Maintenance currently triggers when certain data-size thresholds are
met, such as number of pack-files or loose objects. Users may want to
run certain maintenance tasks based on frequency instead. For example,
a user may want to perform a 'prefetch' task every hour, or 'gc' task
every day. To help these users, update the 'git maintenance run' command
to include a '--schedule=<frequency>' option. The allowed frequencies
are 'hourly', 'daily', and 'weekly'. These values are also allowed in a
new config value 'maintenance.<task>.schedule'.
The 'git maintenance run --schedule=<frequency>' checks the '*.schedule'
config value for each enabled task to see if the configured frequency is
at least as frequent as the frequency from the '--schedule' argument. We
use the following order, for full clarity:
'hourly' > 'daily' > 'weekly'
Use new 'enum schedule_priority' to track these values numerically.
The following cron table would run the scheduled tasks with the correct
frequencies:
0 1-23 * * * git -C <repo> maintenance run --schedule=hourly
0 0 * * 1-6 git -C <repo> maintenance run --schedule=daily
0 0 * * 0 git -C <repo> maintenance run --schedule=weekly
This cron schedule will run --schedule=hourly every hour except at
midnight. This avoids a concurrent run with the --schedule=daily that
runs at midnight every day except the first day of the week. This avoids
a concurrent run with the --schedule=weekly that runs at midnight on
the first day of the week. Since --schedule=daily also runs the
'hourly' tasks and --schedule=weekly runs the 'hourly' and 'daily'
tasks, we will still see all tasks run with the proper frequencies.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The incremental-repack task updates the multi-pack-index by deleting pack-
files that have been replaced with new packs, then repacking a batch of
small pack-files into a larger pack-file. This incremental repack is faster
than rewriting all object data, but is slower than some other
maintenance activities.
The 'maintenance.incremental-repack.auto' config option specifies how many
pack-files should exist outside of the multi-pack-index before running
the step. These pack-files could be created by 'git fetch' commands or
by the loose-objects task. The default value is 10.
Setting the option to zero disables the task with the '--auto' option,
and a negative value makes the task run every time.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When repacking during the 'incremental-repack' task, we use the
--batch-size option in 'git multi-pack-index repack'. The initial setting
used --batch-size=0 to repack everything into a single pack-file. This is
not sustainable for a large repository. The amount of work required is
also likely to use too many system resources for a background job.
Update the 'incremental-repack' task by dynamically computing a
--batch-size option based on the current pack-file structure.
The dynamic default size is computed with this idea in mind for a client
repository that was cloned from a very large remote: there is likely one
"big" pack-file that was created at clone time. Thus, do not try
repacking it as it is likely packed efficiently by the server.
Instead, we select the second-largest pack-file, and create a batch size
that is one larger than that pack-file. If there are three or more
pack-files, then this guarantees that at least two will be combined into
a new pack-file.
Of course, this means that the second-largest pack-file size is likely
to grow over time and may eventually surpass the initially-cloned
pack-file. Recall that the pack-file batch is selected in a greedy
manner: the packs are considered from oldest to newest and are selected
if they have size smaller than the batch size until the total selected
size is larger than the batch size. Thus, that oldest "clone" pack will
be first to repack after the new data creates a pack larger than that.
We also want to place some limits on how large these pack-files become,
in order to bound the amount of time spent repacking. A maximum
batch-size of two gigabytes means that large repositories will never be
packed into a single pack-file using this job, but also that repack is
rather expensive. This is a trade-off that is valuable to have if the
maintenance is being run automatically or in the background. Users who
truly want to optimize for space and performance (and are willing to pay
the upfront cost of a full repack) can use the 'gc' task to do so.
Create a test for this two gigabyte limit by creating an EXPENSIVE test
that generates two pack-files of roughly 2.5 gigabytes in size, then
performs an incremental repack. Check that the --batch-size argument in
the subcommand uses the hard-coded maximum.
Helped-by: Chris Torek <chris.torek@gmail.com>
Reported-by: Son Luong Ngoc <sluongng@gmail.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous change cleaned up loose objects using the
'loose-objects' that can be run safely in the background. Add a
similar job that performs similar cleanups for pack-files.
One issue with running 'git repack' is that it is designed to
repack all pack-files into a single pack-file. While this is the
most space-efficient way to store object data, it is not time or
memory efficient. This becomes extremely important if the repo is
so large that a user struggles to store two copies of the pack on
their disk.
Instead, perform an "incremental" repack by collecting a few small
pack-files into a new pack-file. The multi-pack-index facilitates
this process ever since 'git multi-pack-index expire' was added in
19575c7 (multi-pack-index: implement 'expire' subcommand,
2019-06-10) and 'git multi-pack-index repack' was added in ce1e4a1
(midx: implement midx_repack(), 2019-06-10).
The 'incremental-repack' task runs the following steps:
1. 'git multi-pack-index write' creates a multi-pack-index file if
one did not exist, and otherwise will update the multi-pack-index
with any new pack-files that appeared since the last write. This
is particularly relevant with the background fetch job.
When the multi-pack-index sees two copies of the same object, it
stores the offset data into the newer pack-file. This means that
some old pack-files could become "unreferenced" which I will use
to mean "a pack-file that is in the pack-file list of the
multi-pack-index but none of the objects in the multi-pack-index
reference a location inside that pack-file."
2. 'git multi-pack-index expire' deletes any unreferenced pack-files
and updaes the multi-pack-index to drop those pack-files from the
list. This is safe to do as concurrent Git processes will see the
multi-pack-index and not open those packs when looking for object
contents. (Similar to the 'loose-objects' job, there are some Git
commands that open pack-files regardless of the multi-pack-index,
but they are rarely used. Further, a user that self-selects to
use background operations would likely refrain from using those
commands.)
3. 'git multi-pack-index repack --bacth-size=<size>' collects a set
of pack-files that are listed in the multi-pack-index and creates
a new pack-file containing the objects whose offsets are listed
by the multi-pack-index to be in those objects. The set of pack-
files is selected greedily by sorting the pack-files by modified
time and adding a pack-file to the set if its "expected size" is
smaller than the batch size until the total expected size of the
selected pack-files is at least the batch size. The "expected
size" is calculated by taking the size of the pack-file divided
by the number of objects in the pack-file and multiplied by the
number of objects from the multi-pack-index with offset in that
pack-file. The expected size approximates how much data from that
pack-file will contribute to the resulting pack-file size. The
intention is that the resulting pack-file will be close in size
to the provided batch size.
The next run of the incremental-repack task will delete these
repacked pack-files during the 'expire' step.
In this version, the batch size is set to "0" which ignores the
size restrictions when selecting the pack-files. It instead
selects all pack-files and repacks all packed objects into a
single pack-file. This will be updated in the next change, but
it requires doing some calculations that are better isolated to
a separate change.
These steps are based on a similar background maintenance step in
Scalar (and VFS for Git) [1]. This was incredibly effective for
users of the Windows OS repository. After using the same VFS for Git
repository for over a year, some users had _thousands_ of pack-files
that combined to up to 250 GB of data. We noticed a few users were
running into the open file descriptor limits (due in part to a bug
in the multi-pack-index fixed by af96fe3 (midx: add packs to
packed_git linked list, 2019-04-29).
These pack-files were mostly small since they contained the commits
and trees that were pushed to the origin in a given hour. The GVFS
protocol includes a "prefetch" step that asks for pre-computed pack-
files containing commits and trees by timestamp. These pack-files
were grouped into "daily" pack-files once a day for up to 30 days.
If a user did not request prefetch packs for over 30 days, then they
would get the entire history of commits and trees in a new, large
pack-file. This led to a large number of pack-files that had poor
delta compression.
By running this pack-file maintenance step once per day, these repos
with thousands of packs spanning 200+ GB dropped to dozens of pack-
files spanning 30-50 GB. This was done all without removing objects
from the system and using a constant batch size of two gigabytes.
Once the work was done to reduce the pack-files to small sizes, the
batch size of two gigabytes means that not every run triggers a
repack operation, so the following run will not expire a pack-file.
This has kept these repos in a "clean" state.
[1] https://github.com/microsoft/scalar/blob/master/Scalar.Common/Maintenance/PackfileMaintenanceStep.cs
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The loose-objects task deletes loose objects that already exist in a
pack-file, then place the remaining loose objects into a new pack-file.
If this step runs all the time, then we risk creating pack-files with
very few objects with every 'git commit' process. To prevent
overwhelming the packs directory with small pack-files, place a minimum
number of objects to justify the task.
The 'maintenance.loose-objects.auto' config option specifies a minimum
number of loose objects to justify the task to run under the '--auto'
option. This defaults to 100 loose objects. Setting the value to zero
will prevent the step from running under '--auto' while a negative value
will force it to run every time.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
One goal of background maintenance jobs is to allow a user to
disable auto-gc (gc.auto=0) but keep their repository in a clean
state. Without any cleanup, loose objects will clutter the object
database and slow operations. In addition, the loose objects will
take up extra space because they are not stored with deltas against
similar objects.
Create a 'loose-objects' task for the 'git maintenance run' command.
This helps clean up loose objects without disrupting concurrent Git
commands using the following sequence of events:
1. Run 'git prune-packed' to delete any loose objects that exist
in a pack-file. Concurrent commands will prefer the packed
version of the object to the loose version. (Of course, there
are exceptions for commands that specifically care about the
location of an object. These are rare for a user to run on
purpose, and we hope a user that has selected background
maintenance will not be trying to do foreground maintenance.)
2. Run 'git pack-objects' on a batch of loose objects. These
objects are grouped by scanning the loose object directories in
lexicographic order until listing all loose objects -or-
reaching 50,000 objects. This is more than enough if the loose
objects are created only by a user doing normal development.
We noticed users with _millions_ of loose objects because VFS
for Git downloads blobs on-demand when a file read operation
requires populating a virtual file.
This step is based on a similar step in Scalar [1] and VFS for Git.
[1] https://github.com/microsoft/scalar/blob/master/Scalar.Common/Maintenance/LooseObjectsStep.cs
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When working with very large repositories, an incremental 'git fetch'
command can download a large amount of data. If there are many other
users pushing to a common repo, then this data can rival the initial
pack-file size of a 'git clone' of a medium-size repo.
Users may want to keep the data on their local repos as close as
possible to the data on the remote repos by fetching periodically in
the background. This can break up a large daily fetch into several
smaller hourly fetches.
The task is called "prefetch" because it is work done in advance
of a foreground fetch to make that 'git fetch' command much faster.
However, if we simply ran 'git fetch <remote>' in the background,
then the user running a foreground 'git fetch <remote>' would lose
some important feedback when a new branch appears or an existing
branch updates. This is especially true if a remote branch is
force-updated and this isn't noticed by the user because it occurred
in the background. Further, the functionality of 'git push
--force-with-lease' becomes suspect.
When running 'git fetch <remote> <options>' in the background, use
the following options for careful updating:
1. --no-tags prevents getting a new tag when a user wants to see
the new tags appear in their foreground fetches.
2. --refmap= removes the configured refspec which usually updates
refs/remotes/<remote>/* with the refs advertised by the remote.
While this looks confusing, this was documented and tested by
b40a50264a (fetch: document and test --refmap="", 2020-01-21),
including this sentence in the documentation:
Providing an empty `<refspec>` to the `--refmap` option
causes Git to ignore the configured refspecs and rely
entirely on the refspecs supplied as command-line arguments.
3. By adding a new refspec "+refs/heads/*:refs/prefetch/<remote>/*"
we can ensure that we actually load the new values somewhere in
our refspace while not updating refs/heads or refs/remotes. By
storing these refs here, the commit-graph job will update the
commit-graph with the commits from these hidden refs.
4. --prune will delete the refs/prefetch/<remote> refs that no
longer appear on the remote.
5. --no-write-fetch-head prevents updating FETCH_HEAD.
We've been using this step as a critical background job in Scalar
[1] (and VFS for Git). This solved a pain point that was showing up
in user reports: fetching was a pain! Users do not like waiting to
download the data that was created while they were away from their
machines. After implementing background fetch, the foreground fetch
commands sped up significantly because they mostly just update refs
and download a small amount of new data. The effect is especially
dramatic when paried with --no-show-forced-udpates (through
fetch.showForcedUpdates=false).
[1] https://github.com/microsoft/scalar/blob/master/Scalar.Common/Maintenance/FetchStep.cs
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We already match "committer", and we're about to start
matching more things. Let's use a more neutral variable to
avoid confusion.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 06f5608c14 (bisect--helper: `bisect_start` shell function
partially in C, 2019-01-02), we changed the following shell
code:
- rev=$(git rev-parse -q --verify "$arg^{commit}") || {
- test $has_double_dash -eq 1 &&
- die "$(eval_gettext "'\$arg' does not appear to be a valid revision")"
- break
- }
- revs="$revs $rev"
into:
+ char *commit_id = xstrfmt("%s^{commit}", arg);
+ if (get_oid(commit_id, &oid) && has_double_dash)
+ die(_("'%s' does not appear to be a valid "
+ "revision"), arg);
+
+ string_list_append(&revs, oid_to_hex(&oid));
+ free(commit_id);
In case of an invalid "arg" when "has_double_dash" is false, the old
code would "break" out of the argument loop.
In the new C code though, `oid_to_hex(&oid)` is unconditonally
appended to "revs". This is wrong first because "oid" is junk as
`get_oid(commit_id, &oid)` failed and second because it doesn't break
out of the argument loop.
Not breaking out of the argument loop means that "arg" is then not
treated as a path restriction (which is wrong).
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A user who understands enough to set pull.ff does not need additional
instructions.
Signed-off-by: Alex Henrie <alexhenrie24@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The command reads list of object names to place on the ignore list
either from the command line or from a file, but they are not
checked with their object type (those read from the file are not
even checked for object existence).
Extend the oidset_parse_file() API and allow it to take a callback
that can be used to die (e.g. when an inappropriate input is read)
or modify the object name read (e.g. when a tag pointing at a commit
is read, and the caller wants a commit object name), and use it in
the code that handles ignore list.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This reverts commit f39ad38410.
That commit was trying to silence a type-punning warning on older
versions of gcc. However, its analysis was all wrong. I didn't notice
that we _were_ in fact type-punning because there are two versions of
put_be32(): one that uses casts and unaligned loads, and another that
uses bitshifts. I looked at the latter, but on my platform we were
defaulting to the former.
However, as of the previous commit, we'll always use the bitshift
version. So we can drop this hackery to avoid the warning, making the
code slightly cleaner.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Reimplement the `bisect_next()` and the `bisect_auto_next()` shell functions
in C and add the subcommands to `git bisect--helper` to call them from
git-bisect.sh .
bisect_auto_next() function returns an enum bisect_error type as whole
`git bisect` can exit with an error code when bisect_next() does.
Return an error when `bisect_next()` fails, that fix a bug on shell script
version.
Using `--bisect-next` and `--bisect-auto-next` subcommands is a
temporary measure to port shell function to C so as to use the existing
test suite. As more functions are ported, `--bisect-auto-next`
subcommand will be retired and will be called by some other methods.
Mentored-by: Lars Schneider <larsxschneider@gmail.com>
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>
Reimplement the `bisect_autostart()` shell function in C and add the
C implementation from `bisect_next()` which was previously left
uncovered.
Add `--bisect-autostart` subcommand to be called from git-bisect.sh.
Using `--bisect-autostart` subcommand is a temporary measure to port
the shell function to C so as to use the existing test suite. As more
functions are ported, this subcommand will be retired and
bisect_autostart() will be called directly by `bisect_state()`.
Change behavior of shell script that returned success when user aborted
the bisection.
Mentored-by: Lars Schneider <larsxschneider@gmail.com>
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>
"git fetch --all --ipv4/--ipv6" forgot to pass the protocol options
to instances of the "git fetch" that talk to individual remotes,
which has been corrected.
* ar/fetch-ipversion-in-all:
fetch: pass --ipv4 and --ipv6 options to sub-fetches
"git remote set-head" that failed still said something that hints
the operation went through, which was misleading.
* cs/don-t-pretend-a-failed-remote-set-head-succeeded:
remote: don't show success message when set-head fails
"git for-each-ref" and friends that list refs used to allow only
one --merged or --no-merged to filter them; they learned to take
combination of both kind of filtering.
* al/ref-filter-merged-and-no-merged:
Doc: prefer more specific file name
ref-filter: make internal reachable-filter API more precise
ref-filter: allow merged and no-merged filters
Doc: cover multiple contains/no-contains filters
t3201: test multiple branch filter combinations
The 'meld' backend of the "git mergetool" learned to give the
underlying 'meld' the '--auto-merge' option, which would help
reduce the amount of text that requires manual merging.
* ls/mergetool-meld-auto-merge:
mergetool: allow auto-merge for meld to follow the vim-diff behavior
"git index-pack" learned to resolve deltified objects with greater
parallelism.
* jt/threaded-index-pack:
index-pack: make quantum of work smaller
index-pack: make resolve_delta() assume base data
index-pack: calculate {ref,ofs}_{first,last} early
index-pack: remove redundant child field
index-pack: unify threaded and unthreaded code
index-pack: remove redundant parameter
Documentation: deltaBaseCacheLimit is per-thread
"format-patch --range-diff=<prev> <origin>..HEAD" has been taught
not to ignore <origin> when <prev> is a single version.
* es/format-patch-interdiff-cleanup:
format-patch: use 'origin' as start of current-series-range when known
diff-lib: tighten show_interdiff()'s interface
diff: move show_interdiff() from its own file to diff-lib
If a user is cloning a SHA-1 repository with GIT_DEFAULT_HASH set to
"sha256", then we can end up with a repository where the repository
format version is 0 but the extensions.objectformat key is set to
"sha256". This is both wrong (the user has a SHA-1 repository) and
nonfunctional (because the extension cannot be used in a v0 repository).
This happens because in a clone, we initially set up the repository, and
then change its algorithm based on what the remote side tells us it's
using. We've initially set up the repository as SHA-256 in this case,
and then later on reset the repository version without clearing the
extension.
We could just always set the extension in this case, but that would mean
that our SHA-1 repositories weren't compatible with older Git versions,
even though there's no reason why they shouldn't be. And we also don't
want to initialize the repository as SHA-1 initially, since that means
if we're cloning an empty repository, we'll have failed to honor the
GIT_DEFAULT_HASH variable and will end up with a SHA-1 repository, not a
SHA-256 repository.
Neither of those are appealing, so let's tell the repository
initialization code if we're doing a reinit like this, and if so, to
clear the extension if we're using SHA-1. This makes sure we produce a
valid and functional repository and doesn't break any of our other use
cases.
Reported-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In an ongoing effort to avoid non-inclusive language, let's avoid using
the branch name "master" in a code comment.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous commit introduced ---merge-base a way to take the diff
between the working tree or index and the merge base between an arbitrary
commit and HEAD. It makes sense to extend this option to support the
case where two commits are given too and behave in a manner identical to
`git diff A...B`.
Introduce the --merge-base flag as an alternative to triple-dot
notation. Thus, we would be able to write the above as
`git diff --merge-base A B`.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There is currently no easy way to take the diff between the working tree
or index and the merge base between an arbitrary commit and HEAD. Even
diff's `...` notation doesn't allow this because it only works between
commits. However, the ability to do this would be desirable to a user
who would like to see all the changes they've made on a branch plus
uncommitted changes without taking into account changes made in the
upstream branch.
Teach diff-index and diff (with one commit) the --merge-base option
which allows a user to use the merge base of a commit and HEAD as the
"before" side.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In a future commit, we will teach run_diff_index() to accept more
options via flag bits. For now, change `cached` into a flag in the
`option` bitfield. The behaviour should remain exactly the same.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Unlike "git config --local", "git config --worktree" did not fail
early and cleanly when started outside a git repository.
* mt/config-fail-nongit-early:
config: complain about --worktree outside of a git repo
"git status --short" quoted a path with SP in it when tracked, but
not those that are untracked, ignored or unmerged. They are all
shown quoted consistently.
* jc/quote-path-cleanup:
quote: turn 'nodq' parameter into a set of flags
quote: rename misnamed sq_lookup[] to cq_lookup[]
wt-status: consistently quote paths in "status --short" output
quote_path: code clarification
quote_path: optionally allow quoting a path with SP in it
quote_path: give flags parameter to quote_path()
quote_path: rename quote_path_relative() to quote_path()
"git worktree add" learns that the "-d" is a synonym to "--detach"
option to create a new worktree without being on a branch.
* es/wt-add-detach:
git-worktree.txt: discuss branch-based vs. throwaway worktrees
worktree: teach `add` to recognize -d as shorthand for --detach
git-checkout.txt: document -d short option for --detach
The "add -i/-p" machinery has been written in C but it is not used
by default yet. It is made default to those who are participating
in feature.experimental experiment.
* jc/add-i-use-builtin-experimental:
add -i: use the built-in version when feature.experimental is set
Misc cleanups.
* rs/misc-cleanups:
pack-bitmap-write: use hashwrite_be32() in write_hash_cache()
midx: use hashwrite_u8() in write_midx_header()
fast-import: use write_pack_header()
Introduce a configuration variable to specify a default value for the
recently-introduce '--max-new-filters' option of 'git commit-graph
write'.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Introduce a command-line flag to specify the maximum number of new Bloom
filters that a 'git commit-graph write' is willing to compute from
scratch.
Prior to this patch, a commit-graph write with '--changed-paths' would
compute Bloom filters for all selected commits which haven't already
been computed (i.e., by a previous commit-graph write with '--split'
such that a roll-up or replacement is performed).
This behavior can cause prohibitively-long commit-graph writes for a
variety of reasons:
* There may be lots of filters whose diffs take a long time to
generate (for example, they have close to the maximum number of
changes, diffing itself takes a long time, etc).
* Old-style commit-graphs (which encode filters with too many entries
as not having been computed at all) cause us to waste time
recomputing filters that appear to have not been computed only to
discover that they are too-large.
This can make the upper-bound of the time it takes for 'git commit-graph
write --changed-paths' to be rather unpredictable.
To make this command behave more predictably, introduce
'--max-new-filters=<n>' to allow computing at most '<n>' Bloom filters
from scratch. This lets "computing" already-known filters proceed
quickly, while bounding the number of slow tasks that Git is willing to
do.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the subsequent commit, additional options will be added to the
commit-graph API which have nothing to do with splitting.
Rename the 'split_commit_graph_opts' structure to the more-generic
'commit_graph_opts' to encompass both. Likewise, rename the 'flags'
member to instead be 'split_flags' to clarify that it only has to do
with the behavior implied by '--split'.
Suggested-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Suppress the message 'origin/HEAD set to master' in case of an error.
$ git remote set-head origin -a
error: Not a valid ref: refs/remotes/origin/master
origin/HEAD set to master
Signed-off-by: Christian Schlack <christian@backhub.co>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of writing a new commit-graph in every 'git maintenance run
--auto' process (when maintenance.commit-graph.enalbed is configured to
be true), only write when there are "enough" commits not in a
commit-graph file.
This count is controlled by the maintenance.commit-graph.auto config
option.
To compute the count, use a depth-first search starting at each ref, and
leaving markers using the SEEN flag. If this count reaches the limit,
then terminate early and start the task. Otherwise, this operation will
peel every ref and parse the commit it points to. If these are all in
the commit-graph, then this is typically a very fast operation. Users
with many refs might feel a slow-down, and hence could consider updating
their limit to be very small. A negative value will force the step to
run every time.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'git maintenance run' command has an '--auto' option. This is used
by other Git commands such as 'git commit' or 'git fetch' to check if
maintenance should be run after adding data to the repository.
Previously, this --auto option was only used to add the argument to the
'git gc' command as part of the 'gc' task. We will be expanding the
other tasks to perform a check to see if they should do work as part of
the --auto flag, when they are enabled by config.
First, update the 'gc' task to perform the auto check inside the
maintenance process. This prevents running an extra 'git gc --auto'
command when not needed. It also shows a model for other tasks.
Second, use the 'auto_condition' function pointer as a signal for
whether we enable the maintenance task under '--auto'. For instance, we
do not want to enable the 'fetch' task in '--auto' mode, so that
function pointer will remain NULL.
Now that we are not automatically calling 'git gc', a test in
t5514-fetch-multiple.sh must be changed to watch for 'git maintenance'
instead.
We continue to pass the '--auto' option to the 'git gc' command when
necessary, because of the gc.autoDetach config option changes behavior.
Likely, we will want to absorb the daemonizing behavior implied by
gc.autoDetach as a maintenance.autoDetach config option.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently, a normal run of "git maintenance run" will only run the 'gc'
task, as it is the only one enabled. This is mostly for backwards-
compatible reasons since "git maintenance run --auto" commands replaced
previous "git gc --auto" commands after some Git processes. Users could
manually run specific maintenance tasks by calling "git maintenance run
--task=<task>" directly.
Allow users to customize which steps are run automatically using config.
The 'maintenance.<task>.enabled' option then can turn on these other
tasks (or turn off the 'gc' task).
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Performing maintenance on a Git repository involves writing data to the
.git directory, which is not safe to do with multiple writers attempting
the same operation. Ensure that only one 'git maintenance' process is
running at a time by holding a file-based lock. Simply the presence of
the .git/maintenance.lock file will prevent future maintenance. This
lock is never committed, since it does not represent meaningful data.
Instead, it is only a placeholder.
If the lock file already exists, then no maintenance tasks are
attempted. This will become very important later when we implement the
'prefetch' task, as this is our stop-gap from creating a recursive process
loop between 'git fetch' and 'git maintenance run --auto'.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A user may want to only run certain maintenance tasks in a certain
order. Add the --task=<task> option, which allows a user to specify an
ordered list of tasks to run. These cannot be run multiple times,
however.
Here is where our array of maintenance_task pointers becomes critical.
We can sort the array of pointers based on the task order, but we do not
want to move the struct data itself in order to preserve the hashmap
references. We use the hashmap to match the --task=<task> arguments into
the task struct data.
Keep in mind that the 'enabled' member of the maintenance_task struct is
a placeholder for a future 'maintenance.<task>.enabled' config option.
Thus, we use the 'enabled' member to specify which tasks are run when
the user does not specify any --task=<task> arguments. The 'enabled'
member should be ignored if --task=<task> appears.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The first new task in the 'git maintenance' builtin is the
'commit-graph' task. This updates the commit-graph file
incrementally with the command
git commit-graph write --reachable --split
By writing an incremental commit-graph file using the "--split"
option we minimize the disruption from this operation. The default
behavior is to merge layers until the new "top" layer is less than
half the size of the layer below. This provides quick writes most
of the time, with the longer writes following a power law
distribution.
Most importantly, concurrent Git processes only look at the
commit-graph-chain file for a very short amount of time, so they
will verly likely not be holding a handle to the file when we try
to replace it. (This only matters on Windows.)
If a concurrent process reads the old commit-graph-chain file, but
our job expires some of the .graph files before they can be read,
then those processes will see a warning message (but not fail).
This could be avoided by a future update to use the --expire-time
argument when writing the commit-graph.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In anticipation of implementing multiple maintenance tasks inside the
'maintenance' builtin, use a list of structs to describe the work to be
done.
The struct maintenance_task stores the name of the task (as given by a
future command-line argument) along with a function pointer to its
implementation and a boolean for whether the step is enabled.
A list these structs are initialized with the full list of implemented
tasks along with a default order. For now, this list only contains the
"gc" task. This task is also the only task enabled by default.
The run subcommand will return a nonzero exit code if any task fails.
However, it will attempt all tasks in its loop before returning with the
failure. Also each failed task will print an error message.
Helped-by: Taylor Blau <me@ttaylorr.com>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The run_auto_gc() method is used in several places to trigger a check
for repo maintenance after some Git commands, such as 'git commit' or
'git fetch'.
To allow for extra customization of this maintenance activity, replace
the 'git gc --auto [--quiet]' call with one to 'git maintenance run
--auto [--quiet]'. As we extend the maintenance builtin with other
steps, users will be able to select different maintenance activities.
Rename run_auto_gc() to run_auto_maintenance() to be clearer what is
happening on this call, and to expose all callers in the current diff.
Rewrite the method to use a struct child_process to simplify the calls
slightly.
Since 'git fetch' already allows disabling the 'git gc --auto'
subprocess, add an equivalent option with a different name to be more
descriptive of the new behavior: '--[no-]maintenance'. Update the
documentation to include these options at the same time.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Maintenance activities are commonly used as steps in larger scripts.
Providing a '--quiet' option allows those scripts to be less noisy when
run on a terminal window. Turn this mode on by default when stderr is
not a terminal.
Pipe the option to the 'git gc' child process.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'gc' builtin is our current entrypoint for automatically maintaining
a repository. This one tool does many operations, such as repacking the
repository, packing refs, and rewriting the commit-graph file. The name
implies it performs "garbage collection" which means several different
things, and some users may not want to use this operation that rewrites
the entire object database.
Create a new 'maintenance' builtin that will become a more general-
purpose command. To start, it will only support the 'run' subcommand,
but will later expand to add subcommands for scheduling maintenance in
the background.
For now, the 'maintenance' builtin is a thin shim over the 'gc' builtin.
In fact, the only option is the '--auto' toggle, which is handed
directly to the 'gc' builtin. The current change is isolated to this
simple operation to prevent more interesting logic from being lost in
all of the boilerplate of adding a new builtin.
Use existing builtin/gc.c file because we want to share code between the
two builtins. It is possible that we will have 'maintenance' replace the
'gc' builtin entirely at some point, leaving 'git gc' as an alias for
some specific arguments to 'git maintenance run'.
Create a new test_subcommand helper that allows us to test if a certain
subcommand was run. It requires storing the GIT_TRACE2_EVENT logs in a
file. A negation mode is available that will be used in later tests.
Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Make the mergetool used with "meld" backend behave similarly to "vimdiff" by
telling it to auto-merge non-conflicting parts and highlight the conflicting
parts when `mergetool.meld.useAutoMerge` is configured with `true`, or `auto`
for detecting the `--auto-merge` option automatically.
Helped-by: Đoàn Trần Công Danh <congdanhqx@gmail.com>
Helped-by: David Aguilar <davvid@gmail.com>
Signed-off-by: Lin Sun <lin.sun@zoom.us>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Enable ref-filter to process multiple merged and no-merged filters, and
extend functionality to git branch, git tag and git for-each-ref. This
provides an easy way to check for branches that are "graduation
candidates:"
$ git branch --no-merged master --merged next
If passed more than one merged (or more than one no-merged) filter, refs
must be reachable from any one of the merged commits, and reachable from
none of the no-merged commits.
Signed-off-by: Aaron Lipman <alipman88@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The options indicate user intent for the whole fetch operation, and
ignoring them in sub-fetches (i.e. "--all" and recursive fetching of
submodules) is quite unexpected when, for instance, it is intended
to limit all of the communication to a specific transport protocol
for some reason.
Signed-off-by: Alex Riesen <alexander.riesen@cetitec.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The quote_path() function computes a path (relative to its base
directory) and c-quotes the result if necessary. Teach it to take a
flags parameter to allow its behaviour to be enriched later.
No behaviour change intended.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There is no quote_path_absolute() or anything that causes confusion,
and one of the two large consumers already rename the long name
locally with a preprocessor macro.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git status" has trouble showing where it came from by interpreting
reflog entries that recordcertain events, e.g. "checkout @{u}", and
gives a hard/fatal error. Even though it inherently is impossible
to give a correct answer because the reflog entries lose some
information (e.g. "@{u}" does not record what branch the user was
on hence which branch 'the upstream' needs to be computed, and even
if the record were available, the relationship between branches may
have changed), at least hide the error to allow "status" show its
output.
* jt/interpret-branch-name-fallback:
wt-status: tolerate dangling marks
refs: move dwim_ref() to header file
sha1-name: replace unsigned int with option struct
Fixups to a topic in 'next'.
* ss/submodule-summary-in-c-fixes:
t7421: eliminate 'grep' check in t7421.4 for mingw compatibility
submodule: fix style in function definition
submodule: eliminate unused parameters from print_submodule_summary()
"git worktree" gained a "repair" subcommand to help users recover
after moving the worktrees or repository manually without telling
Git. Also, "git init --separate-git-dir" no longer corrupts
administrative data related to linked worktrees.
* es/worktree-repair:
init: make --separate-git-dir work from within linked worktree
init: teach --separate-git-dir to repair linked worktrees
worktree: teach "repair" to fix outgoing links to worktrees
worktree: teach "repair" to fix worktree back-links to main worktree
worktree: add skeleton "repair" command
When a packfile is removed by "git repack", multi-pack-index gets
cleared; the code was taught to do so less aggressively by first
checking if the midx actually refers to a pack that no longer
exists.
* tb/repack-clearing-midx:
midx: traverse the local MIDX first
builtin/repack.c: invalidate MIDX only when necessary
Yet another subcommand of "git submodule" is getting rewritten in C.
* ss/submodule-summary-in-c:
submodule: port submodule subcommand 'summary' from shell to C
t7421: introduce a test script for verifying 'summary' output
submodule: rename helper functions to avoid ambiguity
submodule: remove extra line feeds between callback struct and macro
In a future commit, some commit-graph internals will want access to
'r->settings', but we only have the 'struct object_directory *'
corresponding to that repository.
Add an additional parameter to pass the repository around in more
places.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Running `git config --worktree` outside of a git repository hits a BUG()
when trying to enumerate the worktrees. Let's catch this error earlier
and die() with a friendlier message.
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently, when index-pack resolves deltas, it does not split up delta
trees into threads: each delta base root (an object that is not a
REF_DELTA or OFS_DELTA) can go into its own thread, but all deltas on
that root (direct or indirect) are processed in the same thread.
This is a problem when a repository contains a large text file (thus,
delta-able) that is modified many times - delta resolution time during
fetching is dominated by processing the deltas corresponding to that
text file.
This patch contains a solution to that. When cloning using
git -c core.deltabasecachelimit=1g clone \
https://fuchsia.googlesource.com/third_party/vulkan-cts
on my laptop, clone time improved from 3m2s to 2m5s (using 3 threads,
which is the default).
The solution is to have a global work stack. This stack contains delta
bases (objects, whether appearing directly in the packfile or generated
by delta resolution, that themselves have delta children) that need to
be processed; whenever a thread needs work, it peeks at the top of the
stack and processes its next unprocessed child. If a thread finds the
stack empty, it will look for more delta base roots to push on the stack
instead.
The main weakness of having a global work stack is that more time is
spent in the mutex, but profiling has shown that most time is spent in
the resolution of the deltas themselves, so this shouldn't be an issue
in practice. In any case, experimentation (as described in the clone
command above) shows that this patch is a net improvement.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When formatting a patch series over `origin..HEAD`, one would expect
that range to be used as the current-series-range when computing a
range-diff between the previous and current versions of a patch series.
However, infer_range_diff_ranges() ignores `origin..HEAD` when
--range-diff=<prev> specifies a single revision rather than a range, and
instead unexpectedly computes the current-series-range based upon
<prev>. Address this anomaly by unconditionally using `origin..HEAD` as
the current-series-range regardless of <prev> as long as `origin` is
known, and only fall back to basing current-series-range on <prev> when
`origin` is not known.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
To compute and show an interdiff, show_interdiff() needs only the two
OID's to compare and a diffopts, yet it expects callers to supply an
entire rev_info. The demand for rev_info is not only overkill, but also
places unnecessary burden on potential future callers which might not
otherwise have a rev_info at hand. Address this by tightening its
signature to require only the items it needs instead of a full rev_info.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
show_interdiff() is a relatively small function and not likely to grow
larger or more complicated. Rather than dedicating an entire source file
to it, relocate it to diff-lib.c which houses other "take two things and
compare them" functions meant to be re-used but not so low-level as to
reside in the core diff implementation.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We have had parallel implementations of "add -i/-p" since 2.25 and
have been using them from various codepaths since 2.26 days, but
never made the built-in version the default.
We have found and fixed a handful of corner case bugs in the
built-in version, and it may be a good time to start switching over
the user base from the scripted version to the built-in version.
Let's enable the built-in version for those who opt into the
feature.experimental guinea-pig program to give wider exposure.
Acked-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Like `git switch` and `git checkout`, `git worktree add` can check out a
branch or set up a detached HEAD. However, unlike those other commands,
`git worktree add` does not understand -d as shorthand for --detach,
which may confound users accustomed to using -d for this purpose.
Address this shortcoming by teaching `add` to recognize -d for --detach,
thus bringing it in line with the other commands.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Call write_pack_header() to hash and write a pack header instead of
open-coding this function. This gets rid of duplicate code and of the
magic version number 2 -- which has been used here since c90be46abd
(Changed fast-import's pack header creation to use pack.h, 2006-08-16)
and in pack.h (again) since 29f049a0c2 (Revert "move pack creation to
version 3", 2006-10-14).
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a function for building a refspec using printf-style formatting. It
frees callers from managing their own buffer. Use it throughout the
tree to shorten and simplify its callers.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
map_refspec() either returns the passed in ref string or a detached
strbuf. This makes it hard for callers to release the possibly
allocated memory, and set_refspecs() consequently leaks it.
Let map_refspec() append any refspecs directly and release its own
strbufs after use. Rename it to refspec_append_mapped() and don't
return anything to reflect its increased responsibility.
set_refspecs() also leaks its strbufs. Do the same here and directly
call refspec_append() in each if branch instead of holding onto a
detached strbuf, then dispose of the allocated memory after use. We
need to add an else branch for the final call because all the other
conditional branches already add their formatted refspec now.
setup_push_upstream() and setup_push_current() forgot to release their
strbufs as well; plug these leaks, too, while at it.
None of these leaks were likely to impact users, because the number
and sizes of refspecs are usually small and the allocations are only
done once per program run. Clean them up nevertheless, as another
step on the long road towards zero memory leaks.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When fetching recursively with submodules, for each ref in the
superproject, we call check_for_new_submodule_commits() which collects all
the objects that have to be checked for submodule changes on
calculate_changed_submodule_paths(). On the first call, it also collects all
the existing refs for excluding them from the scan.
calculate_changed_submodule_paths() creates an argument array with all the
collected new objects, followed by --not and all the old objects. This argv
is passed to setup_revisions, which parses each argument, converts it back
to an oid and resolves the object. The parsing itself also does redundant
work, because it is treated like user input, while in fact it is a full
oid. So it needlessly attempts to look it up as ref (checks if it has ^, ~
etc.), checks if it is a file name etc.
For a repository with many refs, all of this is expensive. But if the fetch
in the superproject did not update the ref (i.e. the objects that are
required to exist in the submodule did not change), there is no need to
include it in the list.
Before commit be76c212 (fetch: ensure submodule objects fetched,
2018-12-06), submodule reference changes were only detected for refs that
were changed, but not for new refs. This commit covered also this case, but
what it did was to just include every ref.
This change should reduce the number of scanned refs by about half (except
the case of a no-op fetch, which will not scan any ref), because all the
existing refs will still be listed after --not.
The regression was reported here:
https://public-inbox.org/git/CAGHpTBKSUJzFSWc=uznSu2zB33qCSmKXM-
iAjxRCpqNK5bnhRg@mail.gmail.com/
Signed-off-by: Orgad Shaneh <orgads@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Updates to on-demand fetching code in lazily cloned repositories.
* jt/lazy-fetch:
fetch: no FETCH_HEAD display if --no-write-fetch-head
fetch-pack: remove no_dependents code
promisor-remote: lazy-fetch objects in subprocess
fetch-pack: do not lazy-fetch during ref iteration
fetch: only populate existing_refs if needed
fetch: avoid reading submodule config until needed
fetch: allow refspecs specified through stdin
negotiator/noop: add noop fetch negotiator
A handful of places in in-tree code still relied on being able to
execute the git subcommands, especially built-ins, in "git-foo"
form, which have been corrected.
* jc/undash-in-tree-git-callers:
credential-cache: use child_process.args
cvsexportcommit: do not run git programs in dashed form
transport-helper: do not run git-remote-ext etc. in dashed form
Trim an unused binary and turn a bunch of commands into built-in.
* jk/slimmed-down:
drop vcs-svn experiment
make git-fast-import a builtin
make git-bugreport a builtin
make credential helpers builtins
Makefile: drop builtins from MSVC pdb list
"git rebase -i" learns a bit more options.
* pw/rebase-i-more-options:
t3436: do not run git-merge-recursive in dashed form
rebase: add --reset-author-date
rebase -i: support --ignore-date
rebase -i: support --committer-date-is-author-date
am: stop exporting GIT_COMMITTER_DATE
rebase -i: add --ignore-whitespace flag
When a user checks out the upstream branch of HEAD, the upstream branch
not being a local branch, and then runs "git status", like this:
git clone $URL client
cd client
git checkout @{u}
git status
no status is printed, but instead an error message:
fatal: HEAD does not point to a branch
(This error message when running "git branch" persists even after
checking out other things - it only stops after checking out a branch.)
This is because "git status" reads the reflog when determining the "HEAD
detached" message, and thus attempts to DWIM "@{u}", but that doesn't
work because HEAD no longer points to a branch.
Therefore, when calculating the status of a worktree, tolerate dangling
marks. This is done by adding an additional parameter to
dwim_ref() and repo_dwim_ref().
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
887952b8c6 ("fetch: optionally allow disabling FETCH_HEAD update",
2020-08-18) introduced the ability to disable writing to FETCH_HEAD
during fetch, but did not suppress the "<source> -> FETCH_HEAD" message
when this ability is used. This message is misleading in this case,
because FETCH_HEAD is not written. Also, because "fetch" is used to
lazy-fetch missing objects in a partial clone, this significantly
clutters up the output in that case since the objects to be fetched are
potentially numerous.
Therefore, suppress this message when --no-write-fetch-head is passed
(but not when --dry-run is set).
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git restore/checkout --no-overlay" with wildcarded pathspec
mistakenly removed matching paths in subdirectories, which has been
corrected.
* rs/checkout-no-overlay-pathspec-fix:
checkout, restore: make pathspec recursive
Accesses to two pseudorefs have been updated to properly use ref
API.
* hn/refs-pseudorefs:
sequencer: treat REVERT_HEAD as a pseudo ref
builtin/commit: suggest update-ref for pseudoref removal
sequencer: treat CHERRY_PICK_HEAD as a pseudo ref
refs: make refs_ref_exists public
Long ago, we decided to use 3 threads by default when running the
index-pack task in parallel, which has been adjusted a bit upwards.
* jk/index-pack-w-more-threads:
index-pack: adjust default threading cap
p5302: count up to online-cpus for thread tests
p5302: disable thread-count parameter tests by default
The intention of `git init --separate-work-dir=<path>` is to move the
.git/ directory to a location outside of the main worktree. When used
within a linked worktree, however, rather than moving the .git/
directory as intended, it instead incorrectly moves the worktree's
.git/worktrees/<id> directory to <path>, thus disconnecting the linked
worktree from its parent repository and breaking the worktree in the
process since its local .git file no longer points at a location at
which it can find the object database. Fix this broken behavior.
An intentional side-effect of this change is that it also closes a
loophole not caught by ccf236a23a (init: disallow --separate-git-dir
with bare repository, 2020-08-09) in which the check to prevent
--separate-git-dir being used in conjunction with a bare repository was
unable to detect the invalid combination when invoked from within a
linked worktree. Therefore, add a test to verify that this loophole is
closed, as well.
Reported-by: Henré Botha <henrebotha@gmail.com>
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A linked worktree's .git file is a "gitfile" pointing at the
.git/worktrees/<id> directory within the repository. When `git init
--separate-git-dir=<path>` is used on an existing repository to relocate
the repository's .git/ directory to a different location, it neglects to
update the .git files of linked worktrees, thus breaking the worktrees
by making it impossible for them to locate the repository. Fix this by
teaching --separate-git-dir to repair the .git file of each linked
worktree to point at the new repository location.
Reported-by: Henré Botha <henrebotha@gmail.com>
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The .git/worktrees/<id>/gitdir file points at the location of a linked
worktree's .git file. Its content must be of the form
/path/to/worktree/.git (from which the location of the worktree itself
can be derived by stripping the "/.git" suffix). If the gitdir file is
deleted or becomes corrupted or outdated, then Git will be unable to
find the linked worktree. An easy way for the gitdir file to become
outdated is for the user to move the worktree manually (without using
"git worktree move"). Although it is possible to manually update the
gitdir file to reflect the new linked worktree location, doing so
requires a level of knowledge about worktree internals beyond what a
user should be expected to know offhand.
Therefore, teach "git worktree repair" how to repair broken or outdated
.git/worktrees/<id>/gitdir files automatically. (For this to work, the
command must either be invoked from within the worktree whose gitdir
file requires repair, or from within the main or any linked worktree by
providing the path of the broken worktree as an argument to "git
worktree repair".)
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The .git file in a linked worktree is a "gitfile" which points back to
the .git/worktrees/<id> entry in the main worktree or bare repository.
If a worktree's .git file is deleted or becomes corrupted or outdated,
then the linked worktree won't know how to find the repository or any of
its own administrative files (such as 'index', 'HEAD', etc.). An easy
way for the .git file to become outdated is for the user to move the
main worktree or bare repository. Although it is possible to manually
update each linked worktree's .git file to reflect the new repository
location, doing so requires a level of knowledge about worktree
internals beyond what a user should be expected to know offhand.
Therefore, teach "git worktree repair" how to repair broken or outdated
worktree .git files automatically. (For this to work, the command must
be invoked from within the main worktree or bare repository, or from
within a worktree which has not become disconnected from the repository
-- such as one which was created after the repository was moved.)
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let's refactor code adding a new `write_in_file()` function
that opens a file for writing a message and closes it and a
wrapper for writing mode.
This helper will be used in later steps and makes the code
simpler and easier to understand.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Miriam Rubio <mirucam@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Following 'enum bisect_error' vocabulary, return variable 'res' is
always non-positive.
Let's use '-res' instead of 'abs(res)' to make the code clearer.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Miriam Rubio <mirucam@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In cmd_bisect__helper() function, if an invalid or no
subcommand is passed there is a BUG.
BUG() out instead of returning an error.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Miriam Rubio <mirucam@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a repository has an alternate object directory configured, callers
can traverse through each alternate's MIDX by walking the '->next'
pointer.
But, when 'prepare_multi_pack_index_one()' loads multiple MIDXs, it
places the new ones at the front of this pointer chain, not at the end.
This can be confusing for callers such as 'git repack -ad', causing test
failures like in t7700.6 with 'GIT_TEST_MULTI_PACK_INDEX=1'.
The occurs when dropping a pack known to the local MIDX with alternates
configured that have their own MIDX. Since the alternate's MIDX is
returned via 'get_multi_pack_index()', 'midx_contains_pack()' returns
true (which is correct, since it traverses through the '->next' pointer
to find the MIDX in the chain that does contain the requested object).
But, we call 'clear_midx_file()' on 'the_repository', which drops the
MIDX at the path of the first MIDX in the chain, which (in the case of
t7700.6 is the one in the alternate).
This patch addresses that by:
- placing the local MIDX first in the chain when calling
'prepare_multi_pack_index_one()', and
- introducing a new 'get_local_multi_pack_index()', which explicitly
returns the repository-local MIDX, if any.
Don't impose an additional order on the MIDX's '->next' pointer beyond
that the first item in the chain must be local if one exists so that we
avoid a quadratic insertion.
Likewise, use 'get_local_multi_pack_index()' in
'remove_redundant_pack()' to fix the formerly broken t7700.6 when run
with 'GIT_TEST_MULTI_PACK_INDEX=1'.
Finally, note that the MIDX ordering invariant is only preserved by the
insertion order in 'prepare_packed_git()', which traverses through the
ODB's '->next' pointer, meaning we visit the local object store first.
This fragility makes this an undesirable long-term solution if more
callers are added, but it is acceptable for now since this is the only
caller.
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The positional arguments are specified in this order: "bad" then "good".
To avoid confusion, the options above the positional arguments
are now specified in the same order. They can still be specified in any
order since they're options, not positional arguments.
Signed-off-by: Hugo Locurcio <hugo.locurcio@hugo.pro>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Code clean-up.
* jk/leakfix:
submodule--helper: fix leak of core.worktree value
config: fix leak in git_config_get_expiry_in_days()
config: drop git_config_get_string_const()
config: fix leaks from git_config_get_string_const()
checkout: fix leak of non-existent branch names
submodule--helper: use strbuf_release() to free strbufs
clear_pattern_list(): clear embedded hashmaps
Add a new multi-valued config variable "receive.procReceiveRefs"
for `receive-pack` command, like the follows:
git config --system --add receive.procReceiveRefs refs/for
git config --system --add receive.procReceiveRefs refs/drafts
If the specific prefix strings given by the config variables match the
reference names of the commands which are sent from git client to
`receive-pack`, these commands will be executed by an external hook
(named "proc-receive"), instead of the internal `execute_commands`
function.
For example, if it is set to "refs/for", pushing to a reference such as
"refs/for/master" will not create or update reference "refs/for/master",
but may create or update a pull request directly by running the hook
"proc-receive".
Optional modifiers can be provided in the beginning of the value to
filter commands for specific actions: create (a), modify (m),
delete (d). A `!` can be included in the modifiers to negate the
reference prefix entry. E.g.:
git config --system --add receive.procReceiveRefs ad:refs/heads
git config --system --add receive.procReceiveRefs !:refs/heads
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The new introduced "proc-receive" hook may handle a command for a
pseudo-reference with a zero-old as its old-oid, while the hook may
create or update a reference with different name, different new-oid,
and different old-oid (the reference may exist already with a non-zero
old-oid). Current "report-status" protocol cannot report the status for
such reference rewrite.
Add new capability "report-status-v2" and new report protocol which is
not backward compatible for report of git-push.
If a user pushes to a pseudo-reference "refs/for/master/topic", and
"receive-pack" creates two new references "refs/changes/23/123/1" and
"refs/changes/24/124/1", for client without the knowledge of
"report-status-v2", "receive-pack" will only send "ok/ng" directives in
the report, such as:
ok ref/for/master/topic
But for client which has the knowledge of "report-status-v2",
"receive-pack" will use "option" directives to report more attributes
for the reference given by the above "ok/ng" directive.
ok refs/for/master/topic
option refname refs/changes/23/123/1
option new-oid <new-oid>
ok refs/for/master/topic
option refname refs/changes/24/124/1
option new-oid <new-oid>
The client will report two new created references to the end user.
Suggested-by: Junio C Hamano <gitster@pobox.com>
Suggested-by: Jeff King <peff@peff.net>
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When commands are fed to the "post-receive" hook, report options will
be parsed and the real old-oid, new-oid, reference name will feed to
the "post-receive" hook.
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Git calls an internal `execute_commands` function to handle commands
sent from client to `git-receive-pack`. Regardless of what references
the user pushes, git creates or updates the corresponding references if
the user has write-permission. A contributor who has no
write-permission, cannot push to the repository directly. So, the
contributor has to write commits to an alternate location, and sends
pull request by emails or by other ways. We call this workflow as a
distributed workflow.
It would be more convenient to work in a centralized workflow like what
Gerrit provided for some cases. For example, a read-only user who
cannot push to a branch directly can run the following `git push`
command to push commits to a pseudo reference (has a prefix "refs/for/",
not "refs/heads/") to create a code review.
git push origin \
HEAD:refs/for/<branch-name>/<session>
The `<branch-name>` in the above example can be as simple as "master",
or a more complicated branch name like "foo/bar". The `<session>` in
the above example command can be the local branch name of the client
side, such as "my/topic".
We cannot implement a centralized workflow elegantly by using
"pre-receive" + "post-receive", because Git will call the internal
function "execute_commands" to create references (even the special
pseudo reference) between these two hooks. Even though we can delete
the temporarily created pseudo reference via the "post-receive" hook,
having a temporary reference is not safe for concurrent pushes.
So, add a filter and a new handler to support this kind of workflow.
The filter will check the prefix of the reference name, and if the
command has a special reference name, the filter will turn a specific
field (`run_proc_receive`) on for the command. Commands with this filed
turned on will be executed by a new handler (a hook named
"proc-receive") instead of the internal `execute_commands` function.
We can use this "proc-receive" command to create pull requests or send
emails for code review.
Suggested by Junio, this "proc-receive" hook reads the commands,
push-options (optional), and send result using a protocol in pkt-line
format. In the following example, the letter "S" stands for
"receive-pack" and letter "H" stands for the hook.
# Version and features negotiation.
S: PKT-LINE(version=1\0push-options atomic...)
S: flush-pkt
H: PKT-LINE(version=1\0push-options...)
H: flush-pkt
# Send commands from server to the hook.
S: PKT-LINE(<old-oid> <new-oid> <ref>)
S: ... ...
S: flush-pkt
# Send push-options only if the 'push-options' feature is enabled.
S: PKT-LINE(push-option)
S: ... ...
S: flush-pkt
# Receive result from the hook.
# OK, run this command successfully.
H: PKT-LINE(ok <ref>)
# NO, I reject it.
H: PKT-LINE(ng <ref> <reason>)
# Fall through, let 'receive-pack' to execute it.
H: PKT-LINE(ok <ref>)
H: PKT-LINE(option fall-through)
# OK, but has an alternate reference. The alternate reference name
# and other status can be given in options
H: PKT-LINE(ok <ref>)
H: PKT-LINE(option refname <refname>)
H: PKT-LINE(option old-oid <old-oid>)
H: PKT-LINE(option new-oid <new-oid>)
H: PKT-LINE(option forced-update)
H: ... ...
H: flush-pkt
After receiving a command, the hook will execute the command, and may
create/update different reference. For example, a command for a pseudo
reference "refs/for/master/topic" may create/update different reference
such as "refs/pull/123/head". The alternate reference name and other
status are given in option lines.
The list of commands returned from "proc-receive" will replace the
relevant commands that are sent from user to "receive-pack", and
"receive-pack" will continue to run the "execute_commands" function and
other routines. Finally, the result of the execution of these commands
will be reported to end user.
The reporting function from "receive-pack" to "send-pack" will be
extended in latter commit just like what the "proc-receive" hook reports
to "receive-pack".
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'grep' check in test 4 of t7421 resulted in the failure of t7421 on
Windows due to a different error message
error: cannot spawn git: No such file or directory
instead of
fatal: exec 'rev-parse': cd to 'my-subm' failed: No such file or directory
Tighten up the check to compute 'src_abbrev' by guarding the
'verify_submodule_committish()' call using `p->status !='D'`, so that
the former isn't called in case of non-existent submodule directory,
consequently, there is no such error message on any execution
environment. The same need not be implemented for 'dst_abbrev' and is
rather redundant since the conditional 'if (S_ISGITLINK(p->mod_dst))'
already guards the 'verify_submodule_committish()' when we have a
status of 'D'.
Therefore, eliminate the 'grep' check in t7421. Instead, verify the
absence of an error message by doing a 'test_must_be_empty' on the
file containing the error.
Reported-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Helped-by: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
Signed-off-by: Shourya Shukla <shouryashukla.oo@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Worktree administrative files can become corrupted or outdated due to
external factors. Although, it is often possible to recover from such
situations by hand-tweaking these files, doing so requires intimate
knowledge of worktree internals. While information necessary to make
such repairs manually can be obtained from git-worktree.txt and
gitrepository-layout.txt, we can assist users more directly by teaching
git-worktree how to repair its administrative files itself (at least to
some extent). Therefore, add a "git worktree repair" command which
attempts to correct common problems which may arise due to factors
beyond Git's control.
At this stage, the "repair" command is a mere skeleton; subsequent
commits will flesh out the functionality.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We allocate a child_env strvec but never free its memory. Instead, let's
just use the strvec that our child_process struct provides, which is
cleaned up automatically when we run the command.
And while we're moving the initialization of the child_process around,
let's switch it to use the official init function (zero-initializing it
works OK, since strvec is happy enough with that, but it sets a bad
example).
Signed-off-by: Jeff King <peff@peff.net>
Reviewed-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 525e18c04b (midx: clear midx on repack, 2018-07-12), 'git repack'
learned to remove a multi-pack-index file if it added or removed a pack
from the object store.
This mechanism is a little over-eager, since it is only necessary to
drop a MIDX if 'git repack' removes a pack that the MIDX references.
Adding a pack outside of the MIDX does not require invalidating the
MIDX, and likewise for removing a pack the MIDX does not know about.
Teach 'git repack' to check for this by loading the MIDX, and checking
whether the to-be-removed pack is known to the MIDX. This requires a
slightly odd alternation to a test in t5319, which is explained with a
comment. A new test is added to show that the MIDX is left alone when
both packs known to it are marked as .keep, but two packs unknown to it
are removed and combined into one new pack.
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The definitions of 'verify_submodule_committish()' and
'print_submodule_summary()' had wrong styling in terms of the asterisk
placement. Amend them.
Also, the warning printed in case of an unexpected file mode printed the
mode in decimal. Print it in octal for enhanced readability.
Reported-by: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Shourya Shukla <shouryashukla.oo@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Eliminate the parameters 'missing_{src,dst}' from the
'print_submodule_summary()' function call since they are not used
anywhere in the function.
Reported-by: Jeff King <peff@peff.net>
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
Signed-off-by: Shourya Shukla <shouryashukla.oo@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The purpose of "git init --separate-git-dir" is to initialize a
new project with the repository separate from the working tree,
or, in the case of an existing project, to move the repository
(the .git/ directory) out of the working tree. It does not make
sense to use --separate-git-dir with a bare repository for which
there is no working tree, so disallow its use with bare
repositories.
* es/init-no-separate-git-dir-in-bare:
init: disallow --separate-git-dir with bare repository
A subsequent commit will make the quantum of work smaller, necessitating
more locking. This commit allows resolve_delta() to be called outside
the lock.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is refactoring 2 of 2 to simplify struct base_data.
Whenever we make a struct base_data, immediately calculate its delta
children. This eliminates confusion as to when the
{ref,ofs}_{first,last} fields are initialized.
Before this patch, the delta children were calculated at the last
possible moment. This allowed the members of struct base_data to be
populated in any order, superficially useful when we have the object
contents before the struct object_entry. But this makes reasoning about
the state of struct base_data more complicated, hence this patch.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is refactoring 1 of 2 to simplify struct base_data.
In index-pack, each thread maintains a doubly-linked list of the delta
chain that it is currently processing (the "base" and "child" pointers
in struct base_data). When a thread exceeds the delta base cache limit
and needs to reclaim memory, it uses the "child" pointers to traverse
the lineage, reclaiming the memory of the eldest delta bases first.
A subsequent patch will perform memory reclaiming in a different way and
will thus no longer need the "child" pointer. Because the "child"
pointer is redundant even now, remove it so that the aforementioned
subsequent patch will be clearer. In the meantime, reclaim memory in the
reverse order of the "base" pointers.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
find_{ref,ofs}_delta_{,children} take an enum object_type parameter, but
the object type is already present in the name of the function. Remove
that parameter from these functions.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The pathspec given to git checkout and git restore is used with both
tree_entry_interesting (via read_tree_recursive) and match_pathspec
(via ce_path_match). The latter effectively only supports recursive
matching regardless of the value of the pathspec flag "recursive",
which is unset here.
That causes different match results for pathspecs with wildcards, and
can lead checkout and restore in no-overlay mode to remove entries
instead of modifying them. Enable recursive matching for both checkout
and restore to make matching consistent.
Setting the flag in checkout_main() technically also affects git switch,
but since that command doesn't accept pathspecs at all this has no
actual consequence.
Reported-by: Sergii Shkarnikov <sergii.shkarnikov@globallogic.com>
Initial-test-by: Sergii Shkarnikov <sergii.shkarnikov@globallogic.com>
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit b8a2486f15 (index-pack: support multithreaded delta resolving,
2012-05-06) describes an experiment that shows that setting the number
of threads for index-pack higher than 3 does not help.
I repeated that experiment using a more modern version of Git and a more
modern CPU and got different results.
Here are timings for p5302 against linux.git run on my laptop, a Core
i9-9880H with 8 cores plus hyperthreading (so online-cpus returns 16):
5302.3: index-pack 0 threads 256.28(253.41+2.79)
5302.4: index-pack 1 threads 257.03(254.03+2.91)
5302.5: index-pack 2 threads 149.39(268.34+3.06)
5302.6: index-pack 4 threads 94.96(294.10+3.23)
5302.7: index-pack 8 threads 68.12(339.26+3.89)
5302.8: index-pack 16 threads 70.90(655.03+7.21)
5302.9: index-pack default number of threads 116.91(290.05+3.21)
You can see that wall-clock times continue to improve dramatically up to
the number of cores, but bumping beyond that (into hyperthreading
territory) does not help (and in fact hurts a little).
Here's the same experiment on a machine with dual Xeon 6230's, totaling
40 cores (80 with hyperthreading):
5302.3: index-pack 0 threads 310.04(302.73+6.90)
5302.4: index-pack 1 threads 310.55(302.68+7.40)
5302.5: index-pack 2 threads 178.17(304.89+8.20)
5302.6: index-pack 5 threads 99.53(315.54+9.56)
5302.7: index-pack 10 threads 72.80(327.37+12.79)
5302.8: index-pack 20 threads 60.68(357.74+21.66)
5302.9: index-pack 40 threads 58.07(454.44+67.96)
5302.10: index-pack 80 threads 59.81(720.45+334.52)
5302.11: index-pack default number of threads 134.18(309.32+7.98)
The results are similar; things stop improving at 40 threads. Curiously,
going from 20 to 40 really doesn't help much, either (and increases CPU
time considerably). So that may represent an actual barrier to
parallelism, where we lose out due to context-switching and loss of
cache locality, but don't reap the wall-clock benefits due to contention
of our coarse-grained locks.
So what's a good default value? It's clear that the current cap of 3 is
too low; our default values are 42% and 57% slower than the best times
on each machine. The results on the 40-core machine imply that 20
threads is an actual barrier regardless of the number of cores, so we'll
take that as a maximum. We get the best results on these machines at
half of the online-cpus value. That's presumably a result of the
hyperthreading. That's common on multi-core Intel processors, but not
necessarily elsewhere. But if we take it as an assumption, we can
perform optimally on hyperthreaded machines and still do much better
than the status quo on other machines, as long as we never half below
the current value of 3.
So that's what this patch does.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When pseudorefs move to a different ref storage mechanism, pseudorefs no longer
can be removed with 'rm'. Instead, suggest a "update-ref -d" command, which will
work regardless of ref storage backend.
Signed-off-by: Han-Wen Nienhuys <hanwen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Check for existence and delete CHERRY_PICK_HEAD through ref functions.
This will help cherry-pick work with alternate ref storage backends.
Signed-off-by: Han-Wen Nienhuys <hanwen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A few end-user facing messages have been updated to be
hash-algorithm agnostic.
* jc/object-names-are-not-sha-1:
messages: avoid SHA-1 in end-user facing messages
The previous commit introduced --ignore-date flag to rebase -i, but the
name is rather vague as it does not say whether the author date or the
committer date is ignored. Add an alias to convey the precise purpose.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Rohit Ashiwal <rohit.ashiwal265@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rebase is implemented with two different backends - 'apply' and
'merge' each of which support a different set of options. In
particular the apply backend supports a number of options implemented
by 'git am' that are not implemented in the merge backend. This means
that the available options are different depending on which backend is
used which is confusing. This patch adds support for the --ignore-date
option to the merge backend. This option uses the current time as the
author date rather than reusing the original author date when
rewriting commits. We take care to handle the combination of
--ignore-date and --committer-date-is-author-date in the same way as
the apply backend.
Original-patch-by: Rohit Ashiwal <rohit.ashiwal265@gmail.com>
Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The dir structure seemed to have a number of leaks and problems around
it. First I noticed that parent_hashmap and recursive_hashmap were
being leaked (though Peff noticed and submitted fixes before me). Then
I noticed in the previous commit that clear_directory() was only taking
responsibility for a subset of fields within dir_struct, despite the
fact that entries[] and ignored[] we allocated internally to dir.c.
That, of course, resulted in many callers either leaking or haphazardly
trying to free these arrays and their contents.
Digging further, I found that despite the pretty clear documentation
near the top of dir.h that folks were supposed to call clear_directory()
when the user no longer needed the dir_struct, there were four callers
that didn't bother doing that at all. However, two of them clearly
thought about leaks since they had an UNLEAK(dir) directive, which to me
suggests that the method to free the data was too unclear. I suspect
the non-obviousness of the API and its holes led folks to avoid it,
which then snowballed into further problems with the entries[],
ignored[], parent_hashmap, and recursive_hashmap problems.
Rename clear_directory() to dir_clear() to be more in line with other
data structures in git, and introduce a dir_init() to handle the
suggested memsetting of dir_struct to all zeroes. I hope that a name
like "dir_clear()" is more clear, and that the presence of dir_init()
will provide a hint to those looking at the code that they need to look
for either a dir_clear() or a dir_free() and lead them to find
dir_clear().
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The calling convention for the dir API is supposed to end with a call to
clear_directory() to free up no longer needed memory. However,
clear_directory() didn't free dir->entries or dir->ignored. I believe
this was an oversight, but a number of callers noticed memory leaks and
started free'ing these. Unfortunately, they did so somewhat haphazardly
(sometimes freeing the entries in the arrays, and sometimes only
free'ing the arrays themselves). This suggests the callers weren't
trying to make sure any possible memory used might be free'd, but just
the memory they noticed their usecase definitely had allocated.
Fix this mess by moving all the duplicated free'ing logic into
clear_directory(). End by resetting dir to a pristine state so it could
be reused if desired.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that Git has switched to using a subprocess to lazy-fetch missing
objects, remove the no_dependents code as it is no longer used.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In "fetch", get_ref_map() iterates over all refs to populate
"existing_refs" in order to populate peer_ref->old_oid in the returned
refmap, even if the refmap has no peer_ref set - which is the case when
only literal hashes (i.e. no refs by name) are fetched.
Iterating over refs causes the targets of those refs to be checked for
existence. Avoiding this is especially important when we use "git fetch"
to perform lazy fetches in a partial clone because a target of such a
ref may need to be itself lazy-fetched (and otherwise causing an
infinite loop).
Therefore, avoid populating "existing_refs" until necessary. With this
patch, because Git lazy-fetches objects by literal hashes (to be done in
a subsequent commit), it will then be able to guarantee avoiding reading
targets of refs.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In "fetch", there are two parameters submodule_fetch_jobs_config and
recurse_submodules that can be set in a variety of ways: through
.gitmodules, through .git/config, and through the command line.
Currently "fetch" handles this by first reading .gitmodules, then
reading .git/config (allowing it to overwrite existing values), then
reading the command line (allowing it to overwrite existing values).
Notice that we can avoid reading .gitmodules if .git/config and/or the
command line already provides us with what we need. In addition, if
recurse_submodules is found to be "no", we do not need the value of
submodule_fetch_jobs_config.
Avoiding reading .gitmodules is especially important when we use "git
fetch" to perform lazy fetches in a partial clone because the
.gitmodules file itself might need to be lazy fetched (and otherwise
causing an infinite loop).
In light of all this, avoid reading .gitmodules until necessary. When
reading it, we may only need one of the two parameters it provides, so
teach fetch_config_from_gitmodules() to support NULL arguments. With
this patch, users (including Git itself when invoking "git fetch" to
lazy-fetch) will be able to guarantee avoiding reading .gitmodules by
passing --recurse-submodules=no.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In a subsequent patch, partial clones will be taught to fetch missing
objects using a "git fetch" subprocess. Because the number of objects
fetched may be too numerous to fit on the command line, teach "fetch" to
accept refspecs passed through stdin.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If you run fetch but record the result in remote-tracking branches,
and either if you do nothing with the fetched refs (e.g. you are
merely mirroring) or if you always work from the remote-tracking
refs (e.g. you fetch and then merge origin/branchname separately),
you can get away with having no FETCH_HEAD at all.
Teach "git fetch" a command line option "--[no-]write-fetch-head".
The default is to write FETCH_HEAD, and the option is primarily
meant to be used with the "--no-" prefix to override this default,
because there is no matching fetch.writeFetchHEAD configuration
variable to flip the default to off (in which case, the positive
form may become necessary to defeat it).
Note that under "--dry-run" mode, FETCH_HEAD is never written;
otherwise you'd see list of objects in the file that you do not
actually have. Passing `--write-fetch-head` does not force `git
fetch` to write the file.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git log --first-parent -p" showed patches only for single-parent
commits on the first-parent chain; the "--first-parent" option has
been made to imply "-m". Use "--no-diff-merges" to restore the
previous behaviour to omit patches for merge commits.
* jk/log-fp-implies-m:
doc/git-log: clarify handling of merge commit diffs
doc/git-log: move "-t" into diff-options list
doc/git-log: drop "-r" diff option
doc/git-log: move "Diff Formatting" from rev-list-options
log: enable "-m" automatically with "--first-parent"
revision: add "--no-diff-merges" option to counteract "-m"
log: drop "--cc implies -m" logic
"git bisect" learns the "--first-parent" option to find the first
breakage along the first-parent chain.
* al/bisect-first-parent:
bisect: combine args passed to find_bisection()
bisect: introduce first-parent flag
cmd_bisect__helper: defer parsing no-checkout flag
rev-list: allow bisect and first-parent flags
t6030: modernize "git bisect run" tests
In the ensure_core_worktree() function, we load the core.worktree value
of the submodule repository using repo_config_get_string(). This
function copies the string, but we never free it, leaking the memory.
We can instead use the "tmp" version of that function to avoid the
allocation at all. We don't have to worry about lifetime issues, since
we never even look at the value (we just want to know if it's set).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rebase is implemented with two different backends - 'apply' and
'merge' each of which support a different set of options. In
particular the apply backend supports a number of options implemented
by 'git am' that are not implemented in the merge backend. This means
that the available options are different depending on which backend is
used which is confusing. This patch adds support for the
--committer-date-is-author-date option to the merge backend. This
option uses the author date of the commit that is being rewritten as
the committer date when the new commit is created.
Original-patch-by: Rohit Ashiwal <rohit.ashiwal265@gmail.com>
Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The implementation of --committer-date-is-author-date exports
GIT_COMMITTER_DATE to override the default committer date but does not
reset GIT_COMMITTER_DATE in the environment after creating the commit
so it is set in the environment of any hooks that get run. We're about
to add the same functionality to the sequencer and do not want to have
GIT_COMMITTER_DATE set when running hooks or exec commands so lets
update commit_tree_extended() to take an explicit committer so we
override the default date without setting GIT_COMMITTER_DATE in the
environment.
Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A couple of functions that used struct refspec_item did not zero out the
structure memory. This can result in unexpected behavior, especially if
additional parameters are ever added to refspec_item in the future. Use
memset to ensure that unset structure members are zero.
It may make sense to convert most of these uses of struct refspec_item
to use either struct initializers or refspec_item_init_or_die. However,
other similar code uses memset. Converting all of these uses has been
left as a future exercise.
Signed-off-by: Jacob Keller <jacob.keller@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are two functions to get a single config string:
- git_config_get_string()
- git_config_get_string_const()
One might naively think that the first one allocates a new string and
the second one just points us to the internal configset storage. But
in fact they both allocate a new copy; the second one exists only to
avoid having to cast when using it with a const global which we never
intend to free.
The documentation for the function explains that clearly, but it seems
I'm not alone in being surprised by this. Of 17 calls to the function,
13 of them leak the resulting value.
We could obviously fix these by adding the appropriate free(). But it
would be simpler still if we actually had a non-allocating way to get
the string. There's git_config_get_value() but that doesn't quite do
what we want. If the config key is present but is a boolean with no
value (e.g., "[foo]bar" in the file), then we'll get NULL (whereas the
string versions will print an error and die).
So let's introduce a new variant, git_config_get_string_tmp(), that
behaves as these callers expect. We need a new name because we have new
semantics but the same function signature (so even if we converted the
four remaining callers, topics in flight might be surprised). The "tmp"
is because this value should only be held onto for a short time. In
practice it's rare for us to clear and refresh the configset,
invalidating the pointer, but hopefully the "tmp" makes callers think
about the lifetime. In each of the converted cases here the value only
needs to last within the local function or its immediate caller.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We unconditionally write a branch name into a newly allocated buffer in
new_branch_info->path, via setup_branch_path(). We then check to see if
the branch exists; if not, we set that field to NULL, leaking the
memory. We should take care to free() it when doing so.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The prepare_to_clone_next_submodule() function has a few local-variable
strbufs. We use strbuf_reset() throughout the function to reuse the
buffers over and over. But at the end of the function we also use
strbuf_reset() as they go out of scope, which means we end up leaking
their heap buffers. This should be strbuf_release() instead.
These were introduced by 48308681b0 (git submodule update: have a
dedicated helper for cloning, 2016-02-29), but it doesn't seem to have
the same mistake elsewhere. Likewise, I looked for other instances of
the pattern in the submodule--helper file but couldn't find any.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are still a handful mentions of SHA-1 when we meant the
(hexadecimal) object names in end-user facing messages. Rewrite
them.
I was hoping that this can mostly be s/SHA-1/object name/, but
a few messages needed rephrasing to keep the result readable.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A new helper function has_object() has been introduced to make it
easier to mark object existence checks that do and don't want to
trigger lazy fetches, and a few such checks are converted using it.
* jt/has_object:
fsck: do not lazy fetch known non-promisor object
pack-objects: no fetch when allow-{any,promisor}
apply: do not lazy fetch when applying binary
sha1-file: introduce no-lazy-fetch has_object()
We UNLEAK() the "sorting" list created by parsing command-line options
(which is essentially used until the program exits). But we do so right
before leaving the cmd_ls_remote() function, which means we have to hit
all of the exits. But the point of UNLEAK() is that it's an annotation
which doesn't impact the variable itself. We can mark it as soon as
we're done writing its value, and then we only have to do so once.
This gives us a minor code reduction, and serves as a better example of
how UNLEAK() can be used.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There's no reason that git-fast-import benefits from being a separate
binary. And as it links against libgit.a, it has a non-trivial disk
footprint. Let's make it a builtin, which reduces the size of a stripped
installation from 22MB to 21MB.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There's no reason that bugreport has to be a separate binary. And since
it links against libgit.a, it has a rather large disk footprint. Let's
make it a builtin, which reduces the size of a stripped installation
from 24MB to 22MB.
This also simplifies our Makefile a bit. And we can take advantage of
builtin niceties like RUN_SETUP_GENTLY.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There's no real reason for credential helpers to be separate binaries. I
did them this way originally under the notion that helper don't _need_
to be part of Git, and so can be built totally separately (and indeed,
the ones in contrib/credential are). But the ones in our main Makefile
build on libgit.a, and the resulting binaries are reasonably large.
We can slim down our total disk footprint by just making them builtins.
This reduces the size of:
make strip install
from 29MB to 24MB on my Debian system.
Note that credential-cache can't operate without support for Unix
sockets. Currently we just don't build it at all when NO_UNIX_SOCKETS is
set. We could continue that with conditionals in the Makefile and our
list of builtins. But instead, let's build a dummy implementation that
dies with an informative message. That has two advantages:
- it's simpler, because the conditional bits are all kept inside
the credential-cache source
- a user who is expecting it to exist will be told _why_ they can't
use it, rather than getting the "credential-cache is not a git
command" error which makes it look like the Git install is broken.
Note that our dummy implementation does still respond to "-h" in order
to appease t0012 (and this may be a little friendlier for users, as
well).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Convert submodule subcommand 'summary' to a builtin and call it via
'git-submodule.sh'.
The shell version had to call $diff_cmd twice, once to find the modified
modules cared by the user and then again, with that list of modules
to do various operations for computing the summary of those modules.
On the other hand, the C version does not need a second call to
$diff_cmd since it reuses the module list from the first call to do the
aforementioned tasks.
In the C version, we use the combination of setting a child process'
working directory to the submodule path and then calling
'prepare_submodule_repo_env()' which also sets the 'GIT_DIR' to '.git',
so that we can be certain that those spawned processes will not access
the superproject's ODB by mistake.
A behavioural difference between the C and the shell version is that the
shell version outputs two line feeds after the 'git log' output when run
outside of the tests while the C version outputs one line feed in any
case. The reason for this is that the shell version calls log with
'--pretty=format:<fmt>' whose output is followed by two echo
calls; 'format' does not have "terminator" semantics like its 'tformat'
counterpart. So, the log output is terminated by a newline only when
invoked by the user and not when invoked from the scripts. This results
in the one & two line feed differences in the shell version.
On the other hand, the C version calls log with '--pretty=<fmt>'
which is equivalent to '--pretty:tformat:<fmt>' which is then
followed by a 'printf("\n")'. Due to its "terminator" semantics the
log output is always terminated by newline and hence one line feed in
any case.
Also, when we try to pass an option-like argument after a non-option
argument, for instance:
git submodule summary HEAD --foo-bar
(or)
git submodule summary HEAD --cached
That argument would be treated like a path to the submodule for which
the user is requesting a summary. So, the option ends up having no
effect. Though, passing '--quiet' is an exception to this:
git submodule summary HEAD --quiet
While 'summary' doesn't support '--quiet', we don't get an output for
the above command as '--quiet' is treated as a path which means we get
an output only if a submodule whose path is '--quiet' exists.
The error message in case of computing a summary for non-existent
submodules in the C version is different from that of the shell version.
Since the new error message is not marked for translation, change the
'test_i18ngrep' in t7421.4 to 'grep'.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Stefan Beller <stefanbeller@gmail.com>
Mentored-by: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
Helped-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Prathamesh Chavan <pc44800@gmail.com>
Signed-off-by: Shourya Shukla <shouryashukla.oo@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Many `submodule--helper` subcommands follow the convention that a struct
defines their callback data, and the declaration of that struct is
followed immediately by a macro to use in static initializers, without
any separating empty line.
Let's align the `init`, `status` and `sync` subcommands with that convention.
Mentored-by: Christian Couder <chriscool@tuxfamily.org>
Mentored-by: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
Helped-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Helped-by: Philip Oakley <philipoakley@iee.email>
Signed-off-by: Shourya Shukla <shouryashukla.oo@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The final leg of SHA-256 transition.
* bc/sha-256-part-3: (39 commits)
t: remove test_oid_init in tests
docs: add documentation for extensions.objectFormat
ci: run tests with SHA-256
t: make SHA1 prerequisite depend on default hash
t: allow testing different hash algorithms via environment
t: add test_oid option to select hash algorithm
repository: enable SHA-256 support by default
setup: add support for reading extensions.objectformat
bundle: add new version for use with SHA-256
builtin/verify-pack: implement an --object-format option
http-fetch: set up git directory before parsing pack hashes
t0410: mark test with SHA1 prerequisite
t5308: make test work with SHA-256
t9700: make hash size independent
t9500: ensure that algorithm info is preserved in config
t9350: make hash size independent
t9301: make hash size independent
t9300: use $ZERO_OID instead of hard-coded object ID
t9300: abstract away SHA-1-specific constants
t8011: make hash size independent
...
Update "git help guides" documentation organization.
* pb/guide-docs:
git.txt: add list of guides
Documentation: don't hardcode command categories twice
help: drop usage of 'common' and 'useful' for guides
command-list.txt: add missing 'gitcredentials' and 'gitremote-helpers'
All "mergy" operations that internally use the merge-recursive
machinery should honor the merge.renormalize configuration, but
many of them didn't.
* en/eol-attrs-gotchas:
checkout: support renormalization with checkout -m <paths>
merge: make merge.renormalize work for all uses of merge machinery
t6038: remove problematic test
t6038: make tests fail for the right reason
The argv_array API is useful for not just managing argv but any
"vector" (NULL-terminated array) of strings, and has seen adoption
to a certain degree. It has been renamed to "strvec" to reduce the
barrier to adoption.
* jk/strvec:
strvec: rename struct fields
strvec: drop argv_array compatibility layer
strvec: update documention to avoid argv_array
strvec: fix indentation in renamed calls
strvec: convert remaining callers away from argv_array name
strvec: convert more callers away from argv_array name
strvec: convert builtin/ callers away from argv_array name
quote: rename sq_dequote_to_argv_array to mention strvec
strvec: rename files from argv-array to strvec
argv-array: rename to strvec
argv-array: use size_t for count and alloc
The purpose of "git init --separate-git-dir" is to separate the
repository from the worktree. This is true even when --separate-git-dir
is used on an existing worktree, in which case, it moves the .git/
subdirectory to a new location outside the worktree.
However, an outright bare repository (such as one created by "git init
--bare"), has no worktree, so using --separate-git-dir to separate it
from its non-existent worktree is nonsensical. Therefore, make it an
error to use --separate-git-dir on a bare repository.
Implementation note: "git init" considers a repository bare if told so
explicitly via --bare or if it guesses it to be so based upon
heuristics. In the explicit --bare case, a conflict with
--separate-git-dir is easy to detect early. In the guessed case,
however, the conflict can only be detected once "bareness" is guessed,
which happens after "git init" has begun creating the repository.
Technically, we can get by with a single late check which would cover
both cases, however, erroring out early, when possible, without leaving
detritus provides a better user experience.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that find_bisection() accepts multiple boolean arguments, these may
be combined into a single unsigned integer in order to declutter some of
the code in bisect.c
Also, rename the existing "flags" bitfield to "commit_flags", to
explicitly differentiate it from the new "bisect_flags" bitfield.
Based-on-patch-by: Harald Nordgren <haraldnordgren@gmail.com>
Signed-off-by: Aaron Lipman <alipman88@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Upon seeing a merge commit when bisecting, this option may be used to
follow only the first parent.
In detecting regressions introduced through the merging of a branch, the
merge commit will be identified as introduction of the bug and its
ancestors will be ignored.
This option is particularly useful in avoiding false positives when a
merged branch contained broken or non-buildable commits, but the merge
itself was OK.
Signed-off-by: Aaron Lipman <alipman88@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
cmd_bisect__helper() is intended as a temporary shim layer serving as an
interface for git-bisect.sh. This function and git-bisect.sh should
eventually be replaced by a C implementation, cmd_bisect(), serving as
an entrypoint for all "git bisect ..." shell commands: cmd_bisect() will
only parse the first token following "git bisect", and dispatch the
remaining args to the appropriate function ["bisect_start()",
"bisect_next()", etc.].
Thus, cmd_bisect__helper() should not be responsible for parsing flags
like --no-checkout. Instead, let the --no-checkout flag remain in the
argv array, so it may be evaluated alongside the other options already
parsed by bisect_start().
Signed-off-by: Aaron Lipman <alipman88@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add first_parent_only parameter to find_bisection(), removing the
barrier that prevented combining the --bisect and --first-parent flags
when using git rev-list
Based-on-patch-by: Tiago Botelho <tiagonbotelho@hotmail.com>
Signed-off-by: Aaron Lipman <alipman88@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There is a call to has_object_file(), which lazily fetches missing
objects in a partial clone, when the object is known to not be
a promisor object. Change that call to has_object(), which does not do
any lazy fetching.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The options --missing=allow-{any,promisor} were introduced in caf3827e2f
("rev-list: add list-objects filtering support", 2017-11-22) with the
following note in the commit message:
This patch introduces handling of missing objects to help
debugging and development of the "partial clone" mechanism,
and once the mechanism is implemented, for a power user to
perform operations that are missing-object aware without
incurring the cost of checking if a missing link is expected.
The idea that these options are missing-object aware (and thus do not
need to lazily fetch objects, unlike unaware commands that assume that
all objects are present) are assumed in later commits such as 07ef3c6604
("fetch test: use more robust test for filtered objects", 2020-01-15).
However, the current implementations of these options use
has_object_file(), which indeed lazily fetches missing objects. Teach
these implementations not to do so. Also, update the documentation of
these options to be clearer.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 1b81d8cb19 (help: use command-list.txt for the source of guides,
2018-05-20), all man5/man7 guides listed in command-list.txt appear in
the output of 'git help -g'.
However, 'git help -g' still prefixes this list with "The common Git
guides are:", which makes one wonder if there are others!
In the same spirit, the man page for 'git help' describes the '--guides'
option as listing 'useful' guides, which is not false per se but can
also be taken to mean that there are other guides that exist but are not
useful.
Instead of 'common' and 'useful', use 'Git concept guides' in both
places. To keep the code in line with this change, rename
help.c::list_common_guides_help to list_guides_help.
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While packing many objects in a repository with a promissor remote,
lazily fetching missing objects from the promissor remote one by
one may be inefficient---the code now attempts to fetch all the
missing objects in batch (obviously this won't work for a lazy
clone that lazily fetches tree objects as you cannot even enumerate
what blobs are missing until you learn which trees are missing).
* jt/pack-objects-prefetch-in-batch:
pack-objects: prefetch objects to be packed
pack-objects: refactor to oid_object_info_extended
The 'merge' command is not the only one that does merges; other commands
like checkout -m or rebase do as well. Unfortunately, the only area of
the code that checked for the "merge.renormalize" config setting was in
builtin/merge.c, meaning it could only affect merges performed by the
"merge" command. Move the handling of this config setting to
merge_recursive_config() so that other commands can benefit from it as
well. Fixes a few tests in t6038.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "argc" and "argv" names made sense when the struct was argv_array,
but now they're just confusing. Let's rename them to "nr" (which we use
for counts elsewhere) and "v" (which is rather terse, but reads well
when combined with typical variable names like "args.v").
Note that we have to update all of the callers immediately. Playing
tricks with the preprocessor is hard here, because we wouldn't want to
rewrite unrelated tokens.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git mv src dst", when src is an unmerged path, errored out
correctly but with an incorrect error message to claim that src is
not tracked, which has been clarified.
* ct/mv-unmerged-path-error:
git-mv: improve error message for conflicted file
Preliminary clean-up of the refs API in preparation for adding a
new refs backend "reftable".
* hn/reftable:
reflog: cleanse messages in the refs.c layer
bisect: treat BISECT_HEAD as a pseudo ref
t3432: use git-reflog to inspect the reflog for HEAD
lib-t6000.sh: write tag using git-update-ref
"git clone --separate-git-dir=$elsewhere" used to stomp on the
contents of the existing directory $elsewhere, which has been
taught to fail when $elsewhere is not an empty directory.
* bw/fail-cloning-into-non-empty:
git clone: don't clone into non-empty directory
Updates to the changed-paths bloom filter.
* ds/commit-graph-bloom-updates:
commit-graph: check all leading directories in changed path Bloom filters
revision: empty pathspecs should not use Bloom filters
revision.c: fix whitespace
commit-graph: check chunk sizes after writing
commit-graph: simplify chunk writes into loop
commit-graph: unify the signatures of all write_graph_chunk_*() functions
commit-graph: persist existence of changed-paths
bloom: fix logic in get_bloom_filter()
commit-graph: change test to die on parse, not load
commit-graph: place bloom_settings in context
Now that we have a complete SHA-256 implementation in Git, let's enable
it so people can use it. Remove the ENABLE_SHA256 define constant
everywhere it's used. Add tests for initializing a repository with
SHA-256.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Reviewed-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently we detect the hash algorithm in use by the length of the
object ID. This is inelegant and prevents us from using a different
hash algorithm that is also 256 bits in length.
Since we cannot extend the v2 format in a backward-compatible way, let's
add a v3 format, which is identical, except for the addition of
capabilities, which are prefixed by an at sign. We add "object-format"
as the only capability and reject unknown capabilities, since we do not
have a network connection and therefore cannot negotiate with the other
side.
For compatibility, default to the v2 format for SHA-1 and require v3
for SHA-256.
In t5510, always use format v3 so we can be sure we produce consistent
results across hash algorithms. Since head -n N lists the top N lines
instead of the Nth line, let's run our output through sed to normalize
it and compare it against a fixed value, which will make sure we get
exactly what we're expecting.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Reviewed-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A recently added test in t5702 started using git verify-pack outside of
a repository. While this poses no problems with SHA-1, with SHA-256 we
implicitly rely on the setup of the repository to initialize our hash
algorithm settings.
Since we're not in a repository here, we need to provide git verify-pack
help to set things up properly. git index-pack already knows an
--object-format option, so let's accept one as well and pass it down to
our git index-pack invocation. Since we're now dynamically adjusting
the elements in argv, let's switch to using struct argv_array to manage
them. Finally, let's make t5702 pass the proper argument on down to its
git verify-pack caller.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Reviewed-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When using "--first-parent" to consider history as a single line of
commits, git-log still defaults to treating merges specially, even
though they could be considered as single commits in the linearized
history (that just introduce all of the changes from the second and
higher parents).
Let's instead have "--first-parent" imply "-m", which makes something
like:
git log --first-parent -p
do what you'd expect. Likewise:
git log --first-parent -Sfoo
will find "foo" in merge commits.
No new test is needed; we'll tweak the output of the existing
"--first-parent -p" test, which now matches the "-m --first-parent -p"
test. The unchanged existing test for "--no-diff-merges" confirms that
the user can get the old behavior if they want.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "-m" option sets revs->ignore_merges to "0", but there's no way to
undo it. This probably isn't something anybody overly cares about, since
"1" is already the default, but it will serve as an escape hatch when we
flip the default for ignore_merges to "0" in more situations.
We'll also add a few extra niceties:
- initialize the value to "-1" to indicate "not set", and then resolve
it to the normal 0/1 bool in setup_revisions(). This lets any tweak
functions, as well as setup_revisions() itself, avoid clobbering the
user's preference (which until now they couldn't actually express).
- since we now have --no-diff-merges, let's add the matching
--diff-merges, which is just a synonym for "-m". Then we don't even
need to document --no-diff-merges separately; it countermands the
long form of "-m" in the usual way.
The new test shows that this behaves just the same as the current
behavior without "-m".
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This was added by 82dee4160c (log: show merge commit when --cc is given,
2015-08-20), which explains why we need it. But that commit failed to
notice that setup_revisions() already does the same thing, since
cd2bdc5309 (Common option parsing for "git log --diff" and friends,
2006-04-14).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
parse_object_or_die() is passed an object ID and a name to show if the
object cannot be parsed. If the name is NULL then it shows the
hexadecimal object ID. Use that feature instead of preparing and
passing the hexadecimal representation to the function proactively.
That's shorter and a bit more efficient.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Code which split an argv_array call across multiple lines, like:
argv_array_pushl(&args, "one argument",
"another argument", "and more",
NULL);
was recently mechanically renamed to use strvec, which results in
mis-matched indentation like:
strvec_pushl(&args, "one argument",
"another argument", "and more",
NULL);
Let's fix these up to align the arguments with the opening paren. I did
this manually by sifting through the results of:
git jump grep 'strvec_.*,$'
and liberally applying my editor's auto-format. Most of the changes are
of the form shown above, though I also normalized a few that had
originally used a single-tab indentation (rather than our usual style of
aligning with the open paren). I also rewrapped a couple of obvious
cases (e.g., where previously too-long lines became short enough to fit
on one), but I wasn't aggressive about it. In cases broken to three or
more lines, the grouping of arguments is sometimes meaningful, and it
wasn't worth my time or reviewer time to ponder each case individually.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We eventually want to drop the argv_array name and just use strvec
consistently. There's no particular reason we have to do it all at once,
or care about interactions between converted and unconverted bits.
Because of our preprocessor compat layer, the names are interchangeable
to the compiler (so even a definition and declaration using different
names is OK).
This patch converts all of the files in builtin/ to keep the diff to a
manageable size.
The conversion was done purely mechanically with:
git ls-files '*.c' '*.h' |
xargs perl -i -pe '
s/ARGV_ARRAY/STRVEC/g;
s/argv_array/strvec/g;
'
and then selectively staging files with "git add builtin/". We'll deal
with any indentation/style fallouts separately.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We want to eventually drop the use of the "argv_array" name in favor of
"strvec." Unlike most other uses of the name, this one is embedded in a
function name, so the definition and all of the callers need to be
updated at the same time.
We don't technically need to update the parameter types here (our
preprocessor compat macros make the two names interchangeable), but
let's do so to keep the site consistent for now.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This requires updating #include lines across the code-base, but that's
all fairly mechanical, and was done with:
git ls-files '*.c' '*.h' |
xargs perl -i -pe 's/argv-array.h/strvec.h/'
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When an object to be packed is noticed to be missing, prefetch all
to-be-packed objects in one batch.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use oid_object_info_extended() instead of oid_object_info() because a
subsequent commit needs to specify an additional flag here.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git mv' has always complained about renaming a conflicted
file, as it cannot handle multiple index entries for one file.
However, the error message it uses has been the same as the
one for an untracked file:
fatal: not under version control, src=...
which is patently wrong. Distinguish the two cases and
add a test to make sure we produce the correct message.
Signed-off-by: Chris Torek <chris.torek@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix to the code to produce progress bar, which is new in the
upcoming release.
* tb/commit-graph-no-check-oids:
commit-graph: fix "Collecting commits from input" progress line
To display a progress line while reading commits from standard input
and looking them up, 5b6653e523 (builtin/commit-graph.c: dereference
tags in builtin, 2020-05-13) should have added a pair of
start_delayed_progress() and stop_progress() calls around the loop
reading stdin. Alas, the stop_progress() call ended up at the wrong
place, after write_commit_graph(), which does all the commit-graph
computation and writing, and has several progress lines of its own.
Consequently, that new
Collecting commits from input: 1234
progress line is overwritten by the first progress line shown by
write_commit_graph(), and its final "done" line is shown last, after
everything is finished:
$ { sleep 3 ; git rev-list -3 HEAD ; sleep 1 ; } | ~/src/git/git commit-graph write --stdin-commits
Expanding reachable commits in commit graph: 873402, done.
Writing out commit graph in 4 passes: 100% (3493608/3493608), done.
Collecting commits from input: 3, done.
Furthermore, that stop_progress() call was added after the 'cleanup'
label, where that loop reading stdin jumps in case of an error. In
case of invalid input this then results in the "done" line shown after
the error message:
$ { sleep 3 ; git rev-list -3 HEAD ; echo junk ; } | ~/src/git/git commit-graph write --stdin-commits
error: unexpected non-hex object ID: junk
Collecting commits from input: 3, done.
Move that stop_progress() call to the right place.
While at it, drop the unnecessary 'if (progress)' condition protecting
the stop_progress() call, because that function is prepared to handle
a NULL progress struct.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rebase is implemented with two different backends - 'apply' and
'merge' each of which support a different set of options. In
particular the apply backend supports a number of options implemented
by 'git am' that are not implemented in the merge backend. This means
that the available options are different depending on which backend is
used which is confusing. This patch adds support for the
--ignore-whitespace option to the merge backend. This option treats
lines with only whitespace changes as unchanged and is implemented in
the merge backend by translating it to -Xignore-space-change.
Signed-off-by: Rohit Ashiwal <rohit.ashiwal265@gmail.com>
Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Both the git-bisect.sh as bisect--helper inspected the file system
directly.
Signed-off-by: Han-Wen Nienhuys <hanwen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When using git clone with --separate-git-dir realgitdir and
realgitdir already exists, it's content is destroyed.
So, make sure we don't clone into an existing non-empty directory.
When d45420c1 (clone: do not clean up directories we didn't create,
2018-01-02) tightened the clean-up procedure after a failed cloning
into an empty directory, it assumed that the existing directory
given is an empty one so it is OK to keep that directory, while
running the clean-up procedure that is designed to remove everything
in it (since there won't be any, anyway). Check and make sure that
the $GIT_DIR is empty even cloning into an existing repository.
Signed-off-by: Ben Wijen <ben@wijen.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Recent update to "git diff" meant as a code clean-up introduced a
bug in its error handling code, which has been corrected.
* ct/diff-with-merge-base-clarification:
diff: check for merge bases before assigning sym->base
In symdiff_prepare(), we iterate over the set of parsed objects to pick
out any symmetric differences, including the left, right, and base
elements. We assign the results into pointers in a "struct symdiff", and
then complain if we didn't find a base, like so:
sym->left = rev->pending.objects[lpos].name;
sym->right = rev->pending.objects[rpos].name;
sym->base = rev->pending.objects[basepos].name;
if (basecount == 0)
die(_("%s...%s: no merge base"), sym->left, sym->right);
But the least lines are backwards. If basecount is 0, then basepos will
be -1, and we will access memory outside of the pending array. This
isn't usually that big a deal, since we don't do anything besides a
single pointer-sized read before exiting anyway, but it does violate the
C standard, and of course memory-checking tools like ASan complain.
Let's put the basecount check first. Note that we haveto split it from
the other assignments, since the die() relies on sym->left and
sym->right having been assigned (this isn't strictly necessary, but is
easier to read than dereferencing the pending array again).
Reported-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git fast-export --anonymize" learned to take customized mapping to
allow its users to tweak its output more usable for debugging.
* jk/fast-export-anonym-alt:
fast-export: use local array to store anonymized oid
fast-export: anonymize "master" refname
fast-export: allow seeding the anonymized mapping
fast-export: add a "data" callback parameter to anonymize_str()
fast-export: move global "idents" anonymize hashmap into function
fast-export: use a flex array to store anonymized entries
fast-export: stop storing lengths in anonymized hashmaps
fast-export: tighten anonymize_mem() interface to handle only strings
fast-export: store anonymized oids as hex strings
fast-export: use xmemdupz() for anonymizing oids
t9351: derive anonymized tree checks from original repo
The name of the primary branch in existing repositories, and the
default name used for the first branch in newly created
repositories, is made configurable, so that we can eventually wean
ourselves off of the hardcoded 'master'.
* js/default-branch-name:
contrib: subtree: adjust test to change in fmt-merge-msg
testsvn: respect `init.defaultBranch`
remote: use the configured default branch name when appropriate
clone: use configured default branch name when appropriate
init: allow setting the default for the initial branch name via the config
init: allow specifying the initial branch name for the new repository
docs: add missing diamond brackets
submodule: fall back to remote's HEAD for missing remote.<name>.branch
send-pack/transport-helper: avoid mentioning a particular branch
fmt-merge-msg: stop treating `master` specially
API cleanup for get_worktrees()
* es/get-worktrees-unsort:
worktree: drop get_worktrees() unused 'flags' argument
worktree: drop get_worktrees() special-purpose sorting option
A few fields in "struct commit" that do not have to always be
present have been moved to commit slabs.
* ak/commit-graph-to-slab:
commit-graph: minimize commit_graph_data_slab access
commit: move members graph_pos, generation to a slab
commit-graph: introduce commit_graph_data_slab
object: drop parsed_object_pool->commit_count
SHA-256 migration work continues.
* bc/sha-256-part-2: (44 commits)
remote-testgit: adapt for object-format
bundle: detect hash algorithm when reading refs
t5300: pass --object-format to git index-pack
t5704: send object-format capability with SHA-256
t5703: use object-format serve option
t5702: offer an object-format capability in the test
t/helper: initialize the repository for test-sha1-array
remote-curl: avoid truncating refs with ls-remote
t1050: pass algorithm to index-pack when outside repo
builtin/index-pack: add option to specify hash algorithm
remote-curl: detect algorithm for dumb HTTP by size
builtin/ls-remote: initialize repository based on fetch
t5500: make hash independent
serve: advertise object-format capability for protocol v2
connect: parse v2 refs with correct hash algorithm
connect: pass full packet reader when parsing v2 refs
Documentation/technical: document object-format for protocol v2
t1302: expect repo format version 1 for SHA-256
builtin/show-index: provide options to determine hash algo
t5302: modernize test formatting
...
When displaying cat-file usage, the fact that a <format> can
be specified is only visible when lookling at the --batch and
--batch-check options which are shown like this:
--batch[=<format>] show info and content of objects fed from the standard input
--batch-check[=<format>]
show info about objects fed from the standard input
It seems more coherent and improves discovery to also show it
on the usage line.
In the documentation the DESCRIPTION tells us that "The output
format can be overridden using the optional <format> argument",
but we can't see the <format> argument in the SYNOPSIS above
the description which is confusing.
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The changed-path Bloom filters were released in v2.27.0, but have a
significant drawback. A user can opt-in to writing the changed-path
filters using the "--changed-paths" option to "git commit-graph write"
but the next write will drop the filters unless that option is
specified.
This becomes even more important when considering the interaction with
gc.writeCommitGraph (on by default) or fetch.writeCommitGraph (part of
features.experimental). These config options trigger commit-graph writes
that the user did not signal, and hence there is no --changed-paths
option available.
Allow a user that opts-in to the changed-path filters to persist the
property of "my commit-graph has changed-path filters" automatically. A
user can drop filters using the --no-changed-paths option.
In the process, we need to be extremely careful to match the Bloom
filter settings as specified by the commit-graph. This will allow future
versions of Git to customize these settings, and the version with this
change will persist those settings as commit-graphs are rewritten on
top.
Use the trace2 API to signal the settings used during the write, and
check that output in a test after manually adjusting the correct bytes
in the commit-graph file.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git diff-files" has been taught to say paths that are marked as
intent-to-add are new files, not modified from an empty blob.
* sk/diff-files-show-i-t-a-as-new:
diff-files: treat "i-t-a" files as "not-in-index"
An in-code comment in "git diff" has been updated.
* dl/diff-usage-comment-update:
builtin/diff: fix botched update of usage comment
builtin/diff: update usage comment
Allow runtime upgrade of the repository format version, which needs
to be done carefully.
There is a rather unpleasant backward compatibility worry with the
last step of this series, but it is the right thing to do in the
longer term.
* xl/upgrade-repo-format:
check_repository_format_gently(): refuse extensions for old repositories
sparse-checkout: upgrade repository to version 1 when enabling extension
fetch: allow adding a filter after initial clone
repository: add a helper function to perform repository format upgrade
Some older versions of gcc complain about this line:
builtin/fast-export.c:412:2: error: dereferencing type-punned pointer
will break strict-aliasing rules [-Werror=strict-aliasing]
put_be32(oid.hash + hashsz - 4, counter++);
^
This seems to be a false positive, as there's no type-punning at all
here. oid.hash is an array of unsigned char; when we pass it to a
function it decays to a pointer to unsigned char. We do take a void
pointer in put_be32(), but it's immediately aliased with another pointer
to unsigned char (and clearly the compiler is looking inside the inlined
put_be32(), since the warning doesn't happen with -O0).
This happens on gcc 4.8 and 4.9, but not later versions (I tested gcc 6,
7, 8, and 9).
We can work around it by using a local array instead of an object_id
struct. This is a little more intimate with the details of object_id,
but for whatever reason doesn't seem to trigger the compiler warning.
We can revert this patch once we decide that those gcc versions are too
old to care about for a warning like this (gcc 4.8 is the default
compiler for Ubuntu Trusty, which is out-of-support but not fully
end-of-life'd until April 2022).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Running "fast-export --anonymize" will leave "refs/heads/master"
untouched in the output, for two reasons:
- it helped to have some known reference point between the original
and anonymized repository
- since it's historically the default branch name, it doesn't leak any
information
Now that we can ask fast-export to retain particular tokens, we have a
much better tool for the first one (because it works for any ref, not
just master).
For the second, the notion of "default branch name" is likely to become
configurable soon, at which point the name _does_ leak information.
Let's drop this special case in preparation.
Note that we have to adjust the test a bit, since it relied on using the
name "master" in the anonymized repos. We could just use
--anonymize-map=master to keep the same output, but then we wouldn't
know if it works because of our hard-coded master or because of the
explicit map.
So let's flip the test a bit, and confirm that we anonymize "master",
but keep "other" in the output.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
After you anonymize a repository, it can be hard to find which commits
correspond between the original and the result, and thus hard to
reproduce commands that triggered bugs in the original.
Let's make it possible to seed the anonymization map. This lets users
either:
- mark names to be retained as-is, if they don't consider them secret
(in which case their original commands would just work)
- map names to new values, which lets them adapt the reproduction
recipe to the new names without revealing the originals
The implementation is fairly straight-forward. We already store each
anonymized token in a hashmap (so that the same token appearing twice is
converted to the same result). We can just introduce a new "seed"
hashmap which is consulted first.
This does make a few more promises to the user about how we'll anonymize
things (e.g., token-splitting pathnames). But it's unlikely that we'd
want to change those rules, even if the actual anonymization of a single
token changes. And it makes things much easier for the user, who can
unblind only a directory name without having to specify each path within
it.
One alternative to this approach would be to anonymize as we see fit,
and then dump the whole refname and pathname mappings to a file. This
does work, but it's a bit awkward to use (you have to manually dig the
items you care about out of the mapping).
Helped-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "fetch/clone" protocol has been updated to allow the server to
instruct the clients to grab pre-packaged packfile(s) in addition
to the packed object data coming over the wire.
* jt/cdn-offload:
upload-pack: fix a sparse '0 as NULL pointer' warning
upload-pack: send part of packfile response as uri
fetch-pack: support more than one pack lockfile
upload-pack: refactor reading of pack-objects out
Documentation: add Packfile URIs design doc
Documentation: order protocol v2 sections
http-fetch: support fetching packfiles by URL
http-fetch: refactor into function
http: refactor finish_http_pack_request()
http: use --stdin when indexing dumb HTTP pack
Rewrite of parts of the scripted "git submodule" Porcelain command
continues; this time it is "git submodule set-branch" subcommand's
turn.
* ss/submodule-set-branch-in-c:
submodule: port subcommand 'set-branch' from shell to C
Code clean-up around "git branch" with a minor bugfix.
* dl/branch-cleanup:
branch: don't mix --edit-description
t3200: test for specific errors
t3200: rename "expected" to "expect"
"git diff" used to take arguments in random and nonsense range
notation, e.g. "git diff A..B C", "git diff A..B C...D", etc.,
which has been cleaned up.
* ct/diff-with-merge-base-clarification:
Documentation: usage for diff combined commits
git diff: improve range handling
t/t3430: avoid undefined git diff behavior
Code clean-up of "git clean" resulted in a fix of recent
performance regression.
* en/clean-cleanups:
clean: optimize and document cases where we recurse into subdirectories
clean: consolidate handling of ignored parameters
dir, clean: avoid disallowed behavior
dir: fix a few confusing comments
When cloning a repository without any branches, Git chooses a default
branch name for the as-yet unborn branch.
As part of the implicit initialization of the local repository, Git just
learned to respect `init.defaultBranch` to choose a different initial
branch name. We now really want that branch name to be used as a
fall-back.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We just introduced the command-line option
`--initial-branch=<branch-name>` to allow initializing a new repository
with a different initial branch than the hard-coded one.
To allow users to override the initial branch name more permanently
(i.e. without having to specify the name manually for each and every
`git init` invocation), let's introduce the `init.defaultBranch` config
setting.
Helped-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Don Goodman-Wilson <don@goodman-wilson.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There is a growing number of projects and companies desiring to change
the main branch name of their repositories (see e.g.
https://twitter.com/mislav/status/1270388510684598272 for background on
this).
To change that branch name for new repositories, currently the only way
to do that automatically is by copying all of Git's template directory,
then hard-coding the desired default branch name into the `.git/HEAD`
file, and then configuring `init.templateDir` to point to those copied
template files.
To make this process much less cumbersome, let's introduce a new option:
`--initial-branch=<branch-name>`.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When `remote.<name>.branch` is not configured, `git submodule update`
currently falls back to using the branch name `master`. A much better
idea, however, is to use the remote `HEAD`: on all Git servers running
reasonably recent Git versions, the symref `HEAD` points to the main
branch.
Note: t7419 demonstrates that there _might_ be use cases out there that
_expect_ `git submodule update --remote` to update submodules to the
remote `master` branch even if the remote `HEAD` points to another
branch. Arguably, this patch makes the behavior more intuitive, but
there is a slight possibility that this might cause regressions in
obscure setups.
Even so, it should be okay to fix this behavior without anything like a
longer transition period:
- The `git submodule update --remote` command is not really common.
- Current Git's behavior when running this command is outright
confusing, unless the remote repository's current branch _is_ `master`
(in which case the proposed behavior matches the old behavior).
- If a user encounters a regression due to the changed behavior, the fix
is actually trivial: setting `submodule.<name>.branch` to `master`
will reinstate the old behavior.
Helped-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The anonymize_str() function takes a generator callback, but there's no
way to pass extra context to it. Let's add the usual "void *data"
parameter to the generator interface and pass it along.
This is mildly annoying for existing callers, all of which pass NULL,
but is necessary to avoid extra globals in some cases we'll add in a
subsequent patch.
While we're touching each of these callbacks, we can further observe
that none of them use the existing orig/len parameters at all. This
makes sense, since the point is for their output to have no discernable
basis in the original (my original version had some notion that we might
use a one-way function to obfuscate the names, but it was never
implemented). So let's drop those extra parameters. If a caller really
wants to do something with them, it can pass a struct through the new
data parameter.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
All of the other anonymization functions keep their static mappings
inside the function to avoid polluting the global namespace. Let's do
the same for "idents", as nobody needs it outside of
anonymize_ident_line().
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that we're using a separate keydata struct for hash lookups, we have
more flexibility in how we allocate anonymized_entry structs. Let's push
the "orig" key into a flex member within the struct. That should save us
a few bytes of memory per entry (a pointer plus any malloc overhead),
and may make lookups a little faster (since it's one less pointer to
chase in the comparison function).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that the anonymize_str() interface is restricted to NUL-terminated
strings, there's no need for us to keep track of the length of each
entry in the hashmap. This simplifies the code and saves a bit of
memory.
Note that we do still need to compare the stored results to partial
strings passed in by the callers. We can do that by using hashmap's
keydata feature to get the ptr/len pair into the comparison function,
and then using strncmp().
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While the anonymize_mem() interface _can_ store arbitrary byte
sequences, none of the callers uses this feature (as of the previous
commit). We'd like to keep it that way, as we'll be exposing the
string-like nature of the anonymization routines to the user. So let's
tighten up the interface a bit:
- don't treat "len" as an out-parameter from anonymize_mem(); this
ensures callers treat the pointer result as a NUL-terminated string
- likewise, don't treat "len" as an out-parameter from generator
functions
- swap out "void *" for "char *" as appropriate to signal that we
don't handle arbitrary memory
- rename the function to anonymize_str()
This will also open up some optimization opportunities in a future
patch.
Note that we can't drop the "len" parameter entirely. Some callers do
pass in partial strings (e.g., "foo/bar", len=3) to avoid copying, and
we need to handle those still.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When fast-export stores anonymized oids, it does so as binary strings.
And while the anonymous mapping storage is binary-clean (at least as of
the previous commit), this will become awkward when we start exposing
more of it to the user. In particular, if we allow a method for
retaining token "foo", then users may want to specify a hex oid as such
a token.
Let's just switch to storing the hex strings. The difference in memory
usage is negligible (especially considering how infrequently we'd
generally store an oid compared to, say, path components).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Our anonymize_mem() function is careful to take a ptr/len pair to allow
storing binary tokens like object ids, as well as partial strings (e.g.,
just "foo" of "foo/bar"). But it duplicates the hash key using
xstrdup()! That means that:
- for a partial string, we'd store all bytes up to the NUL, even
though we'd never look at anything past "len". This didn't produce
wrong behavior, but was wasteful.
- for a binary oid that doesn't contain a zero byte, we'd copy garbage
bytes off the end of the array (though as long as nothing complained
about reading uninitialized bytes, further reads would be limited by
"len", and we'd produce the correct results)
- for a binary oid that does contain a zero byte, we'd copy _fewer_
bytes than intended into the hashmap struct. When we later try to
look up a value, we'd access uninitialized memory and potentially
falsely claim that a particular oid is not present.
The most common reason to store an oid is an anonymized gitlink, but our
test case doesn't have any gitlinks at all. So let's add one whose oid
contains a NUL and is present at two different paths. ASan catches the
memory error, but even without it we can detect the bug because the oid
is not anonymized the same way for both paths.
And of course the fix is to copy the correct number of bytes. We don't
technically need the appended NUL from xmemdupz(), but it doesn't hurt
as an extra protection against anybody treating it like a string (plus a
future patch will push us more in that direction).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the previous commit, an attempt was made to correct the "N=1, M=0"
case. However, the fix was botched and it introduced two half-correct
sections by mistake. Combine these half-correct sections into one fully
correct section.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
d91d6fbf26 (commit-reach: create repo_is_descendant_of(), 2020-06-17)
adds a repository aware version of is_descendant_of() and a backward
compatibility shim that is barely used.
Update all callers to directly use the new repo_is_descendant_of()
function instead; making the codebase simpler and pushing more
the_repository references higher up the stack.
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Reviewed-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The same worktree directory must be registered only once, but
"git worktree move" allowed this invariant to be violated, which
has been corrected.
* es/worktree-duplicate-paths:
worktree: make "move" refuse to move atop missing registered worktree
worktree: generalize candidate worktree path validation
worktree: prune linked worktree referencing main worktree path
worktree: prune duplicate entries referencing same worktree path
worktree: make high-level pruning re-usable
worktree: give "should be pruned?" function more meaningful name
worktree: factor out repeated string literal
The `diff-files' command and related commands which call the function
`cmd_diff_files()', consider the "intent-to-add" files as a part of the
index when comparing the work-tree against it. This was previously
addressed in commits [1] and [2] by turning the option
`--ita-invisible-in-index' (introduced in [3]) on by default.
For `diff-files' (and `add -p' as a consequence) to show the i-t-a
files as as new, `ita_invisible_in_index' will be enabled by default
here as well.
[1] 0231ae71d3 (diff: turn --ita-invisible-in-index on by default,
2018-05-26)
[2] 425a28e0a4 (diff-lib: allow ita entries treated as "not yet exist
in index", 2016-10-24)
[3] b42b451919 (diff: add --ita-[in]visible-in-index, 2016-10-24)
Signed-off-by: Srinidhi Kaushik <shrinidhi.kaushik@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
get_worktrees() accepts a 'flags' argument, however, there are no
existing flags (the lone flag GWT_SORT_LINKED was recently retired) and
no behavior which can be tweaked. Therefore, drop the 'flags' argument.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Of all the clients of get_worktrees(), only "git worktree list" wants
the list sorted in a very specific way; other clients simply don't care
about the order. Rather than imbuing get_worktrees() with special
knowledge about how various clients -- now and in the future -- may want
the list sorted, drop the sorting capability altogether and make it the
client's responsibility to sort the list if needed.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
git index-pack is usually run in a repository, but need not be. Since
packs don't contains information on the algorithm in use, instead
relying on context, add an option to index-pack to tell it which one
we're using in case someone runs it outside of a repository. Since
using --stdin necessarily implies a repository, don't allow specifying
an object format if it's provided to prevent users from passing an
option that won't work. Add documentation for this option.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
cmd_pull() builds a commit_list to pass a single potential ancestor to
is_descendant_of(). The latter leaves the list intact. Release the
allocated memory after the call.
Leaking in cmd_*() isn't a big deal, but sets a bad example for other
users of is_descendant_of().
Signed-off-by: René Scharfe <l.s.r@web.de>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A comment in cmd_diff() states that if one tree-ish and no blobs are
provided, (the "N=1, M=0" case), it will provide a diff between the tree
and the cache. This is incorrect because a diff happens between the
tree-ish and the working tree. Remove the `--cached` in the comment so
that the correct behavior is shown. Add a new section describing the
"N=1, M=0, --cached" behavior.
Next, describe the "N=0, M=0, --cached" case, similar to the above since
it is undocumented.
Finally, fix some spacing issues. Add spaces between each section for
consistency and readability. Also, change tabs within the comment into
spaces.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The behaviour of "sparse-checkout" in the state "git clone
--no-checkout" left was changed accidentally in 2.27, which has
been corrected.
* en/sparse-checkout:
sparse-checkout: avoid staging deletions of all files
The reflog entries for "git clone" and "git fetch" did not
anonymize the URL they operated on.
* js/reflog-anonymize-for-clone-and-fetch:
clone/fetch: anonymize URLs in the reflog
14ba97f8 (alloc: allow arbitrary repositories for alloc functions,
2018-05-15) introduced parsed_object_pool->commit_count to keep count of
commits per repository and was used to assign commit->index.
However, commit-slab code requires commit->index values to be unique
and a global count would be correct, rather than a per-repo count.
Let's introduce a static counter variable, `parsed_commits_count` to
keep track of parsed commits so far.
As commit_count has no use anymore, let's also drop it from the struct.
Signed-off-by: Abhishek Kumar <abhishekkumar8222@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
`git branch` accepts `--edit-description` in conjunction with other
arguments. However, `--edit-description` is its own mode, similar to
`--set-upstream-to`, which is also made mutually exclusive with other
modes. Prevent `--edit-description` from being mixed with other modes.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit 6b1db43109 ("clean: teach clean -d to preserve ignored paths",
2017-05-23) added the following code block (among others) to git-clean:
if (remove_directories)
dir.flags |= DIR_SHOW_IGNORED_TOO | DIR_KEEP_UNTRACKED_CONTENTS;
The reason for these flags is well documented in the commit message, but
isn't obvious just from looking at the code. Add some explanations to
the code to make it clearer.
Further, it appears git-2.26 did not correctly handle this combination
of flags from git-clean. With both these flags and without
DIR_SHOW_IGNORED_TOO_MODE_MATCHING set, git is supposed to recurse into
all untracked AND ignored directories. git-2.26.0 clearly was not doing
that. I don't know the full reasons for that or whether git < 2.27.0
had additional unknown bugs because of that misbehavior, because I don't
feel it's worth digging into. As per the huge changes and craziness
documented in commit 8d92fb2927 ("dir: replace exponential algorithm
with a linear one", 2020-04-01), the old algorithm was a mess and was
thrown out. What I can say is that git-2.27.0 correctly recurses into
untracked AND ignored directories with that combination.
However, in clean's case we don't need to recurse into ignored
directories; that is just a waste of time. Thus, when git-2.27.0
started correctly handling those flags, we got a performance regression
report. Rather than relying on other bugs in fill_directory()'s former
logic to provide the behavior of skipping ignored directories, make use
of the DIR_SHOW_IGNORED_TOO_MODE_MATCHING value specifically added in
commit eec0f7f2b7 ("status: add option to show ignored files
differently", 2017-10-30) for this purpose.
Reported-by: Brian Malehorn <bmalehorn@gmail.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
I spent a long time trying to figure out how and whether the code worked
with different values of ignore, ignore_only, and remove_directories.
After lots of time setting up lots of testcases, sifting through lots of
print statements, and walking through the debugger, I finally realized
that one piece of code related to how it was all setup was found in
clean.c rather than dir.c. Make a change that would have made it easier
for me to do the extra testing by putting this handling in one spot.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
dir.h documented quite clearly that DIR_SHOW_IGNORED and
DIR_SHOW_IGNORED_TOO are mutually exclusive, with a big comment to this
effect by the definition of both enum values. However, a command like
git clean -fx $DIR
would set both values for dir.flags. I _think_ it happened to work
because:
* As dir.h points out, DIR_KEEP_UNTRACKED_CONTENTS only takes effect
if DIR_SHOW_IGNORED_TOO is set.
* As coded, I believe DIR_SHOW_IGNORED would just happen to take
precedence over DIR_SHOW_IGNORED_TOO in the code as currently
constructed.
Which is a long way of saying "we just got lucky".
Fix clean.c to avoid setting these mutually exclusive values at the same
time, and add a check to dir.c that will throw a BUG() to prevent anyone
else from making this mistake.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Document the usage for producing combined commits with "git diff".
This includes updating the synopsis section.
While here, add the three-dot notation to the synopsis.
Make "git diff -h" print the same usage summary as the manual
page synopsis, minus the "A..B" form, which is now discouraged.
Signed-off-by: Chris Torek <chris.torek@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When git diff is given a symmetric difference A...B, it chooses
some merge base from the two specified commits (as documented).
This fails, however, if there is *no* merge base: instead, you
see the differences between A and B, which is certainly not what
is expected.
Moreover, if additional revisions are specified on the command
line ("git diff A...B C"), the results get a bit weird:
* If there is a symmetric difference merge base, this is used
as the left side of the diff. The last final ref is used as
the right side.
* If there is no merge base, the symmetric status is completely
lost. We will produce a combined diff instead.
Similar weirdness occurs if you use, e.g., "git diff C A...B D".
Likewise, using multiple two-dot ranges, or tossing extra
revision specifiers into the command line with two-dot ranges,
or mixing two and three dot ranges, all produce nonsense.
To avoid all this, add a routine to catch the range cases and
verify that that the arguments make sense. As a side effect,
produce a warning showing *which* merge base is being used when
there are multiple choices; die if there is no merge base.
Signed-off-by: Chris Torek <chris.torek@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach upload-pack to send part of its packfile response as URIs.
An administrator may configure a repository with one or more
"uploadpack.blobpackfileuri" lines, each line containing an OID, a pack
hash, and a URI. A client may configure fetch.uriprotocols to be a
comma-separated list of protocols that it is willing to use to fetch
additional packfiles - this list will be sent to the server. Whenever an
object with one of those OIDs would appear in the packfile transmitted
by upload-pack, the server may exclude that object, and instead send the
URI. The client will then download the packs referred to by those URIs
before performing the connectivity check.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Whenever a fetch results in a packfile being downloaded, a .keep file is
generated, so that the packfile can be preserved (from, say, a running
"git repack") until refs are written referring to the contents of the
packfile.
In a subsequent patch, a successful fetch using protocol v2 may result
in more than one .keep file being generated. Therefore, teach
fetch_pack() and the transport mechanism to support multiple .keep
files.
Implementation notes:
- builtin/fetch-pack.c normally does not generate .keep files, and thus
is unaffected by this or future changes. However, it has an
undocumented "--lock-pack" feature, used by remote-curl.c when
implementing the "fetch" remote helper command. In keeping with the
remote helper protocol, only one "lock" line will ever be written;
the rest will result in warnings to stderr. However, in practice,
warnings will never be written because the remote-curl.c "fetch" is
only used for protocol v0/v1 (which will not generate multiple .keep
files). (Protocol v2 uses the "stateless-connect" command, not the
"fetch" command.)
- connected.c has an optimization in that connectivity checks on a ref
need not be done if the target object is in a pack known to be
self-contained and connected. If there are multiple packfiles, this
optimization can no longer be done.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git worktree add" takes special care to avoid creating a new worktree
at a location already registered to an existing worktree even if that
worktree is missing (which can happen, for instance, if the worktree
resides on removable media). "git worktree move", however, is not so
careful when validating the destination location and will happily move
the source worktree atop the location of a missing worktree. This leads
to the anomalous situation of multiple worktrees being associated with
the same path, which is expressly forbidden by design. For example:
$ git clone foo.git
$ cd foo
$ git worktree add ../bar
$ git worktree add ../baz
$ rm -rf ../bar
$ git worktree move ../baz ../bar
$ git worktree list
.../foo beefd00f [master]
.../bar beefd00f [bar]
.../bar beefd00f [baz]
$ git worktree remove ../bar
fatal: validation failed, cannot remove working tree:
'.../bar' does not point back to '.git/worktrees/bar'
Fix this shortcoming by enhancing "git worktree move" to perform the
same additional validation of the destination directory as done by "git
worktree add".
While at it, add a test to verify that "git worktree move" won't move a
worktree atop an existing (non-worktree) path -- a restriction which has
always been in place but was never tested.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git worktree add" checks that the specified path is a valid location
for a new worktree by ensuring that the path does not already exist and
is not already registered to another worktree (a path can be registered
but missing, for instance, if it resides on removable media). Since "git
worktree add" is not the only command which should perform such
validation ("git worktree move" ought to also), generalize the the
validation function for use by other callers, as well.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git worktree prune" detects when multiple entries are associated with
the same path and prunes the duplicates, however, it does not detect
when a linked worktree points at the path of the main worktree.
Although "git worktree add" disallows creating a new worktree with the
same path as the main worktree, such a case can arise outside the
control of Git even without the user mucking with .git/worktree/<id>/
administrative files. For instance:
$ git clone foo.git
$ git -C foo worktree add ../bar
$ rm -rf bar
$ mv foo bar
$ git -C bar worktree list
.../bar deadfeeb [master]
.../bar deadfeeb [bar]
Help the user recover from such corruption by extending "git worktree
prune" to also detect when a linked worktree is associated with the path
of the main worktree.
Reported-by: Jonathan Müller <jonathanmueller.dev@gmail.com>
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A fundamental restriction of linked working trees is that there must
only ever be a single worktree associated with a particular path, thus
"git worktree add" explicitly disallows creation of a new worktree at
the same location as an existing registered worktree. Nevertheless,
users can still "shoot themselves in the foot" by mucking with
administrative files in .git/worktree/<id>/. Worse, "git worktree move"
is careless[1] and allows a worktree to be moved atop a registered but
missing worktree (which can happen, for instance, if the worktree is on
removable media). For instance:
$ git clone foo.git
$ cd foo
$ git worktree add ../bar
$ git worktree add ../baz
$ rm -rf ../bar
$ git worktree move ../baz ../bar
$ git worktree list
.../foo beefd00f [master]
.../bar beefd00f [bar]
.../bar beefd00f [baz]
Help users recover from this form of corruption by teaching "git
worktree prune" to detect when multiple worktrees are associated with
the same path.
[1]: A subsequent commit will fix "git worktree move" validation to be
more strict.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The low-level logic for removing a worktree is well encapsulated in
delete_git_dir(). However, high-level details related to pruning a
worktree -- such as dealing with verbosity and dry-run mode -- are not
encapsulated. Factor out this high-level logic into its own function so
it can be re-used as new worktree corruption detectors are added.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Readers of the name prune_worktree() are likely to expect the function
to actually prune a worktree, however, it only answers the question
"should this worktree be pruned?". Give it a name more reflective of its
true purpose to avoid such confusion.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The code to parse "git bisect start" command line was lax in
validating the arguments.
* cb/bisect-helper-parser-fix:
bisect--helper: avoid segfault with bad syntax in `start --term-*`
On-the-wire protocol v2 easily falls into a deadlock between the
remote-curl helper and the fetch-pack process when the server side
prematurely throws an error and disconnects. The communication has
been updated to make it more robust.
* dl/remote-curl-deadlock-fix:
stateless-connect: send response end packet
pkt-line: define PACKET_READ_RESPONSE_END
remote-curl: error on incomplete packet
pkt-line: extern packet_length()
transport: extract common fetch_pack() call
remote-curl: remove label indentation
remote-curl: fix typo
Code simplification and test coverage enhancement.
* bc/filter-process:
t2060: add a test for switch with --orphan and --discard-changes
builtin/checkout: simplify metadata initialization
For each worktree removed by "git worktree prune", it reports the reason
for the removal. All reasons share the common prefix "Removing
worktrees/%s:". As new removal reasons are added, this prefix needs to
be duplicated, which is error-prone and potentially cumbersome.
Therefore, factor out the common prefix.
Although this change seems to increase the "sentence lego quotient", it
should be reasonably safe, as the reason for removal is a distinct
clause, not strictly related to the prefix. Moreover, the "worktrees" in
"Removing worktrees/%s:" is a path literal which ought not be localized,
so by factoring it out, we can more easily avoid exposing that path
fragment to translators.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'extensions' configuration variable gets special meaning in the new
repository version, so when enabling the extension we should upgrade the
repository to version 1.
Signed-off-by: Xin Li <delphij@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Retroactively adding a filter can be useful for existing shallow clones as
they allow users to see earlier change histories without downloading all
git objects in a regular --unshallow fetch.
Without this patch, users can make a clone partial by editing the
repository configuration to convert the remote into a promisor, like:
git config core.repositoryFormatVersion 1
git config extensions.partialClone origin
git fetch --unshallow --filter=blob:none origin
Since the hard part of making this work is already in place and such
edits can be error-prone, teach Git to perform the required configuration
change automatically instead.
Note that this change does not modify the existing git behavior which
recognizes setting extensions.partialClone without changing
repositoryFormatVersion.
Signed-off-by: Xin Li <delphij@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
sparse-checkout's purpose is to update the working tree to have it
reflect a subset of the tracked files. As such, it shouldn't be
switching branches, making commits, downloading or uploading data, or
staging or unstaging changes. Other than updating the worktree, the
only thing sparse-checkout should touch is the SKIP_WORKTREE bit of the
index. In particular, this sets up a nice invariant: running
sparse-checkout will never change the status of any file in `git status`
(reflecting the fact that we only set the SKIP_WORKTREE bit if the file
is safe to delete, i.e. if the file is unmodified).
Traditionally, we did a _really_ bad job with this goal. The
predecessor to sparse-checkout involved manual editing of
.git/info/sparse-checkout and running `git read-tree -mu HEAD`. That
command would stage and unstage changes and overwrite dirty changes in
the working tree.
The initial implementation of the sparse-checkout command was no better;
it simply invoked `git read-tree -mu HEAD` as a subprocess and had the
same caveats, though this issue came up repeatedly in review comments
and workarounds for the problems were put in place before the feature
was merged[1, 2, 3, 4, 5, 6; especially see 4 & 6].
[1] https://lore.kernel.org/git/CABPp-BFT9A5n=_bx5LsjCvbogqwSjiwgr5amcjgbU1iAk4KLJg@mail.gmail.com/
[2] https://lore.kernel.org/git/CABPp-BEmwSwg4tgJg6nVG8a3Hpn_g-=ZjApZF4EiJO+qVgu4uw@mail.gmail.com/
[3] https://lore.kernel.org/git/CABPp-BFV7TA0qwZCQpHCqx9N+JifyRyuBQ-pZ_oGfe-NOgyh7A@mail.gmail.com/
[4] https://lore.kernel.org/git/CABPp-BHYCCD+Vx5fq35jH82eHc1-P53Lz_aGNpHJNcx9kg2K-A@mail.gmail.com/
[5] https://lore.kernel.org/git/CABPp-BF+JWYZfDqp2Tn4AEKVp4b0YMA=Mbz4Nz62D-gGgiduYQ@mail.gmail.com/
[6] https://lore.kernel.org/git/20191121163706.GV23183@szeder.dev/
However, these workarounds, in addition to disabling the feature in a
number of important cases, also missed one special case. I'll get back
to it later.
In the 2.27.0 cycle, the disabling of the feature was lifted by finally
replacing the internal equivalent of `git read-tree -mu HEAD` with
something that did what we wanted: the new update_sparsity() function in
unpack-trees.c that only ever updates SKIP_WORKTREE bits in the index
and updates the working tree to match. This new function handles all
the cases that were problematic for the old implementation, except that
it breaks the same special case that avoided the workarounds of the old
implementation, but broke it in a different way.
So...that brings us to the special case: a git clone performed with
--no-checkout. As per the meaning of the flag, --no-checkout does not
check out any branch, with the implication that you aren't on one and
need to switch to one after the clone. Implementationally, HEAD is
still set (so in some sense you are partially on a branch), but
* the index is "unborn" (non-existent)
* there are no files in the working tree (other than .git/)
* the next time git switch (or git checkout) is run it will run
unpack_trees with `initial_checkout` flag set to true.
It is not until you run, e.g. `git switch <somebranch>` that the index
will be written and files in the working tree populated.
With this special --no-checkout case, the traditional `read-tree -mu
HEAD` behavior would have done the equivalent of acting like checkout --
switch to the default branch (HEAD), write out an index that matches
HEAD, and update the working tree to match. This special case slipped
through the avoid-making-changes checks in the original sparse-checkout
command and thus continued there.
After update_sparsity() was introduced and used (see commit f56f31af03
("sparse-checkout: use new update_sparsity() function", 2020-03-27)),
the behavior for the --no-checkout case changed: Due to git's
auto-vivification of an empty in-memory index (see do_read_index() and
note that `must_exist` is false), and due to sparse-checkout's
update_working_directory() code to always write out the index after it
was done, we got a new bug. That made it so that sparse-checkout would
switch the repository from a clone with an "unborn" index (i.e. still
needing an initial_checkout), to one that had a recorded index with no
entries. Thus, instead of all the files appearing deleted in `git
status` being known to git as a special artifact of not yet being on a
branch, our recording of an empty index made it suddenly look to git as
though it was definitely on a branch with ALL files staged for deletion!
A subsequent checkout or switch then had to contend with the fact that
it wasn't on an initial_checkout but had a bunch of staged deletions.
Make sure that sparse-checkout changes nothing in the index other than
the SKIP_WORKTREE bit; in particular, when the index is unborn we do not
have any branch checked out so there is no sparsification or
de-sparsification work to do. Simply return from
update_working_directory() early.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Even if we strongly discourage putting credentials into the URLs passed
via the command-line, there _is_ support for that, and users _do_ do
that.
Let's scrub them before writing them to the reflog.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
