Titanium JIRA Archive
Titanium SDK/CLI (TIMOB)

[TIMOB-1524] iOS - flash/redraw issue on 3d transform/animation

GitHub Issuen/a
TypeBug
PriorityTrivial
StatusClosed
ResolutionWon't Fix
Resolution Date2011-05-17T14:40:57.000+0000
Affected Version/sn/a
Fix Version/sBacklog
ComponentsiOS
Labelsdefect, ios, regression
ReporterThomas Huelbert
AssigneeBlain Hamon
Created2011-04-15T02:55:02.000+0000
Updated2011-05-17T14:40:57.000+0000

Description

1.BaseUI>animation>3d transform
results: about 3 seconds in, the image flashes larger than it is, then continues the animation. Not seen in the simulator, seen on a 3.1.2 and 4.0 device

Comments

  1. Thomas Huelbert 2011-04-15

    3.1.2 only, not seen on 4.1

  2. Jeff Haynie 2011-04-15

    (from [b4b1b8e8c742baccd4b6f80472efdc6e9e36cc15]) [#1524 state:fixed-in-qa]: Don't relayout while we're animating. https://github.com/appcelerator/titanium_mobile/commit/b4b1b8e8c742baccd4b6f80472efdc6e9e36cc15"> https://github.com/appcelerator/titanium_mobile/commit/b4b1b8e8c742...

  3. Stephen Tramer 2011-04-15

    Forgot to claim before committing the fix.

  4. Jeff Haynie 2011-04-15

    (from [69900554e38b4bb674cf07cefe8a2d4eaf37c193]) [#2348 state:fixed-in-qa][#1524 state:open] Unfixing #1524. Cure was worse than disease. May have led to other regressions. https://github.com/appcelerator/titanium_mobile/commit/69900554e38b4bb674cf07cefe8a2d4eaf37c193"> https://github.com/appcelerator/titanium_mobile/commit/69900554e38b...

  5. Stephen Tramer 2011-04-15

    More information on this one. It's a VERY SPECIFIC race condition that happens between the animation subsystem and our own layout mechanisms. It looks like it has to do with the OS scheduling a relayout of the view at an odd time in between animation frames (later iOSes probably have different scheduling with animations, as evidenced by the fact that the original solution to this created bug #2348 - under later iOSes).

    Either not fixable as a result, or needs to be scheduled in the run loop in a different way. Either could be bad.

  6. Stephen Tramer 2011-04-15

    May never be fixed. Code would be 100% in support of legacy 3.1.x systems (at best, 3.x) and could complicate matters for later devices.

  7. Thomas Huelbert 2011-05-17

    changed to won't fix based on stramer comments

JSON Source