tag. This is even worse for a shallow clone,
which implies --single-branch, because even tags which are supersets of
each other will be fetched individually.
We can fix this by passing OBJECT_INFO_SKIP_FETCH_OBJECT to the call,
which is what git-fetch does in this case.
Likewise, let's include OBJECT_INFO_QUICK, as that's what git-fetch
does. The rationale is discussed in 5827a03545 (fetch: use "quick"
has_sha1_file for tag following, 2016-10-13), but here the tradeoff
would apply even more so because clone is very unlikely to be racing
with another process repacking our newly-created repository.
This may provide a very small speedup even in the non-partial case case,
as we'd avoid calling reprepare_packed_git() for each tag (though in
practice, we'd only have a single packfile, so that reprepare should be
quite cheap).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We renamed the actual data structure in 910650d2f8 (Rename sha1_array to
oid_array, 2017-03-31), but the file is still called sha1-array. Besides
being slightly confusing, it makes it more annoying to grep for leftover
occurrences of "sha1" in various files, because the header is included
in so many places.
Let's complete the transition by renaming the source and header files
(and fixing up a few comment references).
I kept the "-" in the name, as that seems to be our style; cf.
fc1395f4a4 (sha1_file.c: rename to use dash in file name, 2018-04-10).
We also have oidmap.h and oidset.h without any punctuation, but those
are "struct oidmap" and "struct oidset" in the code. We _could_ make
this "oidarray" to match, but somehow it looks uglier to me because of
the length of "array" (plus it would be a very invasive patch for little
gain).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
With 50033772d5 ("connected: verify promisor-ness of partial clone",
2020-01-30), the fast path (checking promisor packs) in
check_connected() now passes a subset of the slow path (rev-list) - if
all objects to be checked are found in promisor packs, both the fast
path and the slow path will pass; otherwise, the fast path will
definitely not pass. This means that we can always attempt the fast path
whenever we need to do the slow path.
The fast path is currently guarded by a flag; therefore, remove that
flag. Also, make the fast path fallback to the slow path - if the fast
path fails, the failing OID and all remaining OIDs will be passed to
rev-list.
The main user-visible benefit is the performance of fetch from a partial
clone - specifically, the speedup of the connectivity check done before
the fetch. In particular, a no-op fetch into a partial clone on my
computer was sped up from 7 seconds to 0.01 seconds. This is a
complement to the work in 2df1aa239c ("fetch: forgo full
connectivity check if --filter", 2020-01-30), which is the child of the
aforementioned 50033772d5. In that commit, the connectivity check
*after* the fetch was sped up.
The addition of the fast path might cause performance reductions in
these cases:
- If a partial clone or a fetch into a partial clone fails, Git will
fruitlessly run rev-list (it is expected that everything fetched
would go into promisor packs, so if that didn't happen, it is most
likely that rev-list will fail too).
- Any connectivity checks done by receive-pack, in the (in my opinion,
unlikely) event that a partial clone serves receive-pack.
I think that these cases are rare enough, and the performance reduction
in this case minor enough (additional object DB access), that the
benefit of avoiding a flag outweighs these.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Reviewed-by: Josh Steadmon <steadmon@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'pack.useSparse' configuration variable now defaults to 'true',
enabling an optimization that has been experimental since Git 2.21.
* ds/default-pack-use-sparse-to-true:
pack-objects: flip the use of GIT_TEST_PACK_SPARSE
config: set pack.useSparse=true by default
The fetch options --deepen, --negotiation-tip, --server-option,
--shallow-exclude, and --shallow-since are documented for git pull as
well, but are not actually accepted by that command. Pass them on to
make the code match its documentation.
Reported-by: 天几 <muzimuzhi@gmail.com>
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When opt_rebase is true, we still first check if we can fast-forward.
If the branch is fast-forwardable, then we can avoid the rebase and just
use merge to do the fast-forward logic. However, when commit a6d7eb2c7a
("pull: optionally rebase submodules (remote submodule changes only)",
2017-06-23) added the ability to rebase submodules it accidentally
caused us to run BOTH a merge and a rebase. Add a flag to avoid doing
both.
This was found when a user had both pull.rebase and rebase.autosquash
set to true. In such a case, the running of both merge and rebase would
cause ORIG_HEAD to be updated twice (and match HEAD at the end instead
of the commit before the rebase started), against expectation.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If commands like merge or rebase materialize files as part of their work,
or a previous sparse-checkout command failed to update individual files
due to dirty changes, users may want a command to simply 'reapply' the
sparsity rules. Provide one.
Reviewed-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
setup_unpack_trees_porcelain() provides much improved error/warning
messages; instead of a message that assumes that there is only one path
with a given problem despite being used by code that intentionally is
grouping and showing errors together, it uses a message designed to be
used with groups of paths. For example, this transforms
error: Entry ' folder1/a
folder2/a
' not uptodate. Cannot update sparse checkout.
into
error: Cannot update sparse checkout: the following entries are not up to date:
folder1/a
folder2/a
In the past the suboptimal messages were never actually triggered
because we would error out if the working directory wasn't clean before
we even called unpack_trees(). The previous commit changed that,
though, so let's use the better error messages.
Reviewed-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove the equivalent of 'git read-tree -mu HEAD' in the sparse-checkout
codepaths for setting the SKIP_WORKTREE bits and instead use the new
update_sparsity() function.
Note that when an issue is hit, the error message splits 'error' and
'Cannot update sparse checkout' on separate lines. For now, we use two
greps to find both pieces of the error message but subsequent commits
will clean up the messages reported to the user.
Reviewed-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
commit e091228e17 ("sparse-checkout: update working directory
in-process", 2019-11-21) allowed passing a pre-defined set of patterns
to unpack_trees(). However, if o->pl was NULL, it would still read the
existing patterns and use those. If those patterns were read into a
data structure that was allocated, naturally they needed to be free'd.
However, despite the same function being responsible for knowing about
both the allocation and the free'ing, the logic for tracking whether to
free the pattern_list was hoisted to an outer function with an
additional flag in unpack_trees_options. Put the logic back in the
relevant function and discard the now unnecessary flag.
Reviewed-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git pull" learned to warn when no pull.rebase configuration
exists, and neither --[no-]rebase nor --ff-only is given (which
would result a merge).
* ah/force-pull-rebase-configuration:
pull: warn if the user didn't say whether to rebase or to merge
"git stash" has kept an escape hatch to use the scripted version
for a few releases, which got stale. It has been removed.
* tg/retire-scripted-stash:
stash: remove the stash.useBuiltin setting
stash: get git_stash_config at the top level
When "git describe C" finds an annotated tag with tagname A to be
the best name to explain commit C, and the tag is stored in a
"wrong" place in the refs/tags hierarchy, e.g. refs/tags/B, the
command gave a warning message but used A (not B) to describe C.
If C is exactly at the tag, the describe output would be "A", but
"git rev-parse A^0" would not be equal as "git rev-parse C^0". The
behavior of the command has been changed to use the "long" form
i.e. A-0-gOBJECTNAME, which is correctly interpreted by rev-parse.
* jc/describe-misnamed-annotated-tag:
describe: force long format for a name based on a mislocated tag
The "--fork-point" mode of "git rebase" regressed when the command
was rewritten in C back in 2.20 era, which has been corrected.
* at/rebase-fork-point-regression-fix:
rebase: --fork-point regression fix
Provide more information (e.g. the object of the tree-ish in which
the blob being converted appears, in addition to its path, which
has already been given) to smudge/clean conversion filters.
* bc/filter-process:
t0021: test filter metadata for additional cases
builtin/reset: compute checkout metadata for reset
builtin/rebase: compute checkout metadata for rebases
builtin/clone: compute checkout metadata for clones
builtin/checkout: compute checkout metadata for checkouts
convert: provide additional metadata to filters
convert: permit passing additional metadata to filter processes
builtin/checkout: pass branch info down to checkout_worktree
The code to interface with GnuPG has been refactored.
* hi/gpg-prefer-check-signature:
gpg-interface: prefer check_signature() for GPG verification
t: increase test coverage of signature verification output
SHA-256 transition continues.
* bc/sha-256-part-1-of-4: (22 commits)
fast-import: add options for rewriting submodules
fast-import: add a generic function to iterate over marks
fast-import: make find_marks work on any mark set
fast-import: add helper function for inserting mark object entries
fast-import: permit reading multiple marks files
commit: use expected signature header for SHA-256
worktree: allow repository version 1
init-db: move writing repo version into a function
builtin/init-db: add environment variable for new repo hash
builtin/init-db: allow specifying hash algorithm on command line
setup: allow check_repository_format to read repository format
t/helper: make repository tests hash independent
t/helper: initialize repository if necessary
t/helper/test-dump-split-index: initialize git repository
t6300: make hash algorithm independent
t6300: abstract away SHA-1-specific constants
t: use hash-specific lookup tables to define test constants
repository: require a build flag to use SHA-256
hex: add functions to parse hex object IDs in any algorithm
hex: introduce parsing variants taking hash algorithms
...
The mechanism to prevent "git commit" from making an empty commit
or amending during an interrupted cherry-pick was broken during the
rewrite of "git rebase" in C, which has been corrected.
* pw/advise-rebase-skip:
commit: give correct advice for empty commit during a rebase
commit: encapsulate determine_whence() for sequencer
commit: use enum value for multiple cherry-picks
sequencer: write CHERRY_PICK_HEAD for reword and edit
cherry-pick: check commit error messages
cherry-pick: add test for `--skip` advice in `git commit`
t3404: use test_cmp_rev
The real_path() convenience function can easily be misused; with a
bit of code refactoring in the callers' side, its use has been
eliminated.
* am/real-path-fix:
get_superproject_working_tree(): return strbuf
real_path_if_valid(): remove unsafe API
real_path: remove unsafe API
set_git_dir: fix crash when used with real_path()
Revamping of the advise API to allow more systematic enumeration of
advice knobs in the future.
* hw/advise-ng:
tag: use new advice API to check visibility
advice: revamp advise API
advice: change "setupStreamFailure" to "setUpstreamFailure"
advice: extract vadvise() from advise()
In builtin.h, there exists the distinctly lib-ish function
prune_packed_objects(). This function can currently only be called by
built-in commands but, unlike all of the other functions in the header,
it does not make sense to impose this restriction as the functionality
can be logically reused in libgit.
Extract this function into prune-packed.c so that related definitions
can exist clearly in their own header file.
While we're at it, clean up #includes that are unused.
This patch is best viewed with --color-moved.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In builtin.h, there exists the distinctly "lib-ish" function
fmt_merge_msg(). This function can currently only be called by built-in
commands but, unlike most of the other functions in the header, it does
not make sense to impose this restriction as the functionality can be
logically reused in libgit.
Extract this function into fmt-merge-msg.c so that related definitions
can exist clearly in their own header file.
While we're at it, clean up #includes that are unused.
This patch is best viewed with --color-moved.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The environment variable GIT_TEST_PACK_SPARSE was previously used
to allow testing the --sparse option for "git pack-objects" in
the test suite. This allowed interesting cases of "git push" to
also test this algorithm.
Since pack.useSparse is now true by default, we do not need this
variable to _enable_ the --sparse option, but instead to _disable_
it. This flips how we work with the variable a bit.
When checking for the variable, default to a value of -1 for
"unset". If unset, then take the default from the repo settings,
which is currently 1. Then, the --[no-]sparse command-line option
will override either of these settings.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 'submodule--helper.c', the structures and macros for callbacks belonging
to any subcommand are named in the format: 'subcommand_cb' and 'SUBCOMMAND_CB_INIT'
respectively.
This was an exception for the subcommand 'foreach' of the command
'submodule'. Rename the aforementioned structures and macros:
'struct cb_foreach' to 'struct foreach_cb' and 'CB_FOREACH_INIT'
to 'FOREACH_CB_INIT'.
Signed-off-by: Shourya Shukla <shouryashukla.oo@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git merge signed-tag" while lacking the public key started to say
"No signature", which was utterly wrong. This regression has been
reverted.
* hi/gpg-use-check-signature:
Revert "gpg-interface: prefer check_signature() for GPG verification"
Fix for a bug revealed by a recent change to make the protocol v2
the default.
* ds/partial-clone-fixes:
partial-clone: avoid fetching when looking for objects
partial-clone: demonstrate bugs in partial fetch
"git check-ignore" did not work when the given path is explicitly
marked as not ignored with a negative entry in the .gitignore file.
* en/check-ignore:
check-ignore: fix documentation and implementation to match
The index-pack code now diagnoses a bad input packstream that
records the same object twice when it is used as delta base; the
code used to declare a software bug when encountering such an
input, but it is an input error.
* jk/index-pack-dupfix:
index-pack: downgrade twice-resolved REF_DELTA to die()
The option name "--use-mailmap" looks OK, but it becomes awkward
when you have to negate it, i.e. "--no-use-mailmap". I, perhaps
with many other users, always try "--no-mailmap" and become unhappy
to see it fail.
Add an alias "--[no-]mailmap" to remedy this.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous step made an option that is an alias to another option
identify itself as an alias to the latter. Because it is easier to
scan the list when a pointer goes backward to what a reader already
has seen, mention "recurse-submodules" first with its true short
help string, and then "recurse" with the statement that it is a
synonym to "recurse-submodules".
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Pass the commit, and if we have it, the ref to the filters when we
perform a checkout. This should only be the case when we invoke git
reset --hard; the metadata will be unused otherwise.
Signed-off-by: brian m. carlson <bk2204@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When checking out a commit, provide metadata to the filter process
including the ref we're using.
Signed-off-by: brian m. carlson <bk2204@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Provide commit metadata for checkout code paths that use unpack_trees
and friends. When we're checking out a commit, use the commit
information, but don't provide commit information if we're checking out
from the index, since there need not be any particular commit associated
with the index, and even if there is one, we can't know what it is.
Signed-off-by: brian m. carlson <bk2204@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that we have the codebase wired up to pass any additional metadata
to filters, let's collect the additional metadata that we'd like to
pass.
The two main places we pass this metadata are checkouts and archives.
In these two situations, reading HEAD isn't a valid option, since HEAD
isn't updated for checkouts until after the working tree is written and
archives can accept an arbitrary tree. In other situations, HEAD will
usually reflect the refname of the branch in current use.
We pass a smaller amount of data in other cases, such as git cat-file,
where we can really only logically know about the blob.
This commit updates only the parts of the checkout code where we don't
use unpack_trees. That function and callers of it will be handled in a
future commit.
In the archive code, we leak a small amount of memory, since nothing we
pass in the archiver argument structure is freed.
Signed-off-by: brian m. carlson <bk2204@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are a variety of situations where a filter process can make use of
some additional metadata. For example, some people find the ident
filter too limiting and would like to include the commit or the branch
in their smudged files. This information isn't available during
checkout as HEAD hasn't been updated at that point, and it wouldn't be
available in archives either.
Let's add a way to pass this metadata down to the filter. We pass the
blob we're operating on, the treeish (preferring the commit over the
tree if one exists), and the ref we're operating on. Note that we won't
pass this information in all cases, such as when renormalizing or when
we're performing diffs, since it doesn't make sense in those cases.
The data we currently get from the filter process looks like the
following:
command=smudge
pathname=git.c
0000
With this change, we'll get data more like this:
command=smudge
pathname=git.c
refname=refs/tags/v2.25.1
treeish=c522f061d551c9bb8684a7c3859b2ece4499b56b
blob=7be7ad34bd053884ec48923706e70c81719a8660
0000
There are a couple things to note about this approach. For operations
like checkout, treeish will always be a commit, since we cannot check
out individual trees, but for other operations, like archive, we can end
up operating on only a particular tree, so we'll provide only a tree as
the treeish. Similar comments apply for refname, since there are a
variety of cases in which we won't have a ref.
This commit wires up the code to print this information, but doesn't
pass any of it at this point. In a future commit, we'll have various
code paths pass the actual useful data down.
Signed-off-by: brian m. carlson <bk2204@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This commit refactors the use of verify_signed_buffer() outside of
gpg-interface.c to use check_signature() instead. It also turns
verify_signed_buffer() into a file-local function since it's now only
invoked internally by check_signature().
There were previously two globally scoped functions used in different
parts of Git to perform GPG signature verification:
verify_signed_buffer() and check_signature(). Now only
check_signature() is used.
The verify_signed_buffer() function doesn't guard against duplicate
signatures as described by Michał Górny [1]. Instead it only ensures a
non-erroneous exit code from GPG and the presence of at least one
GOODSIG status field. This stands in contrast with check_signature()
that returns an error if more than one signature is encountered.
The lower degree of verification makes the use of verify_signed_buffer()
problematic if callers don't parse and validate the various parts of the
GPG status message themselves. And processing these messages seems like
a task that should be reserved to gpg-interface.c with the function
check_signature().
Furthermore, the use of verify_signed_buffer() makes it difficult to
introduce new functionality that relies on the content of the GPG status
lines.
Now all operations that does signature verification share a single entry
point to gpg-interface.c. This makes it easier to propagate changed or
additional functionality in GPG signature verification to all parts of
Git, without having odd edge-cases that don't perform the same degree of
verification.
[1] https://dev.gentoo.org/~mgorny/articles/attack-on-git-signature-verification.html
Signed-off-by: Hans Jerry Illikainen <hji@dyntopia.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Band-aid fixes for two fallouts from switching the default "rebase"
backend.
* en/rebase-backend:
git-rebase.txt: highlight backend differences with commit rewording
sequencer: clear state upon dropping a become-empty commit
i18n: unmark a message in rebase.c
In the future, we're going to want to use the branch info in
checkout_worktree, so let's pass the whole struct branch_info down, not
just the revision name. We hoist the definition of struct branch_info
so it's in scope.
Signed-off-by: brian m. carlson <bk2204@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit v2.25.0-4-ge98c4269c8 (rebase (interactive-backend): fix handling
of commits that become empty, 2020-02-15) marked "{drop,keep,ask}" for
translation, but this message should not be changed.
Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Often novice Git users forget to say "pull --rebase" and end up with an
unnecessary merge from upstream. What they usually want is either "pull
--rebase" in the simpler cases, or "pull --ff-only" to update the copy
of main integration branches, and rebase their work separately. The
pull.rebase configuration variable exists to help them in the simpler
cases, but there is no mechanism to make these users aware of it.
Issue a warning message when no --[no-]rebase option from the command
line and no pull.rebase configuration variable is given. This will
inconvenience those who never want to "pull --rebase", who haven't had
to do anything special, but the cost of the inconvenience is paid only
once per user, which should be a reasonable cost to help a number of new
users.
Signed-off-by: Alex Henrie <alexhenrie24@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Together with the previous commits, this commit fully fixes the problem
of using shared buffer for `real_path()` in `get_superproject_working_tree()`.
Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Returning a shared buffer invites very subtle bugs due to reentrancy or
multi-threading, as demonstrated by the previous patch.
There was an unfinished effort to abolish this [1].
Let's finally rid of `real_path()`, using `strbuf_realpath()` instead.
This patch uses a local `strbuf` for most places where `real_path()` was
previously called.
However, two places return the value of `real_path()` to the caller. For
them, a `static` local `strbuf` was added, effectively pushing the
problem one level higher:
read_gitfile_gently()
get_superproject_working_tree()
[1] https://lore.kernel.org/git/1480964316-99305-1-git-send-email-bmwill@google.com/
Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git am --short-current-patch" is a way to show the piece of e-mail
for the stopped step, which is not suitable to directly feed "git
apply" (it is designed to be a good "git am" input). It learned a
new option to show only the patch part.
* pb/am-show-current-patch:
am: support --show-current-patch=diff to retrieve .git/rebase-apply/patch
am: support --show-current-patch=raw as a synonym for--show-current-patch
am: convert "resume" variable to a struct
parse-options: convert "command mode" to a flag
parse-options: add testcases for OPT_CMDMODE()
`real_path()` returns result from a shared buffer, inviting subtle
reentrance bugs. One of these bugs occur when invoked this way:
set_git_dir(real_path(git_dir))
In this case, `real_path()` has reentrance:
real_path
read_gitfile_gently
repo_set_gitdir
setup_git_env
set_git_dir_1
set_git_dir
Later, `set_git_dir()` uses its now-dead parameter:
!is_absolute_path(path)
Fix this by using a dedicated `strbuf` to hold `strbuf_realpath()`.
Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove the stash.useBuiltin setting which was added as an escape hatch
to disable the builtin version of stash first released with Git 2.22.
Carrying the legacy version is a maintenance burden, and has in fact
become out of date failing a test since the 2.23 release, without
anyone noticing until now. So users would be getting a hint to fall
back to a potentially buggy version of the tool.
We used to shell out to git config to get the useBuiltin configuration
to avoid changing any global state before spawning legacy-stash.
However that is no longer necessary, so just use the 'git_config'
function to get the setting instead.
Similar to what we've done in d03ebd411c ("rebase: remove the
rebase.useBuiltin setting", 2019-03-18), where we remove the
corresponding setting for rebase, we leave the documentation in place,
so people can refer back to it when searching for it online, and so we
can refer to it in the commit message.
Signed-off-by: Thomas Gummerer <t.gummerer@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git merge signed-tag" while lacking the public key started to say
"No signature", which was utterly wrong. This regression has been
reverted.
* hi/gpg-use-check-signature:
Revert "gpg-interface: prefer check_signature() for GPG verification"
"git describe" in a repository with multiple root commits sometimes
gave up looking for the best tag to describe a given commit with
too early, which has been adjusted.
* be/describe-multiroot:
describe: don't abort too early when searching tags
"git clone --recurse-submodules --single-branch" now uses the same
single-branch option when cloning the submodules.
* es/recursive-single-branch-clone:
clone: pass --single-branch during --recurse-submodules
submodule--helper: use C99 named initializer
Code cleanup to use "struct object_id" more by replacing use of
"char *sha1"
* jk/nth-packed-object-id:
packfile: drop nth_packed_object_sha1()
packed_object_info(): use object_id internally for delta base
packed_object_info(): use object_id for returning delta base
pack-check: push oid lookup into loop
pack-check: convert "internal error" die to a BUG()
pack-bitmap: use object_id when loading on-disk bitmaps
pack-objects: use object_id struct in pack-reuse code
pack-objects: convert oe_set_delta_ext() to use object_id
pack-objects: read delta base oid into object_id struct
nth_packed_object_oid(): use customary integer return
"git rebase BASE BRANCH" rebased/updated the tip of BRANCH and
checked it out, even when the BRANCH is checked out in a different
worktree. This has been corrected.
* es/do-not-let-rebase-switch-to-protected-branch:
rebase: refuse to switch to branch already checked out elsewhere
t3400: make test clean up after itself
"git push" should stop from updating a branch that is checked out
when receive.denyCurrentBranch configuration is set, but it failed
to pay attention to checkouts in secondary worktrees. This has
been corrected.
* hv/receive-denycurrent-everywhere:
t2402: test worktree path when called in .git directory
receive.denyCurrentBranch: respect all worktrees
t5509: use a bare repository for test push target
get_main_worktree(): allow it to be called in the Git directory
In rare cases "git worktree add <path>" could think that <path>
was already a registered worktree even when it wasn't and refuse
to add the new worktree. This has been corrected.
* es/worktree-avoid-duplication-fix:
worktree: don't allow "add" validation to be fooled by suffix matching
worktree: add utility to find worktree by pathname
worktree: improve find_worktree() documentation
Underlying machinery of "git bisect--helper" is being refactored
into pieces that are more easily reused.
* mr/bisect-in-c-1:
bisect: libify `bisect_next_all`
bisect: libify `handle_bad_merge_base` and its dependents
bisect: libify `check_good_are_ancestors_of_bad` and its dependents
bisect: libify `check_merge_bases` and its dependents
bisect: libify `bisect_checkout`
bisect: libify `exit_if_skipped_commits` to `error_if_skipped*` and its dependents
bisect--helper: return error codes from `cmd_bisect__helper()`
bisect: add enum to represent bisect returning codes
bisect--helper: introduce new `decide_next()` function
bisect: use the standard 'if (!var)' way to check for 0
bisect--helper: change `retval` to `res`
bisect--helper: convert `vocab_*` char pointers to char arrays
"git sparse-checkout" learned a new "add" subcommand.
* ds/sparse-add:
sparse-checkout: allow one-character directories in cone mode
sparse-checkout: work with Windows paths
sparse-checkout: create 'add' subcommand
sparse-checkout: extract pattern update from 'set' subcommand
sparse-checkout: extract add_patterns_from_input()
change the advise call in tag library from advise() to
advise_if_enabled() to construct an example of the usage of
the new API.
Signed-off-by: Heba Waly <heba.waly@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the next commit we're adding another config variable to be read
from 'git_stash_config', that is valid for the top level command
instead of just a subset. Move the 'git_config' invocation for
'git_stash_config' to the top-level to prepare for that.
Signed-off-by: Thomas Gummerer <t.gummerer@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix for a bug revealed by a recent change to make the protocol v2
the default.
* ds/partial-clone-fixes:
partial-clone: avoid fetching when looking for objects
partial-clone: demonstrate bugs in partial fetch
"git rebase" has learned to use the merge backend (i.e. the
machinery that drives "rebase -i") by default, while allowing
"--apply" option to use the "apply" backend (e.g. the moral
equivalent of "format-patch piped to am"). The rebase.backend
configuration variable can be set to customize.
* en/rebase-backend:
rebase: rename the two primary rebase backends
rebase: change the default backend from "am" to "merge"
rebase: make the backend configurable via config setting
rebase tests: repeat some tests using the merge backend instead of am
rebase tests: mark tests specific to the am-backend with --am
rebase: drop '-i' from the reflog for interactive-based rebases
git-prompt: change the prompt for interactive-based rebases
rebase: add an --am option
rebase: move incompatibility checks between backend options a bit earlier
git-rebase.txt: add more details about behavioral differences of backends
rebase: allow more types of rebases to fast-forward
t3432: make these tests work with either am or merge backends
rebase: fix handling of restrict_revision
rebase: make sure to pass along the quiet flag to the sequencer
rebase, sequencer: remove the broken GIT_QUIET handling
t3406: simplify an already simple test
rebase (interactive-backend): fix handling of commits that become empty
rebase (interactive-backend): make --keep-empty the default
t3404: directly test the behavior of interest
git-rebase.txt: update description of --allow-empty-message
"git check-ignore" did not work when the given path is explicitly
marked as not ignored with a negative entry in the .gitignore file.
* en/check-ignore:
check-ignore: fix documentation and implementation to match
The object reachability bitmap machinery and the partial cloning
machinery were not prepared to work well together, because some
object-filtering criteria that partial clones use inherently rely
on object traversal, but the bitmap machinery is an optimization
to bypass that object traversal. There however are some cases
where they can work together, and they were taught about them.
* jk/object-filter-with-bitmap:
rev-list --count: comment on the use of count_right++
pack-objects: support filters with bitmaps
pack-bitmap: implement BLOB_LIMIT filtering
pack-bitmap: implement BLOB_NONE filtering
bitmap: add bitmap_unset() function
rev-list: use bitmap filters for traversal
pack-bitmap: basic noop bitmap filter infrastructure
rev-list: allow commit-only bitmap traversals
t5310: factor out bitmap traversal comparison
rev-list: allow bitmaps when counting objects
rev-list: make --count work with --objects
rev-list: factor out bitmap-optimized routines
pack-bitmap: refuse to do a bitmap traversal with pathspecs
rev-list: fallback to non-bitmap traversal when filtering
pack-bitmap: fix leak of haves/wants object lists
pack-bitmap: factor out type iterator initialization
This reverts commit 72b006f4bf, which
breaks the end-user experience when merging a signed tag without
having the public key. We should report "can't check because we
have no public key", but the code with this change claimed that
there was no signature.
When searching the commit graph for tag candidates, `git-describe`
will stop as soon as there is only one active branch left and
it already found an annotated tag as a candidate.
This works well as long as all branches eventually connect back
to a common root, but if the tags are found across branches
with no common ancestor
B
o----.
\
o-----o---o----x
A
it can happen that the search on one branch terminates prematurely
because a tag was found on another, independent branch. This scenario
isn't quite as obscure as it sounds, since cloning with a limited
depth often introduces many independent "dead ends" into the commit
graph.
The help text of `git-describe` states pretty clearly that when
describing a commit D, the number appended to the emitted tag X should
correspond to the number of commits found by `git log X..D`.
Thus, this commit modifies the stopping condition to only abort
the search when only one branch is left to search *and* all current
best candidates are descendants from that branch.
For repositories with a single root, this condition is always
true: When the search is reduced to a single active branch, the
current commit must be an ancestor of *all* tag candidates. This
means that in the common case, this change will have no negative
performance impact since the same number of commits as before will
be traversed.
Signed-off-by: Benno Evers <benno@bmevers.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When `options.switch_to' is set, `options.orig_head' is populated right
after with the object name the ref/commit argument points at.
Therefore, there is no need to parse `switch_to' again.
Signed-off-by: Alban Gruin <alban.gruin@gmail.com>
Acked-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git remote rename X Y" needs to adjust configuration variables
(e.g. branch.<name>.remote) whose value used to be X to Y.
branch.<name>.pushRemote is now also updated.
* bw/remote-rename-update-config:
remote rename/remove: gently handle remote.pushDefault config
config: provide access to the current line number
remote rename/remove: handle branch.<name>.pushRemote config values
remote: clean-up config callback
remote: clean-up by returning early to avoid one indentation
pull --rebase/remote rename: document and honor single-letter abbreviations rebase types
Previously, performing "git clone --recurse-submodules --single-branch"
resulted in submodules cloning all branches even though the superproject
cloned only one branch. Pipe --single-branch through the submodule
helper framework to make it to 'clone' later on.
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Start using a named initializer list for SUBMODULE_UPDATE_CLONE_INIT, as
the struct is becoming cumbersome for a typical struct initializer list.
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git worktree add <path>" performs various checks before approving
<path> as a valid location for the new worktree. Aside from ensuring
that <path> does not already exist, one of the questions it asks is
whether <path> is already a registered worktree. To perform this check,
it queries find_worktree() and disallows the "add" operation if
find_worktree() finds a match for <path>. As a convenience, however,
find_worktree() casts an overly wide net to allow users to identify
worktrees by shorthand in order to keep typing to a minimum. For
instance, it performs suffix matching which, given subtrees "foo/bar"
and "foo/baz", can correctly select the latter when asked only for
"baz".
"add" validation knows the exact path it is interrogating, so this sort
of heuristic-based matching is, at best, questionable for this use-case
and, at worst, may may accidentally interpret <path> as matching an
existing worktree and incorrectly report it as already registered even
when it isn't. (In fact, validate_worktree_add() already contains a
special case to avoid accidentally matching against the main worktree,
precisely due to this problem.)
Avoid the problem of potential accidental matching against an existing
worktree by instead taking advantage of find_worktree_by_path() which
matches paths deterministically, without applying any sort of magic
shorthand matching performed by find_worktree().
Reported-by: Cameron Gunnin <cameron.gunnin@synopsys.com>
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If a caller sets the object_info.delta_base_sha1 to a non-NULL pointer,
we'll write the oid of the object's delta base to it. But we can
increase our type safety by switching this to a real object_id struct.
All of our callers are just pointing into the hash member of an
object_id anyway, so there's no inconvenience.
Note that we do still keep it as a pointer-to-struct, because the NULL
sentinel value tells us whether the caller is even interested in the
information.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the pack-reuse code is dumping an OFS_DELTA entry to a client that
doesn't support it, we re-write it as a REF_DELTA. To do so, we use
nth_packed_object_sha1() to get the oid, but that function is soon going
away in favor of the more type-safe nth_packed_object_id(). Let's switch
now in preparation.
Note that this does incur an extra hash copy (from the pack idx mmap to
the object_id and then to the output, rather than straight from mmap to
the output). But this is not worth worrying about. It's probably not
measurable even when it triggers, and this is fallback code that we
expect to trigger very rarely (since everybody supports OFS_DELTA these
days anyway).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We already store an object_id internally, and now our sole caller also
has one. Let's stop passing around the internal hash array, which adds a
bit of type safety.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we're considering reusing an on-disk delta, we get the oid of the
base as a pointer to unsigned char bytes of the hash, either into the
packfile itself (for REF_DELTA) or into the pack idx (using the revindex
to convert the offset into an index entry).
Instead, we'd prefer to use a more type-safe object_id as much as
possible. We can get the pack idx using nth_packed_object_id() instead.
For the packfile bytes, we can copy them out using oidread().
This doesn't even incur an extra copy overall, since the next thing we'd
always do with that pointer is pass it to can_reuse_delta(), which needs
an object_id anyway (and called oidread() itself). So this patch also
converts that function to take the object_id directly.
Note that we did previously use NULL as a sentinel value when the object
isn't a delta. We could probably get away with using the null oid for
this, but instead we'll use an explicit boolean flag, which should make
things more obvious for people reading the code later.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Our nth_packed_object_sha1() function returns NULL for error. So when we
wrapped it with nth_packed_object_oid(), we kept the same semantics. But
it's a bit funny, because the caller actually passes in an out
parameter, and the pointer we return is just that same struct they
passed to us (or NULL).
It's not too terrible, but it does make the interface a little
non-idiomatic. Let's switch to our usual "0 for success, negative for
error" return value. Most callers either don't check it, or are
trivially converted. The one that requires the biggest change is
actually improved, as we can ditch an extra aliased pointer variable.
Since we are changing the interface in a subtle way that the compiler
wouldn't catch, let's also change the name to catch any topics in
flight. We can drop the 'o' and make it nth_packed_object_id(). That's
slightly shorter, but also less redundant since the 'o' stands for
"object" already.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The invocation "git rebase <upstream> <branch>" switches to <branch>
before performing the rebase operation. However, unlike git-switch,
git-checkout, and git-worktree which all refuse to switch to a branch
that is already checked out in some other worktree, git-rebase switches
to <branch> unconditionally. Curb this careless behavior by making
git-rebase also refuse to switch to a branch checked out elsewhere.
Reported-by: Mike Hommey <mh@glandium.org>
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The receive.denyCurrentBranch config option controls what happens if
you push to a branch that is checked out into a non-bare repository.
By default, it rejects it. It can be disabled via `ignore` or `warn`.
Another yet trickier option is `updateInstead`.
However, this setting was forgotten when the git worktree command was
introduced: only the main worktree's current branch is respected.
With this change, all worktrees are respected.
That change also leads to revealing another bug,
i.e. `receive.denyCurrentBranch = true` was ignored when pushing into a
non-bare repository's unborn current branch using ref namespaces. As
`is_ref_checked_out()` returns 0 which means `receive-pack` does not get
into conditional statement to switch `deny_current_branch` accordingly
(ignore, warn, refuse, unconfigured, updateInstead).
receive.denyCurrentBranch uses the function `refs_resolve_ref_unsafe()`
(called via `resolve_refdup()`) to resolve the symbolic ref HEAD, but
that function fails when HEAD does not point at a valid commit.
As we replace the call to `refs_resolve_ref_unsafe()` with
`find_shared_symref()`, which has no problem finding the worktree for a
given branch even if it is unborn yet, this bug is fixed at the same
time: receive.denyCurrentBranch now also handles worktrees with unborn
branches as intended even while using ref namespaces.
Helped-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Hariom Verma <hariom18599@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The transition plan anticipates that we will allow signatures using
multiple algorithms in a single commit. In order to do so, we need to
use a different header per algorithm so that it will be obvious over
which data to compute the signature.
The transition plan specifies that we should use "gpgsig-sha256", so
wire up the commit code such that it can write and parse the current
algorithm, and it can remove the headers for any algorithm when creating
a new commit. Add tests to ensure that we write using the right header
and that git fsck doesn't reject these commits.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we perform a clone, we won't know the remote side's hash algorithm
until we've read the heads. Consequently, we'll need to rewrite the
repository format version and hash algorithm once we know what the
remote side has. Move the code that does this into its own function so
that we can call it from clone in the future.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
For the foreseeable future, SHA-1 will be the default algorithm for Git.
However, when running the testsuite, we want to be able to test an
arbitrary algorithm. It would be quite burdensome and very untidy to
have to specify the algorithm we'd like to test every time we
initialized a new repository somewhere in the testsuite, so add an
environment variable to allow us to specify the default hash algorithm
for Git.
This has the benefit that we can set it once for the entire testsuite
and not have to think about it. In the future, users can also use it to
set the default for their repositories if they would like to do so.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Allow the user to specify the hash algorithm on the command line by
using the --object-format option to git init. Validate that the user is
not attempting to reinitialize a repository with a different hash
algorithm. Ensure that if we are writing a non-SHA-1 repository that we
set the repository version to 1 and write the objectFormat extension.
Restrict this option to work only when ENABLE_SHA256 is set until the
codebase is in a situation to fully support this.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In some cases, we will want to not only check the repository format, but
extract the information that we've gained. To do so, allow
check_repository_format to take a pointer to struct repository_format.
Allow passing NULL for this argument if we're not interested in the
information, and pass NULL for all existing callers. A future patch
will make use of this information.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Avoid hard-coding a hash size, instead preferring to use the_hash_algo.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We can check if certain characters are present in a string by calling
strchr(3) on each of them, or we can pass them all to a single
strpbrk(3) call. The latter is shorter, less repetitive and slightly
more efficient, so let's do that instead.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When using partial clone, find_non_local_tags() in builtin/fetch.c
checks each remote tag to see if its object also exists locally. There
is no expectation that the object exist locally, but this function
nevertheless triggers a lazy fetch if the object does not exist. This
can be extremely expensive when asking for a commit, as we are
completely removed from the context of the non-existent object and
thus supply no "haves" in the request.
6462d5eb9a (fetch: remove fetch_if_missing=0, 2019-11-05) removed a
global variable that prevented these fetches in favor of a bitflag.
However, some object existence checks were not updated to use this flag.
Update find_non_local_tags() to use OBJECT_INFO_SKIP_FETCH_OBJECT in
addition to OBJECT_INFO_QUICK. The _QUICK option only prevents
repreparing the pack-file structures. We need to be extremely careful
about supplying _SKIP_FETCH_OBJECT when we expect an object to not exist
due to updated refs.
This resolves a broken test in t5616-partial-clone.sh.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
An annotated tag has two names: where it sits in the refs/tags
hierarchy and the tagname recorded in the "tag" field in the object
itself. They usually should match.
Since 212945d4 ("Teach git-describe to verify annotated tag names
before output", 2008-02-28), a commit described using an annotated
tag bases its name on the tagname from the object. While this was a
deliberate design decision to make it easier to converse about tags
with others, even if the tags happen to be fetched to a different
name than it was given by its creator, it had one downside.
The output from "git describe", at least in the modern Git, should
be usable as an object name to name the exact commit given to the
"git describe" command. Using the tagname, when two names differ,
breaks this property, when describing a commit that is directly
pointed at by such a tag. An annotated tag Bob made as "v1.0" may
sit at "refs/tags/v1.0-bob" in the ref hierarchy, and output from
"git describe v1.0-bob^0" would say "v1.0", but there may not be
any tag at "refs/tags/v1.0" locally or there may be another tag that
points at a different object.
Note that this won't be a problem if a commit being described is not
directly pointed at by such a mislocated tag. In the example in the
previous paragraph, describing a commit whose parent is v1.0-bob
would result in "v1.0" (i.e. the tagname taken from the tag object)
followed by "-1-gXXXXX" where XXXXX is the abbreviated object name,
and a string that ends with "-g" followed by a hexadecimal string is
an object name for the object whose name begins with hexadecimal
string (as long as it is unique), so it does not matter if the
leading part is "v1.0" or "v1.0-bob".
Show the name in the long format, i.e. with "-0-gXXXXX" suffix, when
the name we give is based on a mislocated annotated tag to ensure
that the output can be used as the object name for the object
originally given to the command to fix the issue.
While at it, remove an overly cautious dead code to protect against
an annotated tag object without the tagname. Such a tag is filtered
out much earlier in the codeflow, and will not reach this part of
the code.
Helped-by: Matheus Tavares <matheus.bernardino@usp.br>
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When "git am --show-current-patch" was added in commit 984913a210 ("am:
add --show-current-patch", 2018-02-12), "git am" started recommending it
as a replacement for .git/rebase-merge/patch. Unfortunately the suggestion
is somewhat misguided; for example, the output of "git am --show-current-patch"
cannot be passed to "git apply" if it is encoded as quoted-printable
or base64. Add a new mode to "git am --show-current-patch" in order to
straighten the suggestion.
Reported-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When "git am --show-current-patch" was added in commit 984913a210 ("am:
add --show-current-patch", 2018-02-12), "git am" started recommending it
as a replacement for .git/rebase-merge/patch. Unfortunately the suggestion
is somewhat misguided; for example, the output "git am --show-current-patch"
cannot be passed to "git apply" if it is encoded as quoted-printable or
base64. To simplify worktree operations and to avoid that users poke into
.git, it would be better if "git am" also provided a mode that copies
.git/rebase-merge/patch to stdout.
One possibility could be to have completely separate options, introducing
for example --show-current-message (for .git/rebase-apply/NNNN)
and --show-current-diff (for .git/rebase-apply/patch), while possibly
deprecating --show-current-patch.
That would even remove the need for the first two patches in the series.
However, the long common prefix would have prevented using an abbreviated
option such as "--show". Therefore, I chose instead to add a string
argument to --show-current-patch. The new argument is optional, so that
"git am --show-current-patch"'s behavior remains backwards-compatible.
The next choice to make is how to handle multiple --show-current-patch
options. Right now, something like "git am --abort --show-current-patch"
is rejected, and the previous suggestion would likewise have naturally
rejected a command line like
git am --show-current-message --show-current-diff
Therefore, I decided to also reject for example
git am --show-current-patch=diff --show-current-patch=raw
In other words the whole of --show-current-patch=xxx (including the
optional argument) is treated as the command mode. I found this to be
more consistent and intuitive, even though it differs from the usual
"last one wins" semantics of the git command line.
Add the code to parse submodes based on the above design, where for now
"raw" is the only valid submode. "raw" prints the full e-mail message
just like "git am --show-current-patch".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This will allow stashing the submode of --show-current-patch from a
callback function. Using a struct will allow accessing both fields from
outside cmd_am (through container_of).
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Decisions taken for simplicity:
1) For now, `--pathspec-from-file` is declared incompatible with
`--patch`, even when <file> is not `-`. Such use case is not
really expected.
2) It is not allowed to pass pathspec in both args and file.
Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Eliminate crude option parsing and rely on real parsing instead, because
1) Crude parsing is crude, for example it's not capable of
handling things like `git stash -m Message`
2) Adding options in two places is inconvenient and prone to bugs
As a side result, the case of `git stash -m Message` gets fixed.
Also give a good error message instead of just throwing usage at user.
----
Some review of what's been happening to this code:
Before [1], `git-stash.sh` only verified that all args begin with `-` :
# The default command is "push" if nothing but options are given
seen_non_option=
for opt
do
case "$opt" in
--) break ;;
-*) ;;
*) seen_non_option=t; break ;;
esac
done
Later, [1] introduced the duplicate code I'm now removing, also making
the previous test more strict by white-listing options.
----
[1] Commit 40af1468 ("stash: convert `stash--helper.c` into `stash.c`" 2019-02-26)
Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Decisions taken for simplicity:
1) It is not allowed to pass pathspec in both args and file.
Adjustments were needed for `if (!argc)` block:
This code actually means "pathspec is not present". Previously, pathspec
could only come from commandline arguments, so testing for `argc` was a
valid way of testing for the presence of pathspec. But this is no longer
true with `--pathspec-from-file`.
During the entire `--pathspec-from-file` story, I tried to keep its
behavior very close to giving pathspec on commandline, so that switching
from one to another doesn't involve any surprises.
However, throwing usage at user in the case of empty
`--pathspec-from-file` would puzzle because there's nothing wrong with
"usage" (that is, argc/argv array).
On the other hand, throwing usage in the old case also feels bad to me.
While it's less of a puzzle, I (as user) never liked the experience of
comparing my commandline to "usage", trying to spot a difference. Since
it's already known what the error is, it feels a lot better to give that
specific error to user.
Judging from [1] it doesn't seem that showing usage in this case was
important (the patch was to avoid segfault), and it doesn't fit into how
other commands react to empty pathspec (see for example `git add` with a
custom message).
Therefore, I decided to show new error text in both cases. In order to
continue testing for error early, I moved `parse_pathspec()` higher. Now
it happens before `read_cache()` / `hold_locked_index()` /
`setup_work_tree()`, which shouldn't cause any issues.
[1] Commit 7612a1ef ("git-rm: honor -n flag" 2006-06-09)
Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since we want to get rid of git-bisect.sh, it would be necessary to
convert those exit() calls to return statements so that errors can be
reported.
Emulate try catch in C by converting `exit(<positive-value>)` to
`return <negative-value>`. Follow POSIX conventions to return
<negative-value> to indicate error.
Code that turns BISECT_INTERNAL_SUCCESS_MERGE_BASE (-11)
to BISECT_OK (0) from `check_good_are_ancestors_of_bad()` has been moved to
`cmd_bisect__helper()`.
Update all callers to handle the error returns.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Pranit Bauva <pranit.bauva@gmail.com>
Signed-off-by: Tanushree Tumane <tanushreetumane@gmail.com>
Signed-off-by: Miriam Rubio <mirucam@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since we want to get rid of git-bisect.sh, it would be necessary
to convert bisect.c exit() calls to return statements so
that errors can be reported. Let's prepare for that by making
it possible to return different error codes than just 0 or 1.
Different error codes might enable a bisecting script calling the
bisect command that uses this function to do different things
depending on the exit status of the bisect command.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Tanushree Tumane <tanushreetumane@gmail.com>
Signed-off-by: Miriam Rubio <mirucam@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let's refactor code from bisect_next_check() into a new
decide_next() helper function.
This removes some goto statements and makes the code simpler,
clearer and easier to understand.
While at it `bad_ref` and `good_glob` are not const any more
to void casting them inside `free()`.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Tanushree Tumane <tanushreetumane@gmail.com>
Signed-off-by: Miriam Rubio <mirucam@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let's rename variable retval to res, so that variable names
in bisect--helper.c are more consistent.
After this change, there are 110 occurrences of res in the file
and zero of retval, while there were 26 instances of retval before.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Tanushree Tumane <tanushreetumane@gmail.com>
Signed-off-by: Miriam Rubio <mirucam@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of using a pointer that points at a constant string,
just give name directly to the constant string; this way, we
do not have to allocate a pointer variable in addition to
the string we want to use.
Let's convert `vocab_bad` and `vocab_good` char pointers to char arrays.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Miriam Rubio <mirucam@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
check-ignore has two different modes, and neither of these modes has an
implementation that matches the documentation. These modes differ in
whether they just print paths or whether they also print the final
pattern matched by the path. The fix is different for both modes, so
I'll discuss both separately.
=== First (default) mode ===
The first mode is documented as:
For each pathname given via the command-line or from a file via
--stdin, check whether the file is excluded by .gitignore (or other
input files to the exclude mechanism) and output the path if it is
excluded.
However, it fails to do this because it did not account for negated
patterns. Commands other than check-ignore verify exclusion rules via
calling
... -> treat_one_path() -> is_excluded() -> last_matching_pattern()
while check-ignore has a call path of the form:
... -> check_ignore() -> last_matching_pattern()
The fact that the latter does not include the call to is_excluded()
means that it is susceptible to to messing up negated patterns (since
that is the only significant thing is_excluded() adds over
last_matching_pattern()). Unfortunately, we can't make it just call
is_excluded(), because the same codepath is used by the verbose mode
which needs to know the matched pattern in question. This brings us
to...
=== Second (verbose) mode ===
The second mode, known as verbose mode, references the first in the
documentation and says:
Also output details about the matching pattern (if any) for each
given pathname. For precedence rules within and between exclude
sources, see gitignore(5).
The "Also" means it will print patterns that match the exclude rules as
noted for the first mode, and also print which pattern matches. Unless
more information is printed than just pathname and pattern (which is not
done), this definition is somewhat ill-defined and perhaps even
self-contradictory for negated patterns: A path which matches a negated
exclude pattern is NOT excluded and thus shouldn't be printed by the
former logic, while it certainly does match one of the explicit patterns
and thus should be printed by the latter logic.
=== Resolution ==
Since the second mode exists to find out which pattern matches given
paths, and showing the user a pattern that begins with a '!' is
sufficient for them to figure out whether the pattern is excluded, the
existing behavior is desirable -- we just need to update the
documentation to match the implementation (i.e. it is about printing
which pattern is matched by paths, not about showing which paths are
excluded).
For the first or default mode, users just want to know whether a pattern
is excluded. As such, the existing documentation is desirable; change
the implementation to match the documented behavior.
Finally, also adjust a few tests in t0008 that were caught up by this
discrepancy in how negated paths were handled.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git config" learned to show in which "scope", in addition to in
which file, each config setting comes from.
* mr/show-config-scope:
config: add '--show-scope' to print the scope of a config value
submodule-config: add subomdule config scope
config: teach git_config_source to remember its scope
config: preserve scope in do_git_config_sequence
config: clarify meaning of command line scoping
config: split repo scope to local and worktree
config: make scope_name non-static and rename it
t1300: create custom config file without special characters
t1300: fix over-indented HERE-DOCs
config: fix typo in variable name
Memory footprint and performance of "git name-rev" has been
improved.
* rs/name-rev-memsave:
name-rev: sort tip names before applying
name-rev: release unused name strings
name-rev: generate name strings only if they are better
name-rev: pre-size buffer in get_parent_name()
name-rev: factor out get_parent_name()
name-rev: put struct rev_name into commit slab
name-rev: don't _peek() in create_or_update_name()
name-rev: don't leak path copy in name_ref()
name-rev: respect const qualifier
name-rev: remove unused typedef
name-rev: rewrite create_or_update_name()
Two related changes, with separate rationale for each:
Rename the 'interactive' backend to 'merge' because:
* 'interactive' as a name caused confusion; this backend has been used
for many kinds of non-interactive rebases, and will probably be used
in the future for more non-interactive rebases than interactive ones
given that we are making it the default.
* 'interactive' is not the underlying strategy; merging is.
* the directory where state is stored is not called
.git/rebase-interactive but .git/rebase-merge.
Rename the 'am' backend to 'apply' because:
* Few users are familiar with git-am as a reference point.
* Related to the above, the name 'am' makes sentences in the
documentation harder for users to read and comprehend (they may read
it as the verb from "I am"); avoiding this difficult places a large
burden on anyone writing documentation about this backend to be very
careful with quoting and sentence structure and often forces
annoying redundancy to try to avoid such problems.
* Users stumble over pronunciation ("am" as in "I am a person not a
backend" or "am" as in "the first and thirteenth letters in the
alphabet in order are "A-M"); this may drive confusion when one user
tries to explain to another what they are doing.
* While "am" is the tool driving this backend, the tool driving git-am
is git-apply, and since we are driving towards lower-level tools
for the naming of the merge backend we may as well do so here too.
* The directory where state is stored has never been called
.git/rebase-am, it was always called .git/rebase-apply.
For all the reasons listed above:
* Modify the documentation to refer to the backends with the new names
* Provide a brief note in the documentation connecting the new names
to the old names in case users run across the old names anywhere
(e.g. in old release notes or older versions of the documentation)
* Change the (new) --am command line flag to --apply
* Rename some enums, variables, and functions to reinforce the new
backend names for us as well.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A large variety of rebase types are supported by the interactive
machinery, not just the explicitly interactive ones. These all share
the same code and write the same reflog messages, but the "-i" moniker
in those messages doesn't really have much meaning. It also becomes
somewhat distracting once we switch the default from the am-backend to
the interactive one. Just remove the "-i" from these messages.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently, this option doesn't do anything except error out if any
options requiring the interactive-backend are also passed. However,
when we make the default backend configurable later in this series, this
flag will provide a way to override the config setting.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the past, we dis-allowed rebases using the interactive backend from
performing a fast-forward to short-circuit the rebase operation. This
made sense for explicitly interactive rebases and some implicitly
interactive rebases, but certainly became overly stringent when the
merge backend was re-implemented via the interactive backend.
Just as the am-based rebase has always had to disable the fast-forward
based on a variety of conditions or flags (e.g. --signoff, --whitespace,
etc.), we need to do the same but now with a few more options. However,
continuing to use REBASE_FORCE for tracking this is problematic because
the interactive backend used it for a different purpose. (When
REBASE_FORCE wasn't set, the interactive backend would not fast-forward
the whole series but would fast-forward individual "pick" commits at the
beginning of the todo list, and then a squash or something would cause
it to start generating new commits.) So, introduce a new
allow_preemptive_ff flag contained within cmd_rebase() and use it to
track whether we are going to allow a pre-emptive fast-forward that
short-circuits the whole rebase.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
restrict_revision in the original shell script was an excluded revision
range. It is also treated that way by the am-backend. In the
conversion from shell to C (see commit 6ab54d17be ("rebase -i:
implement the logic to initialize $revisions in C", 2018-08-28)), the
interactive-backend accidentally treated it as a positive revision
rather than a negated one.
This was missed as there were no tests in the testsuite that tested an
interactive rebase with fork-point behavior.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The GIT_QUIET environment variable was used to signal the non-am
backends that the rebase should perform quietly. The preserve-merges
backend does not make use of the quiet flag anywhere (other than to
write out its state whenever it writes state), and this mechanism was
broken in the conversion from shell to C. Since this environment
variable was specifically designed for scripts and the only backend that
would still use it is no longer a script, just gut this code.
A subsequent commit will fix --quiet for the interactive/merge backend
in a different way.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As established in the previous commit and commit b00bf1c9a8
(git-rebase: make --allow-empty-message the default, 2018-06-27), the
behavior for rebase with different backends in various edge or corner
cases is often more happenstance than design. This commit addresses
another such corner case: commits which "become empty".
A careful reader may note that there are two types of commits which would
become empty due to a rebase:
* [clean cherry-pick] Commits which are clean cherry-picks of upstream
commits, as determined by `git log --cherry-mark ...`. Re-applying
these commits would result in an empty set of changes and a
duplicative commit message; i.e. these are commits that have
"already been applied" upstream.
* [become empty] Commits which are not empty to start, are not clean
cherry-picks of upstream commits, but which still become empty after
being rebased. This happens e.g. when a commit has changes which
are a strict subset of the changes in an upstream commit, or when
the changes of a commit can be found spread across or among several
upstream commits.
Clearly, in both cases the changes in the commit in question are found
upstream already, but the commit message may not be in the latter case.
When cherry-mark can determine a commit is already upstream, then
because of how cherry-mark works this means the upstream commit message
was about the *exact* same set of changes. Thus, the commit messages
can be assumed to be fully interchangeable (and are in fact likely to be
completely identical). As such, the clean cherry-pick case represents a
case when there is no information to be gained by keeping the extra
commit around. All rebase types have always dropped these commits, and
no one to my knowledge has ever requested that we do otherwise.
For many of the become empty cases (and likely even most), we will also
be able to drop the commit without loss of information -- but this isn't
quite always the case. Since these commits represent cases that were
not clean cherry-picks, there is no upstream commit message explaining
the same set of changes. Projects with good commit message hygiene will
likely have the explanation from our commit message contained within or
spread among the relevant upstream commits, but not all projects run
that way. As such, the commit message of the commit being rebased may
have reasoning that suggests additional changes that should be made to
adapt to the new base, or it may have information that someone wants to
add as a note to another commit, or perhaps someone even wants to create
an empty commit with the commit message as-is.
Junio commented on the "become-empty" types of commits as follows[1]:
WRT a change that ends up being empty (as opposed to a change that
is empty from the beginning), I'd think that the current behaviour
is desireable one. "am" based rebase is solely to transplant an
existing history and want to stop much less than "interactive" one
whose purpose is to polish a series before making it publishable,
and asking for confirmation ("this has become empty--do you want to
drop it?") is more appropriate from the workflow point of view.
[1] https://lore.kernel.org/git/xmqqfu1fswdh.fsf@gitster-ct.c.googlers.com/
I would simply add that his arguments for "am"-based rebases actually
apply to all non-explicitly-interactive rebases. Also, since we are
stating that different cases should have different defaults, it may be
worth providing a flag to allow users to select which behavior they want
for these commits.
Introduce a new command line flag for selecting the desired behavior:
--empty={drop,keep,ask}
with the definitions:
drop: drop commits which become empty
keep: keep commits which become empty
ask: provide the user a chance to interact and pick what to do with
commits which become empty on a case-by-case basis
In line with Junio's suggestion, if the --empty flag is not specified,
pick defaults as follows:
explicitly interactive: ask
otherwise: drop
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Different rebase backends have different treatment for commits which
start empty (i.e. have no changes relative to their parent), and the
--keep-empty option was added at some point to allow adjusting behavior.
The handling of commits which start empty is actually quite similar to
commit b00bf1c9a8 (git-rebase: make --allow-empty-message the default,
2018-06-27), which pointed out that the behavior for various backends is
often more happenstance than design. The specific change made in that
commit is actually quite relevant as well and much of the logic there
directly applies here.
It makes a lot of sense in 'git commit' to error out on the creation of
empty commits, unless an override flag is provided. However, once
someone determines that there is a rare case that merits using the
manual override to create such a commit, it is somewhere between
annoying and harmful to have to take extra steps to keep such
intentional commits around. Granted, empty commits are quite rare,
which is why handling of them doesn't get considered much and folks tend
to defer to existing (accidental) behavior and assume there was a reason
for it, leading them to just add flags (--keep-empty in this case) that
allow them to override the bad defaults. Fix the interactive backend so
that --keep-empty is the default, much like we did with
--allow-empty-message. The am backend should also be fixed to have
--keep-empty semantics for commits that start empty, but that is not
included in this patch other than a testcase documenting the failure.
Note that there was one test in t3421 which appears to have been written
expecting --keep-empty to not be the default as correct behavior. This
test was introduced in commit 00b8be5a4d ("add tests for rebasing of
empty commits", 2013-06-06), which was part of a series focusing on
rebase topology and which had an interesting original cover letter at
https://lore.kernel.org/git/1347949878-12578-1-git-send-email-martinvonz@gmail.com/
which noted
Your input especially appreciated on whether you agree with the
intent of the test cases.
and then went into a long example about how one of the many tests added
had several questions about whether it was correct. As such, I believe
most the tests in that series were about testing rebase topology with as
many different flags as possible and were not trying to state in general
how those flags should behave otherwise.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The code to compute the commit-graph has been taught to use a more
robust way to tell if two object directories refer to the same
thing.
* tb/commit-graph-object-dir:
commit-graph.h: use odb in 'load_commit_graph_one_fd_st'
commit-graph.c: remove path normalization, comparison
commit-graph.h: store object directory in 'struct commit_graph'
commit-graph.h: store an odb in 'struct write_commit_graph_context'
t5318: don't pass non-object directory to '--object-dir'
The index-pack code now diagnoses a bad input packstream that
records the same object twice when it is used as delta base; the
code used to declare a software bug when encountering such an
input, but it is an input error.
* jk/index-pack-dupfix:
index-pack: downgrade twice-resolved REF_DELTA to die()
The way "git submodule status" reports an initialized but not yet
populated submodule has not been reimplemented correctly when a
part of the "git submodule" command was rewritten in C, which has
been corrected.
* pk/status-of-uncloned-submodule:
t7400: testcase for submodule status on unregistered inner git repos
submodule: fix status of initialized but not cloned submodules
t7400: add a testcase for submodule status on empty dirs
Some codepaths were given a repository instance as a parameter to
work in the repository, but passed the_repository instance to its
callees, which has been cleaned up (somewhat).
* mt/use-passed-repo-more-in-funcs:
sha1-file: allow check_object_signature() to handle any repo
sha1-file: pass git_hash_algo to hash_object_file()
sha1-file: pass git_hash_algo to write_object_file_prepare()
streaming: allow open_istream() to handle any repo
pack-check: use given repo's hash_algo at verify_packfile()
cache-tree: use given repo's hash_algo at verify_one()
diff: make diff_populate_filespec() honor its repo argument
Some rough edges in the sparse-checkout feature, especially around
the cone mode, have been cleaned up.
* ds/sparse-checkout-harden:
sparse-checkout: fix cone mode behavior mismatch
sparse-checkout: improve docs around 'set' in cone mode
sparse-checkout: escape all glob characters on write
sparse-checkout: use C-style quotes in 'list' subcommand
sparse-checkout: unquote C-style strings over --stdin
sparse-checkout: write escaped patterns in cone mode
sparse-checkout: properly match escaped characters
sparse-checkout: warn on globs in cone patterns
sparse-checkout: detect short patterns
sparse-checkout: cone mode does not recognize "**"
sparse-checkout: fix documentation typo for core.sparseCheckoutCone
clone: fix --sparse option with URLs
sparse-checkout: create leading directories
t1091: improve here-docs
t1091: use check_files to reduce boilerplate
Unneeded connectivity check is now disabled in a partial clone when
fetching into it.
* jt/connectivity-check-optim-in-partial-clone:
fetch: forgo full connectivity check if --filter
connected: verify promisor-ness of partial clone
"git rebase -i" (and friends) used to unnecessarily check out the
tip of the branch to be rebased, which has been corrected.
* ag/rebase-avoid-unneeded-checkout:
rebase -i: stop checking out the tip of the branch to rebase
Traditionally, we avoided threaded grep while searching in objects
(as opposed to files in the working tree) as accesses to the object
layer is not thread-safe. This limitation is getting lifted.
* mt/threaded-grep-in-object-store:
grep: use no. of cores as the default no. of threads
grep: move driver pre-load out of critical section
grep: re-enable threads in non-worktree case
grep: protect packed_git [re-]initialization
grep: allow submodule functions to run in parallel
submodule-config: add skip_if_read option to repo_read_gitmodules()
grep: replace grep_read_mutex by internal obj read lock
object-store: allow threaded access to object reading
replace-object: make replace operations thread-safe
grep: fix racy calls in grep_objects()
grep: fix race conditions at grep_submodule()
grep: fix race conditions on userdiff calls
Two help messages given when "git add" notices the user gave it
nothing to add have been updated to use advise() API.
* hw/advice-add-nothing:
add: change advice config variables used by the add API
add: use advise function to display hints
"git grep --no-index" should not get affected by the contents of
the .gitmodules file but when "--recurse-submodules" is given or
the "submodule.recurse" variable is set, it did. Now these
settings are ignored in the "--no-index" mode.
* pb/do-not-recurse-grep-no-index:
grep: ignore --recurse-submodules if --no-index is given
"git restore --staged" did not correctly update the cache-tree
structure, resulting in bogus trees to be written afterwards, which
has been corrected.
* nd/switch-and-restore:
restore: invalidate cache-tree when removing entries with --staged
"git commit" gives output similar to "git status" when there is
nothing to commit, but without honoring the advise.statusHints
configuration variable, which has been corrected.
* hw/commit-advise-while-rejecting:
commit: honor advice.statusHints when rejecting an empty commit
Just as rev-list recently learned to combine filters and bitmaps, let's
do the same for pack-objects. The infrastructure is all there; we just
need to pass along our filter options, and the pack-bitmap code will
decide to use bitmaps or not.
This unsurprisingly makes things faster for partial clones of large
repositories (here we're cloning linux.git):
Test HEAD^ HEAD
------------------------------------------------------------------------------
5310.11: simulated partial clone 38.94(37.28+5.87) 11.06(11.27+4.07) -71.6%
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This just passes the filter-options struct to prepare_bitmap_walk().
Since the bitmap code doesn't actually support any filters yet, it will
fallback to the non-bitmap code if any --filter is specified. But this
lets us exercise that rejection code path, as well as getting us ready
to test filters via rev-list when we _do_ support them.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently you can't use object filters with bitmaps, but we plan to
support at least some filters with bitmaps. Let's introduce some
infrastructure that will help us do that:
- prepare_bitmap_walk() now accepts a list_objects_filter_options
parameter (which can be NULL for no filtering; all the current
callers pass this)
- we'll bail early if the filter is incompatible with bitmaps (just as
we would if there were no bitmaps at all). Currently all filters are
incompatible.
- we'll filter the resulting bitmap; since there are no supported
filters yet, this is always a noop.
There should be no behavior change yet, but we'll support some actual
filters in a future patch.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Ever since we added reachability bitmap support, we've been able to use
it with rev-list to get the full list of objects, like:
git rev-list --objects --use-bitmap-index --all
But you can't do so without --objects, since we weren't ready to just
show the commits. However, the internals of the bitmap code are mostly
ready for this: they avoid opening up trees when walking to fill in the
bitmaps. We just need to actually pass in the rev_info to
traverse_bitmap_commit_list() so it knows which types to bother
triggering our callback for.
For completeness, the perf test now covers both the existing --objects
case, as well as the new commits-only behavior (the objects one got way
faster when we introduced bitmaps, but obviously isn't improved now).
Here are numbers for linux.git:
Test HEAD^ HEAD
------------------------------------------------------------------------
5310.7: rev-list (commits) 8.29(8.10+0.19) 1.76(1.72+0.04) -78.8%
5310.8: rev-list (objects) 8.06(7.94+0.12) 8.14(7.94+0.13) +1.0%
That run was cheating a little, as I didn't have any commit-graph in the
repository, and we'd built it by default these days when running git-gc.
Here are numbers with a commit-graph:
Test HEAD^ HEAD
------------------------------------------------------------------------
5310.7: rev-list (commits) 0.70(0.58+0.12) 0.51(0.46+0.04) -27.1%
5310.8: rev-list (objects) 6.20(6.09+0.10) 6.27(6.16+0.11) +1.1%
Still an improvement, but a lot less impressive.
We could have the perf script remove any commit-graph to show the
out-sized effect, but it probably makes sense to leave it in what would
be a more typical setup.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The prior commit taught "--count --objects" to work without bitmaps. We
should be able to get the same answer much more quickly with bitmaps.
Note that we punt on the max_count case here. This perhaps _could_ be
made to work if we find all of the boundary commits and treat them as
UNINTERESTING, subtracting them (and their reachable objects) from the
set we return. That implies an actual commit traversal, but we'd still
be faster due to avoiding opening up any trees. Given the complexity and
the fact that anyone is unlikely to want this, it makes sense to just
fall back to the non-bitmap case for now.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The current behavior from "rev-list --count --objects" is nonsensical:
we enumerate all of the objects except commits, but then give a count of
commits. This wasn't planned, and is just what the code happens to do.
Instead, let's give the answer the user almost certainly wanted: the
full count of objects.
Note that there are more complicated cases around cherry-marking, etc.
We'll punt on those for now, but let the user know that we can't produce
an answer (rather than giving them something useless).
We'll test both the new feature as well as a vanilla --count of commits,
since that surprisingly doesn't seem to be covered in the existing
tests.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are a few operations in rev-list that are optimized for bitmaps.
Rather than having the code inline in cmd_rev_list(), let's move them
into helpers. This not only makes the flow of the main function simpler,
but it lets us replace the complex "can we do the optimization?"
conditionals with a series of early returns from the functions. That
also makes it easy to add comments explaining those conditions.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
rev-list has refused to use bitmaps with pathspec limiting since
c8a70d3509 (rev-list: disable --use-bitmap-index when pruning commits,
2015-07-01). But this is true not just for rev-list, but for anyone who
calls prepare_bitmap_walk(); the code isn't equipped to handle this
case. We never noticed because the only other callers would never pass
a pathspec limiter.
But let's push the check down into prepare_bitmap_walk() anyway. That's
a more logical place for it to live, as callers shouldn't need to know
the details (and must be prepared to fall back to a regular traversal
anyway, since there might not be bitmaps in the repository).
It would also prepare us for a day where this case _is_ handled, but
that's pretty unlikely. E.g., we could use bitmaps to generate the set
of commits, and then diff each commit to see if it matches the pathspec.
That would be slightly faster than a naive traversal that actually walks
the commits. But you'd probably do better still to make use of the newer
commit-graph feature to make walking the commits very cheap.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "--use-bitmap-index" option is usually aspirational: if we have
bitmaps and the request can be fulfilled more quickly using them we'll
do so, but otherwise fall back to a non-bitmap traversal.
The exception is object filtering, which explicitly dies if the two
options are combined. Let's convert this to the usual fallback behavior.
This is a minor convenience for now (since the caller can easily know
that --filter and --use-bitmap-index don't combine), but will become
much more useful as we start to support _some_ filters with bitmaps, but
not others.
The test infrastructure here is bigger than necessary for checking this
one small feature. But it will serve as the basis for more filtering
bitmap tests in future patches.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git grep --no-index" should not get affected by the contents of
the .gitmodules file but when "--recurse-submodules" is given or
the "submodule.recurse" variable is set, it did. Now these
settings are ignored in the "--no-index" mode.
* pb/do-not-recurse-grep-no-index:
grep: ignore --recurse-submodules if --no-index is given
"git rebase --fork-point master" used to work OK, as it internally
called "git merge-base --fork-point" that knew how to handle short
refname and dwim it to the full refname before calling the
underlying get_fork_point() function.
This is no longer true after the command was rewritten in C, as its
internall call made directly to get_fork_point() does not dwim a
short ref.
Move the "dwim the refname argument to the full refname" logic that
is used in "git merge-base" to the underlying get_fork_point()
function, so that the other caller of the function in the
implementation of "git rebase" behaves the same way to fix this
regression.
Signed-off-by: Alex Torok <alext9@gmail.com>
[jc: revamped the fix and used Alex's tests]
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When using Windows, a user may run 'git sparse-checkout set A\B\C'
to add the Unix-style path A/B/C to their sparse-checkout patterns.
Normalizing the input path converts the backslashes to slashes before we
add the string 'A/B/C' to the recursive hashset.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When using the sparse-checkout feature, a user may want to incrementally
grow their sparse-checkout pattern set. Allow adding patterns using a
new 'add' subcommand. This is not much different from the 'set'
subcommand, because we still want to allow the '--stdin' option and
interpret inputs as directories when in cone mode and patterns
otherwise.
When in cone mode, we are growing the cone. This may actually reduce the
set of patterns when adding directory A when A/B is already a directory
in the cone. Test the different cases: siblings, parents, ancestors.
When not in cone mode, we can only assume the patterns should be
appended to the sparse-checkout file.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In anticipation of adding "add" and "remove" subcommands to the
sparse-checkout builtin, extract a modify_pattern_list() method from the
sparse_checkout_set() method. This command will read input from the
command-line or stdin to construct a set of patterns, then modify the
existing sparse-checkout patterns after a successful update of the
working directory.
Currently, the only way to modify the patterns is to replace all of the
patterns. This will be extended in a later update.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In anticipation of extending the sparse-checkout builtin with "add"
and "remove" subcommands, extract the code that fills a pattern list
based on the input values. The input changes depending on the
presence of "--stdin" or the value of core.sparseCheckoutCone.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When renaming a remote with
git remote rename X Y
git remote remove X
Git already renames or removes any branch.<name>.remote and
branch.<name>.pushRemote configurations if their value is X.
However remote.pushDefault needs a more gentle approach, as this may be
set in a non-repo configuration file. In such a case only a warning is
printed, such as:
warning: The global configuration remote.pushDefault in:
$HOME/.gitconfig:35
now names the non-existent remote origin
It is changed to remote.pushDefault = Y or removed when set in a repo
configuration though.
Signed-off-by: Bert Wesarg <bert.wesarg@googlemail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When renaming or removing a remote with
git remote rename X Y
git remote remove X
Git already renames/removes any config values from
branch.<name>.remote = X
to
branch.<name>.remote = Y
As branch.<name>.pushRemote also names a remote, it now also renames
or removes these config values from
branch.<name>.pushRemote = X
to
branch.<name>.pushRemote = Y
Signed-off-by: Bert Wesarg <bert.wesarg@googlemail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some minor clean-ups in function `config_read_branches`:
* remove hardcoded length in `key += 7`
* call `xmemdupz` only once
* use a switch to handle the configuration type and add a `BUG()`
Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Bert Wesarg <bert.wesarg@googlemail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When 46af44b07d (pull --rebase=<type>: allow single-letter abbreviations
for the type, 2018-08-04) landed in Git, it had the side effect that
not only 'pull --rebase=<type>' accepted the single-letter abbreviations
but also the 'pull.rebase' and 'branch.<name>.rebase' configurations.
However, 'git remote rename' did not honor these single-letter
abbreviations when reading the 'branch.*.rebase' configurations.
We now document the single-letter abbreviations and both code places
share a common function to parse the values of 'git pull --rebase=*',
'pull.rebase', and 'branches.*.rebase'.
The only functional change is the handling of the `branch_info::rebase`
value. Before it was an unsigned enum, thus the truth value could be
checked with `branch_info::rebase != 0`. But `enum rebase_type` is
signed, thus the truth value must now be checked with
`branch_info::rebase >= REBASE_TRUE`
Signed-off-by: Bert Wesarg <bert.wesarg@googlemail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a user queries config values with --show-origin, often it's
difficult to determine what the actual "scope" (local, global, etc.) of
a given value is based on just the origin file.
Teach 'git config' the '--show-scope' option to print the scope of all
displayed config values. Note that we should never see anything of
"submodule" scope as that is only ever used by submodule-config.c when
parsing the '.gitmodules' file.
Signed-off-by: Matthew Rogers <mattr94@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are many situations where the scope of a config command is known
beforehand, such as passing of '--local', '--file', etc. to an
invocation of git config. However, this information is lost when moving
from builtin/config.c to /config.c. This historically hasn't been a big
deal, but to prepare for the upcoming --show-scope option we teach
git_config_source to keep track of the source and the config machinery
to use that information to set current_parsing_scope appropriately.
Signed-off-by: Matthew Rogers <mattr94@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a function for inserting a C string into a strbuf. Use it
throughout the source to get rid of magic string length constants and
explicit strlen() calls.
Like strbuf_addstr(), implement it as an inline function to avoid the
implicit strlen() calls to cause runtime overhead.
Helped-by: Taylor Blau <me@ttaylorr.com>
Helped-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
advice.addNothing config variable is used to control the visibility of
two advice messages in the add library. This config variable is
replaced by two new variables, whose names are more clear and relevant
to the two cases.
Also add the two new variables to the documentation.
Signed-off-by: Heba Waly <heba.waly@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git checkout X" did not correctly fail when X is not a local
branch but could name more than one remote-tracking branches
(i.e. to be dwimmed as the starting point to create a corresponding
local branch), which has been corrected.
* am/checkout-file-and-ref-ref-ambiguity:
checkout: don't revert file on ambiguous tracking branches
parse_branchname_arg(): extract part as new function
The effort to move "git-add--interactive" to C continues.
* js/patch-mode-in-others-in-c:
commit --interactive: make it work with the built-in `add -i`
built-in add -p: implement the "worktree" patch modes
built-in add -p: implement the "checkout" patch modes
built-in stash: use the built-in `git add -p` if so configured
legacy stash -p: respect the add.interactive.usebuiltin setting
built-in add -p: implement the "stash" and "reset" patch modes
built-in add -p: prepare for patch modes other than "stage"
name_ref() is called for each ref and checks if its a better name for
the referenced commit. If that's the case it remembers it and checks if
a name based on it is better for its ancestors as well. This in done in
the the order for_each_ref() imposes on us.
That might not be optimal. If bad names happen to be encountered first
(as defined by is_better_name()), names derived from them may spread to
a lot of commits, only to be replaced by better names later. Setting
better names first can avoid that.
is_better_name() prefers tags, short distances and old references. The
distance is a measure that we need to calculate for each candidate
commit, but the other two properties are not dependent on the
relationships of commits. Sorting the refs by them should yield better
performance than the essentially random order we currently use.
And applying older references first should also help to reduce rework
due to the fact that older commits have less ancestors than newer ones.
So add all details of names to the tip table first, then sort them
to prefer tags and older references and then apply them in this order.
Here's the performance as measures by hyperfine for the Linux repo
before:
Benchmark #1: ./git -C ../linux/ name-rev --all
Time (mean ± σ): 851.1 ms ± 4.5 ms [User: 806.7 ms, System: 44.4 ms]
Range (min … max): 845.9 ms … 859.5 ms 10 runs
... and with this patch:
Benchmark #1: ./git -C ../linux/ name-rev --all
Time (mean ± σ): 736.2 ms ± 8.7 ms [User: 688.4 ms, System: 47.5 ms]
Range (min … max): 726.0 ms … 755.2 ms 10 runs
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
name_rev() assigns a name to a commit and its parents and grandparents
and so on. Commits share their name string with their first parent,
which in turn does the same, recursively to the root. That saves a lot
of allocations. When a better name is found, the old name is replaced,
but its memory is not released. That leakage can become significant.
Can we release these old strings exactly once even though they are
referenced multiple times? Yes, indeed -- we can make use of the fact
that name_rev() visits the ancestors of a commit after it set a new name
for it and tries to update their names as well.
Members of the first ancestral line have the same taggerdate and
from_tag values, but a higher distance value than their child commit at
generation 0. These are the only criteria used by is_better_name().
Lower distance values are considered better, so a name that is better
for a child will also be better for its parent and grandparent etc.
That means we can free(3) an inferior name at generation 0 and rely on
name_rev() to replace all references in ancestors as well.
If we do that then we need to stop using the string pointer alone to
distinguish new empty rev_name slots from initialized ones, though, as
it technically becomes invalid after the free(3) call -- even though its
value is still different from NULL.
We can check the generation value first, as empty slots will have it
initialized to 0, and for the actual generation 0 we'll set a new valid
name right after the create_or_update_name() call that releases the
string.
For the Chromium repo, releasing superceded names reduces the memory
footprint of name-rev --all significantly. Here's the output of GNU
time before:
0.98user 0.48system 0:01.46elapsed 99%CPU (0avgtext+0avgdata 2601812maxresident)k
0inputs+0outputs (0major+571470minor)pagefaults 0swaps
... and with this patch:
1.01user 0.26system 0:01.28elapsed 100%CPU (0avgtext+0avgdata 1559196maxresident)k
0inputs+0outputs (0major+314370minor)pagefaults 0swaps
It also gets faster; hyperfine before:
Benchmark #1: ./git -C ../chromium/src name-rev --all
Time (mean ± σ): 1.534 s ± 0.006 s [User: 1.039 s, System: 0.494 s]
Range (min … max): 1.522 s … 1.542 s 10 runs
... and with this patch:
Benchmark #1: ./git -C ../chromium/src name-rev --all
Time (mean ± σ): 1.338 s ± 0.006 s [User: 1.047 s, System: 0.291 s]
Range (min … max): 1.327 s … 1.346 s 10 runs
For the Linux repo it doesn't pay off; memory usage only gets down from:
0.76user 0.03system 0:00.80elapsed 99%CPU (0avgtext+0avgdata 292848maxresident)k
0inputs+0outputs (0major+44579minor)pagefaults 0swaps
... to:
0.78user 0.03system 0:00.81elapsed 100%CPU (0avgtext+0avgdata 284696maxresident)k
0inputs+0outputs (0major+44892minor)pagefaults 0swaps
The runtime actually increases slightly from:
Benchmark #1: ./git -C ../linux/ name-rev --all
Time (mean ± σ): 828.8 ms ± 5.0 ms [User: 797.2 ms, System: 31.6 ms]
Range (min … max): 824.1 ms … 838.9 ms 10 runs
... to:
Benchmark #1: ./git -C ../linux/ name-rev --all
Time (mean ± σ): 847.6 ms ± 3.4 ms [User: 807.9 ms, System: 39.6 ms]
Range (min … max): 843.4 ms … 854.3 ms 10 runs
Why is that? In the Chromium repo, ca. 44000 free(3) calls in
create_or_update_name() release almost 1GB, while in the Linux repo
240000+ calls release a bit more than 5MB, so the average discarded
name is ca. 1000x longer in the latter.
Overall I think it's the right tradeoff to make, as it helps curb the
memory usage in repositories with big discarded names, and the added
overhead is small.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Leave setting the tip_name member of struct rev_name to callers of
create_or_update_name(). This avoids allocations for names that are
rejected by that function. Here's how this affects the runtime when
working with a fresh clone of Git's own repository; performance numbers
by hyperfine before:
Benchmark #1: ./git -C ../git-pristine/ name-rev --all
Time (mean ± σ): 437.8 ms ± 4.0 ms [User: 422.5 ms, System: 15.2 ms]
Range (min … max): 432.8 ms … 446.3 ms 10 runs
... and with this patch:
Benchmark #1: ./git -C ../git-pristine/ name-rev --all
Time (mean ± σ): 408.5 ms ± 1.4 ms [User: 387.2 ms, System: 21.2 ms]
Range (min … max): 407.1 ms … 411.7 ms 10 runs
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We can calculate the size of new name easily and precisely. Open-code
the xstrfmt() calls and grow the buffers as needed before filling them.
This provides a surprisingly large benefit when working with the
Chromium repository; here are the numbers measured using hyperfine
before:
Benchmark #1: ./git -C ../chromium/src name-rev --all
Time (mean ± σ): 5.822 s ± 0.013 s [User: 5.304 s, System: 0.516 s]
Range (min … max): 5.803 s … 5.837 s 10 runs
... and with this patch:
Benchmark #1: ./git -C ../chromium/src name-rev --all
Time (mean ± σ): 1.527 s ± 0.003 s [User: 1.015 s, System: 0.511 s]
Range (min … max): 1.524 s … 1.535 s 10 runs
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>