$FDO_CI_CONCURRENT is provided by in the CI environment by the fd.o
GitLab runners to tell us how many parallel processes would be 'good' to
use.
Use this to override the default Ninja invocation which uses as many
CPUs as available, and instead tell it to use as many parallel processes
as the runner thinks we should during the build process.
Tests are invoked using `meson test` inside a virtme/QEmu VM; whilst
Meson's test backend will use as many processors as availble, virtme
will by default create a single-CPU VM. So if we create a VM with as
many CPUs as we should have parallel processes, we can let it use all of
them. This also requires quadrupling the requested RAM so ASan doesn't
force us straight into OOM.
Suggested by @daenzer.
Signed-off-by: Daniel Stone <daniels@collabora.com>
The old job names were really long, so long that Gitlab web UI had to
almost always ellipsize them, showing only the beginning which is the
same fof the two and hiding the part that actually matters.
Rename the templates and the jobs to be shorter and more descriptive.
Meson is the only tool we build with. All builds are native. Default
options is a long way to say full-featured and besides libseat was not
even default yet we build it.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
I'm tired of scrolling through that flood when looking at CI logs. It's
just listing every gcno, gcda and c file name as it parses them.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Change BUILDDIR and PREFIX from exported shell variables into CI
variables. This way they can be used in CI configuration, e.g.
artifacts.
The main reason for this change is that it makes it possible to use
these variables in after_script, which is not possible with shell
variables as the values do not carry over there.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Use address sanitizer to catch use-after-free and other errors when
running the test suite.
Leak detection is disabled, because currently there are too many leaks,
making almost all tests fail otherwise.
The atexit=1 is for verifying that ASan was actually used.
The default 128 MB of RAM in the qemu machine leads to oom-killer
killing most tests, so bump the memory size to 1 GB.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Predictably, there was one thing I forgot with !600, which was to fix up
the CI rules for the transition. Oops.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Most of our stages take just a single minute to complete, while the standard
timeout on gitlab CI is 60 minutes.
Set a 5 minute timeout on quick stages, and a 30 minute timeout on the image
build step to ensure we fail fast and don't tie up CI resources.
Signed-off-by: Kenny Levinsen <kl@kl.wtf>
The standard style of LCOV HTML reports is a bit harsh to look at. This
commit replaces it with a new one.
The new CSS was written from scratch by looking at the HTML source code
of a generated LCOV report. The original gcov.css file was not used.
The color scheme is neutral, trying to avoid a Christmas tree effect.
The colors are intended to be calm while also distinguishable, and not
hamper text readability.
The font lists were taken from Gitlab with the hope that it will blend
in a little better when viewing from MR artifacts.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This installs libseat in the debian image build from source in order to enable,
build and test weston with libseat support.
Signed-off-by: Kenny Levinsen <kl@kl.wtf>
In CI we should never get a skip, so turn them into failures to make
sure we notice.
This is enabled only for the configuration where we build everything. If
anything is disabled, skips are expected.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This should not be necessary, but my testing with fd.o Gitlab 13.9.1
shows this is needed. Otherwise the coverage markings will not appear in
a MR diff view.
Apparently Gitlab has some problem with 'filename' attribute containing
"../" in Cobertura XML files, even when that does result in a correct
path. Or maybe the problem is is with the <source> path referring to the
build dir which from Gitlab perspective does not exist in the project,
even though builddir/../ is a good path.
This sed hack removes the "../" part and the last element in the
<source> path correspondingly.
See https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/567
for my testing.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This runs the coverage tools to produce HTML pages listing the code lines /
functions / branches hit/totalled by the test suite.
Nowadays Gitlab has some Cobertura support itself:
https://docs.gitlab.com/ee/user/project/merge_requests/test_coverage_visualization.html
lcov is needed for the HTML report, gcovr is needed for the Cobertura
report. 'ninja clean' must be removed, otherwise it deletes the coverage
files before they are analysed.
Seeing the test suite code coverage is really interesting. It can guide
designing tests. If Gitlab MRs show the coverage in diff view, it shows
if new code actually gets executed in CI.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This support is added so that the XYUV shader variant can be tested with
wl_shm from the test suite.
Libwayland version requirement is bumped to get WL_SHM_FORMAT_XYUV8888.
Libwayland is bumped to 1.18 too in the CI image. libwayland-dev package
is dropped, because we build wayland anyway.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
The version we used does not advertise GL_EXT_color_buffer_half_float
with llvmpipe even though the functionality seems to work.
If the extension is not advertised, the future commit
"tests: extend output-damage to GL shadow framebuffer"
will result in a test failure.
Upgrade Mesa, this gets the extension advertised and the test is happy.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This allows Gitlab to show a detailed test report in a human friendly
manner on a merge request page.
The junit output depends on Meson 0.55, but Meson in CI is bumped to the
latest release on Feb 15th. It is beneficial to use the newest
possible Meson in CI even if we do not require it, so that we benefit
from fixes, e.g. new warnings about accidentally using more recent
Meson features than our minimum Meson version supports.
Meson 0.57 is required for proper test names in the Gitlab report after
switching the tests to TAP. See:
https://github.com/mesonbuild/meson/issues/8316
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This helps identify if we have actually have a build issue or if we have
a documentation build failure. Notifications sent out are specifying the
name of the build, making things much easier to figure out what actually
failed.
Uses another job w/o the need to run the tests (again).
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
If a screenshooting test fails, it quite likely writes not only the screenshot
but also a diff image highlighting the failed pixels. These would be good to
have in the CI artifacts for postmortem.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This adds a link in any MR to the documentation from the CI build, making it
easy to preview the documentation.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
The stable FreeRDP 2.x branch has been released, so let's rely on that maintained
version and drop all the hacks for older versions. That makes the code and build
cleaner.
Signed-off-by: David Fort <contact@hardening-consulting.com>
The pipewire plugin uses this API as well, not just the remoting plugin. So
enable it if either is enabled.
And disable pipewire in the no-gl-renderer CI build. The virtual outputs don't
work without it.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
In order to run DRM-backend tests, a DRM-device is needed. As we
do not necessarily have control of the hardware that is going to
run our tests in GitLab CI, DRM-backend tests were being skipped.
This patch add support to run the tests using VKMS (virtual KMS).
To achieve this, virtualization is needed, as we need to run a
custom kernel during the CI job. We've decided to go with virtme,
as it is simpler to setup and works good for our use case.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
Make use of the templating structure the templates provide. No
functional changes in the end, container-build's default behavior is the
previously called container-if-not-exists template.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The project was moved a while ago to make it look less waylandy. Same
sha, so no actual changes here.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The reason the Xwayland test skips is that Xwayland was not installed. We
should exercise Xwayland as well in CI, so install it.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
wayland/ci-templates!18 allows the $HOME environment variable to
propagate into the execution of creating our build container. Without
that, buildah would leave us with an empty $HOME, which would have pip
install into /.local/bin.
As $HOME is preserved during our actual native builds, we would try to
find meson in /root/.local/bin and fail, since it had been stored to a
different path.
Bump the ci-templates dependency to one with the fix so we can build new
images again.
Signed-off-by: Daniel Stone <daniels@collabora.com>
This client contains driver-specific code to allocate buffers. However clients
shouldn't contain driver-specific code and should rely on e.g. mesa to allocate
buffers via standard interfaces.
Additionally, because the build system always tries to enable all features, some
experimental drivers and drivers that aren't included in amd64 distribution
packages were required. Users would need to manually disable some drivers.
Releasers would need to install libdrm from source (because the release process
forbids adding custom build flags). Dropping simple-dmabuf-drm simplifies both
building and releasing.
The functionality previously tested via simple-dmabuf-drm can now be tested with
simple-dmabuf-egl.
Signed-off-by: Simon Ser <contact@emersion.fr>
The next commit bumps the required Wayland version beyond what the repo
version supports.
Build wayland 1.17 into the docker image.
Signed-off-by: Sebastian Wick <sebastian@sebastianwick.net>
This hopefully fixes a failure introduced in
5a0706b238987bf3dcf7dde54556850ee88ccdba:
cp: cannot stat '/builds/wayland/weston/prefix-weston-build-native-meson/share/doc/weston/*': No such file or directory
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Building without gl-renderer seems to fail rather often. Introduce a
specific build job which tests this configuration.
Make use of the Folded Block Scalar in YAML to concatenate the build
option string with spaces. This will make diffing build option
changes very readable.
Signed-off-by: Stefan Agner <stefan@agner.ch>
Fixes 0a13641ff: "gitlab-ci: Enable gitlab page for publishing documentation",
to point to correct sphinx documentation.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
While adding pages, remove CI_JOB_ID and CI_JOB_SHA from PREFIX, as this
apparently is not needed. Make build-native-meson a dependency for the
pages stage.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Weston 6.0.0 was released with both autotools and Meson build systems. That
should be enough for downstream to migrate to Meson build on their on pace.
Maintaining two build systems is a hassle, keep the one that is easier to work
with and let the other one go.
doc/dozygen/tool*.doxygen.in are not deleted, because they have not been
integrated with Meson yet.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Reuse the templates from wayland/ci-templates:
whenever DEBIAN_TAG is changed, this will rebuild a new container.
This adds two things:
- better reliability (we do not randomly pull packages whenever the CI
runs and we can reproduce with this particular environment)
- faster builds, as we do not need to pull the universe at each run
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
[Pekka: bump DEBIAN_TAG]
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Autotools is going away. Break the autotools build so that people are
guaranteed to notice before it is gone. If they have problems with Meson, they
can still use --enable-autotools to build with autotools, but we really want to
hear about any problems.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This is so that, for instance, people using weston as their main Wayland
compositor can invert the sense of two finger scrolling or change
pointer acceleration using weston.ini, rather than having to edit C
code.
All of the options that libinput itself exposes through its API are now
exposed in weston.ini. The new options are called `tap-and-drag`,
`tap-and-drag-lock`, `disable-while-typing`, `middle-emulation`,
`left-handed`, `rotation`, `accel-profile`, `accel-speed`,
`scroll-method`, `natural-scroll`, and `scroll-button`. I have
successfully tested everything except for `rotation`, out of a lack of
hardware support.
weston now depends directly on libevdev for turning button name strings into
kernel input codes. This was needed for the `scroll-button` config
option. (weston already depends indirectly on libevdev through
libinput, so I figured people would be OK with this.) As a practical
matter for debian-style packagers, weston now has a build dependency on
libevdev-dev.
Right now, the code applies the same options to all attached devices
that a given option is relevant for. There are plans for multiple
[libinput] sections, each with different device filters, for users who
need more control here.
Signed-off-by: Eric Toombs <3672-ewtoombs@users.noreply.gitlab.freedesktop.org>
Correct the path to the build directory so we can capture Meson logs;
especially useful when tests fail like in #184.
An example of this change having been run with a deliberately-failing
test, capturing the Meson logs, can be found at:
https://gitlab.freedesktop.org/daniels/weston/-/jobs/94623
Signed-off-by: Daniel Stone <daniels@collabora.com>
Otherwise CI might fail due to
https://github.com/mesonbuild/meson/issues/4718 but the fix isn't
included in any release yet, so install meson from the 0.49 branch.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Daniel Stone
Pekka Paalanen
Kenny Levinsen
Marius Vlad
David Fort
James Hilliard
Michael Olbrich
Leandro Ribeiro
Peter Hutterer
Simon Ser
Sebastian Wick
Stefan Agner
Harish Krupo
Benjamin Tissoires
Eric Toombs
Jonas Ådahl
Pekka Paalanen