Closed Bug 813934 Opened 12 years ago Closed 11 years ago

[B2G] Marionette touch swipe is inconsistent

Categories

(Remote Protocol :: Marionette, defect)

x86_64
Windows 7
defect
Not set
major

Tracking

(Not tracked)

RESOLVED WORKSFORME

People

(Reporter: zcampbell, Assigned: jlal)

References

Details

Attachments

(1 file, 1 obsolete file)

Attached file Gallery flick test case (obsolete) —
I am having not much luck using marionette flick method to swipe navigate through gallery items in the Gaia gallery.

I've tried several different variations of duration, swipe distance (back/forward) and so on but the flick action always starts but the new image will get 'stuck' somewhere between the transition. It's as if the transform process is cancelled and the image gets stuck halfway through the transform.

It doesn't get stuck in the same place every time, either.
Severity: normal → major
Yes I agree this is a valid major severity.

We'll need flick a lot in the future.
I have renamed this to be more generic and hopefully catch all cases.

We're suffering this in several apps/test cases:
Homescreen swipe
Cards view (ie kill app)
Gallery
Summary: [B2G] Marionette touch swipe is inconsistent on gallery nav → [B2G] Marionette touch swipe is inconsistent
fyi, I've been pulled away from this bug, but when I was investigating it, the 'mouseup' event wouldn't be triggered from the synthetic gestures lib. Right before mouseup code (https://hg.mozilla.org/mozilla-central/file/4be7dbc93bba/testing/marionette/client/marionette/touch/synthetic_gestures.js#l507), we stop execution. I didn't have a chance to investigate further.
I will give this a shot this week.
Assignee: nobody → jlal
OK, I had a very similar issue. Synthetic gesture takes an element but looks up whatever its in the path of that element to emit events on (as it changes coordinates it changes the target element). I suspect the coordinates are off a bit in the test-case or there is a lag between the test and the element itself.

If we where to literally "drag" the element as we move (as I suspect is happening in gallery) it probably makes this worse.
(In reply to James Lal [:lightsofapollo] from comment #5)
> OK, I had a very similar issue. Synthetic gesture takes an element but looks
> up whatever its in the path of that element to emit events on (as it changes
> coordinates it changes the target element). I suspect the coordinates are
> off a bit in the test-case or there is a lag between the test and the
> element itself.
> 

Hmm. For touch events, the target shoudl always be the same as the target of the touchstart event.  So if it is doing this for touch events, that might be the bug.
Try setting touchSupported to true on line 583, so that it sends touch events. Right now it is sending mouse events for all swipes, not just mouseswipe()
Is this bug still valid? I was able to get the attached test case to pass.
Attachment #683964 - Attachment is obsolete: true
Flags: needinfo?(zcampbell)
That test case works and so does another one I have for the home page. It may not work for everything but it's a big improvement so far.

I wish we could say when this started working.
Status: NEW → RESOLVED
Closed: 11 years ago
Flags: needinfo?(zcampbell)
Resolution: --- → FIXED
Resolution: FIXED → WORKSFORME
Product: Testing → Remote Protocol
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: