| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
"git apply -R" did not handle patches that touch the same path
twice correctly, which has been corrected. This is most relevant
in a patch that changes a path from a regular file to a symbolic
link (and vice versa).
* jt/apply-reverse-twice:
apply: when -R, also reverse list of sections
|
|
"git rebase --rebase-merges" did not correctly pass --gpg-sign
command line option to underlying "git merge" when replaying a merge
using non-default merge strategy or when replaying an octopus merge
(because replaying a two-head merge with the default strategy was
done in a separate codepath, the problem did not trigger for most
users), which has been corrected.
* sc/sequencer-gpg-octopus:
t3435: add tests for rebase -r GPG signing
sequencer: pass explicit --no-gpg-sign to merge
sequencer: fix gpg option passed to merge subcommand
|
|
Our test scripts can be told to run only individual pieces while
skipping others with the "--run=..." option; they were taught to
take a substring of test title, in addition to numbers, to name the
test pieces to run.
* en/test-selector:
test-lib: reduce verbosity of skipped tests
t6006, t6012: adjust tests to use 'setup' instead of synonyms
test-lib: allow selecting tests by substring/glob with --run
|
|
"git credential' didn't honor the core.askPass configuration
variable (among other things), which has been corrected.
* tk/credential-config:
credential: load default config
|
|
"git diff A...B" learned "git diff --merge-base A B", which is a
longer short-hand to say the same thing.
* dl/diff-merge-base:
contrib/completion: complete `git diff --merge-base`
builtin/diff-tree: learn --merge-base
builtin/diff-index: learn --merge-base
t4068: add --merge-base tests
diff-lib: define diff_get_merge_base()
diff-lib: accept option flags in run_diff_index()
contrib/completion: extract common diff/difftool options
git-diff.txt: backtick quote command text
git-diff-index.txt: make --cached description a proper sentence
t4068: remove unnecessary >tmp
|
|
Test-coverage enhancement of running commit-graph task "git
maintenance" as needed led to discovery and fix of a bug.
* ds/maintenance-commit-graph-auto-fix:
maintenance: core.commitGraph=false prevents writes
maintenance: test commit-graph auto condition
|
|
When "git commit-graph" detects the same commit recorded more than
once while it is merging the layers, it used to die. The code now
ignores all but one of them and continues.
* ds/commit-graph-merging-fix:
commit-graph: don't write commit-graph when disabled
commit-graph: ignore duplicates when merging layers
|
|
A test helper "test_cmp A B" was taught to diagnose missing files A
or B as a bug in test, but some tests legitimately wanted to notice
a failure to even create file B as an error, in addition to leaving
the expected result in it, and were misdiagnosed as a bug. This
has been corrected.
* es/test-cmp-typocatcher:
Revert "test_cmp: diagnose incorrect arguments"
|
|
"git fast-import" wasted a lot of memory when many marks were in use.
* jk/fast-import-marks-alloc-fix:
fast-import: fix over-allocation of marks storage
|
|
The side-band status report can be sent at the same time as the
primary payload multiplexed, but the demultiplexer on the receiving
end incorrectly split a single status report into two, which has
been corrected.
* js/avoid-split-sideband-message:
test-pkt-line: drop colon from sideband identity
sideband: report unhandled incomplete sideband messages as bugs
sideband: avoid reporting incomplete sideband messages
|
|
Micro clean-up.
* cm/t7xxx-cleanup:
t7102: prepare expected output inside test_expect_* block
t7201: put each command on a separate line
t7201: use 'git -C' to avoid subshell
t7102,t7201: remove whitespace after redirect operator
t7102,t7201: remove unnecessary blank spaces in test body
t7101,t7102,t7201: modernize test formatting
|
|
Micro clean-up of a test script.
* ct/t0000-use-test-path-is-file:
t0000: use test_path_is_file instead of "test -f"
|
|
Work around flakiness in a test.
* en/t7518-unflake:
t7518: fix flaky grep invocation
|
|
"git checkout" learned to use checkout.guess configuration variable
and enable/disable its "--[no-]guess" option accordingly.
* dl/checkout-guess:
checkout: learn to respect checkout.guess
Documentation/config/checkout: replace sq with backticks
|
|
"git checkout -p A...B [-- <path>]" did not work, even though the
same command without "-p" correctly used the merge-base between
commits A and B.
* dl/checkout-p-merge-base:
t2016: add a NEEDSWORK about the PERL prerequisite
add-patch: add NEEDSWORK about comparing commits
Doc: document "A...B" form for <tree-ish> in checkout and switch
builtin/checkout: fix `git checkout -p HEAD...` bug
|
|
"git clone" learned clone.defaultremotename configuration variable
to customize what nickname to use to call the remote the repository
was cloned from.
* sb/clone-origin:
clone: allow configurable default for `-o`/`--origin`
clone: read new remote name from remote_name instead of option_origin
clone: validate --origin option before use
refs: consolidate remote name validation
remote: add tests for add and rename with invalid names
clone: use more conventional config/option layering
clone: add tests for --template and some disallowed option pairs
|
|
"git push --force-with-lease[=<ref>]" can easily be misused to lose
commits unless the user takes good care of their own "git fetch".
A new option "--force-if-includes" attempts to ensure that what is
being force-pushed was created after examining the commit at the
tip of the remote ref that is about to be force-replaced.
* sk/force-if-includes:
t, doc: update tests, reference for "--force-if-includes"
push: parse and set flag for "--force-if-includes"
push: add reflog check for "--force-if-includes"
|
|
"git maintenance", an extended big brother of "git gc", continues
to evolve.
* ds/maintenance-part-2:
maintenance: add incremental-repack auto condition
maintenance: auto-size incremental-repack batch
maintenance: add incremental-repack task
midx: use start_delayed_progress()
midx: enable core.multiPackIndex by default
maintenance: create auto condition for loose-objects
maintenance: add loose-objects task
maintenance: add prefetch task
|
|
"git worktree list" now shows if each worktree is locked. This
possibly may open us to show other kinds of states in the future.
* rs/worktree-list-show-locked:
worktree: teach `list` to annotate locked worktree
|
|
Userdiff for CSS update.
* sd/userdiff-css-update:
userdiff: expand detected chunk headers for css
|
|
Userdiff for Rust update.
* kb/userdiff-rust-macro-rules:
userdiff: recognize 'macro_rules!' as starting a Rust function block
|
|
Userdiff for PHP update.
* js/userdiff-php:
userdiff: PHP: catch "abstract" and "final" functions
|
|
We pass "sideband: " as our identity for errors to recv_sideband(). But
it already adds the trailing colon and space. This doesn't invalidate
any tests, but it looks funny when you examine the test output.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
In 2.29, "--committer-date-is-author-date" option of "rebase" and
"am" subcommands lost the e-mail address by mistake, which has been
corrected.
* jk/committer-date-is-author-date-fix:
rebase: fix broken email with --committer-date-is-author-date
am: fix broken email with --committer-date-is-author-date
t3436: check --committer-date-is-author-date result more carefully
|
|
Commit 7573cec52c (rebase -i: support --committer-date-is-author-date,
2020-08-17) copied the committer ident-parsing code from builtin/am.c.
And in doing so, it copied a bug in which we always set the email to an
empty string. We fixed the version in git-am in the previous commit;
this commit fixes the copied code.
Reported-by: VenomVendor <info@venomvendor.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Commit e8cbe2118a (am: stop exporting GIT_COMMITTER_DATE, 2020-08-17)
rewrote the code for setting the committer date to use fmt_ident(),
rather than setting an environment variable and letting commit_tree()
handle it. But it introduced two bugs:
- we use the author email string instead of the committer email
- when parsing the committer ident, we used the wrong variable to
compute the length of the email, resulting in it always being a
zero-length string
This commit fixes both, which causes our test of this option via the
rebase "apply" backend to now succeed.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
After running "rebase --committer-date-is-author-date", we confirm that
the committer date is the same as the author date. However, we don't
look at any other parts of the committer ident line to make sure we
didn't screw them up. And indeed, there are a few bugs here. Depending
on the rebase backend in use, we may accidentally use the author email
instead of the committer's, or even an empty string.
Let's teach our test_ctime_is_atime helper to check the committer name
and email, which reveals several failing tests.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
That way we can notice if there is a breakage/bug in the parts of
the test that prepare the expected outcome, which is how modern
tests are arranged.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Modern practice is to avoid multiple commands per line,
and instead place each command on its own line.
Signed-off-by: Charvi Mendiratta <charvi077@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Signed-off-by: Charvi Mendiratta <charvi077@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
According to Documentation/CodingGuidelines, redirect
operator is written with space before, but no space
after them.
Let's remove these whitespaces after redirect operators.
Signed-off-by: Charvi Mendiratta <charvi077@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
A patch changing a symlink into a file is written with 2 sections (in
the code, represented as "struct patch"): firstly, the deletion of the
symlink, and secondly, the creation of the file. When applying that
patch with -R, the sections are reversed, so we get:
(1) creation of a symlink, then
(2) deletion of a file.
This causes an issue when the "deletion of a file" section is checked,
because Git observes that the so-called file is not a file but a
symlink, resulting in a "wrong type" error message.
What we want is:
(1) deletion of a file, then
(2) creation of a symlink.
In the code, this is reflected in the behavior of previous_patch() when
invoked from check_preimage() when the deletion is checked. Creation
then deletion means that when the deletion is checked, previous_patch()
returns the creation section, triggering a mode conflict resulting in
the "wrong type" error message. But deletion then creation means that
when the deletion is checked, previous_patch() returns NULL, so the
deletion mode is checked against lstat, which is what we want.
There are also other ways a patch can contain 2 sections referencing the
same file, for example, in 7a07841c0b ("git-apply: handle a patch that
touches the same path more than once better", 2008-06-27). "git apply
-R" fails in the same way, and this commit makes this case succeed.
Therefore, when building the list of sections, build them in reverse
order (by adding to the front of the list instead of the back) when -R
is passed.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
In 2b695ecd74d (t5500: count objects through stderr, not trace,
2020-05-06) we tried to ensure that the "Total 3" message could be
grepped in Git's output, even if it sometimes got chopped up into
multiple lines in the trace machinery.
However, the first instance where this mattered now goes through the
sideband machinery, where it is _still_ possible for messages to get
chopped up: it *is* possible for the standard error stream to be sent
byte-for-byte and hence it can be easily interrupted. Meaning: it is
possible for the single line that we're looking for to be chopped up
into multiple sideband packets, with a primary packet being delivered
between them.
This seems to happen occasionally in the `vs-test` part of our CI
builds, i.e. with binaries built using Visual C, but not when building
with GCC or clang; The symptom is that t5500.43 fails to find a line
matching `remote: Total 3` in the `log` file, which ends in something
along these lines:
remote: Tota
remote: l 3 (delta 0), reused 0 (delta 0), pack-reused 0
This should not happen, though: we have code in `demultiplex_sideband()`
_specifically_ to stitch back together lines that were delivered in
separate sideband packets.
However, this stitching was broken in a subtle way in fbd76cd450
(sideband: reverse its dependency on pkt-line, 2019-01-16): before that
change, incomplete sideband lines would not be flushed upon receiving a
primary packet, but after that patch, they would be.
The subtleness of this bug comes from the fact that it is easy to get
confused by the ambiguous meaning of the `break` keyword: after writing
the primary packet contents, the `break;` in the original version of
`recv_sideband()` does _not_ break out of the `while` loop, but instead
only ends the `switch` case:
while (!retval) {
[...]
switch (band) {
[...]
case 1:
/* Write the contents of the primary packet */
write_or_die(out, buf + 1, len);
/* Here, we do *not* break out of the loop, `retval` is unchanged */
break;
[...]
}
if (outbuf.len) {
/* Write any remaining sideband messages lacking a trailing LF */
strbuf_addch(&outbuf, '\n');
xwrite(2, outbuf.buf, outbuf.len);
}
In contrast, after fbd76cd450 (sideband: reverse its dependency on
pkt-line, 2019-01-16), the body of the `while` loop was extracted into
`demultiplex_sideband()`, crucially _including_ the logic to write
incomplete sideband messages:
switch (band) {
[...]
case 1:
*sideband_type = SIDEBAND_PRIMARY;
/* This does not break out of the loop: the loop is in the caller */
break;
[...]
}
cleanup:
[...]
/* This logic is now no longer _outside_ the loop but _inside_ */
if (scratch->len) {
strbuf_addch(scratch, '\n');
xwrite(2, scratch->buf, scratch->len);
}
The correct way to fix this is to return from `demultiplex_sideband()`
early. The caller will then write out the contents of the primary packet
and continue looping. The `scratch` buffer for incomplete sideband
messages is owned by that caller, and will continue to accumulate the
remainder(s) of those messages. The loop will only end once
`demultiplex_sideband()` returned non-zero _and_ did not indicate a
primary packet, which is the case only when we hit the `cleanup:` path,
in which we take care of flushing any unfinished sideband messages and
release the `scratch` buffer.
To ensure that this does not get broken again, we introduce a pair of
subcommands of the `pkt-line` test helper that specifically chop up the
sideband message and squeeze a primary packet into the middle.
Final note: The other test case touched by 2b695ecd74d (t5500: count
objects through stderr, not trace, 2020-05-06) is not affected by this
issue because the sideband machinery is not involved there.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
t7102 and t7201 still follow the old style of having blank
lines around test body, which is not consistence with our
current practice.
Let's remove those unnecessary blank lines.
Signed-off-by: Charvi Mendiratta <charvi077@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Some tests in these scripts are formatted using a very old style:
test_expect_success \
'title' \
'body line 1 &&
body line 2'
Updating the formatting to the modern style:
test_expect_success 'title' '
body line 1 &&
body line 2
'
Signed-off-by: Charvi Mendiratta <charvi077@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
This reverts commit d572f52a64c6a69990f72ad6a09504b9b615d2e4; the
idea to detect that "test_cmp expect actual" was fed a misspelt
filename meant well, but when the version of Git tested exhibits a
bug, the reason why these two files do not match may be because one
of them did not get created as expected, in which case missing file
is not a sign of misspelt filename but is a genuine test failure.
Acked-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When using the --run flag to run just two or three tests from a test
file which contains several dozen tests, having every skipped test print
out dozens of lines of output for the test code for that skipped test
(in addition to the TAP output line) adds up to hundreds or thousands of
lines of irrelevant output that make it very hard to fish out the
relevant results you were looking for. Simplify the output for skipped
tests to remove this extra output, leaving only the TAP output line
(i.e. the line reading "ok <number> # skip <test-description>", which
already mentions that the test was "skip"ped).
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
With the new ability to pass --run=setup to select which tests to run,
it is more convenient if tests use the term "setup" instead of synonyms
like 'prepare' or 'rebuild'. There are undoubtedly many other tests in
our testsuite that could be changed over too, these are just a couple
that I ran into.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Many of our test scripts have several "setup" tests. It's a lot easier
to say
./t0050-filesystem.sh --run=setup,9
in order to run all the setup tests as well as test #9, than it is to
track down what all the setup tests are and enter all their numbers in
the list. Also, I often find myself wanting to run just one or a couple
tests from the test file, but I don't know the numbering of any of the
tests -- to get it I either have to first run the whole test file (or
start counting by hand or figure out some other clever but non-obvious
tricks). It's really convenient to be able to just look at the test
description(s) and then run
./t6416-recursive-corner-cases.sh --run=symlink
or
./t6402-merge-rename.sh --run='setup,unnecessary update'
Add such an ability to test selection which relies on merely matching
against the test description.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
t7518.1 added in commit 862e80a413 ("ident: handle NULL email when
complaining of empty name", 2017-02-23), was trying to make sure that
the test with an empty ident did not segfault and did not result in
glibc quiety translating a NULL pointer into a name of "(null)". It did
the latter by ensuring that a grep for "null" didn't appear in the
output, but on one automatic CI run I observed the following output:
fatal: empty ident name (for <runner@fv-az128-670.gcliasfzo2nullsdbrimjtbyhg.cx.internal.cloudapp.net>) not allowed
Note that 'null' appears as a substring of the domain name, found
within 'gcliasfzo2nullsdbrimjtbyhg'. Tighten the test by searching for
"(null)" rather than "null".
Signed-off-by: Elijah Newren <newren@gmail.com>
Reviewed-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Add test cases of various combinations of the commit.gpgsign option and
--gpg-sign, --no-gpg-sign flags with rebase -r with the default merge
strategy. This excercises a different code-path from those with octopus
merges or overridden merge strategy with rebase -s.
Signed-off-by: Samuel Čavoj <samuel@cavoj.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The merge subcommand launched for merges with non-default strategy would
use its own default behaviour to decide how to sign commits, regardless
of what opts->gpg_sign was set to. For example the --no-gpg-sign flag
given to rebase explicitly would get ignored, if commit.gpgsign was set
to true.
Fix the issue and add a test case excercising this behaviour.
Signed-off-by: Samuel Čavoj <samuel@cavoj.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When performing a rebase with --rebase-merges using either a custom
strategy specified with -s or an octopus merge, and at the same time
having gpgsign enabled (either rebase -S or config commit.gpgsign), the
operation would fail on making the merge commit. Instead of "-S%s" with
the key id substituted, only the bare key id would get passed to the
underlying merge command, which tried to interpret it as a ref.
Fix the issue and add test cases as suggested by Johannes Schindelin and
Junio C Hamano.
Signed-off-by: Samuel Čavoj <samuel@cavoj.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Signed-off-by: Caleb Tillman <caleb.tillman@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Make `git credential fill` honour the core.askPass variable.
Signed-off-by: Thomas Koutcher <thomas.koutcher@online.fr>
[jk: added test]
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Fast-import stores its marks in a trie-like structure made of mark_set
structs. Each struct has a fixed size (1024). If our id number is too
large to fit in the struct, then we allocate a new struct which shifts
the id number by 10 bits. Our original struct becomes a child node
of this new layer, and the new struct becomes the top level of the trie.
This scheme was broken by ddddf8d7e2 (fast-import: permit reading
multiple marks files, 2020-02-22). Before then, we had a top-level
"marks" pointer, and the push-down worked by assigning the new top-level
struct to "marks". But after that commit, insert_mark() takes a pointer
to the mark_set, rather than using the global "marks". It continued to
assign to the global "marks" variable during the push down, which was
wrong for two reasons:
- we added a call in option_rewrite_submodules() which uses a separate
mark set; pushing down on "marks" is outright wrong here. We'd
corrupt the "marks" set, and we'd fail to correctly store any
submodule mappings with an id over 1024.
- the other callers passed "marks", but the push-down was still wrong.
In read_mark_file(), we take the pointer to the mark_set as a
parameter. So even though insert_mark() was updating the global
"marks", the local pointer we had in read_mark_file() was not
updated. As a result, we'd add a new level when needed, but then the
next call to insert_mark() wouldn't see it! It would then allocate a
new layer, which would also not be seen, and so on. Lookups for the
lost layers obviously wouldn't work, but before we even hit any
lookup stage, we'd generally run out of memory and die.
Our tests didn't notice either of these cases because they didn't have
enough marks to trigger the push-down behavior. The new tests in t9304
cover both cases (and fail without this patch).
We can solve the problem by having insert_mark() take a pointer-to-pointer
of the top-level of the set. Then our push down can assign to it in a
way that the caller actually sees. Note the subtle reordering in
option_rewrite_submodules(). Our call to read_mark_file() may modify our
top-level set pointer, so we have to wait until after it returns to
assign its value into the string_list.
Reported-by: Sergey Brester <serg.brester@sebres.de>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The "git worktree list" shows the absolute path to the working tree,
the commit that is checked out and the name of the branch. It is not
immediately obvious which of the worktrees, if any, are locked.
"git worktree remove" refuses to remove a locked worktree with
an error message. If "git worktree list" told which worktrees
are locked in its output, the user would not even attempt to
remove such a worktree, or would realize that
"git worktree remove -f -f <path>" is required.
Teach "git worktree list" to append "locked" to its output.
The output from the command becomes like so:
$ git worktree list
/path/to/main abc123 [master]
/path/to/worktree 456def (detached HEAD)
/path/to/locked-worktree 123abc (detached HEAD) locked
Helped-by: Junio C Hamano <gitster@pobox.com>
Helped-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Rafael Silva <rafaeloliveira.cs@gmail.com>
Reviewed-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Recently, a user had an issue due to combining
fetch.writeCommitGraph=true with core.commitGraph=false. The root bug
has been resolved by preventing commit-graph writes when
core.commitGraph is disabled. This happens inside the 'git commit-graph
write' command, but we can be more aware of this situation and prevent
that process from ever starting in the 'commit-graph' maintenance task.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The core.commitGraph config setting can be set to 'false' to prevent
parsing commits from the commit-graph file(s). This causes an issue when
trying to write with "--split" which needs to distinguish between
commits that are in the existing commit-graph layers and commits that
are not. The existing mechanism uses parse_commit() and follows by
checking if there is a 'graph_pos' that shows the commit was parsed from
the commit-graph file.
When core.commitGraph=false, we do not parse the commits from the
commit-graph and 'graph_pos' indicates that no commits are in the
existing file. The --split logic moves forward creating a new layer on
top that holds all reachable commits, then possibly merges down into
those layers, resulting in duplicate commits. The previous change makes
that merging process more robust to such a situation in case it happens
in the written commit-graph data.
The easy answer here is to avoid writing a commit-graph if reading the
commit-graph is disabled. Since the resulting commit-graph will would not
be read by subsequent Git processes. This is more natural than forcing
core.commitGraph to be true for the 'write' process.
Reported-by: Thomas Braun <thomas.braun@virtuell-zuhause.de>
Helped-by: Jeff King <peff@peff.net>
Helped-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Thomas reported [1] that a "git fetch" command was failing with an error
saying "unexpected duplicate commit id". The root cause is that they had
fetch.writeCommitGraph enabled which generates commit-graph chains, and
this instance was merging two layers that both contained the same commit
ID.
[1] https://lore.kernel.org/git/55f8f00c-a61c-67d4-889e-a9501c596c39@virtuell-zuhause.de/
The initial assumption is that Git would not write a commit ID into a
commit-graph layer if it already exists in a lower commit-graph layer.
Somehow, this specific case did get into that situation, leading to this
error.
While unexpected, this isn't actually invalid (as long as the two layers
agree on the metadata for the commit). When we parse a commit that does
not have a graph_pos in the commit_graph_data_slab, we use binary search
in the commit-graph layers to find the commit and set graph_pos. That
position is never used again in this case. However, when we parse a
commit from the commit-graph file, we load its parents from the
commit-graph and assign graph_pos at that point. If those parents were
already parsed from the commit-graph, then nothing needs to be done.
Otherwise, this graph_pos is a valid position in the commit-graph so we
can parse the parents, when necessary.
Thus, this die() is too aggressive. The easiest thing to do would be to
ignore the duplicates.
If we only ignore the duplicates, then we will produce a commit-graph
that has identical commit IDs listed in adjacent positions. This excess
data will never be removed from the commit-graph, which could cascade
into significantly bloated file sizes.
Thankfully, we can collapse the list to erase the duplicate commit
pointers. This allows us to get the end result we want without extra
memory costs and minimal CPU time.
The root cause is due to disabling core.commitGraph, which prevents
parsing commits from the lower layers during a 'git commit-graph write
--split' command. Since we use the 'graph_pos' value to determine
whether a commit is in a lower layer, we never discover that those
commits are already in the commit-graph chain and add them to the top
layer. This layer is then merged down, creating duplicates.
The test added in t5324-split-commit-graph.sh fails without this change.
However, we still have not completely removed the need for this
duplicate check. That will come in a follow-up change.
Reported-by: Thomas Braun <thomas.braun@virtuell-zuhause.de>
Helped-by: Taylor Blau <me@ttaylorr.com>
Co-authored-by: Jeff King <peff@peff.net>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
Junio C Hamano
Jeff King
Charvi Mendiratta
Jonathan Tan
Johannes Schindelin
Elijah Newren
Samuel Čavoj
Caleb Tillman
Thomas Koutcher
Rafael Silva
Derrick Stolee