Tai-Hsu Lin [Sat, 18 May 2013 08:39:53 +0000 (16:39 +0800)]
Add GestureMetrics for collecting UMA touchpad noise stats
This patch adds the metrics gesture (GestureMetrics) which is used
to deliver interesting patterns that we would like to collect with
the UMA system. Due to the deadlock problem, we need to pump the
found instances up into Chrome and send them with the UMA utilities
there. A new filter, MetricsFilterInterpreter, is added accordingly
to catch interested patterns and produce metrics gestures. We
currently catches only the touchpad ground noise pattern in the
filter, which is defined as two consecutive big moves in opposite
directions within a short period of time.
BUG=chromium:237683
TEST=Pass regression tests, no regression.
To test it on real devices, one needs the following patch on CMT:
https://gerrit.chromium.org/gerrit/55729/
On link:
$ DISPLAY=:0.0 xinput list-props 11 | grep "Metrics"
One should see:
Metrics Noisy Ground Distance (403): 10.000000
Metrics Noisy Ground Time (404): 0.100000
To watch for the metrics gesture, use command:
$ DISPLAY=:0.0 xinput test-xi2 11 | grep "9:"
With the default distance threshold (10.0), it should be impossible
to trigger the check. Try to quickly move finger on the touchpad and
make sure no line shows up. Now, change the threshold value to 0.1 with
the following line:
$ DISPLAY=:0.0 xinput set-float-prop 11 403 0.1
Quickly move finger back and forth on the touchpad again, one should
now see various lines coming up with the above watch command:
$ DISPLAY=:0.0 xinput test-xi2 11 | grep "9:"
Detail for Valuator 9:
9: 0.00 (0.00)
9: 0.00 (0.00)
9: 0.00 (0.00)
9: 0.00 (0.00)
...
Change-Id: I5e6c9bd5489f4e5aeb90e0d7d7f48cd3e068d08f
Reviewed-on: https://gerrit.chromium.org/gerrit/55727
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Tai-Hsu Lin <sheckylin@chromium.org>
Tested-by: Tai-Hsu Lin <sheckylin@chromium.org>
Tai-Hsu Lin [Sun, 12 May 2013 09:32:56 +0000 (17:32 +0800)]
Track pressure and touch_major in TrendClassifyingFilter
The patch extends the TrendClassifyingFilterInterpreter to track
both the pressure and the touch_major of a finger. The classification
result is further used in the ImmediateFilterInterpreter to suppress
unintentional cursor move caused by palm-like contacts.
BUG=chromium:239683
TEST=Passed regression test, improved the result upon
https://gerrit.chromium.org/gerrit/50747/
| spring1/move_while_typing | +1.0000 |
Change-Id: I270454611331f2fc5199b2f3b3396f391e032d7b
Reviewed-on: https://gerrit.chromium.org/gerrit/50961
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Tai-Hsu Lin <sheckylin@chromium.org>
Tested-by: Tai-Hsu Lin <sheckylin@chromium.org>
Andrew de los Reyes [Wed, 8 May 2013 23:42:52 +0000 (16:42 -0700)]
ActivityReplay: Replay logs correct
Updates ActivityReplay to use the new interpreter API (which can
produce multiple gestures per SyncInterpret/HandleTimer call). Also,
updates ActivityReplay and ActivityLog to store/load ordinal values.
With this, I can replay logs and have them match exactly.
BUG=chromium:239295
TEST=replayed log from same version w/o differences.
Change-Id: I798e4e0ec3e12240c1c3a8bca961fa1a8d9d7678
Reviewed-on: https://gerrit.chromium.org/gerrit/51212
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Tai-Hsu Lin [Sat, 11 May 2013 10:58:29 +0000 (18:58 +0800)]
Seperate the logic for 2f gesturing and left/right click
The function TwoFingersGesturing is used both for deciding
2f gesture type and for deciding whether it is a left or a
right click. This sometimes causes the problem because the
ideal logic for both use cases may not always be the same.
The patch adds an option to the function to distinguish the
two different cases.
BUG=chromium:240044
TEST=pass regression test
| link/left_click_with_resting_thumb | +1.0000 |
Change-Id: I5efb6d0d3a0bcfb867e47eeef781557baae79c13
Reviewed-on: https://gerrit.chromium.org/gerrit/50953
Commit-Queue: Tai-Hsu Lin <sheckylin@chromium.org>
Reviewed-by: Tai-Hsu Lin <sheckylin@chromium.org>
Tested-by: Tai-Hsu Lin <sheckylin@chromium.org>
Andrew de los Reyes [Fri, 10 May 2013 00:16:09 +0000 (17:16 -0700)]
gestures.cc: Use persistent HardwareProperties
BUG=chromium:239678
TEST=No regression test regressions; unittests pass
Change-Id: Ife97534d99747d6ca70cfb1353c1dd6b83df7fae
Reviewed-on: https://gerrit.chromium.org/gerrit/50743
Reviewed-by: Chung-yih Wang <cywang@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Fri, 10 May 2013 00:15:13 +0000 (17:15 -0700)]
Revert "Fix weird touchpad behavior on Cr48"
This reverts commit
daf12af728625c8b89565896b7e63d8a280ee3d8.
BUG=chromium:239678
TEST=No regression test regressions; unittests pass
Change-Id: Iebaf4373baf7da02939b363474d743cd4e77d1ce
Reviewed-on: https://gerrit.chromium.org/gerrit/50742
Reviewed-by: Chung-yih Wang <cywang@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Tai-Hsu Lin [Tue, 16 Apr 2013 09:34:26 +0000 (17:34 +0800)]
Add TrendClassifyingFilterInterpreter
The patch adds a filter that tries to distinguish between resting
thumbs and fingers that the user intentionally moved. This can
eliminate the un-intended initial cursor movement during the 2f
scroll gesture when one of the finger is in the bottom zone.
BUG=chromium:219126
TEST=pass regression test
Improve the result upon https://gerrit.chromium.org/gerrit/#/c/48362/
| link/scrolling_with_low_press_fingers_10 | +0.9807 |
| link/scrolling_with_low_press_fingers_2 | +0.9716 |
| link/scrolling_with_low_press_fingers_3 | +0.9886 |
| link/scrolling_with_low_press_fingers_4 | +0.9920 |
| link/scrolling_with_low_press_fingers_5 | +0.9562 |
| link/scrolling_with_low_press_fingers_6 | +0.9808 |
| link/scrolling_with_low_press_fingers_7 | +0.9657 |
| link/scrolling_with_low_press_fingers_8 | +0.9607 |
| link/scrolling_with_low_press_fingers_9 | +0.9739 |
Change-Id: I01d1668d63853be1ceb4a58f71c1f74427482f83
Reviewed-on: https://gerrit.chromium.org/gerrit/48218
Commit-Queue: Tai-Hsu Lin <sheckylin@chromium.org>
Tested-by: Tai-Hsu Lin <sheckylin@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Chung-yih Wang [Wed, 8 May 2013 05:59:34 +0000 (13:59 +0800)]
Fix weird touchpad behavior on Cr48
HardwareProperties messed up on Cr48 which causes weird behaviors such
as stumble cursor move, right click when tap-to-click intended. The root
cause is incorrect backup of HardwareProperties which points to
somewhere in stack.
BUG=chromium:238778
TEST=on Cr48, cursor move smoothly, no right click when ttc intended
Change-Id: I395862e885b8950ed37fbc1e6ee0c6583974b367
Reviewed-on: https://gerrit.chromium.org/gerrit/50428
Reviewed-by: Joseph Shyh-In Hwang <josephsih@chromium.org>
Commit-Queue: Chung-yih Wang <cywang@chromium.org>
Tested-by: Chung-yih Wang <cywang@chromium.org>
Charlie Mooney [Mon, 29 Apr 2013 17:49:40 +0000 (10:49 -0700)]
Increase vertical range for 2finger taps
Per a missed click report where two fingers were slightly too vertically
alligned to register as a 2f click this CL increases the range a little
in the Y direction from 30mm to 35mm.
For two fingers to be tapping together the second finger has to fall
within an elipse of the first finger on the pad. This elipse is wider
than it is tall, so in this report the fingers were skewed vertically
more than horizontally and it interpreted it as a left click instead.
BUG=chromium:236446
TEST=regression tests (no regressions detected)
Change-Id: Icbf8d455d8668c5a95a3dc9fb0ba4dabe14e39d4
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/49491
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Andrew de los Reyes [Fri, 3 May 2013 00:31:17 +0000 (17:31 -0700)]
MultitouchMouseInterpreter: Prevent unintended small scrolls
This CL introduces a new technique to prevent unintended small scrolls
on a multitouch mouse. First, it requires that a finger deviate from
it's initial position before it can cause gestures. Also, when the
mouse is moving significantly, the start positions are reset to their
current positions.
In practice, this does a good job of limiting accidental scrolls.
BUG=chromium:237692
TEST=Manually tested on device; magicmouse/no_scrolls test
Change-Id: I0c4ef55beb77074d832d8e38db39cdc08e904d41
Reviewed-on: https://gerrit.chromium.org/gerrit/49994
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Fri, 3 May 2013 00:07:17 +0000 (17:07 -0700)]
ActivityLog/Replay: Log/replay mouse relative events.
BUG=chromium:237685
TEST=Saw rel events in log
Change-Id: I6f5564e58ecd82c03d0b633d710b04c9654a6cce
Reviewed-on: https://gerrit.chromium.org/gerrit/49993
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Sat, 4 May 2013 01:47:11 +0000 (18:47 -0700)]
UnittestUtils: Fix use of uninitialized memory
mprops_ holds properties, which hold a reference to their
PropRegistry. By resetting prop_reg_ before mprops_, this caused
mprops_ (in the props' dtors) to call into random memory for where the
old prop registry was.
BUG=chromium:238032
TEST=unittest don't have double-free/segfaults anymore.
Change-Id: If80f7cd9eec4a1c5e0d8b2ee1d05fc0a136e4c3f
Reviewed-on: https://gerrit.chromium.org/gerrit/50094
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Che-Liang Chiou [Wed, 1 May 2013 00:25:02 +0000 (17:25 -0700)]
ScrollManager: Do not report high pressure change when fast swiping
A fast-swiping finger may generate rapidly changing pressure and this
should not be confused with a finger leaving the touchpad; instead, we
consider this finger not stationary and do not report high pressure
changes in such cases.
BUG=chromium:235968
TEST=manual
* Pass regression test generated from the logs of
https://feedback.corp.google.com/#/Report/
812524363?context=ri
* No regressions of regression tests from the logs of
https://feedback.corp.google.com/#/Report/
786686567?context=ri
https://feedback.corp.google.com/#/Report/
786691579?context=ri
* Pass unit tests
Change-Id: I8b545790d78214131d6f35a2842a11b7e0d218bc
Reviewed-on: https://gerrit.chromium.org/gerrit/49695
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Yunlian Jiang [Wed, 1 May 2013 00:03:51 +0000 (17:03 -0700)]
gestures: fix conflict between struct and class
This fixes the conflict of definition and declaration of struct and class.
This also fixes the warning of "variable length array of non-POD element type"
BUG=chromium:237039
TEST=FEATURES="test" CFLAGS="-clang" emerge-lumpy gestures passes.
Change-Id: I8512793c9c27bfbeb987352a88b18f2635ccc9c0
Reviewed-on: https://gerrit.chromium.org/gerrit/49724
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Yunlian Jiang <yunlian@chromium.org>
Commit-Queue: Yunlian Jiang <yunlian@chromium.org>
Che-Liang Chiou [Mon, 29 Apr 2013 21:54:03 +0000 (14:54 -0700)]
box_filter_interpreter: Add box height
So now boxes not only have widths but also have heights, and they could
be rectangular shaped.
BUG=chromium:235971
TEST=make test && ./test # unit tests passed
Change-Id: I242f3003577938e791b8f453141aaf036222a694
Reviewed-on: https://gerrit.chromium.org/gerrit/49533
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Dennis Kempin [Mon, 29 Apr 2013 22:35:41 +0000 (15:35 -0700)]
new class: FingerMetrics/Metrics
Metrics is a class to persistently store information on fingers and
derive common metrics from this information.
Each interpreter can get access to a Metrics class by setting
requires_metrics_ = true, if it cannot get Metrics from the previous
interpreter in the filter chain it will create one.
Note that some interpreter (scaling, lookahead) invalidate existing
metrics information and pass NULL to their next_ to have it create a new
instance if it needs one.
As properties can only be registered once, we require a separate
MetricsProperties class that has to be passed to the filter chain.
BUG=chromium:226699
TEST=check for regressions
Change-Id: I0a3bcd3fdb4b1706f86e3eedcddb726335b103c4
Reviewed-on: https://gerrit.chromium.org/gerrit/49539
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Sat, 27 Apr 2013 01:00:54 +0000 (18:00 -0700)]
Refactoring: Initialize instead of SetHwProps/SetGestureConsumer
The initialize method is called after the filter chain has been set up.
This allows easier extension of the filter chain setup process without
having to update every single unittest.
Also to make unittest independent of the way a filter chain is set up
this CL updates all unittests to use TestInterpreterWrapper for creating
the test interpreter.
BUG=chromium:226699
TEST=unit tests pass / touchtests shows no regressions / test by
gmerging onto a device.
Change-Id: Ib8ed36bf08cc65a4f03de3ca9f0ba292f6607273
Reviewed-on: https://gerrit.chromium.org/gerrit/49403
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Che-Liang Chiou [Thu, 25 Apr 2013 00:26:30 +0000 (17:26 -0700)]
ScalingFilterInterpreter: Remove zero area fingers
Sometimes magic mouse reports fingers with zero area, which should be
removed.
BUG=none
TEST=Feed log from feedback report and see that zero area fingers are
removed
Change-Id: If71ecd8dc2597a412a46cba4dcdf655089336267
Reviewed-on: https://gerrit.chromium.org/gerrit/49128
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Che-Liang Chiou [Thu, 11 Apr 2013 23:32:16 +0000 (16:32 -0700)]
MultitouchMouseInterpreter: No small scroll after click
In general, finger movements are interpreted as scroll, but as clicks
and scrolls on multi-touch mice are both single-finger gesture, we have
to recognize and separate clicks from scrolls, when a user is actually
clicking.
This is how we do for now: We look for characteristic patterns of
clicks, and if we find one, we hold off emitting scroll gesture for a
few time frames to prevent premature scrolls.
The patterns we look for:
* Small finger movements when button is down
* Finger movements after button goes up
BUG=none
TEST=manual, see no small scrolls in 9 out of 10 clicks
Change-Id: I96ee96ab2edfb545d603dad74e029b6331c2e381
Reviewed-on: https://gerrit.chromium.org/gerrit/47947
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Che-Liang Chiou <clchiou@chromium.org>
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Andrew de los Reyes [Thu, 18 Apr 2013 23:58:14 +0000 (16:58 -0700)]
ImmediateInterpreter: In buttons timer callback, check if valid button
A recent change went in (
6864bbd60695b7807984b11e5683e86044071e4a)
that relaxed the checks in the button handling timer callback to allow
for any button. With this CL, we make it be any valid button.
BUG=chromium:231984
TEST=lumpy/2f_scroll_fail
Change-Id: Ieb56d7811cbc1891e55aae17907e876b5eb3d6f7
Reviewed-on: https://gerrit.chromium.org/gerrit/48580
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Dennis Kempin [Thu, 11 Apr 2013 01:55:11 +0000 (18:55 -0700)]
new class: vector and FilteredRange
Adds a std::vector-like class that works on a fixed-size array similar
to gestures::set.
Also adds a FilteredRange class that can be used to create iterable
containers that skip certain elements. This can be used for a helper
method to iterate over all pointing fingers by skipping fingers with
certain flags. In combination with c++11 range-based for loops this
will reduce a lot of repetitive code in the filter interpreter.
Note: This CL changes the compilation standard to GNU C++11.
BUG=chromium:226699
TEST=src/vector_unittest.cc
Change-Id: I4118c92b878923d398513c5d2645da41ab84a395
Reviewed-on: https://gerrit.chromium.org/gerrit/47828
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Thu, 11 Apr 2013 21:48:42 +0000 (14:48 -0700)]
gestures: allow callback to be set and reset
The previous refactoring expected the callback to be set only once
during init. However when switching VT's the device is turned off by CMT
which will reset the callback to NULL.
This caused assertion failures.
BUG=chromium:230293
TEST=gmerge gestures
switch VT between terminal and X multiple times. check the logs for
assertion failed errors.
Change-Id: I59f3826fd3984273fad8db0b7191956856fe8159
Reviewed-on: https://gerrit.chromium.org/gerrit/47913
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Chung-yih Wang [Thu, 11 Apr 2013 09:57:46 +0000 (17:57 +0800)]
PalmClassifyingFilterInterpreter: fix false positive
We have false positive case especially for a sensor jump with a
finger merge case where the merged finger doubles its pressure
only for that event. The patch tries to skip UpdatePalmState()
for the state of the number of finger decreases.
BUG=chrome-os-partner:12358
TEST=touchtests snow2.3/scrolling_with_sensor_jump
and no regression on others
Change-Id: Icd712d1a4264cedaf23dc40b15c0f0c117db31a4
Reviewed-on: https://gerrit.chromium.org/gerrit/47847
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Chung-yih Wang <cywang@chromium.org>
Tested-by: Chung-yih Wang <cywang@chromium.org>
Andrew de los Reyes [Wed, 10 Apr 2013 19:54:40 +0000 (12:54 -0700)]
ImmediateInterpreter: Don't warn delaying button down on non-left button.
We used to only delay deciding to send a button down if the current
button would be a left click, but now we may wait even on 2-finger
click (right click) in case a third finger arrives. Thus, this old
warning doesn't make any sense anymore.
Also, when the timer does fire and it's time to send a button down,
send the proper one.
BUG=chromium:225994
TEST=Replayed the log and saw no error messages.
Change-Id: I2fe81ea95a83338e285d1d25d6f8d0ba84b8c0da
Reviewed-on: https://gerrit.chromium.org/gerrit/47763
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Chung-yih Wang [Tue, 9 Apr 2013 09:21:46 +0000 (17:21 +0800)]
ImmediateInterpreter: Fix right when left click intended
When a moving finger moves back to its starting point and then clicks
the pad, current logic can not detect it is a moving finger as
previous moving indicator is cleared right after the click happened.
This patch reserves previous moving finger set with the variable
'prev_moving_', so it could be used to detect if a finger is 'hot' one.
BUG=chromium:219399
TEST=link/right_when_left_click_intended
CQ-DEPEND=I69d48303cfb8ac30add6ce08480c18bbca1bcc35
Change-Id: I9346cc53a351366fc226ff5fa423e2b885def531
Reviewed-on: https://gerrit.chromium.org/gerrit/47634
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Chung-yih Wang <cywang@chromium.org>
Tested-by: Chung-yih Wang <cywang@chromium.org>
Dennis Kempin [Sat, 6 Apr 2013 01:09:38 +0000 (18:09 -0700)]
Refactoring: Updated unit tests to work with the new API
After refactoring the interpreter api to use a GestureConsumer instead
of returning gestures, the unit tests had to be updated to use the same
API.
BUG=chromium:226699
TEST=make test && ./test
All pass
Change-Id: I4740185889f26afe1c5e2e78b48c9ecb76454719
Reviewed-on: https://gerrit.chromium.org/gerrit/47508
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Fri, 5 Apr 2013 21:33:21 +0000 (14:33 -0700)]
Refactoring: Removed Gesture* return value from API
All filters are nore exclusively using Produce/Consume gesture.
No regressions.
BUG=chromium:226699
TEST=check for regressions
Change-Id: Iae981c90af0e305d6b8dca29c7fcc42100f7aef4
Reviewed-on: https://gerrit.chromium.org/gerrit/47474
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Fri, 5 Apr 2013 20:24:05 +0000 (13:24 -0700)]
Refactoring: Updated rest of filter chain to use GestureConsumer
The remaining filters have been updated to use the GestureConsumer API.
All filters now return NULL in sync interpret, but still catch possible
sync interpret return values until the API is removed.
Note: The changes to LookAheadFilter introduce slightly lower scores on
3 test cases (score delta < 0.01) due to gestures that are being
produced with the new API that had to be dropped before. This is an
improvment.
BUG=chromium:26699
TEST=check for regressions: touchtests -r before
Change-Id: I4e96a17e8698bdaa35d9c6730c7a1b7f9676a031
Reviewed-on: https://gerrit.chromium.org/gerrit/47450
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Thu, 4 Apr 2013 23:55:09 +0000 (16:55 -0700)]
Refactoring: Transition of filter chain to GestureConsumer
Updated all filters before the LookAheadFilter to use the
new GestureConsumer API for returning gestures.
Every filter has temporary code to convert a Gesture that might be
returned by SyncInterpret into ConsumeGesture calls.
This code will be removed when the return value of SyncInterPret is
removed from the API.
BUG=chromium:226699
TEST=run touchtests to check for regressions
Change-Id: Iee1d5fe42138adda055a0e128693932f403433cc
Reviewed-on: https://gerrit.chromium.org/gerrit/47383
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Thu, 4 Apr 2013 21:13:16 +0000 (14:13 -0700)]
New interface: GestureConsumer
The gesture consumer interface describes an object that can receive
gestures for processing. All interpreters gain a method ProduceGesture
to send a gesture to a GestureConsumer, and all FilterInterpreters
implement this interface to pass it backwards through the filter chain.
The end point of this backwards chain is a private class
GestureInterpreterConsumer that sends the gesture to the GestureReady
callback.
This CL does not make use of this new way of generating gestures yet,
but only provides the interface and methods to do so.
BUG=chromium:26699
TEST=touchtests should not show any regressions!
Change-Id: I9d9e629d7c8ae37cb4a6d0a9923e5641b1a0998b
Reviewed-on: https://gerrit.chromium.org/gerrit/47362
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Thu, 4 Apr 2013 20:44:53 +0000 (13:44 -0700)]
gestures: New Assert/AssertReturn macros
These macros are used for checking for programming errors to allow
more compact sanity checking during execution.
BUG=chromium:226699
TEST=none
Change-Id: I6a7a5b94e1d43eed14d4fce651b1b5b0e81ab7a5
Reviewed-on: https://gerrit.chromium.org/gerrit/47361
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Yunlian Jiang [Mon, 8 Apr 2013 16:56:04 +0000 (09:56 -0700)]
Fix clang syntax checking error.
Removed unused functions.
Added __attribute__ ((noreturn)) to a non-return function.
BUG=chromium:220278
TEST=USE="chrome_internal" CFLAGS="-clang -print-cmdline"
CXXFLAGS="-clang -print-cmdline" emerge-x86-alex gestures
passed.
Change-Id: I9e3b5e7aab8366feeca76dfe77d262aaa21d6aa0
Reviewed-on: https://gerrit.chromium.org/gerrit/47561
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Yunlian Jiang <yunlian@chromium.org>
Tested-by: Yunlian Jiang <yunlian@chromium.org>
Yufeng Shen [Thu, 4 Apr 2013 21:11:12 +0000 (17:11 -0400)]
palm_classifying_filter: Simple logic to detect fat finger
We add a simple logic for reclassifying a previous palm as a fat finger:
if the finger's lifetime max pressure is less than
palm_pressure_ * fat_finger_pressure_ratio_ (default 1.4) and
the finger's lifetime max width is less than
palm_width_ * fat_finger_width_ratio_ (default 1.3)
and the finger has moved more than fat_finger_min_dist_
This is mainly for making a flat thumb to move the cursor. But we
set the controlling parameters very conservative so that the real
palm detection does not regress.
BUG=chrome-os-partner:13144
TEST=1. Manual test on Link & Snow, no obvious regression in palm
detection, and flat thumb can move the cursor (to a certain
degree)
2. Regression test snow2.3/flat_thumb_move passes.
Change-Id: Idf442807b95730cafb4698c5ee1eded39997e6c7
Reviewed-on: https://gerrit.chromium.org/gerrit/47369
Commit-Queue: Yufeng Shen <miletus@chromium.org>
Reviewed-by: Yufeng Shen <miletus@chromium.org>
Tested-by: Yufeng Shen <miletus@chromium.org>
Andrew de los Reyes [Mon, 1 Apr 2013 22:38:41 +0000 (15:38 -0700)]
ImmediateInterpreter: Disable pinch-zoom
We've found that pinch is catching some normal motion. So for now,
we'll disable it.
BUG=chromium:219398
TEST=link/resting_thumb regression test
Change-Id: I751d790ed1daa4604d795bb3a24cacaa3da78550
Reviewed-on: https://gerrit.chromium.org/gerrit/47062
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Charlie Mooney [Mon, 1 Apr 2013 23:46:36 +0000 (16:46 -0700)]
Variable timeout when waiting for FlingStop
We have a report from a user who is flinging and re-flinging very
rapidly, but it gets some stutteriness thrown into it due to a couple of
rouge flingstop events. The user is having one finger hit the pad
before the other and so the fling stop is getting sent out before it
sees the other finger, realizes its a scroll and stops itself.
This CL adds the functionality to add an extra bit of time onto the
timeout when there are 2 fingers on the pad even if they haven't started
to scroll yet. This allows us to push the base timeout lower so it
speeds up the response in single-finger cases and still lets the window
be open enough that re-flinging works.
BUG=chromium:2173333
TEST=manual testing on Daisy and regression tests + Unit Tests
Change-Id: I9e45422b67ee8d9e2b10951073284fadc509c112
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/47123
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 3 Apr 2013 20:10:30 +0000 (13:10 -0700)]
SplitCorrectingFilterInterpreter: Fix tests for invalid tracking numbers.
This code assumed (falsely) that any negative tracking number was
invalid. Actually, 0xffff (aka -1) is the only invalid number. Because
of this mixup, fingers with a high bit set in their tracking ID
weren't considered for finger merging.
BUG=chromium:218293
TEST=lumpy/thumb_split_* regression tests
Change-Id: I3f633d9b02076cb93c2156180bd6da6154b95e9c
Reviewed-on: https://gerrit.chromium.org/gerrit/47265
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Fri, 29 Mar 2013 23:25:06 +0000 (16:25 -0700)]
gestures.cc: Replay all gestures generated in timer callback.
BUG=chromium:225128
TEST=mario/light_ttc regression test
Change-Id: I3b9b1034fb876aac88ff6dff1fb0f0a1a24fd3a6
Reviewed-on: https://gerrit.chromium.org/gerrit/46946
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Dennis Kempin [Fri, 29 Mar 2013 23:34:50 +0000 (16:34 -0700)]
immediate_interpreter: lock onto moving finger
instead of using the finger closest to the keyboard, we now lock onto a
finger once it has been identified for movement. This way the same
finger will be used for movement even if it moves below another finger.
BUG=chromium:217820
TEST=lumpy/high_thumb_dragging
Change-Id: I499d81cdccb5c4677182790c58163910a58cd850
Reviewed-on: https://gerrit.chromium.org/gerrit/46949
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Fri, 29 Mar 2013 21:48:27 +0000 (14:48 -0700)]
immediate_interpreter: more reliable 2F right click
When both fingers touch down at the same time we know that they are
probably not finger and thumb. So we send a right click in this case.
BUG=chromium:217358
TEST=lumpy/left_when_right_click_intended
Change-Id: I1b408d787fe907c40ae869ad393aa29d1b74c27b
Reviewed-on: https://gerrit.chromium.org/gerrit/46913
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Gaurav Shah [Thu, 28 Mar 2013 21:19:12 +0000 (14:19 -0700)]
activity_log: value initialize hwprops_ members
Without this, the hwprops_ structure is default initialized (with
an indeterminate initial value), which can cause ActivityLogTest.SimpleTest
to fail. EncodeHardwareProperties() will complain and abort if any of
the hwprops_ doubles are NaN values.
BUG=chromium:211445
TEST=FEATURES=test emerge-x86-generic gestures
Change-Id: I7c7210ce48d255da55337cfb9c9f7d4ea27b1534
Reviewed-on: https://gerrit.chromium.org/gerrit/46802
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Gaurav Shah <gauravsh@chromium.org>
Tested-by: Gaurav Shah <gauravsh@chromium.org>
Andrew de los Reyes [Wed, 27 Mar 2013 22:31:59 +0000 (15:31 -0700)]
replay_log: Support for new .zip report logs
BUG=chromium:224451
TEST=Manually ran tool to fetch logs
Change-Id: Id5319d37e8dd6cefa5e1a69b3e84f27d84c0fa7c
Reviewed-on: https://gerrit.chromium.org/gerrit/46704
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Dennis Kempin [Tue, 26 Mar 2013 22:15:27 +0000 (15:15 -0700)]
click_wiggle_filter: Don't look at pressure
We used to look at pressure to only block wiggle from the pressing
finger, however resting fingers are causing wiggling too.
This CL removes the pressure logic.
This change introduces minor regressions as it is influencing the
movement distance in drag situations. This is not noticable in day to
day use.
BUG=chromium:222228
TEST=link/right_click_no_wiggle
Change-Id: I72fcfe27e9b9273942f00ed4e4a90ee04337ce43
Reviewed-on: https://gerrit.chromium.org/gerrit/46563
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Tue, 26 Mar 2013 21:22:35 +0000 (14:22 -0700)]
immediate_interpreter: more solid clicking with thumb
clicking with a thumb and a pointing finger now more reliably causes
left clicks instead of right clicks.
BUG=chromium:22837
TEST=link/left_click_with_resting_thumb
Change-Id: I2240a1529ffb108eca0fdfbe4af0681ed1ecdf37
Reviewed-on: https://gerrit.chromium.org/gerrit/46551
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Tue, 26 Mar 2013 18:42:33 +0000 (11:42 -0700)]
immediate_interpreter: use T5R2 logic for semi MT as well
SemiMT touchpads will use the same logic as T5R2 touchpads and use
simple counting of touches to determine the type of click.
To make things work we need a workaround to not
filter low pressure contacts when the button is pressed. The pressing
fingers are often reported as low (or even 0) pressure contacts.
BUG=chromium:223990
TEST=mario/3_finger_click
Change-Id: I5180cffb4ae3c2ab6e1a13883bc69e082b30843a
Reviewed-on: https://gerrit.chromium.org/gerrit/46539
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Tue, 26 Mar 2013 01:18:58 +0000 (18:18 -0700)]
immediate_interpreter: Try to interpret 2F gestures when 3F failed
In case we missed a thumb we might end up with 3 fingers when only
2 fingers are gesturing. If the 3F gesture recognition cannot find a
gesture and one of the fingers is in the dampened zone, try to
recognize a 2F gesture.
This fixes a problem with 2F scrolls with resting thumb when the
pressure differences between the fingers are not strong enough.
BUG=chromium:223677
TEST=link/scroll_false_negatives
Change-Id: I44222b36381686ed1cb4cd80b67ab45204a60b91
Reviewed-on: https://gerrit.chromium.org/gerrit/46474
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Mon, 25 Mar 2013 23:36:03 +0000 (16:36 -0700)]
immediate_interpreter: updated pinch logic to make it harder to lock
In some rare cases the pinch logic is treating resting thumb + pointing
finger as a pinch gesture. This CL makes it harder to enter the gesture
to prevent these issues. However entering pinch-to-zoom is a little
flaky with this logic. As pinch is not used yet, it will have to do for
now.
BUG=chromium:221098
TEST=link/pinch_false_positive_1-2
Change-Id: I0aaaf1e767eb8a5fc9a13667a57090ffadd27043
Reviewed-on: https://gerrit.chromium.org/gerrit/46458
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Andrew de los Reyes [Sat, 23 Mar 2013 00:54:46 +0000 (17:54 -0700)]
MultitouchMouseInterpreter: Scrolling fixes
Return both types of gesture from this device.
BUG=chromium:218453
TEST=Manually tested on device. Will add regression tests.
Change-Id: I2de16f370cf5a107dc654b7725973fc97a0d1455
Reviewed-on: https://gerrit.chromium.org/gerrit/46345
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Che-Liang Chiou <clchiou@chromium.org>
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Andrew de los Reyes [Fri, 22 Mar 2013 21:51:35 +0000 (14:51 -0700)]
AccelFilterInterpreter: Ability to scale pointing with mouse curves
This is needed as we bring mouse support (starting with Magic Mouse)
into this codebase.
BUG=chromium:218453
TEST=tested manually; will add regression tests
Change-Id: I80f0309209b95bba5335f3969c71590f3dcee053
Reviewed-on: https://gerrit.chromium.org/gerrit/46316
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Che-Liang Chiou <clchiou@chromium.org>
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Che-Liang Chiou [Fri, 22 Mar 2013 01:59:42 +0000 (18:59 -0700)]
AccelFilterInterpreter: Distinguish sensitivities of multitouch mice
For multitouch mice, they have two sensitivities: One for pointer and
one for scroll. (For touchpads, these two are the same.)
BUG=chromium-os:36322
TEST=Connect Magic Mouse, manually set 'Touchpad Speed' and 'Mouse
Speed' on chrome://settings, and make sure it works.
Change-Id: I322af697cea0114b686e58c047b62b996e57f760
Reviewed-on: https://gerrit.chromium.org/gerrit/46214
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Che-Liang Chiou <clchiou@chromium.org>
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Dennis Kempin [Fri, 22 Mar 2013 23:37:06 +0000 (16:37 -0700)]
refactoring: Allow multiple gestures to be returned on every sync
The Gesture structure is updated to have a next pointer to build a list
of gestures. All filters have been updated to provess all gestures in
the chain. This change introduces an alternative to CombineGestures
called AppendGestures, which appends a gesture instead of discarding the
lowest priority one.
This behavior is tested by adjusting the fling stop filter to append
gestures instead of combining them. This allows it to send fling stops
ahead of button events.
BUG=chromium:219870
TEST=link/fling_stop_order
also as this is a refactoring make sure to check for regressions.
Change-Id: I106816d7f8601d5b08bbda4d3247e8bee5236ca6
Reviewed-on: https://gerrit.chromium.org/gerrit/46327
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Che-Liang Chiou <clchiou@chromium.org>
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Charlie Mooney [Mon, 18 Mar 2013 21:20:25 +0000 (14:20 -0700)]
Slow down scroll acceleration for high-sensitivity
If you turn up the pointer sensitivity, the scrolling speed would go
pretty crazy, making it almost unusable because it would just fling all
the way to the bottom of whatever page you were using no matter what.
The idea that the scroll needs to go faster for people who turn up their
sensitivity may be a mistake entirely, but for now this just limits the
speedup a lot.
The default sensitivity (3) is untouched, but the speedup when you set
it to 4 or 5 has been reduced.
BUG=chromium-os:38142
TEST=trybots, gmerged onto a Link and manually tested and it feels far
less ridiculous than before
Change-Id: Ifa4c0156d2866e194609b90c13ad82fdf7bfd140
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/45783
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Dennis Kempin [Tue, 19 Mar 2013 22:56:04 +0000 (15:56 -0700)]
immediate_interpreter: Enable 3F click per default
It's working well enough now.
BUG=chromium:221468
TEST=none
Change-Id: Ie4377e47cfff1dfd4584ffa8627ad9e30688567d
Reviewed-on: https://gerrit.chromium.org/gerrit/45906
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Tue, 19 Mar 2013 22:25:45 +0000 (15:25 -0700)]
immediate_interpreter: Button Timeout logic updated
To make the button click detection more responsive the timeout has been
split into two different timeouts. First a timeout after the button down
to wait for fingers to arrive (which is a system caused delay) and for
every change in fingers another timeout to wait for more fingers to arrive
(which is a user caused delay).
This makes button clicks more responsive and solves the issue of
temporary finger merging, as we wait after each change of fingers.
BUG=chromium:196321
TEST=link/button_click_merge_issue
Change-Id: I74aee12742d085fa61bbed4369507b150187983a
Reviewed-on: https://gerrit.chromium.org/gerrit/45903
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Che-Liang Chiou [Wed, 13 Mar 2013 23:02:54 +0000 (16:02 -0700)]
MultitouchMouseInterpreter: Reuse scroll and fling logic
Let MultitouchMouseInterpreter reuse the scroll and fling logic of
ImmediateInterpreter.
BUG=chromium-os:36322
TEST=All unit tests are passed, and we may fling Magic Mouse on Snow
NOTE: You need to set 'Fling Buffer Suppress Zero Length Scrolls'
to 1 to test Magic Mouse
Change-Id: I087a8192a5c80225b96e5ac2efdf7b5e89b81ace
Reviewed-on: https://gerrit.chromium.org/gerrit/45363
Reviewed-by: Che-Liang Chiou <clchiou@chromium.org>
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Che-Liang Chiou [Wed, 13 Mar 2013 18:41:38 +0000 (11:41 -0700)]
ImmediateInterpreter: Factor out scroll and fling logic
Factor out the scroll and fling logic from ImmediateInterpreter as we
are trying to reuse them in MultitouchMouseInterpreter.
We add a manager class that are "basically" stateless, meaning that it
does not have "real" mutable states, as all its members are properties.
(The only exception is prev_result_high_pressure_change_.)
We also add a HardwareStateBuffer class, which acts as a common data
object that is passed from ImmediateInterpreter to the manager class.
BUG=chromium-os:36322
TEST=pass all unit tests
Change-Id: I444bbdfa6d290f9f5e386cf8b2eccb18aab2722b
Reviewed-on: https://gerrit.chromium.org/gerrit/45327
Reviewed-by: Che-Liang Chiou <clchiou@chromium.org>
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Dennis Kempin [Wed, 13 Mar 2013 21:26:57 +0000 (14:26 -0700)]
immediate_interpreter: Reduced TTC log verbosity
The Tap-To-Click log verbosity is very high. Barely any other part of
immediate_interpreter is printing log messages except tap to click.
Even when taps are not used, the code is sending multiple log messages
every update. This CL reduces the log output to print only relevant
messages when things have changed or the TTC state machine is leaving
the idle state. This reduces the amont of log messages by a factor of
approx 30 which makes it easier to find relevant log messages for
debugging.
BUG=chromium:190502
TEST=run touchtests with -vgl to check the updated gestures log
Change-Id: I9d70e4bd730941cfebb4b2c510aa179b462e4829
Reviewed-on: https://gerrit.chromium.org/gerrit/45346
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Mon, 11 Mar 2013 23:44:08 +0000 (16:44 -0700)]
ImmediateInterpreter: New logic for EvaluateButtonType
Implemented a new logic in order to support 3 finger = middle clicks
as well as more variations of left and right clicks.
The new logic is largely based on determining the button type by looking
at wether the fingers recently touched down, have been on the touchpad
for a while but not moved or have been moved.
For unclear cases it falls back to a location based separation of the
fingers to determine if they belong to the same group of fingers.
This CL does not introduce any regressions in the test suite, but as a
side-effect improves the scoring of some existing button tests.
BUG=chromium-os:39368
TEST=added test cases to touchtests:
link/button_middle_click_variations
link/button_right_click_variations
Change-Id: I75eef2ddaf5427c46d396188aafae329340f5166
Reviewed-on: https://gerrit.chromium.org/gerrit/45152
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Andrew de los Reyes [Mon, 18 Mar 2013 21:07:13 +0000 (14:07 -0700)]
LoggingFilterInterpreter: Add 'Integrated Touchpad' property.
We don't have a place to put properties that aren't used by our code,
so I've stuck this property in the logger, which is always used.
This property is used by X (in another patch) to see if a touchpad is
integrated, as that can affect pointer motion on a rotated display.
BUG=chromium:196434
TEST=Tested manually on Link
Change-Id: If4d90d46bc9e972dab38aeae28777d8e890540dc
Reviewed-on: https://gerrit.chromium.org/gerrit/45800
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Charlie Mooney [Mon, 18 Mar 2013 15:17:13 +0000 (08:17 -0700)]
Gestures: Update Fling Buffer Depth to 10
We ran another Salsa experiment to see if increasing the fling buffer
depth even further would make people even happier with it. It looks
like they are. The results aren't as stark as last time but with a
p value around 0.25 we can be pretty sure that a depth of 10 is better
than 7, and even in the unlikely case that it's not most people who
don't think it's better can't tell the difference between them.
Additionally, the scroll buffer needs to now track 20 events since Mario
requires a double-long buffer. For it to be able to have the same
performance, it's fling depth will be set to 20 in it's custom config,
but if the buffer doesn't even hold 20 values it won't work. Having
this larger discrepancy exposed a bug in the way scroll speed was
computed which was fixed. It assumed the entire buffer was used which
is not always the case.
Basically this should just change the default depth from 7 to 10,
slightly increasing the consistency of flings on ChromeOS.
BUG=chromium:216790
TEST=Salsa experiment showed that this was prefferable to the current
default value, regression tests showed that it didn't break anything. A
pair of tests that checked the fling distance did have to be updated
though to reflect the new fling config.
CQ-DEPEND:I0e50560fa8f3ba2374a20d123902588e924b3640
I87a4eb4e42945be68edfa511c4098230eabe004f
Change-Id: Ib985b987d65325140b088a844ef5f8c0ad2b88f7
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/45744
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Chung-yih Wang [Tue, 5 Mar 2013 06:14:56 +0000 (14:14 +0800)]
fix incorrect pointer references
Incorrect pointer usage may crash system.
BUG=chromium-os:37272
TEST=make no difference
Change-Id: I999383f3505510db4762ff7a090bff348c898f80
Reviewed-on: https://gerrit.chromium.org/gerrit/44618
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Chung-yih Wang <cywang@chromium.org>
Tested-by: Chung-yih Wang <cywang@chromium.org>
Charlie Mooney [Mon, 4 Mar 2013 23:59:33 +0000 (15:59 -0800)]
Remove Salsa from the gestures project
Salsa has its own repo now in platform/salsa so it shouldn't be here
in gestures. This CL removes the salsa files from this repo, because
they're all copied over to their new home.
TEST=emerged gestures, it still made correctly
BUG=chromium-os:39493
Change-Id: I9d30e467f1244b644258294bc2d2e532b72212fb
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/44584
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Mon, 4 Mar 2013 17:05:24 +0000 (09:05 -0800)]
Revert "ImmediateInterpreter: Handle TTC w/ possible palms present"
This reverts commit
e6726e872fc08b6c2f8afd49929f28d0d05372d0
Looks like this has a crasher in it
Change-Id: Idf44cbcefce3bc9d210c0bec4bfc03dee7a806d4
Reviewed-on: https://gerrit.chromium.org/gerrit/44526
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 19:54:13 +0000 (11:54 -0800)]
PalmClassifyingInterpreter: Note why a contact is not a palm, avoid bad moves.
With this change, when a contact is declared to not be a palm, we
record why that is. Also, if a contact was marked non-palm because
another finger was close by, but then that finger leaves, we suppress
further motion from that palm until it moves significantly, which is
the same behavior we would have had if the other finger had never
shown up.
This needed because we have cases where a real palm is close to
another finger (since we allow up to 50mm of separation, at the time
of writing), and if a finger shows up, the palm can get classified as
non-palm and then cause movement.
BUG=chromium-os:39231
TEST=link/ttc_with_palm regression test
Change-Id: Idd60aef6d26b5410730c772db42886cb6af9f14e
Reviewed-on: https://gerrit.chromium.org/gerrit/44234
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 19:51:52 +0000 (11:51 -0800)]
ImmediateInterpreter: Only set last_movement_timestamp_ on real movement.
Historically we only looked for the currently set gesture mode, not
actual movement. This broke a tap-to-click case where a stationary
palm was keeping the gesture type as movement, but no movement occured
while another finger did a tap click.
BUG=chromium-os:39231
TEST=link/ttc_with_palm regression test
Change-Id: I9710c6762b283f6e88d3167b1aeadc995328712a
Reviewed-on: https://gerrit.chromium.org/gerrit/44233
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 21:21:27 +0000 (13:21 -0800)]
ImmediateInterpreter test SemiMtNoPinchTest: Handle NULL gesture.
BUG=chromium-os:39320
TEST=This is the unittest
Change-Id: I5438d47605322402602932e98b56324973cde2e2
Reviewed-on: https://gerrit.chromium.org/gerrit/44232
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 21:37:27 +0000 (13:37 -0800)]
ImmediateInterpreter unittest: Relax AvoidAccidentalPinchTest
We were requiring Move gestures, even when such moves would have been
0-magnitude. Releax our requirements for these instances.
BUG=chromium-os:39319
TEST=This is the test fix
Change-Id: Icb71dc78d0287a00da258337c04cf4d65cdb9c01
Reviewed-on: https://gerrit.chromium.org/gerrit/44231
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 21:33:20 +0000 (13:33 -0800)]
ImmediateInterpreter: Favor non-possible-palm for pointing
During other cleanup, I realized unittest AmbiguousPalmCoScrollTest
was broken.
BUG=chromium-os:39318
TEST=This is the test fix
Change-Id: Ibc7cdb74eedba5ef4d335758593479a4104022fc
Reviewed-on: https://gerrit.chromium.org/gerrit/44230
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 19:48:53 +0000 (11:48 -0800)]
ImmediateInterpreter: Handle TTC w/ possible palms present
When a possible palm arrives at the same time as a tapping finger,
then the finger departs but the palm stays, we historically would have
considered the possible palm as tapping and waited for a possible
two-finger tap. That's fine if the possible palm is not a palm, but
when it is, this logic breaks.
With this change, we can timeout after the one finger leaves and do a
left click with that finger.
BUG=chromium-os:39231
TEST=link/ttc_with_palm regression test
Change-Id: Ice78c556cba31ed70db46ab126d1a1a048b75dc9
Reviewed-on: https://gerrit.chromium.org/gerrit/44229
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Che-Liang Chiou [Tue, 26 Feb 2013 19:26:39 +0000 (11:26 -0800)]
MultitouchMouseInterpreter: Implement basic gestures
The basic gestures include mouse movement, button clicking, and single
finger scrolling.
A problem we found for multi-touch mice is that when you are clicking,
your finger might slide a little bit on the surface, which is then
interpreted as a single finger scrolling. To alleviate this problem,
ClickWiggleFilterInterpreter and BoxFilterInterpreter are added to the
chain of filters.
BUG=chromium-os:36322
TEST=Test Magic Mouse: mouse movement, button clicking, single finger
scrolling
Change-Id: Ief71a5a3b54896f68e2d078ca366171113755f40
Reviewed-on: https://gerrit.chromium.org/gerrit/44071
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Charlie Mooney [Wed, 27 Feb 2013 20:36:55 +0000 (12:36 -0800)]
Gestures: Improve fling velocity consistency
The fling velocity calculation is quite inconsistent. When the Touchbot
was programmed to reproduce the same exact fling, the results were all
over the place.
This increases the fling buffer depth from 3 to 7, so now it uses the
last 7 scroll events instead of 3 to estimate the fling trajectory. The
robot produced much more clumped results with this, and a Salsa
experiment was preformed to compare 3, 5, 7, and 15 to see which people
preferred. The support for 7 over 3 was overwhelming, 15 got a few bad
reviews and it pretty extreme, so 7 is a safe and effective compromise.
raw results 3 5 7 15
------------------------------
participant #1 3 2 0 1
participant #2 3 2 1 0
participant #3 3 0 0 2
participant #4 1 3 0 1
participant #5 3 2 1 0
participant #6 2 1 0 3
Sign-test P Values (a low number indicates that the treatment in that
row is likely better than the treatment in its column)
5 3 15 7
5 0.109375 0.890625 1.0
3 0.984375 0.96875 1.0
15 0.34375 0.1875 0.890625
7 0.03125 0.015625 0.34375
BUG=chromium-os:39302
TEST=Salsa experiment and robot tests
Change-Id: I5f876b862e5118647c4e75c1b8480f6ad23cd57e
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/44205
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 19:47:34 +0000 (11:47 -0800)]
ImmediateInterpreter: Merging fingers cancel tapping.
During other fixes, merged fingers on Link started causing tap
clicks. This change prohibits them.
BUG=chromium-os:39314
TEST=link/finger_merge* regression tests
Change-Id: I652401b2427381eda6999f1b941a3e787d18cb98
Reviewed-on: https://gerrit.chromium.org/gerrit/44228
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 19:34:22 +0000 (11:34 -0800)]
ImmediateInterpreter: Don't let clicking fingers ever tap.
During other cleanup, I've run into a case where a finger is
classified as palm during a physical click, then it is un-flagged,
then does a tap.
With this change, we record the IDs of all contacts that are present
during a physical button being down. We never let these tap.
BUG=chromium-os:39313
TEST=mario/base_click regression test
Change-Id: I59bc575b6844399623ce1f7113dd0825df33c2d8
Reviewed-on: https://gerrit.chromium.org/gerrit/44227
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 27 Feb 2013 19:27:06 +0000 (11:27 -0800)]
gestures.cc: Fix logging for MERGE finger flag
BUG=chromium-os:39312
TEST=While logging, saw this value was properly logged.
Change-Id: Ieed72923f19218c1b82ab7b9c621cb580e6d18e6
Reviewed-on: https://gerrit.chromium.org/gerrit/44226
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Mike Frysinger [Wed, 27 Feb 2013 23:01:15 +0000 (18:01 -0500)]
drop pixman usage
I can't find any references to pixman in the current tree, nor can I see
how the commit that added pixman (
1b5f1d55edc81ec9384652a1da42d93ee668eb7a)
used it either -- that was focused on logging. Punt it.
BUG=None
TEST=`emerge-daisy gestures` still works
TEST=`cros_run_unit_tests --board=stumpy -p gestures` still works
Change-Id: If3ef41947d64ac906cc357eaaeea823fa759d3cf
Reviewed-on: https://gerrit.chromium.org/gerrit/44237
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Mike Frysinger <vapier@chromium.org>
Tested-by: Mike Frysinger <vapier@chromium.org>
Dennis Kempin [Wed, 27 Feb 2013 19:34:45 +0000 (11:34 -0800)]
immediate_interpreter: improved detection of left/right button click
The logic for determining ambiguous cases of physical button clicks
has been updated to support both one-handed right clicks with two
fingers as well as two handed left clicks with two fingers.
This CL does not introduce any regressions in touchtests but
fixes 2 right click regressions from previous CLs.
BUG=chromium-os:38276
TEST=3 new test cases available in touchpad-tests
Unit tests pass
Change-Id: Ifb6b97fa43f45bb68a9ccd21e77ac1f0f1c1b013
Reviewed-on: https://gerrit.chromium.org/gerrit/44213
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Charlie Mooney [Wed, 27 Feb 2013 18:42:41 +0000 (10:42 -0800)]
Remove a stray debugging Log command from Gestures
This Log() command accidentally got left in, but it doesn't belong
there. It was just for debugging purposes while developing.
BUG=none
TEST=non-functional change, just removing a Log()
Change-Id: I93fd18a7cd742e24043cc36910a085c0c7da4dbc
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/44187
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Che-Liang Chiou [Tue, 26 Feb 2013 22:09:57 +0000 (14:09 -0800)]
MouseInterpreter: Scroll in correct direction
When your finger is moving downward on scroll wheel, it should be
interpreted as scrolling down.
BUG=chromium-os:36322
TEST=mouse scroll in correction direction
Change-Id: I131078c4d1ed7558930756b0bf36629f59bbe3c2
Reviewed-on: https://gerrit.chromium.org/gerrit/44106
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Dennis Kempin [Tue, 26 Feb 2013 19:16:44 +0000 (11:16 -0800)]
Makefile: in-place now installs locally into in-place folder
instead of installing into the global chroot folders. This keeps things
cleaner.
Note: The in-place folder is necessary as the gestures.h file is
searched for at the path gestures/gestures.h, but resides in
include/gestures.h.
BUG=chromium-os:39158
TEST=remake touchtests
Change-Id: I57c3679ca7e602fe3ad1d312e168d9f2e4ff1717
Reviewed-on: https://gerrit.chromium.org/gerrit/44067
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Mon, 25 Feb 2013 18:49:19 +0000 (10:49 -0800)]
immediate_interpreter: More generous tap-to-drag properties
Tweaked the tap-to-drag properties to make it easier to enter
the tap-to-drag state.
Zero regression in the test suite.
BUG=chromium-os:39072
TEST=added test case to touchtests
Change-Id: Iec99e4cd1a08ae10cd0ac90b7e92ed139365285a
Reviewed-on: https://gerrit.chromium.org/gerrit/43914
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Che-Liang Chiou [Tue, 26 Feb 2013 01:31:12 +0000 (17:31 -0800)]
LookaheadFilterInterpreter: Copy mouse states
This filter forgot to copy mouse states to its queue; so copy it!
BUG=chromium-os:36322
TEST=Magic Mouse works on Snow
Change-Id: I178e9301aee6473ea9f9cd05bb4606ef517857f5
Reviewed-on: https://gerrit.chromium.org/gerrit/44004
Tested-by: Che-Liang Chiou <clchiou@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Che-Liang Chiou <clchiou@chromium.org>
Charlie Mooney [Mon, 25 Feb 2013 16:32:36 +0000 (08:32 -0800)]
Salsa: Fix feedback form to accept long strings
The old free form feedback box could only accept strings up to 500
characters. This changes things up so it can accept upto 1mb per
feedback form, which is hopefully far more than anyone would bother to
type for us. It also adds a counter so they can see how many characters
they have left. I've capped it at 800, so participants don't get out of
hand with their feedback.
BUG=chromium-os:39172
TEST=manually tested using a dev server. I was able to append new
feedback onto existing lists, including feedback >500 bytes and the
count gets updated automatically as I typed, capping me at 800.
Change-Id: If2eddcf0fdc95da579e368fb0ee36effedf47f10
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/43909
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Fri, 18 Jan 2013 20:38:40 +0000 (12:38 -0800)]
LookaheadFilterInterpreter: Look for possible finger liftoff
This is particularly needed on Butterly, where we often see the cursor
jump when a finger lifts off. Normally such a motion would be blocked
by a high-pressure-change filter, but since the pressure values are so
noisy, we turned that off on Butterfly.
With this change, we add a new change to the lookahead filter: If a
finger looks like it may be lifting off, it can add a bit of delay. If
the finger indeed does, the motion will be squashed, otherwise it will
be tolerated.
BUG=chromium-os:36699
TEST=butterfly2.4/move_jumped_at_end regression test
Change-Id: I10d3b8cbaad9503dd472182f34ef0e29daa90fb4
Reviewed-on: https://gerrit.chromium.org/gerrit/41683
Reviewed-by: Benson Leung <bleung@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Tue, 12 Feb 2013 21:53:54 +0000 (13:53 -0800)]
SplitCorrectingFilterInterpreter: Enable based on property
BUG=chromium-os:38848
TEST=unittest pass; no change in regression test scores
Change-Id: Ie7ad4adbb284d6d24e420b452c7c67c0a079c692
Reviewed-on: https://gerrit.chromium.org/gerrit/43126
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Charlie Mooney [Wed, 6 Feb 2013 19:23:29 +0000 (11:23 -0800)]
Salsa: Update the try_touch_experiment UI
This changes the way the participants actually apply the treatments.
Previously, they had to copy paste each treatment and it became clear
that this was a huge pain. Similarly, the UI once you had copied it
into Crosh was confusing.
This change replaces the individual treatment commands with one giant
command you copy for the entire experiment. This command includes
information on *all* of the treatments and if you pass it into the new
try_touch_experiment program it will allow you to switch back and
forth between the treatments however you please with a simple curses
interface. Beyond UI, it checks repeatedly now to make sure that
everything was applied correctly, hopefully getting rid of any errors in
how the treatments are viewed.
Hopefully this makes it easier for users to compare the experiments and
flip back and forth between them, etc.
try_touch_experiment will replace the command that is currently just
a function in Crosh. Crosh will be patched so that the command now just
runs this instead.
BUG=chromium-os:38637
TEST=For the server-side changes, a dev server is running at
cmooney.mtv:8081 where you can see the differences. Manual testing of
the new try_touch_experiment was done on a Link to confirm it works there.
Change-Id: Ia0090629e8d988e8f517ff83ceac03a9bf808c19
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/42772
Dennis Kempin [Fri, 22 Feb 2013 21:03:28 +0000 (13:03 -0800)]
Added in-place makefile target
Gestures can now be built 'in-place' too. Setup-in-place will place
symlinks to the sourcefolder and in-place will rebuild the library.
On order to be able to use the local dynamic library the library needs
to link to -lgconv. This only affects local install, not ebuild
installs.
BUG=chromium-os:38304
TEST=see touchpad tests
Change-Id: If759d4e9261d5359c974246324e72fb2944bd422
Reviewed-on: https://gerrit.chromium.org/gerrit/43818
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Charlie Mooney [Thu, 21 Feb 2013 17:15:37 +0000 (09:15 -0800)]
Salsa: Fix array-out-of-bounds on malformed models
Just ran into a problem where the software assumes that everyone has
submitted feedback. This would be true since everyone would at least
get an empty string if the experiment had been created when free-form
feedback was implemented. However, since the "fling" experiment was
created before that, some people don't even have that which causes an
error when trying to submit now.
This patch just checks for this and fills out the list with empty
strings if it needs to.
TEST=Manually tested on Dev server by making malformed experiments and
submitting to them, and it worked.
BUG=chromium-os:39101
Change-Id: Ib877ece2937466a380a2529eb1b8f45560ff7553
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/43720
Reviewed-by: Yufeng Shen <miletus@chromium.org>
Chris Masone [Fri, 15 Feb 2013 00:51:04 +0000 (16:51 -0800)]
Update gestures to build against libchrome-180609
The interface to JSONWriter::Write() changed, so to get pretty-printed
JSON we had to change to WriteWithOptions() in some places.
Also, FilePath is in the base:: namespace now.
CQ-DEPEND=If8868bd339480f044263c02f59126d335318ab7a
BUG=chromium-os:38938
TEST=FEATURE=test emerge-x86-mario gestures
STATUS=Fixed
Change-Id: I6b193cf098e3bfa9b109e457a4a04f843357aecf
Reviewed-on: https://gerrit.chromium.org/gerrit/43363
Commit-Queue: Chris Masone <cmasone@chromium.org>
Tested-by: Chris Masone <cmasone@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 13 Feb 2013 22:00:23 +0000 (14:00 -0800)]
Activity Logs: Don't leak memory.
Dave Moore noticed X was leaking memory when logs were generated, and
valgrind identified the culprits.
BUG=chromium-os:38880
TEST=valgrind said leaks were happening, but now it isn't
Change-Id: I6c2c5ff3a44f1904b0921672153ff22110da1815
Reviewed-on: https://gerrit.chromium.org/gerrit/43211
Reviewed-by: Dave Moore <davemoore@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Charlie Mooney [Tue, 12 Feb 2013 22:06:06 +0000 (14:06 -0800)]
Salsa: Making the ranking widget easier to use
Previously several people said they were initially confused that they
needed to drag the treatments over into the gray box next to them to
preform the rankings. This adds a little arrow and some formatting to
make this more obvious. The arrow dissapears once the last treatment
has been added to the ranking box indicating that they have finished
ranking all the tasks. Hopefully this is more clear to the users.
BUG=chromium-os:38638
TEST=A dev server with this change is running at cmooney.mtv:8082 and if
you go to the /participate page for an experiment it works well and
looks clearer to me.
Change-Id: Idfec41cf37bb79336f62a77255fdaa298865d615
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/43178
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Tue, 12 Feb 2013 17:52:02 +0000 (09:52 -0800)]
FingerMergeFilterInterpreter: Make stricter
These changes make the filter much stricter, so that they don't
trigger on normal movement.
BUG=chromium-os:37272
TEST=Regression tests link/merge_{1..5} pass. Manual testing on device.
Change-Id: Ib248196e6db1642055c5fe88ca237a6b005662d1
Reviewed-on: https://gerrit.chromium.org/gerrit/43098
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Dennis Kempin [Fri, 8 Feb 2013 21:41:06 +0000 (13:41 -0800)]
TPView now lives in platform/mttools/mtedit
Removed all files of tpview.
To use tpview again:
$ cd platform/mttools
$ make
$ sudo make install
$ mtedit -s
BUG=chromium-os:38686
TEST=none
Change-Id: I576ff86a47ca5ba2ae510460fb90472ac4d6d270
Reviewed-on: https://gerrit.chromium.org/gerrit/42948
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Tue, 5 Feb 2013 22:20:07 +0000 (14:20 -0800)]
TPView server: Don't chdir
The SimpleHTTPServer resolves all files in the local working directory.
This CL overrides this behavior to allow the server to be run
from any directory.
This fixes the -o parameter that was writing results to the tpview
folder instead of the local folder.
BUG=chromium-os:38304
TEST=execute:
$ cd somewhere
$ tpview http:// -d -o test.log
$ ls test.log
Change-Id: I35e2fd9b3bb54329878f2d5910f3e36b8eaf4e35
Reviewed-on: https://gerrit.chromium.org/gerrit/42876
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Andrew de los Reyes [Thu, 29 Nov 2012 20:40:27 +0000 (12:40 -0800)]
Movement gestures: Keep ordinal (non-accelerated) values
Generally, accelerating movements in a good idea, but sometimes it
makes sense to use unaccelerated values. With this CL, we keep both
types of values in the movement gestures.
BUG=chromium-os:36802
TEST=Tested on device. Regression tests and unittests pass
Change-Id: I7f54bb9d55e51186d499d12beddafc270766fd5f
Reviewed-on: https://gerrit.chromium.org/gerrit/38946
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Dennis Kempin [Mon, 4 Feb 2013 22:36:11 +0000 (14:36 -0800)]
TPView: Created TPView command line tool
The tpview command line tool can be used for viewing and editing
touchpad log file. These log files can be local files or automatically
downloaded from a chromebook or obtained from a feedback report.
BUG=chromium-os:38304
TEST=run
$ make install-tpview
$ tpview -h
Change-Id: Iba0c89ff953a3bafa207168e4b995bba2c0712ff
Reviewed-on: https://gerrit.chromium.org/gerrit/42563
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Benson Leung [Wed, 6 Feb 2013 20:42:09 +0000 (12:42 -0800)]
Gestures : Remove Touch Noise Monitor from Gestures.
This includes the following reverts:
Revert "touch_noise_monitor: periodically send touch event crash reports"
This reverts commit
92764020c29c4d6c36704a4ef722ac4799cda9cc.
Revert "Add touch_noise_monitor"
This reverts commit
151228fb750c5b6679e72e7ef450e4e47cdf02b7.
Signed-off-by: Benson Leung <bleung@chromium.org>
BUG=chrome-os-partner:16955
TEST=check that /usr/sbin/touch_noise_monitor does not exist
Change-Id: I8b935219fbf14b4455919dead0ab16f5ebe81e73
Reviewed-on: https://gerrit.chromium.org/gerrit/42763
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Charlie Mooney [Tue, 5 Feb 2013 23:02:06 +0000 (15:02 -0800)]
Salsa: Adding a free-form feedback box
This adds a little text area on the submission form for people taking
the experiments so that they can send freeform feedback about the
treatments. If they noticed and odd behavior, etc.
BUG=chromium-os:38634
TEST=Running on a dev server on my computer at cmooney.mtv:8081
Change-Id: I9f97b35d4af4a0e7ce2b16e33c2f778836d419af
Signed-off-by: Charlie Mooney <charliemooney@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/42659
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Andrew de los Reyes [Wed, 6 Feb 2013 01:51:36 +0000 (17:51 -0800)]
FingerMergeFilterInterpreter: Disable by default
Too much real motion is getting caught by this filter.
BUG=chromium-os:38651
TEST=Manually replayed problematic logs. Will upload regression test
soon.
Change-Id: Ic3db69904bce5a71b785790d1c4eaa66918f4b81
Reviewed-on: https://gerrit.chromium.org/gerrit/42682
Reviewed-by: Benson Leung <bleung@chromium.org>
Commit-Queue: Andrew de los Reyes <adlr@chromium.org>
Tested-by: Andrew de los Reyes <adlr@chromium.org>
Dennis Kempin [Mon, 28 Jan 2013 23:38:22 +0000 (15:38 -0800)]
TPView: Python server for loading/saving logs on the workstation
This CL extends tpview by a python server which can read/save log files
on the workstation. By accessing /edit/logname TPView will load the log
file tpview/logs/logfile.log and after shrinking save it as
/tpview/logs/logfile.log.trimmed.
TEST=python server.py
BUG=chromium-os:38304
Change-Id: I219f8950273e4027f1501d223362a25c02a9674a
Reviewed-on: https://gerrit.chromium.org/gerrit/42180
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Mon, 28 Jan 2013 21:13:52 +0000 (13:13 -0800)]
TPView: Markup cleanup
A little bit of HTML/CSS cleanup. Rearranged elements to make better use
of space. Cleanup before adding UI for saving and loading server-side
logs.
TEST=./tpview/serve.sh look! it's pretty! Not really, but let's settle
on "prettier".
BUG=chromium-os:38304
Change-Id: Ib77517ae226ef83ae58ae5036d5d885f4f25841e
Reviewed-on: https://gerrit.chromium.org/gerrit/42145
Reviewed-by: Dennis Kempin <denniskempin@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
Dennis Kempin [Sat, 26 Jan 2013 01:17:04 +0000 (17:17 -0800)]
palm_classifying_filter: Don't mark as palm if fingers are close
An exception is made to palm detection. Whenever a gesturing finger
is close to the supposed palm, the palm detection will not mark it as a
palm.
Test case in separate CL, 0 regressions in the test suite.
BUG=chromium-os:38126
TEST=touchtests run butterfly2.4/scroll_expected_but_motion
Change-Id: Ibe2b027cb59da650e58b2223356bca1ef834441f
Reviewed-on: https://gerrit.chromium.org/gerrit/42073
Reviewed-by: Andrew de los Reyes <adlr@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@chromium.org>
Tested-by: Dennis Kempin <denniskempin@chromium.org>
projects / chromiumos / platform / gestures.git / log