This will only catch a segfault or some other bug since we don't actually look
at the output. But that's still better than not running it at all.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
My kingdome for a compiler warning. Or a scan-build warning. Or a coverity
warning. Or anything... But no, nothing.
Also make the open_restricted() more robust to a NULL userdata, because
effectively that's what we were passing here.
Fixes https://gitlab.freedesktop.org/libinput/libinput/issues/50
Introduced in 0a13223c39
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
According to multiple sources, referenced in
https://engineering.facile.it/blog/eng/continuous-deployment-from-gitlab-ci-to-k8s-using-docker-in-docker/
The garbage collector of the registry won't clean up docker images that
still have blob references. We should clean up the manifests instead
of simply overwriting the tag.
Note: this requires to set up a personal token with api access from the
maintainers in the form of (for instance): "PERSONAL_TOKEN_bentiss"
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
There is no point in login in to the registry if there is no need to
create a new docker image.
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Leftover from a previous iteration of this code - having a static but
unnecessarily large size for dirname results in:
../test/litest.c:1251:38: warning: ‘snprintf’ output may be truncated before the last format character [-Wformat-truncation=]
snprintf(dest, sizeof(dest), "%s/%s", dirname, filename);
^
../test/litest.c:1251:3: note: ‘snprintf’ output 2 or more bytes (assuming 4097) into a destination of size 4096
snprintf(dest, sizeof(dest), "%s/%s", dirname, filename);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The values are the same and this is ABI so they will never change. Make the
cast explicit for coverity's benefit.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This has been replaced with the GitLab CI in the repository proper. Circle CI
was ony ever run on a private github repo.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
These (probably) all disable the mechanical keyboard anyway, so let's keep it
enabled to be able to access the screen keys, if any.
https://gitlab.freedesktop.org/libinput/libinput/issues/39
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The main tasks it does is build on a few different distros as well as build
with the various build options to make sure they work.It doesn't (yet) run the
test suite runner because that one mostly requires device nodes to operate on.
Most of the fancy is to get the docker images ready. A dnf update takes
forever, so we don't want to do that on 10 different machines. So instead we
build docker images with all the bits pre-installed, push that to the registry
and use those images for testing.
To speed things up, we only do that when the current image is older than a
week. And we only do that when we push to libinput proper, so a merge request
or pushing to your private gitlab repo will never trigger a docker image
update - it will trigger the tests and use the docker images tough.
Co-authored-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Ignore motion when pressure/touch size fell below the threshold, thus
ending the touch.
Real world significance: subjectively scrolling/cursor positioning with
a touchpad now a bit better on SAMSUNG NP305V5A laptop.
https://gitlab.freedesktop.org/libinput/libinput/merge_requests/4
Signed-off-by: Konstantin Kharlamov <Hi-Angel@yandex.ru>
Make the meson -Dtests=false option only apply to the libinput test suite
itself which has extra dependencies, etc. The build tests and symbol leak
tests should always run.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The &grab pointer we used to pass as userdata was the address of the function
argument which goes out of scope at the end of the function. This works fine
for devices immediately opened but when a device connects later, the address
may have been re-used since and it's content is undefined. If not NULL, we
end up grabbing the device.
Instead pass the grab option in which is guaranteed to live until the end of
main.
https://gitlab.freedesktop.org/libinput/libinput/issues/26
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Nothing in libinput needs large buffers, so if we ever get something that
large, we probably passed a negative number to zalloc.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Make it return NULL for a string array in the form of [ NULL ], like the docs
say. This also adds an extra safety check for the joiner to be of a reasonable
length to avoid overflows.
Found in
https://gitlab.freedesktop.org/libinput/libinput/issues/26#note_6320
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
All the tests fill fail anyway if the validation fails but this is a quick way
to fail everything early.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
A bit quirky (haha), because we cannot do this during context creation - we
really want any parsing error messages to show up in the right log file and
the log handler isn't set up during context creation. So we do it on the first
real call to the backend - path_add_device or udev_assign_seat.
Also, failure to initialize the quirks subsystem just means we continue as
normal. This shouldn't be a hard failure, it just means a lot of devices won't
work properly.
If the LIBINPUT_DATA_DIR environment variable is set, that directory is used
for the data file. Only that directory, no custom override file in that case.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
