When reachability bitmap coverage exists in a repository, Git will use a
different (and hopefully faster) traversal to compute revision walks.
Consider a set of positive and negative tips (which we'll refer to with
their standard bitmap parlance by "wants", and "haves"). In order to
figure out what objects exist between the tips, the existing traversal
in `prepare_bitmap_walk()` does something like:
1. Consider if we can even compute the set of objects with bitmaps,
and fall back to the usual traversal if we cannot. For example,
pathspec limiting traversals can't be computed using bitmaps (since
they don't know which objects are at which paths). The same is true
of certain kinds of non-trivial object filters.
2. If we can compute the traversal with bitmaps, partition the
(dereferenced) tips into two object lists, "haves", and "wants",
based on whether or not the objects have the UNINTERESTING flag,
respectively.
3. Fall back to the ordinary object traversal if either (a) there are
more than zero haves, none of which are in the bitmapped pack or
MIDX, or (b) there are no wants.
4. Construct a reachability bitmap for the "haves" side by walking
from the revision tips down to any existing bitmaps, OR-ing in any
bitmaps as they are found.
5. Then do the same for the "wants" side, stopping at any objects that
appear in the "haves" bitmap.
6. Filter the results if any object filter (that can be easily
computed with bitmaps alone) was given, and then return back to the
caller.
When there is good bitmap coverage relative to the traversal tips, this
walk is often significantly faster than an ordinary object traversal
because it can visit far fewer objects.
But in certain cases, it can be significantly *slower* than the usual
object traversal. Why? Because we need to compute complete bitmaps on
either side of the walk. If either one (or both) of the sides require
walking many (or all!) objects before they get to an existing bitmap,
the extra bitmap machinery is mostly or all overhead.
One of the benefits, however, is that even if the walk is slower, bitmap
traversals are guaranteed to provide an *exact* answer. Unlike the
traditional object traversal algorithm, which can over-count the results
by not opening trees for older commits, the bitmap walk builds an exact
reachability bitmap for either side, meaning the results are never
over-counted.
But producing non-exact results is OK for our traversal here (both in
the bitmap case and not), as long as the results are over-counted, not
under.
Relaxing the bitmap traversal to allow it to produce over-counted
results gives us the opportunity to make some significant improvements.
Instead of the above, the new algorithm only has to walk from the
*boundary* down to the nearest bitmap, instead of from each of the
UNINTERESTING tips.
The boundary-based approach still has degenerate cases, but we'll show
in a moment that it is often a significant improvement.
The new algorithm works as follows:
1. Build a (partial) bitmap of the haves side by first OR-ing any
bitmap(s) that already exist for UNINTERESTING commits between the
haves and the boundary.
2. For each commit along the boundary, add it as a fill-in traversal
tip (where the traversal terminates once an existing bitmap is
found), and perform fill-in traversal.
3. Build up a complete bitmap of the wants side as usual, stopping any
time we intersect the (partial) haves side.
4. Return the results.
And is more-or-less equivalent to using the *old* algorithm with this
invocation:
$ git rev-list --objects --use-bitmap-index $WANTS --not \
$(git rev-list --objects --boundary $WANTS --not $HAVES |
perl -lne 'print $1 if /^-(.*)/')
The new result performs significantly better in many cases, particularly
when the distance from the boundary commit(s) to an existing bitmap is
shorter than the distance from (all of) the have tips to the nearest
bitmapped commit.
Note that when using the old bitmap traversal algorithm, the results can
be *slower* than without bitmaps! Under the new algorithm, the result is
computed faster with bitmaps than without (at the cost of over-counting
the true number of objects in a similar fashion as the non-bitmap
traversal):
# (Computing the number of tagged objects not on any branches
# without bitmaps).
$ time git rev-list --count --objects --tags --not --branches
20
real 0m1.388s
user 0m1.092s
sys 0m0.296s
# (Computing the same query using the old bitmap traversal).
$ time git rev-list --count --objects --tags --not --branches --use-bitmap-index
19
real 0m22.709s
user 0m21.628s
sys 0m1.076s
# (this commit)
$ time git.compile rev-list --count --objects --tags --not --branches --use-bitmap-index
19
real 0m1.518s
user 0m1.234s
sys 0m0.284s
The new algorithm is still slower than not using bitmaps at all, but it
is nearly a 15-fold improvement over the existing traversal.
In a more realistic setting (using my local copy of git.git), I can
observe a similar (if more modest) speed-up:
$ argv="--count --objects --branches --not --tags"
hyperfine \
-n 'no bitmaps' "git.compile rev-list $argv" \
-n 'existing traversal' "git.compile rev-list --use-bitmap-index $argv" \
-n 'boundary traversal' "git.compile -c pack.useBitmapBoundaryTraversal=true rev-list --use-bitmap-index $argv"
Benchmark 1: no bitmaps
Time (mean ± σ): 124.6 ms ± 2.1 ms [User: 103.7 ms, System: 20.8 ms]
Range (min … max): 122.6 ms … 133.1 ms 22 runs
Benchmark 2: existing traversal
Time (mean ± σ): 368.6 ms ± 3.0 ms [User: 325.3 ms, System: 43.1 ms]
Range (min … max): 365.1 ms … 374.8 ms 10 runs
Benchmark 3: boundary traversal
Time (mean ± σ): 167.6 ms ± 0.9 ms [User: 139.5 ms, System: 27.9 ms]
Range (min … max): 166.1 ms … 169.2 ms 17 runs
Summary
'no bitmaps' ran
1.34 ± 0.02 times faster than 'boundary traversal'
2.96 ± 0.05 times faster than 'existing traversal'
Here, the new algorithm is also still slower than not using bitmaps, but
represents a more than 2-fold improvement over the existing traversal in
a more modest example.
Since this algorithm was originally written (nearly a year and a half
ago, at the time of writing), the bitmap lookup table shipped, making
the new algorithm's result more competitive. A few other future
directions for improving bitmap traversal times beyond not using bitmaps
at all:
- Decrease the cost to decompress and OR together many bitmaps
together (particularly when enumerating the uninteresting side of
the walk). Here we could explore more efficient bitmap storage
techniques, like Roaring+Run and/or use SIMD instructions to speed
up ORing them together.
- Store pseudo-merge bitmaps, which could allow us to OR together
fewer "summary" bitmaps (which would also help with the above).
Helped-by: Jeff King <peff@peff.net>
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Back in e8c58f894b (t: support GIT_TEST_WRITE_REV_INDEX, 2021-01-25), we
added a test knob to conditionally enable writing a ".rev" file when
indexing a pack. At the time, this was used to ensure that the test
suite worked even when ".rev" files were written, which served as a
stress-test for the on-disk reverse index implementation.
Now that reading from on-disk ".rev" files is enabled by default, the
test knob `GIT_TEST_WRITE_REV_INDEX` no longer has any meaning.
We could get rid of the option entirely, but there would be no
convenient way to test Git when ".rev" files *aren't* in place.
Instead of getting rid of the option, invert its meaning to instead
disable writing ".rev" files, thereby running the test suite in a mode
where the reverse index is generated from scratch.
This ensures that, when GIT_TEST_NO_WRITE_REV_INDEX is set to some
spelling of "true", we are still running and exercising Git's behavior
when forced to generate reverse indexes from scratch. Do so by setting
it in the linux-TEST-vars CI run to ensure that we are maintaining good
coverage of this now-legacy code.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since [1] first released with Git v2.37.0 the built-in version of "add
-i" has been the default. That built-in implementation was added in
[2], first released with Git v2.25.0.
At this point enough time has passed to allow for finding any
remaining bugs in this new implementation, so let's remove the
fallback code.
As with similar migrations for "stash"[3] and "rebase"[4] we're
keeping a mention of "add.interactive.useBuiltin" in the
documentation, but adding a warning() to notify any outstanding users
that the built-in is now the default. As with [5] and [6] we should
follow-up in the future and eventually remove that warning.
1. 0527ccb1b5 (add -i: default to the built-in implementation,
2021-11-30)
2. f83dff60a7 (Start to implement a built-in version of `git add
--interactive`, 2019-11-13)
3. 8a2cd3f512 (stash: remove the stash.useBuiltin setting,
2020-03-03)
4. d03ebd411c (rebase: remove the rebase.useBuiltin setting,
2019-03-18)
5. deeaf5ee07 (stash: remove documentation for `stash.useBuiltin`,
2022-01-27)
6. 9bcde4d531 (rebase: remove transitory rebase.useBuiltin setting &
env, 2021-03-23)
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Update the GitHub workflow support to make it quicker to get to the
failing test.
* js/ci-github-workflow-markup:
ci: call `finalize_test_case_output` a little later
ci(github): mention where the full logs can be found
ci: use `--github-workflow-markup` in the GitHub workflow
ci(github): avoid printing test case preamble twice
ci(github): skip the logs of the successful test cases
ci: optionally mark up output in the GitHub workflow
ci/run-build-and-tests: add some structure to the GitHub workflow output
ci: make it easier to find failed tests' logs in the GitHub workflow
ci/run-build-and-tests: take a more high-level view
test(junit): avoid line feeds in XML attributes
tests: refactor --write-junit-xml code
ci: fix code style
"git add -i" was rewritten in C some time ago and has been in
testing; the reimplementation is now exposed to general public by
default.
* js/use-builtin-add-i:
add -i: default to the built-in implementation
t2016: require the PERL prereq only when necessary
The current output of Git's GitHub workflow can be quite confusing,
especially for contributors new to the project.
To make it more helpful, let's introduce some collapsible grouping.
Initially, readers will see the high-level view of what actually
happened (did the build fail, or the test suite?). To drill down, the
respective group can be expanded.
Note: sadly, workflow output currently cannot contain any nested groups
(see https://github.com/actions/runner/issues/802 for details),
therefore we take pains to ensure to end any previous group before
starting a new one.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When investigating a test failure, the time that matters most is the
time it takes from getting aware of the failure to displaying the output
of the failing test case.
You currently have to know a lot of implementation details when
investigating test failures in the CI runs. The first step is easy: the
failed job is marked quite clearly, but when opening it, the failed step
is expanded, which in our case is the one running
`ci/run-build-and-tests.sh`. This step, most notably, only offers a
high-level view of what went wrong: it prints the output of `prove`
which merely tells the reader which test script failed.
The actually interesting part is in the detailed log of said failed
test script. But that log is shown in the CI run's step that runs
`ci/print-test-failures.sh`. And that step is _not_ expanded in the web
UI by default. It is even marked as "successful", which makes it very
easy to miss that there is useful information hidden in there.
Let's help the reader by showing the failed tests' detailed logs in the
step that is expanded automatically, i.e. directly after the test suite
failed.
This also helps the situation where the _build_ failed and the
`print-test-failures` step was executed under the assumption that the
_test suite_ failed, and consequently failed to find any failed tests.
An alternative way to implement this patch would be to source
`ci/print-test-failures.sh` in the `handle_test_failures` function to
show these logs. However, over the course of the next few commits, we
want to introduce some grouping which would be harder to achieve that
way (for example, we do want a leaner, and colored, preamble for each
failed test script, and it would be trickier to accommodate the lack of
nested groupings in GitHub workflows' output).
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the web UI of GitHub workflows, failed runs are presented with the
job step that failed auto-expanded. In the current setup, this is not
helpful at all because that shows only the output of `prove`, which says
which test failed, but not in what way.
What would help understand the reader what went wrong is the verbose
test output of the failed test.
The logs of the failed runs do contain that verbose test output, but it
is shown in the _next_ step (which is marked as succeeding, and is
therefore _not_ auto-expanded). Anyone not intimately familiar with this
would completely miss the verbose test output, being left mostly
puzzled with the test failures.
We are about to show the failed test cases' output in the _same_ step,
so that the user has a much easier time to figure out what was going
wrong.
But first, we must partially revert the change that tried to improve the
CI runs by combining the `Makefile` targets to build into a single
`make` invocation. That might have sounded like a good idea at the time,
but it does make it rather impossible for the CI script to determine
whether the _build_ failed, or the _tests_. If the tests were run at
all, that is.
So let's go back to calling `make` for the build, and call `make test`
separately so that we can easily detect that _that_ invocation failed,
and react appropriately.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 9a5315edfd (Merge branch 'js/patch-mode-in-others-in-c',
2020-02-05), Git acquired a built-in implementation of `git add`'s
interactive mode that could be turned on via the config option
`add.interactive.useBuiltin`.
The first official Git version to support this knob was v2.26.0.
In 2df2d81ddd (add -i: use the built-in version when
feature.experimental is set, 2020-09-08), this built-in implementation
was also enabled via `feature.experimental`. The first version with this
change was v2.29.0.
More than a year (and very few bug reports) later, it is time to declare
the built-in implementation mature and to turn it on by default.
We specifically leave the `add.interactive.useBuiltin` configuration in
place, to give users an "escape hatch" in the unexpected case should
they encounter a previously undetected bug in that implementation.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "linux-clang" and "linux-gcc" jobs both run "make test" twice, but
with different environment variables. Running these in sequence seems
to have been done to work around some constraint on Travis, see
ae59a4e44f (travis: run tests with GIT_TEST_SPLIT_INDEX, 2018-01-07).
By having these run in parallel we'll get jobs that finish much sooner
than they otherwise would have.
We can also simplify the control flow in "ci/run-build-and-tests.sh"
as a result, since we won't run "make test" twice we don't need to run
"make" twice at all, let's default to "make all test" after setting
the variables, and then override it to just "all" for the compile-only
tests.
Add a comment to clarify that new "test" targets should adjust
$MAKE_TARGETS rather than being added after the "case/esac". This
should avoid future confusion where e.g. the compilation-only
"pedantic" target will unexpectedly start running tests. See [1] and
[2].
1. https://lore.kernel.org/git/211122.86ee78yxts.gmgdl@evledraar.gmail.com/
2. https://lore.kernel.org/git/211123.86ilwjujmd.gmgdl@evledraar.gmail.com/
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove support for running the CI in travis. The last builds in it are
from 5 months ago[1] (as of 2021-11-19), and our documentation has
referred to GitHub CI instead since f003a91f5c (SubmittingPatches:
replace discussion of Travis with GitHub Actions, 2021-07-22).
We'll now run the "t9810 t9816" and tests on OSX. We didn't before, as
we'd carried the Travis exclusion of them forward from
522354d70f (Add Travis CI support, 2015-11-27). Let's hope whatever
issue there was with them was either Travis specific, or fixed since
then (I'm not sure).
The "apt-add-repository" invocation (which we were doing in GitHub CI)
isn't needed, it was another Travis-only case that was carried forward
into more general code. See 0f0c51181d (travis-ci: install packages
in 'ci/install-dependencies.sh', 2018-11-01).
Remove the "linux-gcc-4.8" job added in fb9d7431cf (travis-ci: build
with GCC 4.8 as well, 2019-07-18), it only ran in Travis CI.
1. https://travis-ci.org/github/git/git/builds
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The reachability bitmap file used to be generated only for a single
pack, but now we've learned to generate bitmaps for history that
span across multiple packfiles.
* tb/multi-pack-bitmaps: (29 commits)
pack-bitmap: drop bitmap_index argument from try_partial_reuse()
pack-bitmap: drop repository argument from prepare_midx_bitmap_git()
p5326: perf tests for MIDX bitmaps
p5310: extract full and partial bitmap tests
midx: respect 'GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP'
t7700: update to work with MIDX bitmap test knob
t5319: don't write MIDX bitmaps in t5319
t5310: disable GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP
t0410: disable GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP
t5326: test multi-pack bitmap behavior
t/helper/test-read-midx.c: add --checksum mode
t5310: move some tests to lib-bitmap.sh
pack-bitmap: write multi-pack bitmaps
pack-bitmap: read multi-pack bitmaps
pack-bitmap.c: avoid redundant calls to try_partial_reuse
pack-bitmap.c: introduce 'bitmap_is_preferred_refname()'
pack-bitmap.c: introduce 'nth_bitmap_object_oid()'
pack-bitmap.c: introduce 'bitmap_num_objects()'
midx: avoid opening multiple MIDXs when writing
midx: close linked MIDXs, avoid leaking memory
...
Introduce a new 'GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP' environment
variable to also write a multi-pack bitmap when
'GIT_TEST_MULTI_PACK_INDEX' is set.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
similar to the recently added sparse task, it is nice to know as early
as possible.
add a dockerized build using fedora (that usually has the latest gcc)
to be ahead of the curve and avoid older ISO C issues at the same time.
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The final part of "parallel checkout".
* mt/parallel-checkout-part-3:
ci: run test round with parallel-checkout enabled
parallel-checkout: add tests related to .gitattributes
t0028: extract encoding helpers to lib-encoding.sh
parallel-checkout: add tests related to path collisions
parallel-checkout: add tests for basic operations
checkout-index: add parallel checkout support
builtin/checkout.c: complete parallel checkout support
make_transient_cache_entry(): optionally alloc from mem_pool
We already have tests for the basic parallel-checkout operations. But
this code can also run be executed by other commands, such as
git-read-tree and git-sparse-checkout, which are currently not tested
with multiple workers. To promote a wider test coverage without
duplicating tests:
1. Add the GIT_TEST_CHECKOUT_WORKERS environment variable, to optionally
force parallel-checkout execution during the whole test suite.
2. Set this variable (with a value of 2) in the second test round of our
linux-gcc CI job. This round runs `make test` again with some
optional GIT_TEST_* variables enabled, so there is no additional
overhead in exercising the parallel-checkout code here.
Note that tests checking out less than two parallel-eligible entries
will fall back to the sequential mode. Nevertheless, it's still a good
exercise for the parallel-checkout framework as the fallback codepath
also writes the queued entries using the parallel-checkout functions
(only without spawning any worker).
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Plug the ort merge backend throughout the rest of the system, and
start testing it as a replacement for the recursive backend.
* en/ort-readiness:
Add testing with merge-ort merge strategy
t6423: mark remaining expected failure under merge-ort as such
Revert "merge-ort: ignore the directory rename split conflict for now"
merge-recursive: add a bunch of FIXME comments documenting known bugs
merge-ort: write $GIT_DIR/AUTO_MERGE whenever we hit a conflict
t: mark several submodule merging tests as fixed under merge-ort
merge-ort: implement CE_SKIP_WORKTREE handling with conflicted entries
t6428: new test for SKIP_WORKTREE handling and conflicts
merge-ort: support subtree shifting
merge-ort: let renormalization change modify/delete into clean delete
merge-ort: have ll_merge() use a special attr_index for renormalization
merge-ort: add a special minimal index just for renormalization
merge-ort: use STABLE_QSORT instead of QSORT where required
In preparation for switching from merge-recursive to merge-ort as the
default strategy, have the testsuite default to running with merge-ort.
Keep coverage of the recursive backend by having the linux-gcc job run
with it.
Signed-off-by: Elijah Newren <newren@gmail.com>
Reviewed-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a new option that unconditionally enables the pack.writeReverseIndex
setting in order to run the whole test suite in a mode that generates
on-disk reverse indexes. Additionally, enable this mode in the second
run of tests under linux-gcc in 'ci/run-build-and-tests.sh'.
Once on-disk reverse indexes are proven out over several releases, we
can change the default value of that configuration to 'true', and drop
this patch.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In addition to the manual adjustment to let the `linux-gcc` CI job run
the test suite with `master` and then with `main`, this patch makes sure
that GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME is set in all test scripts
that currently rely on the initial branch name being `master by default.
To determine which test scripts to mark up, the first step was to
force-set the default branch name to `master` in
- all test scripts that contain the keyword `master`,
- t4211, which expects `t/t4211/history.export` with a hard-coded ref to
initialize the default branch,
- t5560 because it sources `t/t556x_common` which uses `master`,
- t8002 and t8012 because both source `t/annotate-tests.sh` which also
uses `master`)
This trick was performed by this command:
$ sed -i '/^ *\. \.\/\(test-lib\|lib-\(bash\|cvs\|git-svn\)\|gitweb-lib\)\.sh$/i\
GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME=master\
export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME\
' $(git grep -l master t/t[0-9]*.sh) \
t/t4211*.sh t/t5560*.sh t/t8002*.sh t/t8012*.sh
After that, careful, manual inspection revealed that some of the test
scripts containing the needle `master` do not actually rely on a
specific default branch name: either they mention `master` only in a
comment, or they initialize that branch specificially, or they do not
actually refer to the current default branch. Therefore, the
aforementioned modification was undone in those test scripts thusly:
$ git checkout HEAD -- \
t/t0027-auto-crlf.sh t/t0060-path-utils.sh \
t/t1011-read-tree-sparse-checkout.sh \
t/t1305-config-include.sh t/t1309-early-config.sh \
t/t1402-check-ref-format.sh t/t1450-fsck.sh \
t/t2024-checkout-dwim.sh \
t/t2106-update-index-assume-unchanged.sh \
t/t3040-subprojects-basic.sh t/t3301-notes.sh \
t/t3308-notes-merge.sh t/t3423-rebase-reword.sh \
t/t3436-rebase-more-options.sh \
t/t4015-diff-whitespace.sh t/t4257-am-interactive.sh \
t/t5323-pack-redundant.sh t/t5401-update-hooks.sh \
t/t5511-refspec.sh t/t5526-fetch-submodules.sh \
t/t5529-push-errors.sh t/t5530-upload-pack-error.sh \
t/t5548-push-porcelain.sh \
t/t5552-skipping-fetch-negotiator.sh \
t/t5572-pull-submodule.sh t/t5608-clone-2gb.sh \
t/t5614-clone-submodules-shallow.sh \
t/t7508-status.sh t/t7606-merge-custom.sh \
t/t9302-fast-import-unpack-limit.sh
We excluded one set of test scripts in these commands, though: the range
of `git p4` tests. The reason? `git p4` stores the (foreign) remote
branch in the branch called `p4/master`, which is obviously not the
default branch. Manual analysis revealed that only five of these tests
actually require a specific default branch name to pass; They were
modified thusly:
$ sed -i '/^ *\. \.\/lib-git-p4\.sh$/i\
GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME=master\
export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME\
' t/t980[0167]*.sh t/t9811*.sh
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that we have Git supporting SHA-256, we'd like to make sure that we
don't regress that state. Unfortunately, it's easy to do so, so to
help, let's add code to run one of our CI jobs with SHA-256 as the
default hash. This will help us detect any problems that may occur.
We pick the linux-clang job because it's relatively fast and the
linux-gcc job already runs the testsuite twice. We want our tests to
run as fast as possible, so we wouldn't want to add a third run to the
linux-gcc job. To make sure we properly exercise the code, let's run
the tests in the default mode (SHA-1) first and then run a second time
with SHA-256. We explicitly specify SHA-1 for the first run so that if
we change the default in the future, we make sure to test both cases.
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>
Add GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS test flag to the test setup suite
in order to toggle writing Bloom filters when running any of the git tests.
If set to true, we will compute and write Bloom filters every time a test
calls `git commit-graph write`, as if the `--changed-paths` option was
passed in.
The test suite passes when GIT_TEST_COMMIT_GRAPH and
GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS are enabled.
Helped-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Garima Singh <garima.singh@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This job runs the test suite twice, once in regular mode, and once with
a whole slew of `GIT_TEST_*` variables set.
Now that the built-in version of `git add --interactive` is
feature-complete, let's also throw `GIT_TEST_ADD_I_USE_BUILTIN` into
that fray.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
C99 'for' loop initial declaration, i.e. 'for (int i = 0; i < n; i++)',
is not allowed in Git's codebase yet, to maintain compatibility with
some older compilers.
Our Travis CI builds used to catch 'for' loop initial declarations,
because the GETTEXT_POISON job has always built Git with the default
'cc', which in Travis CI's previous default Linux image (based on
Ubuntu 14.04 Trusty) is GCC 4.8, and that GCC version errors out on
this construct (not only with DEVELOPER=1, but with our default CFLAGS
as well). Alas, that's not the case anymore, becase after 14.04's EOL
Travis CI's current default Linux image is based on Ubuntu 16.04
Xenial [1] and its default 'cc' is now GCC 5.4, which, just like all
later GCC and Clang versions, simply accepts this construct, even if
we don't explicitly specify '-std=c99'.
Ideally we would adjust our CFLAGS used with DEVELOPER=1 to catch this
undesired construct already when contributors build Git on their own
machines. Unfortunately, however, there seems to be no compiler
option that would catch only this particular construct without choking
on many other things, e.g. while a later compiler with '-std=c90'
and/or '-ansi' does catch this construct, it can't build Git because
of several screenfulls of other errors.
Add the 'linux-gcc-4.8' job to Travis CI, in order to build Git with
GCC 4.8, and thus to timely catch any 'for' loop initial declarations.
To catch those it's sufficient to only build Git with GCC 4.8, so
don't run the test suite in this job, because 'make test' takes rather
long [2], and it's already run five times in other jobs, so we
wouldn't get our time's worth.
[1] The Azure Pipelines builds have been using Ubuntu 16.04 images
from the start, so I belive they never caught 'for' loop initial
declarations.
[2] On Travis CI 'make test' alone would take about 9 minutes in this
new job (without running httpd, Subversion, and P4 tests). For
comparison, starting the job and building Git with GCC 4.8 takes
only about 2 minutes.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Prepare to run test suite on Azure Pipeline.
* js/vsts-ci: (22 commits)
test-date: drop unused parameter to getnanos()
ci: parallelize testing on Windows
ci: speed up Windows phase
tests: optionally skip bin-wrappers/
t0061: workaround issues with --with-dashes and RUNTIME_PREFIX
tests: add t/helper/ to the PATH with --with-dashes
mingw: try to work around issues with the test cleanup
tests: include detailed trace logs with --write-junit-xml upon failure
tests: avoid calling Perl just to determine file sizes
README: add a build badge (status of the Azure Pipelines build)
mingw: be more generous when wrapping up the setitimer() emulation
ci: use git-sdk-64-minimal build artifact
ci: add a Windows job to the Azure Pipelines definition
Add a build definition for Azure DevOps
ci/lib.sh: add support for Azure Pipelines
tests: optionally write results as JUnit-style .xml
test-date: add a subcommand to measure times in shell scripts
ci: use a junction on Windows instead of a symlink
ci: inherit --jobs via MAKEFLAGS in run-build-and-tests
ci/lib.sh: encapsulate Travis-specific things
...
Symbolic links are still not quite as easy to use on Windows as on Linux
(for example, on versions older than Windows 10, only administrators can
create symlinks, and on Windows 10 you still need to be in developer
mode for regular users to have permission), but NTFS junctions can give
us a way out.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let's not decide in the generic ci/ part how many jobs to run in
parallel; different CI configurations would favor a different number of
parallel jobs, and it is easy enough to hand that information down via
the `MAKEFLAGS` variable.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The name is hard-coded to reflect that we use Travis CI for continuous
testing.
In the next commits, we will extend this to be able use Azure DevOps,
too.
So let's adjust the name to make it more generic.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
All Travis CI build jobs run the test suite with 'make --quiet test'.
On one hand, being quiet doesn't save us from much clutter in the
output:
$ make test |wc -l
861
$ make --quiet test |wc -l
848
It only spares 13 lines, mostly the output of entering the 't/'
directory and the pre- and post-cleanup commands, which is negligible
compared to the ~700 lines printed while building Git and the ~850
lines of 'prove' output.
On the other hand, it's asking for trouble. In our CI build scripts
we build Git and run the test suite in two separate 'make'
invocations. In a prelimiary version of one of the later patches in
this series, to explicitly specify which compiler to use, I changed
them to basically run:
make CC=$CC
make --quiet test
naively thinking that it should Just Work... but then that 'make
--quiet test' got all clever on me, noticed the changed build flags,
and then proceeded to rebuild everything with the default 'cc'. And
because of that '--quiet' option, it did so, well, quietly, only
saying "* new build flags", and it was by mere luck that I happened to
notice that something is amiss.
Let's just drop that '--quiet' option when running the test suite in
all build scripts.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The commit-graph and multi-pack-index features introduce optional
data structures that are not required for normal Git operations.
It is important to run the normal test suite without them enabled,
but it is helpful to also run the test suite using them.
Our continuous integration scripts include a second test stage that
runs with optional GIT_TEST_* variables enabled. Add the following
two variables to that stage:
GIT_TEST_COMMIT_GRAPH
GIT_TEST_MULTI_PACK_INDEX
This will slow down the operation, as we build a commit-graph file
after every 'git commit' operation and build a multi-pack-index
during every 'git repack' operation. However, it is important that
future changes are compatible with these features.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let's start with some background about oe_delta_size() and
oe_set_delta_size(). If you already know, skip the next paragraph.
These two are added in 0aca34e826 (pack-objects: shrink delta_size
field in struct object_entry - 2018-04-14) to help reduce 'struct
object_entry' size. The delta size field in this struct is reduced to
only contain max 1MB. So if any new delta is produced and larger than
1MB, it's dropped because we can't really save such a large size
anywhere. Fallback is provided in case existing packfiles already have
large deltas, then we can retrieve it from the pack.
While this should help small machines repacking large repos without
large deltas (i.e. less memory pressure), dropping large deltas during
the delta selection process could end up with worse pack files. And if
existing packfiles already have >1MB delta and pack-objects is
instructed to not reuse deltas, all of them will be dropped on the
floor, and the resulting pack would be definitely bigger.
There is also a regression in terms of CPU/IO if we have large on-disk
deltas because fallback code needs to parse the pack every time the
delta size is needed and just access to the mmap'd pack data is enough
for extra page faults when memory is under pressure.
Both of these issues were reported on the mailing list. Here's some
numbers for comparison.
Version Pack (MB) MaxRSS(kB) Time (s)
------- --------- ---------- --------
2.17.0 5498 43513628 2494.85
2.18.0 10531 40449596 4168.94
This patch provides a better fallback that is
- cheaper in terms of cpu and io because we won't have to read
existing pack files as much
- better in terms of pack size because the pack heuristics is back to
2.17.0 time, we do not drop large deltas at all
If we encounter any delta (on-disk or created during try_delta phase)
that is larger than the 1MB limit, we stop using delta_size_ field for
this because it can't contain such size anyway. A new array of delta
size is dynamically allocated and can hold all the deltas that 2.17.0
can. This array only contains delta sizes that delta_size_ can't
contain.
With this, we do not have to drop deltas in try_delta() anymore. Of
course the downside is we use slightly more memory, even compared to
2.17.0. But since this is considered an uncommon case, a bit more
memory consumption should not be a problem.
Delta size limit is also raised from 1MB to 16MB to better cover
common case and avoid that extra memory consumption (99.999% deltas in
this reported repo are under 12MB; Jeff noted binary artifacts topped
out at about 3MB in some other private repos). Other fields are
shuffled around to keep this struct packed tight. We don't use more
memory in common case even with this limit update.
A note about thread synchronization. Since this code can be run in
parallel during delta searching phase, we need a mutex. The realloc
part in packlist_alloc() is not protected because it only happens
during the object counting phase, which is always single-threaded.
Access to e->delta_size_ (and by extension
pack->delta_size[e - pack->objects]) is unprotected as before, the
thread scheduler in pack-objects must make sure "e" is never updated
by two different threads.
The area under the new lock is as small as possible, avoiding locking
at all in common case, since lock contention with high thread count
could be expensive (most blobs are small enough that delta compute
time is short and we end up taking the lock very often). The previous
attempt to always hold a lock in oe_delta_size() and
oe_set_delta_size() increases execution time by 33% when repacking
linux.git with with 40 threads.
Reported-by: Elijah Newren <newren@gmail.com>
Helped-by: Elijah Newren <newren@gmail.com>
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some recent optimizations have been added to pack-objects to reduce
memory usage and some code paths are split into two: one for common
use cases and one for rare ones. Make sure the rare cases are tested
with Travis since it requires manual test configuration that is
unlikely to be done by developers.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While at there, document about this special mode when running the test
suite.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Build the executable in 'script' phase in Travis CI integration, to
follow the established practice, rather than during 'before_script'
phase. This allows the CI categorize the failures better ('failed'
is project's fault, 'errored' is build environment's).
* sg/travis-build-during-script-phase:
travis-ci: build Git during the 'script' phase
Ever since we started building and testing Git on Travis CI (522354d70
(Add Travis CI support, 2015-11-27)), we build Git in the
'before_script' phase and run the test suite in the 'script' phase
(except in the later introduced 32 bit Linux and Windows build jobs,
where we build in the 'script' phase').
Contrarily, the Travis CI practice is to build and test in the
'script' phase; indeed Travis CI's default build command for the
'script' phase of C/C++ projects is:
./configure && make && make test
The reason why Travis CI does it this way and why it's a better
approach than ours lies in how unsuccessful build jobs are
categorized. After something went wrong in a build job, its state can
be:
- 'failed', if a command in the 'script' phase returned an error.
This is indicated by a red 'X' on the Travis CI web interface.
- 'errored', if a command in the 'before_install', 'install', or
'before_script' phase returned an error, or the build job exceeded
the time limit. This is shown as a red '!' on the web interface.
This makes it easier, both for humans looking at the Travis CI web
interface and for automated tools querying the Travis CI API, to
decide when an unsuccessful build is our responsibility requiring
human attention, i.e. when a build job 'failed' because of a compiler
error or a test failure, and when it's caused by something beyond our
control and might be fixed by restarting the build job, e.g. when a
build job 'errored' because a dependency couldn't be installed due to
a temporary network error or because the OSX build job exceeded its
time limit.
The drawback of building Git in the 'before_script' phase is that one
has to check the trace log of all 'errored' build jobs, too, to see
what caused the error, as it might have been caused by a compiler
error. This requires additional clicks and page loads on the web
interface and additional complexity and API requests in automated
tools.
Therefore, move building Git from the 'before_script' phase to the
'script' phase, updating the script's name accordingly as well.
'ci/run-builds.sh' now becomes basically empty, remove it. Several of
our build job configurations override our default 'before_script' to
do nothing; with this change our default 'before_script' won't do
anything, either, so remove those overriding directives as well.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>