| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Some lockfile code called free() in signal-death code path, which
has been corrected.
* ps/lockfile-cleanup-fix:
fetch: fix deadlock when cleaning up lockfiles in async signals
|
|
Similar message templates have been consolidated so that
translators need to work on fewer number of messages.
* ja/i18n-similar-messages:
i18n: turn even more messages into "cannot be used together" ones
i18n: ref-filter: factorize "%(foo) atom used without %(bar) atom"
i18n: factorize "--foo outside a repository"
i18n: refactor "unrecognized %(foo) argument" strings
i18n: factorize "no directory given for --foo"
i18n: factorize "--foo requires --bar" and the like
i18n: tag.c factorize i18n strings
i18n: standardize "cannot open" and "cannot read"
i18n: turn "options are incompatible" into "cannot be used together"
i18n: refactor "%s, %s and %s are mutually exclusive"
i18n: refactor "foo and bar are mutually exclusive"
|
|
"git fetch" and "git pull" are now declared sparse-index clean.
Also "git ls-files" learns the "--sparse" option to help debugging.
* ds/fetch-pull-with-sparse-index:
test-read-cache: remove --table, --expand options
t1091/t3705: remove 'test-tool read-cache --table'
t1092: replace 'read-cache --table' with 'ls-files --sparse'
ls-files: add --sparse option
fetch/pull: use the sparse index
|
|
When fetching packfiles, we write a bunch of lockfiles for the packfiles
we're writing into the repository. In order to not leave behind any
cruft in case we exit or receive a signal, we register both an exit
handler as well as signal handlers for common signals like SIGINT. These
handlers will then unlink the locks and free the data structure tracking
them. We have observed a deadlock in this logic though:
(gdb) bt
#0 __lll_lock_wait_private () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:95
#1 0x00007f4932bea2cd in _int_free (av=0x7f4932f2eb20 <main_arena>, p=0x3e3e4200, have_lock=0) at malloc.c:3969
#2 0x00007f4932bee58c in __GI___libc_free (mem=<optimized out>) at malloc.c:2975
#3 0x0000000000662ab1 in string_list_clear ()
#4 0x000000000044f5bc in unlock_pack_on_signal ()
#5 <signal handler called>
#6 _int_free (av=0x7f4932f2eb20 <main_arena>, p=<optimized out>, have_lock=0) at malloc.c:4024
#7 0x00007f4932bee58c in __GI___libc_free (mem=<optimized out>) at malloc.c:2975
#8 0x000000000065afd5 in strbuf_release ()
#9 0x000000000066ddb9 in delete_tempfile ()
#10 0x0000000000610d0b in files_transaction_cleanup.isra ()
#11 0x0000000000611718 in files_transaction_abort ()
#12 0x000000000060d2ef in ref_transaction_abort ()
#13 0x000000000060d441 in ref_transaction_prepare ()
#14 0x000000000060e0b5 in ref_transaction_commit ()
#15 0x00000000004511c2 in fetch_and_consume_refs ()
#16 0x000000000045279a in cmd_fetch ()
#17 0x0000000000407c48 in handle_builtin ()
#18 0x0000000000408df2 in cmd_main ()
#19 0x00000000004078b5 in main ()
The process was killed with a signal, which caused the signal handler to
kick in and try free the data structures after we have unlinked the
locks. It then deadlocks while calling free(3P).
The root cause of this is that it is not allowed to call certain
functions in async-signal handlers, as specified by signal-safety(7).
Next to most I/O functions, this list of disallowed functions also
includes memory-handling functions like malloc(3P) and free(3P) because
they may not be reentrant. As a result, if we execute such functions in
the signal handler, then they may operate on inconistent state and fail
in unexpected ways.
Fix this bug by not calling non-async-signal-safe functions when running
in the signal handler. We're about to re-raise the signal anyway and
will thus exit, so it's not much of a problem to keep the string list of
lockfiles untouched. Note that it's fine though to call unlink(2), so
we'll still clean up the lockfiles correctly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Signed-off-by: Jean-Noël Avila <jn.avila@free.fr>
Reviewed-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Use static strings for constant parts of the sentences. They are all
turned into "cannot be used together".
Signed-off-by: Jean-Noël Avila <jn.avila@free.fr>
Reviewed-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
"git fetch --set-upstream" did not check if there is a current
branch, leading to a segfault when it is run on a detached HEAD,
which has been corrected.
* ab/fetch-set-upstream-while-detached:
pull, fetch: fix segfault in --set-upstream option
|
|
The 'git fetch' and 'git pull' commands parse the index in order to
determine if submodules exist. Without command_requires_full_index=0,
this will expand a sparse index, causing slow performance even when
there is no new data to fetch.
The .gitmodules file will never be inside a sparse directory entry, and
even if it was, the index_name_pos() method would expand the sparse
index if needed as we search for the path by name. These commands do not
iterate over the index, which is the typical thing we are careful about
when integrating with the sparse index.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Fix a segfault in the --set-upstream option added in
24bc1a12926 (pull, fetch: add --set-upstream option, 2019-08-19) added
in v2.24.0.
The code added there did not do the same checking we do for "git
branch" itself since 8efb8899cfe (branch: segfault fixes and
validation, 2013-02-23), which in turn fixed the same sort of segfault
I'm fixing now in "git branch --set-upstream-to", see
6183d826ba6 (branch: introduce --set-upstream-to, 2012-08-20).
The warning message I'm adding here is an amalgamation of the error
added for "git branch" in 8efb8899cfe, and the error output
install_branch_config() itself emits, i.e. it trims "refs/heads/" from
the name and says "branch X on remote", not "branch refs/heads/X on
remote".
I think it would make more sense to simply die() here, but in the
other checks for --set-upstream added in 24bc1a12926 we issue a
warning() instead. Let's do the same here for consistency for now.
There was an earlier submitted alternate way of fixing this in [1],
due to that patch breaking threading with the original report at [2] I
didn't notice it before authoring this version. I think the more
detailed warning message here is better, and we should also have tests
for this behavior.
The --no-rebase option to "git pull" is needed as of the recently
merged 7d0daf3f12f (Merge branch 'en/pull-conflicting-options',
2021-08-30).
1. https://lore.kernel.org/git/20210706162238.575988-1-clemens@endorphin.org/
2. https://lore.kernel.org/git/CAG6gW_uHhfNiHGQDgGmb1byMqBA7xa8kuH1mP-wAPEe5Tmi2Ew@mail.gmail.com/
Reported-by: Clemens Fruhwirth <clemens@endorphin.org>
Reported-by: Jan Pokorný <poki@fnusa.cz>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Refuse to fetch into the currently checked out branch of any working
tree, not just the current one.
Fixes this previously reported bug:
https://lore.kernel.org/git/cb957174-5e9a-5603-ea9e-ac9b58a2eaad@mathema.de/
As a side effect of using find_shared_symref, we’ll also refuse the
fetch when we’re on a detached HEAD because we’re rebasing or bisecting
on the branch in question. This seems like a sensible change.
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Documentation/CodingGuidelines says “do not end error messages with a
full stop” and “do not capitalize the first word”. Clean up existing
messages, some of which we will be touching in later steps in the
series, that deviate from these rules in this file, as a preparation for
the main part of the topic.
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The run-command API has been updated so that the callers can easily
ask the file descriptors open for packfiles to be closed immediately
before spawning commands that may trigger auto-gc.
* js/run-command-close-packs:
Close object store closer to spawning child processes
run_auto_maintenance(): implicitly close the object store
run-command: offer to close the object store before running
run-command: prettify the `RUN_COMMAND_*` flags
pull: release packs before fetching
commit-graph: when closing the graph, also release the slab
|
|
Optimize code that handles large number of refs in the "git fetch"
code path.
* ps/fetch-optim:
fetch: avoid second connectivity check if we already have all objects
fetch: merge fetching and consuming refs
fetch: refactor fetch refs to be more extendable
fetch-pack: optimize loading of refs via commit graph
connected: refactor iterator to return next object ID directly
fetch: avoid unpacking headers in object existence check
fetch: speed up lookup of want refs via commit-graph
|
|
Code clean up to migrate callers from older advice_config[] based
API to newer advice_if_enabled() and advice_enabled() API.
* ab/retire-advice-config:
advice: move advice.graftFileDeprecated squashing to commit.[ch]
advice: remove use of global advice_add_embedded_repo
advice: remove read uses of most global `advice_` variables
advice: add enum variants for missing advice variables
|
|
"git fetch --quiet" optimization to avoid useless computation of
info that will never be displayed.
* ps/fetch-omit-formatting-under-quiet:
fetch: skip formatting updated refs with `--quiet`
|
|
Before spawning the auto maintenance, we need to make sure that we
release all open file handles to all the `.pack` files (and MIDX files
and commit-graph files and...) so that the maintenance process has the
freedom to delete those files.
So far, we did this manually every time before calling
`run_auto_maintenance()`. With the new `close_object_store` flag, we can
do that implicitly in that function, which is more robust because future
callers won't be able to forget to close the object store.
Note: this changes behavior slightly, as we previously _always_ closed
the object store, but now we only close the object store when actually
running the auto maintenance. In practice, this should not matter (if
anything, it might speed up operations where auto maintenance is
disabled).
Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When fetching refs, we are doing two connectivity checks:
- The first one is done such that we can skip fetching refs in the
case where we already have all objects referenced by the updated
set of refs.
- The second one verifies that we have all objects after we have
fetched objects.
We always execute both connectivity checks, but this is wasteful in case
the first connectivity check already notices that we have all objects
locally available.
Skip the second connectivity check in case we already had all objects
available. This gives us a nice speedup when doing a mirror-fetch in a
repository with about 2.3M refs where the fetching repo already has all
objects:
Benchmark #1: HEAD~: git-fetch
Time (mean ± σ): 30.025 s ± 0.081 s [User: 27.070 s, System: 4.933 s]
Range (min … max): 29.900 s … 30.111 s 5 runs
Benchmark #2: HEAD: git-fetch
Time (mean ± σ): 25.574 s ± 0.177 s [User: 22.855 s, System: 4.683 s]
Range (min … max): 25.399 s … 25.765 s 5 runs
Summary
'HEAD: git-fetch' ran
1.17 ± 0.01 times faster than 'HEAD~: git-fetch'
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The functions `fetch_refs()` and `consume_refs()` must always be called
together such that we first obtain all missing objects and then update
our local refs to match the remote refs. In a subsequent patch, we'll
further require that `fetch_refs()` must always be called before
`consume_refs()` such that it can correctly assert that we have all
objects after the fetch given that we're about to move the connectivity
check.
Make this requirement explicit by merging both functions into a single
`fetch_and_consume_refs()` function.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Refactor `fetch_refs()` code to make it more extendable by explicitly
handling error cases. The refactored code should behave the same.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The object ID iterator used by the connectivity checks returns the next
object ID via an out-parameter and then uses a return code to indicate
whether an item was found. This is a bit roundabout: instead of a
separate error code, we can just return the next object ID directly and
use `NULL` pointers as indicator that the iterator got no items left.
Furthermore, this avoids a copy of the object ID.
Refactor the iterator and all its implementations to return object IDs
directly. This brings a tiny performance improvement when doing a mirror-fetch of a repository with about 2.3M refs:
Benchmark #1: 328dc58b49919c43897240f2eabfa30be2ce32a4~: git-fetch
Time (mean ± σ): 30.110 s ± 0.148 s [User: 27.161 s, System: 5.075 s]
Range (min … max): 29.934 s … 30.406 s 10 runs
Benchmark #2: 328dc58b49919c43897240f2eabfa30be2ce32a4: git-fetch
Time (mean ± σ): 29.899 s ± 0.109 s [User: 26.916 s, System: 5.104 s]
Range (min … max): 29.696 s … 29.996 s 10 runs
Summary
'328dc58b49919c43897240f2eabfa30be2ce32a4: git-fetch' ran
1.01 ± 0.01 times faster than '328dc58b49919c43897240f2eabfa30be2ce32a4~: git-fetch'
While this 1% speedup could be labelled as statistically insignificant,
the speedup is consistent on my machine. Furthermore, this is an end to
end test, so it is expected that the improvement in the connectivity
check itself is more significant.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When updating local refs after the fetch has transferred all objects, we
do an object existence test as a safety guard to avoid updating a ref to
an object which we don't have. We do so via `oid_object_info()`: if it
returns an error, then we know the object does not exist.
One side effect of `oid_object_info()` is that it parses the object's
type, and to do so it must unpack the object header. This is completely
pointless: we don't care for the type, but only want to assert that the
object exists.
Refactor the code to use `repo_has_object_file()`, which both makes the
code's intent clearer and is also faster because it does not unpack
object headers. In a real-world repo with 2.3M refs, this results in a
small speedup when doing a mirror-fetch:
Benchmark #1: HEAD~: git-fetch
Time (mean ± σ): 33.686 s ± 0.176 s [User: 30.119 s, System: 5.262 s]
Range (min … max): 33.512 s … 33.944 s 5 runs
Benchmark #2: HEAD: git-fetch
Time (mean ± σ): 31.247 s ± 0.195 s [User: 28.135 s, System: 5.066 s]
Range (min … max): 30.948 s … 31.472 s 5 runs
Summary
'HEAD: git-fetch' ran
1.08 ± 0.01 times faster than 'HEAD~: git-fetch'
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When updating our local refs based on the refs fetched from the remote,
we need to iterate through all requested refs and load their respective
commits such that we can determine whether they need to be appended to
FETCH_HEAD or not. In cases where we're fetching from a remote with
exceedingly many refs, resolving these refs can be quite expensive given
that we repeatedly need to unpack object headers for each of the
referenced objects.
Speed this up by opportunistically trying to resolve object IDs via the
commit graph. We only do so for any refs which are not in "refs/tags":
more likely than not, these are going to be a commit anyway, and this
lets us avoid having to unpack object headers completely in case the
object is a commit that is part of the commit-graph. This significantly
speeds up mirror-fetches in a real-world repository with
2.3M refs:
Benchmark #1: HEAD~: git-fetch
Time (mean ± σ): 56.482 s ± 0.384 s [User: 53.340 s, System: 5.365 s]
Range (min … max): 56.050 s … 57.045 s 5 runs
Benchmark #2: HEAD: git-fetch
Time (mean ± σ): 33.727 s ± 0.170 s [User: 30.252 s, System: 5.194 s]
Range (min … max): 33.452 s … 33.871 s 5 runs
Summary
'HEAD: git-fetch' ran
1.67 ± 0.01 times faster than 'HEAD~: git-fetch'
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When fetching, Git will by default print a list of all updated refs in a
nicely formatted table. In order to come up with this table, Git needs
to iterate refs twice: first to determine the maximum column width, and
a second time to actually format these changed refs.
While this table will not be printed in case the user passes `--quiet`,
we still go out of our way and do all these steps. In fact, we even do
more work compared to not passing `--quiet`: without the flag, we will
skip all references in the column width computation which have not been
updated, but if it is set we will now compute widths for all refs.
Fix this issue by completely skipping both preparation of the format and
formatting data for display in case the user passes `--quiet`, improving
performance especially with many refs. The following benchmark shows a
nice speedup for a quiet mirror-fetch in a repository with 2.3M refs:
Benchmark #1: HEAD~: git-fetch
Time (mean ± σ): 26.929 s ± 0.145 s [User: 24.194 s, System: 4.656 s]
Range (min … max): 26.692 s … 27.068 s 5 runs
Benchmark #2: HEAD: git-fetch
Time (mean ± σ): 25.189 s ± 0.094 s [User: 22.556 s, System: 4.606 s]
Range (min … max): 25.070 s … 25.314 s 5 runs
Summary
'HEAD: git-fetch' ran
1.07 ± 0.01 times faster than 'HEAD~: git-fetch'
While at it, this patch also fixes `adjust_refcol_width()` such that it
skips unchanged refs in case the user passed `--quiet`, where verbosity
will be negative. While this function won't be called anymore if so,
this brings the comment in line with actual code. Furthermore, needless
`verbosity >= 0` checks are now removed in `store_updated_refs()`: we
never print to the `note` buffer anymore in case `verbosity < 0`, so we
won't end up in that code block anyway.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
In c4a09cc9ccb (Merge branch 'hw/advise-ng', 2020-03-25), a new API for
accessing advice variables was introduced and deprecated `advice_config`
in favor of a new array, `advice_setting`.
This patch ports all but two uses which read the status of the global
`advice_` variables over to the new `advice_enabled` API. We'll deal
with advice_add_embedded_repo and advice_graft_file_deprecated
separately.
Signed-off-by: Ben Boeckel <mathstuf@gmail.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Bugfix for common ancestor negotiation recently introduced in "git
push" code path.
* jt/push-negotiation-fixes:
fetch: die on invalid --negotiation-tip hash
send-pack: fix push nego. when remote has refs
send-pack: fix push.negotiate with remote helper
|
|
Code recently added to support common ancestry negotiation during
"git push" did not sanity check its arguments carefully enough.
* ab/fetch-negotiate-segv-fix:
fetch: fix segfault in --negotiate-only without --negotiation-tip=*
fetch: document the --negotiate-only option
send-pack.c: move "no refs in common" abort earlier
|
|
If a full hexadecimal hash is given as a --negotiation-tip to "git
fetch", and that hash does not correspond to an object, "git fetch" will
segfault if --negotiate-only is given and will silently ignore that hash
otherwise. Make these cases fatal errors, just like the case when an
invalid ref name or abbreviated hash is given.
While at it, mark the error messages as translatable.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The recent --negotiate-only option would segfault in the call to
oid_array_for_each() in negotiate_using_fetch() unless one or more
--negotiation-tip=* options were provided.
All of the other tests for the feature combine both, but nothing was
checking this assumption, let's do that and add a test for it. Fixes a
bug in 9c1e657a8fd (fetch: teach independent negotiation (no
packfile), 2021-05-04).
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Signed-off-by: Alex Henrie <alexhenrie24@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
"git push" learns to discover common ancestor with the receiving
end over protocol v2.
* jt/push-negotiation:
send-pack: support push negotiation
fetch: teach independent negotiation (no packfile)
fetch-pack: refactor command and capability write
fetch-pack: refactor add_haves()
fetch-pack: refactor process_acks()
|
|
Currently, the packfile negotiation step within a Git fetch cannot be
done independent of sending the packfile, even though there is at least
one application wherein this is useful. Therefore, make it possible for
this negotiation step to be done independently. A subsequent commit will
use this for one such application - push negotiation.
This feature is for protocol v2 only. (An implementation for protocol v0
would require a separate implementation in the fetch, transport, and
transport helper code.)
In the protocol, the main hindrance towards independent negotiation is
that the server can unilaterally decide to send the packfile. This is
solved by a "wait-for-done" argument: the server will then wait for the
client to say "done". In practice, the client will never say it; instead
it will cease requests once it is satisfied.
In the client, the main change lies in the transport and transport
helper code. fetch_refs_via_pack() performs everything needed - protocol
version and capability checks, and the negotiation itself.
There are 2 code paths that do not go through fetch_refs_via_pack() that
needed to be individually excluded: the bundle transport (excluded
through requiring smart_options, which the bundle transport doesn't
support) and transport helpers that do not support takeover. If or when
we support independent negotiation for protocol v0, we will need to
modify these 2 code paths to support it. But for now, report failure if
independent negotiation is requested in these cases.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The --prefetch option will be used by the 'prefetch' maintenance task
instead of sending refspecs explicitly across the command-line. The
intention is to modify the refspec to place all results in
refs/prefetch/ instead of anywhere else.
Create helper method filter_prefetch_refspec() to modify a given refspec
to fit the rules expected of the prefetch task:
* Negative refspecs are preserved.
* Refspecs without a destination are removed.
* Refspecs whose source starts with "refs/tags/" are removed.
* Other refspecs are placed within "refs/prefetch/".
Finally, we add the 'force' option to ensure that prefetch refs are
replaced as necessary.
There are some interesting cases that are worth testing.
An earlier version of this change dropped the "i--" from the loop that
deletes a refspec item and shifts the remaining entries down. This
allowed some refspecs to not be modified. The subtle part about the
first --prefetch test is that the "refs/tags/*" refspec appears directly
before the "refs/heads/bogus/*" refspec. Without that "i--", this
ordering would remove the "refs/tags/*" refspec and leave the last one
unmodified, placing the result in "refs/heads/*".
It is possible to have an empty refspec. This is typically the case for
remotes other than the origin, where users want to fetch a specific tag
or branch. To correctly test this case, we need to further remove the
upstream remote for the local branch. Thus, we are testing a refspec
that will be deleted, leaving nothing to fetch.
Helped-by: Tom Saeger <tom.saeger@oracle.com>
Helped-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
"git clone" tries to locally check out the branch pointed at by
HEAD of the remote repository after it is done, but the protocol
did not convey the information necessary to do so when copying an
empty repository. The protocol v2 learned how to do so.
* jt/clone-unborn-head:
clone: respect remote unborn HEAD
connect, transport: encapsulate arg in struct
ls-refs: report unborn targets of symrefs
|
|
In a future patch we plan to return the name of an unborn current branch
from deep in the callchain to a caller via a new pointer parameter that
points at a variable in the caller when the caller calls
get_remote_refs() and transport_get_remote_refs().
In preparation for that, encapsulate the existing ref_prefixes
parameter into a struct. The aforementioned unborn current branch will
go into this new struct in the future patch.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When executing a fetch, then git will currently allocate one reference
transaction per reference update and directly commit it. This means that
fetches are non-atomic: even if some of the reference updates fail,
others may still succeed and modify local references.
This is fine in many scenarios, but this strategy has its downsides.
- The view of remote references may be inconsistent and may show a
bastardized state of the remote repository.
- Batching together updates may improve performance in certain
scenarios. While the impact probably isn't as pronounced with loose
references, the upcoming reftable backend may benefit as it needs to
write less files in case the update is batched.
- The reference-update hook is currently being executed twice per
updated reference. While this doesn't matter when there is no such
hook, we have seen severe performance regressions when doing a
git-fetch(1) with reference-transaction hook when the remote
repository has hundreds of thousands of references.
Similar to `git push --atomic`, this commit thus introduces atomic
fetches. Instead of allocating one reference transaction per updated
reference, it causes us to only allocate a single transaction and commit
it as soon as all updates were received. If locking of any reference
fails, then we abort the complete transaction and don't update any
reference, which gives us an all-or-nothing fetch.
Note that this may not completely fix the first of above downsides, as
the consistent view also depends on the server-side. If the server
doesn't have a consistent view of its own references during the
reference negotiation phase, then the client would get the same
inconsistent view the server has. This is a separate problem though and,
if it actually exists, can be fixed at a later point.
This commit also changes the way we write FETCH_HEAD in case `--atomic`
is passed. Instead of writing changes as we go, we need to accumulate
all changes first and only commit them at the end when we know that all
reference updates succeeded. Ideally, we'd just do so via a temporary
file so that we don't need to carry all updates in-memory. This isn't
trivially doable though considering the `--append` mode, where we do not
truncate the file but simply append to it. And given that we support
concurrent processes appending to FETCH_HEAD at the same time without
any loss of data, seeding the temporary file with current contents of
FETCH_HEAD initially and then doing a rename wouldn't work either. So
this commit implements the simple strategy of buffering all changes and
appending them to the file on commit.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The handling of ref updates is completely handled by `s_update_ref()`,
which will manage the complete lifecycle of the reference transaction.
This is fine right now given that git-fetch(1) does not support atomic
fetches, so each reference gets its own transaction. It is quite
inflexible though, as `s_update_ref()` only knows about a single
reference update at a time, so it doesn't allow us to alter the
strategy.
This commit prepares `s_update_ref()` and its only caller
`update_local_ref()` to allow passing an external transaction. If none
is given, then the existing behaviour is triggered which creates a new
transaction and directly commits it. Otherwise, if the caller provides a
transaction, then we only queue the update but don't commit it. This
optionally allows the caller to manage when a transaction will be
committed.
Given that `update_local_ref()` is always called with a `NULL`
transaction for now, no change in behaviour is expected from this
change.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The cleanup code in `s_update_ref()` is currently duplicated for both
succesful and erroneous exit paths. This commit refactors the function
to have a shared exit path for both cases to remove the duplication.
Suggested-by: Christian Couder <christian.couder@gmail.com>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
This commit refactors `append_fetch_head()` to use a `struct strbuf` for
formatting the update which we're about to append to the FETCH_HEAD
file. While the refactoring doesn't have much of a benefit right now, it
serves as a preparatory step to implement atomic fetches where we need
to buffer all updates to FETCH_HEAD and only flush them out if all
reference updates succeeded.
No change in behaviour is expected from this commit.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When performing a fetch with the default `--write-fetch-head` option, we
write all updated references to FETCH_HEAD while the updates are
performed. Given that updates are not performed atomically, it means
that we we write to FETCH_HEAD even if some or all of the reference
updates fail.
Given that we simply update FETCH_HEAD ad-hoc with each reference, the
logic is completely contained in `store_update_refs` and thus quite hard
to extend. This can already be seen by the way we skip writing to the
FETCH_HEAD: instead of having a conditional which simply skips writing,
we instead open "/dev/null" and needlessly write all updates there.
We are about to extend git-fetch(1) to accept an `--atomic` flag which
will make the fetch an all-or-nothing operation with regards to the
reference updates. This will also require us to make the updates to
FETCH_HEAD an all-or-nothing operation, but as explained doing so is not
easy with the current layout. This commit thus refactors the wa we write
to FETCH_HEAD and pulls out the logic to open, append to, commit and
close the file. While this may seem rather over-the top at first,
pulling out this logic will make it a lot easier to update the code in a
subsequent commit. It also allows us to easily skip writing completely
in case `--no-write-fetch-head` was passed.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
hashmap_free(), hashmap_free_entries(), and hashmap_free_() have existed
for a while, but aren't necessarily the clearest names, especially with
hashmap_partial_clear() being added to the mix and lazy-initialization
now being supported. Peff suggested we adopt the following names[1]:
- hashmap_clear() - remove all entries and de-allocate any
hashmap-specific data, but be ready for reuse
- hashmap_clear_and_free() - ditto, but free the entries themselves
- hashmap_partial_clear() - remove all entries but don't deallocate
table
- hashmap_partial_clear_and_free() - ditto, but free the entries
This patch provides the new names and converts all existing callers over
to the new naming scheme.
[1] https://lore.kernel.org/git/20201030125059.GA3277724@coredump.intra.peff.net/
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
"git fetch" and "git push" support negative refspecs.
* jk/refspecs-negative:
refspec: add support for negative refspecs
|
|
The lazy fetching done internally to make missing objects available
in a partial clone incorrectly made permanent damage to the partial
clone filter in the repository, which has been corrected.
* jt/keep-partial-clone-filter-upon-lazy-fetch:
fetch: do not override partial clone filter
promisor-remote: remove unused variable
|
|
Both fetch and push support pattern refspecs which allow fetching or
pushing references that match a specific pattern. Because these patterns
are globs, they have somewhat limited ability to express more complex
situations.
For example, suppose you wish to fetch all branches from a remote except
for a specific one. To allow this, you must setup a set of refspecs
which match only the branches you want. Because refspecs are either
explicit name matches, or simple globs, many patterns cannot be
expressed.
Add support for a new type of refspec, referred to as "negative"
refspecs. These are prefixed with a '^' and mean "exclude any ref
matching this refspec". They can only have one "side" which always
refers to the source. During a fetch, this refers to the name of the ref
on the remote. During a push, this refers to the name of the ref on the
local side.
With negative refspecs, users can express more complex patterns. For
example:
git fetch origin refs/heads/*:refs/remotes/origin/* ^refs/heads/dontwant
will fetch all branches on origin into remotes/origin, but will exclude
fetching the branch named dontwant.
Refspecs today are commutative, meaning that order doesn't expressly
matter. Rather than forcing an implied order, negative refspecs will
always be applied last. That is, in order to match, a ref must match at
least one positive refspec, and match none of the negative refspecs.
This is similar to how negative pathspecs work.
Signed-off-by: Jacob Keller <jacob.keller@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When a fetch with the --filter argument is made, the configured default
filter is set even if one already exists. This change was made in
5e46139376 ("builtin/fetch: remove unique promisor remote limitation |
Junio C Hamano
Patrick Steinhardt
Jean-Noël Avila
Derrick Stolee
Ævar Arnfjörð Bjarmason
Anders Kaseorg
Johannes Schindelin
Ben Boeckel
Jonathan Tan
Alex Henrie
Elijah Newren
Jacob Keller