1254834 - [e10s] Intermittent fractional-transform-2.html | failed reftest-no-paint

Status: RESOLVED FIXED

(Core :: Layout, defect)

Description

[Ryan VanderMeulen [:RyanVM]]

Not a permafail, but definitely a new Win8 e10s reftest failure.

https://treeherder.mozilla.org/logviewer.html#?job_id=17687187&repo=try#L66691

17:21:40     INFO -  REFTEST TEST-START | file:///C:/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/fractional-transform-2.html
17:21:40     INFO -  REFTEST TEST-LOAD | file:///C:/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/fractional-transform-2.html | 13272 / 13292 (99%)
17:21:41     INFO -  REFTEST TEST-UNEXPECTED-FAIL | file:///C:/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/fractional-transform-2.html | failed reftest-no-paint
17:21:41     INFO -  REFTEST TEST-END | file:///C:/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/fractional-transform-2.html
17:21:41     INFO -  REFTEST INFO | Saved log: START file:///C:/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/fractional-transform-2.html
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] OnDocumentLoad triggering WaitForTestEnd
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] WaitForTestEnd: Adding listeners
17:21:41     INFO -  REFTEST INFO | Saved log: Initializing canvas snapshot
17:21:41     INFO -  REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FIRE_INVALIDATE_EVENT
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: waiting for MozAfterPaint
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] AfterPaintListener in file:///C:/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/fractional-transform-2.html
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] SendUpdateCanvasForEvent with 1 rects
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] Rect: 0 0 800 1000
17:21:41     INFO -  REFTEST INFO | Saved log: Updating canvas for invalidation
17:21:41     INFO -  REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FIRE_INVALIDATE_EVENT
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: dispatching MozReftestInvalidate
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] AttrModifiedListener fired
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_REFTEST_WAIT_REMOVAL
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_SPELL_CHECKS
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_APZ_FLUSH
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: done requesting APZ flush
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_APZ_FLUSH
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: apz-repaints-flushed fired
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FINISH
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: waiting for MozAfterPaint
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] AfterPaintListener in file:///C:/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/fractional-transform-2.html
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] SendUpdateCanvasForEvent with 1 rects
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] Rect: 0 -197 800 1083
17:21:41     INFO -  REFTEST INFO | Saved log: Updating canvas for invalidation
17:21:41     INFO -  REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FINISH
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: waiting for MozAfterPaint
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] AfterPaintListener in file:///C:/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/fractional-transform-2.html
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] SendUpdateCanvasForEvent with 1 rects
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] Rect: 0 -617 783 3095
17:21:41     INFO -  REFTEST INFO | Saved log: Updating canvas for invalidation
17:21:41     INFO -  REFTEST INFO | Saved log: DoDrawWindow 0,0,783,1000
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FINISH
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] MakeProgress: Completed
17:21:41     INFO -  REFTEST INFO | Saved log: [CONTENT] RecordResult fired
17:21:41     INFO -  REFTEST INFO | Saved log: RecordResult fired
17:21:41     INFO -  REFTEST INFO | Saved log: RecordResult fired

Comment 1

[Markus Stange [:mstange]]

I think we need to go through all layout/reftests/invalidation/ tests that adjust the scroll position and set display ports on all of them.

Comment 2

[Kartikaya Gupta (email:[e-mail])]

Yeah I was just working on a patch to do that :)

There's some tests which adjust the scroll position on a div rather than the top-level document, would those be affected too?

Comment 3

[Kartikaya Gupta (email:[e-mail])]

Attachment: Patch

Need to test it still. I put a displayport on the thing in masklayer-2 but not the one in paintedlayer-recycling-2.html because in that one the thing that's scrolling is a sibling of the no-paint so it shouldn't affect it.

Comment 4

[Kartikaya Gupta (email:[e-mail])]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=14e1a94e1d3b

Comment 5

[Markus Stange [:mstange]]

(In reply to Kartikaya Gupta (email:[e-mail]) from comment #2)
> There's some tests which adjust the scroll position on a div rather than the
> top-level document, would those be affected too?

I think so!

Though now I remember that the invalidations I was fixing were caused by two different things: Some were caused by the fact that a display item moved relative to its reference frame, and some were caused by the change in the visible region. The former can be fixed by having a constant display port. But for the latter, a constant display port will make the test not test what it was intended to test any more. For those, what we really want is to check whether a specific change of the display port dimensions causes invalidations. Not sure how to do that. Oh wait - of course it will! The display port dimensions are relative to the current scroll position! So we should just be able to set the display port to the scroll port size and we should be good. Unless the display port gets enlarged to nearest tile boundaries...

Comment 6

[Kartikaya Gupta (email:[e-mail])]

Setting the displayport-* properties sets the displayport to be relative to the scrollport, and skips the tile alignment stuff [1]. So that should basically simulate what happens with APZ off.

Although thinking about it, we should probably reverse the order of the first two branches at [2] so that the rect data takes priority over displayport suppression. I'll file a bug for that.

[1] http://mxr.mozilla.org/mozilla-central/source/layout/base/nsLayoutUtils.cpp?rev=565778f033af#875
[2] http://mxr.mozilla.org/mozilla-central/source/layout/base/nsLayoutUtils.cpp?rev=565778f033af#1072

Comment 7

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/4123969ee94c

Comment 8

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/4123969ee94c

Comment 9

[Markus Stange [:mstange]]

Attachment: MozReview Request: Bug 1254834 - Set the display port on the right element in fast-scrolling.html. r?kats

Review commit: https://reviewboard.mozilla.org/r/42877/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/42877/

Comment 10

[Kartikaya Gupta (email:[e-mail])]

Comment on attachment 8735624 [details]
MozReview Request: Bug 1254834 - Set the display port on the right element in fast-scrolling.html. r?kats

https://reviewboard.mozilla.org/r/42877/#review39787

Comment 11

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/4f29fd865fad

Comment 12

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/4f29fd865fad