This commit fixes a bug that eats one tap event after the clickpad button
has been pressed and released with no fingers lifted, turned into palm,
or palms lifted, e.g. by pressing it with a non-finger object or with an
otherwise ignored palm that stays down.
After one tap is eaten, the state machine would get un-stuck from its DEAD
state and end up in IDLE as it is supposed to, permitting subsequent taps to
have their intended effect. However, that would even happen when the
clickpad button was still held down, which is also wrong.
The correct solution is to exit to IDLE when both the button and all touches
are up. This is achieved by splitting the DEAD tapping state into one state
for the button down (BUTTON) and one for the button up (DEAD).
Working around this (in the case of a button pressed by a palm lifted
afterwards) was the only remaining justification for the palm up
event, which this commit also removes.
Signed-off-by: satrmb <10471-satrmb@users.noreply.gitlab.freedesktop.org>
This commit is intended to hopefully make reviewing the following commit
a little easier, because it leads to a much more useful diff.
No executed code was actually changed.
Reordering details: moved tp_tap_tapped_handle_event and all
tp_tap_dragging*_handle_event functions up to just above
tp_tapped_idle_handle_event.
They are also renamed from tp_tap_* to tp_drag_*, and not-yet-used functions
named tp_drag_idle_handle_event, tp_drag_dead_handle_event, and
tp_drag_handle_event are added with dummy function bodies.
Another set of renames affects the functions and states connected to
drag-lock, which are renamed to reflect this association more clearly:
- {1|2|3}FGTAP_DRAGGING_WAIT to {1|2|3}FGTAP_DRAGLOCK_WAIT,
- {1|2|3}FGTAP_DRAGGING_OR_TAP to {1|2|3}FGTAP_DRAGLOCK_CONTINUE,
- tp_tap_dragging_wait_handle_event to tp_drag_draglock_wait_handle_event,
- tp_tap_dragging_tap_handle_event to tp_drag_draglock_continue_handle_event.
Signed-off-by: satrmb <10471-satrmb@users.noreply.gitlab.freedesktop.org>
Some Lenovo ThinkBook 14 G7+ ASP models ship with pressure pads (nominally
"Force Pad"). However, they do not appear to be declared as such by the
firmware.
Add a quirk to make them work.
Signed-off-by: Mingcong Bai <jeffbai@aosc.io>
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1079>
If a tablet has an area configured and the pen goes into proximity
outside this area, ignore all events from this sequence. This truly
deactivates that area so it can even be used for e.g. placing a pen
there.
For simplicity, a sequence that starts outside the configured area will
be completely ignored, i.e. moving into the tablet area will not trigger
any fake proximity events as we cross into the allowed area. This
requires quite a bit of effort and it's unclear if it's really needed by
users - we can reconsider when we get complaints.
We do however accept a proximity event within within 3% of the
configured area. This gives us 6mm on a 200mm tablet where we can move
in from the area and still have events work, i.e. some error margin for
where a user needs both an area and work closes to the edge of that
area.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1013>
For external tablets like the Intuos series we now expose the area
rectangle configuration and the (minimum) implementation required to
make this work.
Because an area configuration may apply late and tablet events usually
get scaled by the compositor we need to store the current axis extents
in each event. This is to behave correctly in this events sequence:
1. tool proximity in
2. caller changes config, config is pending
3. tool moves, generates events
4. tool goes out of prox, new config applies
5. caller processes motion events from step 3
If the caller in step five uses any of the get_x_transformed calls these
need to be scaled relative to the original area, not the one set in
step 2.
The current implementation merely clips into the area so moving a stylus
outside the area will be equivalent to moving it along the respective
edge of the area. It's not a true dead zone yet.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1013>
This adds the configuration option to define a rectangle that serves as
an input area on external tablets such as an Intuos.
The intention behind this is to make this input area behave as if it was
the only physical input area on this tablet with libinput emulating
proximity events as required for where the tools moves in and out
of this area.
This could also be achieved with the existing calibration setting but
area configuration is not calibration and we don't want to expose other
side-effects of the matrix (e.g. scaling and rotation) for these
devices.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1013>
device->abs.absinfo_x/y points to the x/y axis we want to use and
that axis is used in all the evdev helper function. For consistency
use that one here too.
This is for consistency only and has no effect on tablet devices, the
only time this isn't ABS_X/Y is on multitouch devices where that points to
ABS_MT_POSITION_X/Y.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1013>
The vast majority of devices that libwacom doesn't know about are the
various built-in ones. Since the only effect in our code here is that we
enable the calibration matrix, let's default to built-in if we don't
know any better - better to have the matrix and not use it than to not
be able to calibrate a tablet.
Note that libwacom 2.11 and later also now default to a built-in tablet.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1074>
This test does a prox in/out and immediately drains events.
Where we are slow enough we may drain only one (or none) of the
proximity events which causes a failure later in the test.
Similar issue with the second test where we may get a delayed tip event.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1074>
Instead of re-creating the the libwacom device from the database every
time we need it let's create it once during tablet|pad_init and pass it
down to the functions.
This allows us to have one point per tablet/pad where we can log an
error if the device is not supported by libwacom - previously this was
printed during the left-handed setup.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1073>
This was historically based on the LEDs on the device: we'd loop through
the LEDs and assign them to pad mode groups, then figure out which
button is the mode toggle for that group and finally which buttons
are in the same position as that toggle button.
Devices like the XP Pen ACK05 Remote don't have LEDs though but they do
have a mode toggle button [1] inside the dial. Let's support those by
switching the initialization on its head: search for mode toggle
buttons, create a mode group per toggle button, then associate
LEDs with that group.
The outcome should be functionally the same for devices with LEDs but
allows us to create mode groups where no LEDs exist.
Closes#1045
[1] As per Windows default button behavior
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1070>
Since our implementation uses the sysfs LED files and we don't have
those, all our test devices have the single fallback mode and no
more. Add a comment to all these tests to make that clear and enforce a
single mode only so it's more obvious.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1070>
This replaces check. The code is a copy of pwtest which I wrote years
ago for pipewire but adjusted for us here the last few days.
There are a few advantages over check:
- Ability to SKIP tests or mark them as NOT_APPLICABLE, the latter
of which is used for early checks if a device doesn't meet
requirements.
- it captures stdout/stderr separately
- colors!
- YAML output format makes it a lot easier to read the results and
eventually parse them for e.g. "restart failed tests"
Less abstraction: we set up the tests, pass them to the runner and run
them with the given number of forks. This is an improvement over before
where we forked into N test suites which each called check which then
forked again. Since we're now keeping track of those processes
ourselves we can also write tests that are expected to fail with
signals.
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1067>
