Grail

Merge lp://staging/~oif-team/grail/trunk.fixes into lp://staging/grail

  1. trunk.fixes
  2. Merge into trunk
Proposed by Henrik Rydberg
Status: Superseded
Proposed branch: lp://staging/~oif-team/grail/trunk.fixes
Merge into: lp://staging/grail
Diff against target: 613 lines (+162/-202)
11 files modified
src/evbuf.h (+1/-1)
src/gestures-drag.c (+14/-0)
src/gestures-pinch.c (+15/-1)
src/gestures-rotate.c (+15/-1)
src/gestures-tapping.c (+6/-2)
src/grail-api.c (+90/-162)
src/grail-event.c (+2/-16)
src/grail-gestures.c (+14/-7)
src/grail-gestures.h (+3/-2)
src/grail-impl.h (+2/-0)
src/grail-recognizer.h (+0/-10)
To merge this branch: bzr merge lp://staging/~oif-team/grail/trunk.fixes
Reviewer Review Type Date Requested Status
Chase Douglas Pending
Stephen M. Webb Pending
Review via email: mp+51210@code.staging.launchpad.net

This proposal supersedes a proposal from 2011-02-24.

This proposal has been superseded by a proposal from 2011-02-24.

Description of the change

A bunch of patches aiming to integrate the ideas of v1.0.19 more seamlessly into the grail code. This will fix the remaining tap bug and help further development. The only part not addressed by this branch is the sysflag, and a discussion around it is likely enough to see how to form a patch for it.

To post a comment you must log in.
Revision history for this message
Stephen M. Webb (bregma) wrote : Posted in a previous version of this proposal

A more elegant and nuanced way of handling the various timeout issues. Compiles and runs OK.

review: Approve
Revision history for this message
Chase Douglas (chasedouglas) wrote : Posted in a previous version of this proposal

In line 371 of the patch below, there's an ambiguity in the if statement "if (a || b && c)". This should be cleaned up.

I need to test this branch, but I think the changes look reasonable. However, I can't approve the request until it also handles the global flag. This is something I think we will really need in natty, since we're going to have more applications using gestures.

review: Needs Fixing
Revision history for this message
Chase Douglas (chasedouglas) wrote : Posted in a previous version of this proposal

Test results:

1. Regression: If you use a physical trackpad button to click while there's no movement on the trackpad, it's held up until a movement is performed.

2. Timeouts seem ok. I was able to request all 3, 4, and 5 finger gestures and still perform two finger multitouch reasonably. I will perform more thorough testing when the regressions are resolved.

3. Regression: I often get a tap event at the end of a drag gesture sequence. The tap touch count matches the drag touch count.

review: Needs Fixing
Revision history for this message
Henrik Rydberg (rydberg) wrote : Posted in a previous version of this proposal

Thank you very much for the tests! I have updated the branch with fixes for
the two problems found. I will add something for the global flag shortly.

Revision history for this message
Henrik Rydberg (rydberg) wrote :

Fix of the fix - the first regression was pretty complex, and revealed some
additional assumptions that were not true. Hopefully this version will do better.

lp://staging/~oif-team/grail/trunk.fixes updated
115. By Henrik Rydberg

Revert "Add button passthrough"

116. By Henrik Rydberg

Revert "Add global vs local gesture handling"

117. By Henrik Rydberg

Introduce gesture timeouts

The timeout of a gesture is integral to the gesture type.
This patch introduces timeout per gesture type, potentially
enabling more control over gesture timing aspects. The timeout
logic replaces the recently added recognition state.

118. By Henrik Rydberg

Introduce activation hold time

Do not activate a drag, pinch or rotate immediately, even
if the activation threshold is met. This allows for a smoother
transition between fingers.

119. By Henrik Rydberg

Wait for the right number of fingers before gesture timeout

A client listening only for multi-finger gestures will not get
the chance to recognize unless there is a gesture timeout. This
patch waits until the drag hold time has passed before releasing
the touch.

120. By Henrik Rydberg

Skip events selectively

Buffer all events as planned, but strip the stream of all events
involved with pointer and MT, rather than bluntly removing all
events. In practise, this means button and switch events are let
through, undistorted although possibly slightly delayed.

121. By Henrik Rydberg

Adjust gesture buffer sizes and timings

Adjust timings to put drag, pinch and rotate on more equal footing,
with regard to gesture timeouts. Also make sure events are not lost
during hold, now that all MT events are buffered as well.

122. By Henrik Rydberg

Reset tap state after timeout

A touch and hold will not properly reset the tap state, such that
the next tap fails to register. Fixed with this patch.

123. By Henrik Rydberg

Remove unused tapping state

This patch removes some cruft that was left over from the
removal of the short tap time logic.

124. By Henrik Rydberg

Make reset logic of pinch and drag same as for drag

There is no direct reason for one-fingered pinch/rotate gestures
to be discarded immediately, so only reset gestures on touch down,
as for drags.

125. By Henrik Rydberg

Make grabbing client exclusive

When a possible gesture is started for a global client (sysflag set),
it is desirable that gestures to other clients are disregarded until
all contacts have left the surface. In principle this could be refined
further, such that if gesture are not detected, they are passed on to
other clients, but that also affects how touches are passed onto lower
layers like XI2.1, so simply making sure a global system client gets
all events potentially leading to gestures is good enough for now.

Also note that in the current combination of synaptics/evdev X servers
and grail, there can be no overlapping potential gestures between two
global clients, or the resulting event recipient is not correct.

126. By Henrik Rydberg

Bump to v1.0.20

127. By Henrik Rydberg

Disable rotation gestures for semi-mt devices

To avoid poor user experience with semi-mt devices, disable
rotation detection, such that only drag and scale is emitted.

Signed-off-by: Chase Douglas <email address hidden>
Signed-off-by: Henrik Rydberg <email address hidden>

Unmerged revisions

Preview Diff

[H/L] Next/Prev Comment, [J/K] Next/Prev File, [N/P] Next/Prev Hunk
The diff is not available at this time. You can reload the page or download it.

Subscribers

People subscribed via source and target branches

to all changes: