If a single log message is composed of multiple calls (as are all from
evdev_log_*), don't prefix multiple times.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Taking the tablet events as-is produces the occasional wobble in what should
be a straight line. Bug 99961 has a jpg attachment to illustrate that.
Emulate the wacom driver behavior and average x/y across the last 4 values to
smoothen out these dents.
https://bugs.freedesktop.org/show_bug.cgi?id=99961
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
We don't have frame events for tablets so we must take care to send the
axis change notifications only once and leave the others as-is. Most of the
axes are absolute so it doesn't really matter, but we need to reset the delta
to make sure clients don't receive the same delta twice.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
The tablet axis struct has a delta field that's only useful for the events,
not for our internal axis handling. Make sure we never set it to anything
nonzero.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
Handle the delta in the end once we've updated the device state for all axes.
This requires us to use the device history rather than the current state
delta, and it also requires us to update both x and y whenever an axis change
comes in.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
This is a bit hard to follow:
- tilt is handled first and if either tilt axis is set we fetch *both* tilt axes
into tablet->axes.tilt
- rotation is handled second but it only triggers if either tilt axis is
flagged. as we now guarantee to have both axes in tablet->axes.tilt, we
can continue with the rotation conversion without needing some other state
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
This is prep work for axis smoothing. Modify the various helper functions to
just update the state in the tablet and then grab the state later for better
grouping.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
No functional changes, part of the grouping of tablet axis manipulation vs.
processing of that manipulated state.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
No functional changes, this is just to group the calls that modify tablet axis
state together and move the bits that rely on this state (but don't modify it)
to the bottom.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
Stores the processed axes values in a history 4 events deep. Currently unused
but will be used to smoothen out axis values to avoid transducer-caused axis
wobbles.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
The current code modifies a bit of state inside the proximity_tip_down
function which makes for confusing reading. Clean this up by having a bunch of
helper functions for the various events.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
This cannot trigger because we'd never get here if out-of-proximity is set,
tablet_flush() will return early.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
The only code path that leads here would see the changed_axes array zeroed out
in tablet_send_axis_proximity_tip_down_events(), zeroing again is unlikely to
make it more zero.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Ping Cheng <pingc@wacom.com>
Print the axis values on proximity out because it a) ensures we have the right
values and b) makes the output better aligned with the proximity in, so it's
easier to spot in a log file. But don't print the tool capabilities because
they're unrelated to the prox out anyway and again it makes the output easier
to spot.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
No effect since we don't care about the mouse itself. But when running
on kernels without uinput's UI_GET_SYSNAME this can cause misdetection of
the uinput device and test case failures. Simply picking a differently named
device avoids that.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If not set, libtool will search directories up to ../.. for an install-sh and
then dump the aux files there. This caused a couple of problems with the xorg
release.sh script that now uses worktrees but is generally bad behaviour
because we can't guarantee that we're not inside some other repository.
Set AC_CONFIG_AUX_DIR to avoid this behavior.
See https://lists.freedesktop.org/archives/xorg-devel/2017-March/053006.html
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
If a touch started hovering in the main area, the button state would start
with AREA and never move to the real button state, despite the finger
triggering the pressure thresholds correctly in one of the areas.
This could even happen across touch sequences if a touch went below pressure
in the software button area, it changed to hovering and the button state
changed to NONE. On the next event, the touch is still hovering and the
current position of the touch is taken for the button state machine.
https://bugs.freedesktop.org/show_bug.cgi?id=99976
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Introduced in commit 8e7f99c27a we only allowed horizontal edge scrolling
on devices larger than 50mm to leave enough reactive space on the touchpad.
Looking at a ruler, a 50mm high touchpad is still large enough to leave the
bottom 7mm as an horizontal edge scroll area. Reduce the minimum size to 40mm
instead, that's closer to where it starts to get a bit iffy.
https://bugzilla.redhat.com/show_bug.cgi?id=1422221
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
We don't cater for the special case of groups having a different number of
modes, there is no hardware right now that does that.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Supresses any printf statements from the tool itself, i.e. it skips printing
any of the events.
Makes it easier to debug the internal state since it's not intermixed with a
whole lot of messages about the events that are generated. Best combined with
--verbose (yes, hilarious, isn't it...)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Prefix device log messages with the device's sysname so it's more obvious
where the messages are coming from. This makes it much easier to grep for a
specific device's messages but also adds some identifier to messages that
were previously without any identifier (e.g. all the state machine debugging)
All info and error messages also automatically prefix the device name, so
those messages are standardised too, e.g
an info message now:
event4 - SynPS/2 Synaptics TouchPad: is tagged by udev as: Touchpad
a debug message now:
event4 - using pressure-based touch detection
And since this required changing a lot of the strings in messages anyway,
polish a few minor things too.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Preparation work for standardizing log messages better
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Easier to track if we leave the original files alone. Actual changes to the
doxygen style:
* indent <dd> blocks
* hide the navigation sub-items. Our current style expands the full navigation
menu but File-list, etc. is mostly useless and just wastes space.
* force some space below the main bar
* change the header sizes around a bit. Primary goal here: making <h1> smaller
than the title
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
