Summary: | Switch WebGL test suites from 1.0.3, 2.0.0 to 1.0.x, 2.0.y | ||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Kimmo Kinnunen <kkinnunen> | ||||||||||||||||||||||
Component: | WebGL | Assignee: | Kimmo Kinnunen <kkinnunen> | ||||||||||||||||||||||
Status: | ASSIGNED --- | ||||||||||||||||||||||||
Severity: | Normal | CC: | amir_mark, dino, gman, kbr, kkinnunen, webkit-bug-importer | ||||||||||||||||||||||
Priority: | P2 | Keywords: | InRadar | ||||||||||||||||||||||
Version: | WebKit Local Build | ||||||||||||||||||||||||
Hardware: | Unspecified | ||||||||||||||||||||||||
OS: | Unspecified | ||||||||||||||||||||||||
Bug Depends on: | |||||||||||||||||||||||||
Bug Blocks: | 214392, 222964, 214393 | ||||||||||||||||||||||||
Attachments: |
|
Description
Kimmo Kinnunen
2021-03-30 09:02:50 PDT
Created attachment 424648 [details]
Patch
Created attachment 437261 [details]
Patch
Created attachment 437401 [details]
Patch
Created attachment 437494 [details]
Patch
Comment on attachment 437494 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=437494&action=review Good work with this difficult test suite upgrade. Overall it looks good; a couple of small comments about some tests marked flaky and a block of preexisting suppressions. r+ > LayoutTests/TestExpectations:3701 > +webkit.org/b/229941 webgl/2.0.y/conformance/extensions/webgl-compressed-texture-s3tc-srgb.html [ Pass Failure ] Unfortunate that the s3tc-srgb test's failing on some configurations, but it's good it's tracked. > LayoutTests/platform/mac-wk2/TestExpectations:329 > +[ Monterey+ ] webgl/2.0.y/conformance/textures/video/tex-2d-rgba-rgba-unsigned_byte.html [ Skip ] These video-related tests have been skipped for a while on Monterey? This is concerning - they used to pass, and the tex-2d-rgba-rgba-unsigned-byte path is core to many video related WebGL applications. Possible to follow on with a bug for this? > LayoutTests/platform/mac-wk2/TestExpectations:332 > +[ Monterey+ ] webgl/2.0.y/conformance2/textures/image_bitmap_from_video/tex-2d-r11f_g11f_b10f-rgb-float.html [ Skip ] I'm pretty sure all of these image_bitmap related tests used to pass earlier in the WebGL 2.0 development cycle. Possible to investigate why they're failing now? (In reply to Kenneth Russell from comment #6) > Good work with this difficult test suite upgrade. Overall it looks good; a > couple of small comments about some tests marked flaky and a block of > preexisting suppressions. r+ Thanks for looking! > > LayoutTests/TestExpectations:3701 > > +webkit.org/b/229941 webgl/2.0.y/conformance/extensions/webgl-compressed-texture-s3tc-srgb.html [ Pass Failure ] > > Unfortunate that the s3tc-srgb test's failing on some configurations, but > it's good it's tracked. I think it's about Metal backend leaving textures uninitialised in general... If this is the case, probably some other tests suffer from this, too. > These video-related tests have been skipped for a while on Monterey? This is > concerning - they used to pass, and the tex-2d-rgba-rgba-unsigned-byte path > is core to many video related WebGL applications. Possible to follow on with > a bug for this? > > I'm pretty sure all of these image_bitmap related tests used to pass earlier > in the WebGL 2.0 development cycle. Possible to investigate why they're > failing now? Yeah, this should be solved but it's a bit complicated. There are few interacting internal platform reasons. We can use for example bug 229837 to track these. (In reply to Kimmo Kinnunen from comment #7) > (In reply to Kenneth Russell from comment #6) > > Good work with this difficult test suite upgrade. Overall it looks good; a > > couple of small comments about some tests marked flaky and a block of > > preexisting suppressions. r+ > > Thanks for looking! > > > > LayoutTests/TestExpectations:3701 > > > +webkit.org/b/229941 webgl/2.0.y/conformance/extensions/webgl-compressed-texture-s3tc-srgb.html [ Pass Failure ] > > > > Unfortunate that the s3tc-srgb test's failing on some configurations, but > > it's good it's tracked. > > I think it's about Metal backend leaving textures uninitialised in general... > If this is the case, probably some other tests suffer from this, too. That would be concerning. Started a thread on the Slack instance about that possibility. We should try to confirm this ASAP. *** Bug 223586 has been marked as a duplicate of this bug. *** Created attachment 453717 [details]
Patch for EWS
Created attachment 453723 [details]
Patch for EWS
Created attachment 453832 [details]
Patch for EWS
Created attachment 453945 [details]
Patch for EWS
Created attachment 454077 [details]
For EWS
Created attachment 454110 [details]
For EWS
|