platform/upstream/libinput.git
5 years agodoc/user: correct some small typos
Matt Mayfield [Sun, 19 Aug 2018 16:54:52 +0000 (11:54 -0500)]
doc/user: correct some small typos

5 years agodoc/user: move the trackpoint multiplier debugging to Troubleshooting
Peter Hutterer [Thu, 16 Aug 2018 02:39:15 +0000 (12:39 +1000)]
doc/user: move the trackpoint multiplier debugging to Troubleshooting

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
5 years agodoc/user: fix two links for sphinx syntax
Peter Hutterer [Thu, 16 Aug 2018 02:26:27 +0000 (12:26 +1000)]
doc/user: fix two links for sphinx syntax

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
5 years agolibinput 1.11.902 1.11.902
Peter Hutterer [Mon, 13 Aug 2018 23:01:46 +0000 (09:01 +1000)]
libinput 1.11.902

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
5 years agotest: when testing for a touch sequence, assert it's the same slot's sequence
Peter Hutterer [Mon, 17 Jul 2017 04:26:34 +0000 (14:26 +1000)]
test: when testing for a touch sequence, assert it's the same slot's sequence

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
5 years agotouchpad: drop check for left button areas
Peter Hutterer [Mon, 13 Aug 2018 04:23:28 +0000 (14:23 +1000)]
touchpad: drop check for left button areas

No functional changes, anything that's in the top/bottom area but not in the
respective middle/right area is a left button.

Introduced by 13bda5adcb2f81f8ed9de762910ac9ecfa3b250b

Fixes coverity complaint about use of uninitialized variable.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
5 years agotouchpad: change the min vector for the scroll lock to 0.15
Peter Hutterer [Mon, 13 Aug 2018 04:00:54 +0000 (14:00 +1000)]
touchpad: change the min vector for the scroll lock to 0.15

This makes the difference between noticable delay and unnoticable while having
virtually no false positives (for me).

https://gitlab.freedesktop.org/libinput/libinput/issues/101

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
5 years agotouchpad: if a finger in the button area moves by more than 5mm, release it
Peter Hutterer [Tue, 7 Aug 2018 04:49:27 +0000 (14:49 +1000)]
touchpad: if a finger in the button area moves by more than 5mm, release it

The software button area is currently a partially-dead area. If the finger
moves into or out of the area pointer motion works. Finger motion within the
area however does not generate motion.

The main motivation for this was to avoid accidental pointer motion when a
button is pressed. This is required for stationary fingers but once you move a
significant distance, those bets are off.

So if the finger moves by more than 5mm from where it was put down, release it
and let it move the pointer.

The full impact is largely limited to horizontal movements within the button
area because:
- leaving the finger at the bottom area for 300ms without movement triggers
  the thumb identification, so it won't move anyway.
- moving the finger north is likely to go off the button area before we
  trigger this threshold.

https://gitlab.freedesktop.org/libinput/libinput/issues/86

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
5 years agotouchpad: coding style fixes
Peter Hutterer [Tue, 7 Aug 2018 05:28:11 +0000 (15:28 +1000)]
touchpad: coding style fixes

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
5 years agotouchpad: rename 'curr' to 'current'
Peter Hutterer [Tue, 7 Aug 2018 04:33:47 +0000 (14:33 +1000)]
touchpad: rename 'curr' to 'current'

We can affort the extra 3 bytes storage.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoquirks: don't allow single quotes for values
Peter Hutterer [Fri, 10 Aug 2018 02:55:13 +0000 (12:55 +1000)]
quirks: don't allow single quotes for values

At least not opening single quotes, same as the double quotes we already have.
Add the tests for both.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotests: more tests for quirks handling
Peter Hutterer [Fri, 10 Aug 2018 02:11:15 +0000 (12:11 +1000)]
tests: more tests for quirks handling

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: add tests for successful/failing parsing of various quirks
Peter Hutterer [Thu, 9 Aug 2018 23:23:43 +0000 (09:23 +1000)]
test: add tests for successful/failing parsing of various quirks

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: fix quirks test for invalid bus type matches
Peter Hutterer [Fri, 10 Aug 2018 02:17:07 +0000 (12:17 +1000)]
test: fix quirks test for invalid bus type matches

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoutil: tighten safe_atod parsing to only parse 'normal' numbers
Peter Hutterer [Fri, 10 Aug 2018 00:44:30 +0000 (10:44 +1000)]
util: tighten safe_atod parsing to only parse 'normal' numbers

We don't have a sensible use case where we want hex to double, or INF to
double, or any of that. So check the strings for invalid characters and bail
out early. Invalid characters include 'e' and whitespaces too, we don't need
those.

Small chance of things breaking: if the user-exposed calibration matrix
property was specified using hex numbers this will stop working now. I'll take
that risk.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoutil: switch to a check for isnormal for safe_atod
Peter Hutterer [Fri, 10 Aug 2018 00:50:39 +0000 (10:50 +1000)]
util: switch to a check for isnormal for safe_atod

Effectively the same check as before but this should also encompass
FP_SUBNORMAL.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoutil: fail property parsing if the dimensions are 0
Peter Hutterer [Thu, 9 Aug 2018 23:18:47 +0000 (09:18 +1000)]
util: fail property parsing if the dimensions are 0

There is no use-case for a zero width/height in anything using that property.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: replace hand-rolled backtrace function with gstack
Peter Hutterer [Wed, 8 Aug 2018 06:32:06 +0000 (16:32 +1000)]
test: replace hand-rolled backtrace function with gstack

Let's use something that specializes in that task and does a better job of it
than whatever we'll come up with. Due to how it's implemented the stacktrace
will always show waitpid() as frame 0 now but we can live with that.

gstack prints to stdout but litest_log() uses stderr, so we cannot just call
system(), we have do do the pipe/fork/exec/waitpid/read dance.
We could use that to filter the #0 frame showing waidpid() from gstack but
meh.

This drops the libunwind and addr2line dependency and replaces it with gstack
instead.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agomeson.build: the selftest doesn't need all litest sources
Peter Hutterer [Fri, 10 Aug 2018 01:54:24 +0000 (11:54 +1000)]
meson.build: the selftest doesn't need all litest sources

Drops about 70 ninja targets.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: remove the deviceless test runner, make it a commandline arg instead
Peter Hutterer [Fri, 10 Aug 2018 01:39:04 +0000 (11:39 +1000)]
test: remove the deviceless test runner, make it a commandline arg instead

No need to rebuild everything with an ifdef, we can just use meson to pass an
argument to the test itself and filter based on that. This drops about 100
ninja targets.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: switch two ints to bools
Peter Hutterer [Fri, 10 Aug 2018 01:32:31 +0000 (11:32 +1000)]
test: switch two ints to bools

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: move touchpad jitter to troubleshooting
Peter Hutterer [Fri, 10 Aug 2018 10:40:44 +0000 (20:40 +1000)]
doc/user: move touchpad jitter to troubleshooting

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoFix doc links to use dashes, not underscores
Peter Hutterer [Fri, 10 Aug 2018 10:39:49 +0000 (20:39 +1000)]
Fix doc links to use dashes, not underscores

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: move the jumping cursor page to troubleshooting
Peter Hutterer [Fri, 10 Aug 2018 10:38:23 +0000 (20:38 +1000)]
doc/user: move the jumping cursor page to troubleshooting

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: ensure we write something during litest_sendfile
Peter Hutterer [Wed, 8 Aug 2018 10:55:58 +0000 (20:55 +1000)]
test: ensure we write something during litest_sendfile

This mostly shuts up coverity about potentially using a negative size to
write.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: remove an obsolete comment
Peter Hutterer [Wed, 8 Aug 2018 05:39:18 +0000 (15:39 +1000)]
test: remove an obsolete comment

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agomeson.build: drop the now-unused liblitest
Peter Hutterer [Wed, 8 Aug 2018 04:02:15 +0000 (14:02 +1000)]
meson.build: drop the now-unused liblitest

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: use configure_file()'s @PLAINNAME@ instead of calling install
Peter Hutterer [Wed, 8 Aug 2018 01:18:20 +0000 (11:18 +1000)]
doc: use configure_file()'s @PLAINNAME@ instead of calling install

Instead of calling out to install on every ninja call, use @PLAINNAME@ as
substitution for just copying the file over. This gets around the "no
directory allowed in output file" limitation.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/api: drop git-version.dox
Peter Hutterer [Wed, 8 Aug 2018 01:01:38 +0000 (11:01 +1000)]
doc/api: drop git-version.dox

No longer needed

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agomeson.build: process the various tools as a single file list
Peter Hutterer [Wed, 8 Aug 2018 01:21:12 +0000 (11:21 +1000)]
meson.build: process the various tools as a single file list

All these tools just copy the input file over, we can do that in a loop

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agomeson.build: silence meson warning that we should use configure_file's copy
Peter Hutterer [Wed, 8 Aug 2018 02:03:13 +0000 (12:03 +1000)]
meson.build: silence meson warning that we should use configure_file's copy

Yes, we should use 'copy'. But that requires 0.47 and that's not an option
right now.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agomeson.build: fix typo in version to compare to
Peter Hutterer [Wed, 8 Aug 2018 01:36:18 +0000 (11:36 +1000)]
meson.build: fix typo in version to compare to

Introduced in e428e5e87a83127db90476e5a7d7e48546fafd6f

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotouchpad: 90-degree scroll helper
Matt Mayfield [Wed, 8 Aug 2018 16:31:57 +0000 (11:31 -0500)]
touchpad: 90-degree scroll helper

This makes two-finger scrolling in straight lines easier, while still
allowing free/diagonal movement. It works in three stages:

1) Initial movement
   - For the first few millimeters, scroll movements within 30 degrees
     of horizontal or vertical are straightened to 90-degree angles.
   - Scroll movements close to 45 degree diagonals are unchanged.
   - If movement continues very close to straight horizontal or
     vertical, stage 2 begins and the axis lock engages.
   - If movement continues along a diagonal, stage 2 is skipped and
     free scrolling is immediately enabled.
2) Axis lock
   - If the user scrolls fairly closely to straight vertical, no
     horizontal movement will happen at all, and vice versa.
   - It is possible to switch between straight vertical and straight
     horizontal, and the axis lock will automatically change.
   - If deliberate diagonal movement is detected at any point, stage
     3 begins and the axis lock disengages.
3) Free scrolling
   - Scrolling is unconstrained until the fingers are lifted.

6 years agotest: touchpad: add tests for 90 degree scroll (axis lock)
Matt Mayfield [Wed, 8 Aug 2018 16:00:22 +0000 (11:00 -0500)]
test: touchpad: add tests for 90 degree scroll (axis lock)

6 years agodoc/user: add 90-degree scroll behavior
Matt Mayfield [Tue, 7 Aug 2018 14:34:28 +0000 (09:34 -0500)]
doc/user: add 90-degree scroll behavior

6 years agotools: debug-*: show unaccelerated deltas for pointer events
Peter Hutterer [Tue, 7 Aug 2018 02:42:40 +0000 (12:42 +1000)]
tools: debug-*: show unaccelerated deltas for pointer events

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: add a 10ms delay for scrolling tests
Peter Hutterer [Mon, 6 Aug 2018 23:34:49 +0000 (09:34 +1000)]
test: add a 10ms delay for scrolling tests

No touchpad gives us these events with a 0 delay, so let's not test for that.

This is required for adding timing-sensitive scroll code, see
https://gitlab.freedesktop.org/libinput/libinput/issues/101.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: touchpad: swap an int for a bool
Peter Hutterer [Mon, 6 Aug 2018 23:53:28 +0000 (09:53 +1000)]
test: touchpad: swap an int for a bool

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: rename the x220 clickpad to something more telling
Peter Hutterer [Mon, 6 Aug 2018 23:56:54 +0000 (09:56 +1000)]
test: rename the x220 clickpad to something more telling

Having this as the generic "synaptics" touchpad in the tests is not helpful,
this touchpad is tiny and quite special these days.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotablet: unify the license text with what we have in COPYING
Peter Hutterer [Mon, 6 Aug 2018 22:26:53 +0000 (08:26 +1000)]
tablet: unify the license text with what we have in COPYING

Use the same blurb everywhere, changing from the old style MIT to the Expat
license we're using everywhere else.

Similar to bc9f16b40e23a1d25c105a4207b97d65253f0d98

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agofilter: always init a delta smoothener for 10/10ms on trackpoints
Peter Hutterer [Mon, 6 Aug 2018 05:24:04 +0000 (15:24 +1000)]
filter: always init a delta smoothener for 10/10ms on trackpoints

If the trackpoint gives us deltas with less than 10ms intervals, something is
wrong. Could be bad hardware, a glitch in the matrix or a discontinuity in
the otherwise appropriately named time-space continuum. Usually it's the
first.

Let's always set up trackpoint delta smoothening for 10ms to improve the
pointer speed calculation and avoid jerky behaviors. i.e. if a trackpoint
delta comes in below 10ms, pretend it came in with a 10ms interval for
calculating the speed.

Fixes https://gitlab.freedesktop.org/libinput/libinput/issues/104

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: drop the sphinx -a flag
Peter Hutterer [Mon, 6 Aug 2018 05:19:57 +0000 (15:19 +1000)]
doc/user: drop the sphinx -a flag

Let's trust sphinx to know what to rebuild

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotools: record: don't bother setting last_ms to 0 on the first event
Peter Hutterer [Mon, 6 Aug 2018 04:06:48 +0000 (14:06 +1000)]
tools: record: don't bother setting last_ms to 0 on the first event

The first event we receive is set to a 0ms offset anyway. Setting last_ms to 0
on the first event means the first two events have +0ms offset printed to the
log. Skip it, so the second event has the right offset.

This is human-readable data only, no effect on the recording file itself.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: add touchpad pressure debugging as separate page
Peter Hutterer [Mon, 6 Aug 2018 03:37:30 +0000 (13:37 +1000)]
doc/user: add touchpad pressure debugging as separate page

This way we can put it under Troubleshooting

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: add the quirks tool to the tool list
Peter Hutterer [Mon, 6 Aug 2018 03:37:00 +0000 (13:37 +1000)]
doc/user: add the quirks tool to the tool list

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: match the quirks list output with the code
Peter Hutterer [Mon, 6 Aug 2018 03:31:34 +0000 (13:31 +1000)]
doc/user: match the quirks list output with the code

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoRemove seemingly extra pasted line from license
Matt Mayfield [Sun, 5 Aug 2018 17:55:02 +0000 (12:55 -0500)]
Remove seemingly extra pasted line from license

6 years agotouchpad: improve finger counting for synaptics serial touchpads
Peter Hutterer [Thu, 2 Aug 2018 06:58:58 +0000 (16:58 +1000)]
touchpad: improve finger counting for synaptics serial touchpads

A three-finger touch may cause slot N to end, in a frame after the
BTN_TOOL_TRIPLETAP. This causes tp->nfinger_down to be decremented to 2 as the
touch switches to MAYBE_END - which happens to be our num_slots. We exit early
and never restore the touch correctly.

Fix this by checking that the number of fake touches is equal to the slots, if
it is higher then we need to check for recovery.

Fixes https://gitlab.freedesktop.org/libinput/libinput/issues/99

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agofallback: add support for ABS_MT_TOOL_TYPE for touch screens
Peter Hutterer [Fri, 13 Jul 2018 06:08:54 +0000 (16:08 +1000)]
fallback: add support for ABS_MT_TOOL_TYPE for touch screens

Cancel any touches that trigger MT_TOOL_PALM.

Fixes https://gitlab.freedesktop.org/libinput/libinput/issues/25

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: handle ABS_MT_TOOL_TYPE from the litest devices
Peter Hutterer [Fri, 13 Jul 2018 06:24:04 +0000 (16:24 +1000)]
test: handle ABS_MT_TOOL_TYPE from the litest devices

We want to autoreplace this value where possible but not scale it to min/max,
this is effectively an enum. The same is true for slot/tracking id, so let's
add it here too.

Because it's an enum and 99% of the cases require MT_TOOL_FINGER, we always
fall back to that instead of using the axis_defaults that we use for other
axes.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotest: add test helpers for touch sequence parts
Peter Hutterer [Fri, 13 Jul 2018 06:04:35 +0000 (16:04 +1000)]
test: add test helpers for touch sequence parts

The necessary helpers to test for a touch event + one touch frame and the
extra case for the TOUCH_CANCEL in is_touch_event

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agofallback: move flushing MT events to a helper function
Peter Hutterer [Fri, 13 Jul 2018 05:45:22 +0000 (15:45 +1000)]
fallback: move flushing MT events to a helper function

No functional changes

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agofallback: change a list of if conditions to a switch statement
Peter Hutterer [Fri, 13 Jul 2018 05:41:59 +0000 (15:41 +1000)]
fallback: change a list of if conditions to a switch statement

This used to do a lot more but now it can be handled as simple switch
statement. Bonus: we get to log a bug if we ever get here in NONE state.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agofallback: move the mt slot state struct/enum here
Peter Hutterer [Fri, 13 Jul 2018 05:22:23 +0000 (15:22 +1000)]
fallback: move the mt slot state struct/enum here

Not needed by the more generic evdev header

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: mark the touchpads page as orphan
Peter Hutterer [Fri, 3 Aug 2018 01:15:31 +0000 (11:15 +1000)]
doc/user: mark the touchpads page as orphan

Only one link leads to it and it doesnt (right now) fit into the hierarchy.
Let's get rid of the sphinx warning.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: add custom 404 pages for the old links
Peter Hutterer [Fri, 3 Aug 2018 01:04:18 +0000 (11:04 +1000)]
doc/user: add custom 404 pages for the old links

Switching from doxygen to sphinx broke a bunch of links because doxygen used
whatever the argument to @page was - and that usually had underscores. Sphinx
uses filenames (which use dashes) so now we have a bunch of old links going to
a 404. For the transition period at least, insert a custom 404 page for each
of those to tell users this doesn't exist anymore.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoRevert "doc/user: add a custom 404 page"
Peter Hutterer [Fri, 3 Aug 2018 00:13:09 +0000 (10:13 +1000)]
Revert "doc/user: add a custom 404 page"

Nope, doesn't work, our server doesn't support htaccess.

This reverts commit 548e0eb98b46e328a64215323962c11df17b49ce.

6 years agoutil: remove now-unused helper functions
Peter Hutterer [Thu, 2 Aug 2018 05:02:18 +0000 (15:02 +1000)]
util: remove now-unused helper functions

Obsolete with the switch to the device quirks

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: add a custom 404 page
Peter Hutterer [Thu, 2 Aug 2018 01:01:50 +0000 (11:01 +1000)]
doc/user: add a custom 404 page

Things moved around, so let's have a custom 404 page where we can put
information in. This ist the barebones version, not sure if .htaccess is
supported.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: drop the doxygen conversion script
Peter Hutterer [Wed, 1 Aug 2018 23:52:59 +0000 (09:52 +1000)]
doc: drop the doxygen conversion script

It's really not that good, so no point to keep it around

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: rename filenames with underscores to dashes
Peter Hutterer [Wed, 1 Aug 2018 23:44:01 +0000 (09:44 +1000)]
doc/user: rename filenames with underscores to dashes

Everything else is, let's be consistent here.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: swap some note out with hint
Peter Hutterer [Wed, 1 Aug 2018 23:37:54 +0000 (09:37 +1000)]
doc/user: swap some note out with hint

Because why not...

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: add a hack to get to the git version
Peter Hutterer [Wed, 1 Aug 2018 23:30:57 +0000 (09:30 +1000)]
doc/user: add a hack to get to the git version

meson doesn't have configuration_data() in vcs_tag so we can only replace one
string. sphinx cannot include things in-line.

Since we want the git version to be replaced in random places, we need to put
it into rst_prolog in conf.py - but that's where we neet to replace other
things too. Work around this by generating a mini python module that returns
the git version, then call that in conf.py.

Side-bonus: we now have access to the full commit and the abbreviated commit.
Not that anything actually uses this...

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: drop the old page-hierarchy page
Peter Hutterer [Wed, 1 Aug 2018 22:47:06 +0000 (08:47 +1000)]
doc/user: drop the old page-hierarchy page

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: use extlink for a common commit links
Peter Hutterer [Wed, 1 Aug 2018 07:12:38 +0000 (17:12 +1000)]
doc/user: use extlink for a common commit links

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: more fixes including adding a device-types section
Peter Hutterer [Wed, 1 Aug 2018 00:34:22 +0000 (10:34 +1000)]
doc/user: more fixes including adding a device-types section

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: tidy up the development page
Peter Hutterer [Wed, 1 Aug 2018 00:16:17 +0000 (10:16 +1000)]
doc/user: tidy up the development page

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoREADME: drop a @ref in favor of the full link
Peter Hutterer [Wed, 1 Aug 2018 23:42:57 +0000 (09:42 +1000)]
README: drop a @ref in favor of the full link

No doxygen leftovers in README anymore, exspecially now that we don't use it
for doxygen anymore.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agofallback: add another debug message for tablet-mode keyboard suspend
Peter Hutterer [Wed, 1 Aug 2018 05:47:03 +0000 (15:47 +1000)]
fallback: add another debug message for tablet-mode keyboard suspend

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agofallback: change a debug message to be less ambiguous
Peter Hutterer [Wed, 1 Aug 2018 05:28:58 +0000 (15:28 +1000)]
fallback: change a debug message to be less ambiguous

"activated" sounds like we're disabling things here, but we're just pairing
the two devices

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: more documentation cleanups and fixes
Peter Hutterer [Tue, 31 Jul 2018 08:17:48 +0000 (18:17 +1000)]
doc/user: more documentation cleanups and fixes

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: some rewording and improvements
Peter Hutterer [Tue, 31 Jul 2018 06:03:49 +0000 (16:03 +1000)]
doc/user: some rewording and improvements

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agomeson.build: fix name of the measure touchpad-tap script
Peter Hutterer [Tue, 31 Jul 2018 03:00:38 +0000 (13:00 +1000)]
meson.build: fix name of the measure touchpad-tap script

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: update the 'no package found' link to the meson version post
Peter Hutterer [Tue, 31 Jul 2018 00:50:57 +0000 (10:50 +1000)]
doc/user: update the 'no package found' link to the meson version post

Now that we dropped autotools anyway, it's better to link to the meson version
of that blog post.

Related #96

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: building - promote dependencies to a higher-level section
Peter Hutterer [Tue, 31 Jul 2018 00:49:47 +0000 (10:49 +1000)]
doc/user: building - promote dependencies to a higher-level section

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: the SELinux section doesn't apply to modern meson
Peter Hutterer [Tue, 31 Jul 2018 00:48:34 +0000 (10:48 +1000)]
doc/user: the SELinux section doesn't apply to modern meson

Related #96

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: re-do the conditional build section
Peter Hutterer [Tue, 31 Jul 2018 00:42:06 +0000 (10:42 +1000)]
doc/user: re-do the conditional build section

Related #96

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: drop the autotools builds
Peter Hutterer [Tue, 31 Jul 2018 00:25:01 +0000 (10:25 +1000)]
doc/user: drop the autotools builds

Anyone who's still on <= 1.8 either knows how to build it already or relies on
a distribution to do that for them. Drop the section.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: add a note on meson's dependency() in the building docs
Peter Hutterer [Tue, 31 Jul 2018 00:22:31 +0000 (10:22 +1000)]
doc/user: add a note on meson's dependency() in the building docs

See #96

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: mesonconf is meson configure
Peter Hutterer [Tue, 31 Jul 2018 00:17:59 +0000 (10:17 +1000)]
doc/user: mesonconf is meson configure

Listed in #96

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotablet: clear tablet history on tool contact change
jeff [Mon, 30 Jul 2018 17:03:03 +0000 (13:03 -0400)]
tablet: clear tablet history on tool contact change

On some ELAN tablets we get a coordinate jump in the same frame that we put
the tip down. The existing axis smoothing causes that jump to be somewhere in
the middle between the previous and the next coordinates, causing a small
stroke from the smoothed position to the next. Prevent this by resetting the
history on tip down/up so we always take that coordinate.

Fixes #94

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agolibinput 1.11.901 1.11.901
Peter Hutterer [Mon, 30 Jul 2018 05:57:08 +0000 (15:57 +1000)]
libinput 1.11.901

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc/user: add a page detailing what configuration toggles are available
Peter Hutterer [Sat, 28 Jul 2018 09:35:36 +0000 (19:35 +1000)]
doc/user: add a page detailing what configuration toggles are available

High level view only but at least it's a link we can point people to.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodox: switch to sphinx for the user-visible documentation
Peter Hutterer [Tue, 10 Jul 2018 07:32:35 +0000 (17:32 +1000)]
dox: switch to sphinx for the user-visible documentation

This is a large commit because it's difficult to split this up and we don't
care about bisecting here anyway.

doxygen is going to produce the API documentation only
sphinx is going to produce the prose user (and a bit of developer) documentation.

The source split is doc/api and doc/user.

Steps performed:
- run the doxygen-to-sphinx.sh script to convert all .dox sources to .rst
- manually fixed the .rst to render correctly
- add a few extra .rst documents to generate the right hierarchy
- hook up sphinx-build in meson
- add a new @mainpage for doxygen more aimed at developers

For the build directory:
- sphinx produces /Documentation
- doxygen now produces /api/

These need to be manually combined in the wayland-web repo, meson doesn't
support subdirectories as output paths within the build dir and the
documentation doesn't need to be installed anywhere.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoGitLab CI: print out the image age in seconds
Peter Hutterer [Mon, 30 Jul 2018 02:04:05 +0000 (12:04 +1000)]
GitLab CI: print out the image age in seconds

Helps with debugging purposes

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoGitlab CI: set DEBIAN_FRONTEND noninteractive for the Ubuntu images
Peter Hutterer [Mon, 30 Jul 2018 00:51:36 +0000 (10:51 +1000)]
Gitlab CI: set DEBIAN_FRONTEND noninteractive for the Ubuntu images

Because Godot doesn't care about tzdata, so waiting for him is pointless.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoAdd the buildreqs for sphinx-build
Peter Hutterer [Fri, 27 Jul 2018 07:31:14 +0000 (17:31 +1000)]
Add the buildreqs for sphinx-build

Separate commit so we can prep the containers for the real PR and thus test if
something break.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: split the graphs into a separate array for re-use
Peter Hutterer [Wed, 25 Jul 2018 22:32:31 +0000 (08:32 +1000)]
doc: split the graphs into a separate array for re-use

We need this in the sphinx documentation.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: move the @verbatim start to the list item
Peter Hutterer [Tue, 24 Jul 2018 05:01:39 +0000 (15:01 +1000)]
doc: move the @verbatim start to the list item

This is mostly for sphinx' parsing benefit, doxygen doesn't need it and
renders the same either way.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: device-quirks: redo the model quirks <dl> for sphinx' benefit
Peter Hutterer [Thu, 26 Jul 2018 01:22:17 +0000 (11:22 +1000)]
doc: device-quirks: redo the model quirks <dl> for sphinx' benefit

sphinx only supports one line of <dt> content, so we have to compress these.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: a few replacements to make sphinx happy/prettier
Peter Hutterer [Tue, 24 Jul 2018 01:25:16 +0000 (11:25 +1000)]
doc: a few replacements to make sphinx happy/prettier

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: refer to "the libinput documentation" in a few @see cases
Peter Hutterer [Fri, 27 Jul 2018 04:01:35 +0000 (14:01 +1000)]
doc: refer to "the libinput documentation" in a few @see cases

This is in preparation for a change to sphinx as the user-visible
documentation. Ideally we could cross-link between the two but that's tricky
to do automatically. Linking to the html pages/anchors directly works fine but
risks the links going stale, especially while the documentation is still in
flux.

Having a generic "refer to the libinput documentation" is a bit of a cop-out
least this way the links cannot go stale.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: drop a few unnecessary @ref
Peter Hutterer [Fri, 27 Jul 2018 05:51:31 +0000 (15:51 +1000)]
doc: drop a few unnecessary @ref

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: enable warnings in doxygen
Peter Hutterer [Wed, 25 Jul 2018 22:50:13 +0000 (08:50 +1000)]
doc: enable warnings in doxygen

We want to see missed links

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: copy some bits of the tablet docs to doxygen
Peter Hutterer [Fri, 27 Jul 2018 03:58:40 +0000 (13:58 +1000)]
doc: copy some bits of the tablet docs to doxygen

So it shows up in the doxygen contents. These are the bits that are mostly of
interest to developers, we might remove the user documentation entry later,
but let's leave it there for now.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agogitlab-ci: use skopeo to check on container images
Benjamin Tissoires [Mon, 23 Jul 2018 11:01:17 +0000 (13:01 +0200)]
gitlab-ci: use skopeo to check on container images

this allows to replace complex curl queries with simpler commands.

We need a newer minimalist image with skopeo in addition to jq and
curl.

Also, I am currently not relying on skopeo to delete the image as I am
not so sure we will get the same cleanup than with the current sha method
and also:
https://bugzilla.redhat.com/show_bug.cgi?id=1481196

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
6 years agodoc: move the Architecture blurb to the 'what is libinput' page
Peter Hutterer [Wed, 25 Jul 2018 23:18:29 +0000 (09:18 +1000)]
doc: move the Architecture blurb to the 'what is libinput' page

The vast majority of ppl reading the README is unlikely to care about this,
but they do care about the Wayland vs X.Org situation so let's split this up.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agotools: drop two unused imports
Peter Hutterer [Thu, 26 Jul 2018 05:05:40 +0000 (15:05 +1000)]
tools: drop two unused imports

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agodoc: we don't need the Namespaces doxygen page
Peter Hutterer [Thu, 26 Jul 2018 03:56:15 +0000 (13:56 +1000)]
doc: we don't need the Namespaces doxygen page

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
6 years agoutil: fix a ubsan complaint about undefined left-shift
Peter Hutterer [Thu, 26 Jul 2018 02:06:10 +0000 (12:06 +1000)]
util: fix a ubsan complaint about undefined left-shift

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>