Closed Bug 1157090 Opened 9 years ago Closed 9 years ago

"Intermittent" failure in reftest-no-accel scroll-rounding.html,iframe-scrolling-attr-1.html,iframe-scrolling-attr-2.html,image-scrolling-zoom-1.html with Windows APZ enabled

Categories

(Core :: Panning and Zooming, defect)

x86
Windows
defect
Not set
normal

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: philor, Unassigned)

References

Details

(Whiteboard: [gfx-noted])

      No description provided.
Summary: "Intermittent" reftest-no-accel scroll-rounding.html | image comparison (==), max difference: 102, number of differing pixels: 1770 with Windows APZ enabled → "Intermittent" failure in reftest-no-accel scroll-rounding.html,iframe-scrolling-attr-1.html,iframe-scrolling-attr-2.html with Windows APZ enabled
This test failure happens whenever we make any changes to layerization. I have another patch in my queue that also needs to fuzz this test.

Adding a fuzzy-if(!layersGPUAccelerated,102,1770) to it is totally warranted.
Summary: "Intermittent" failure in reftest-no-accel scroll-rounding.html,iframe-scrolling-attr-1.html,iframe-scrolling-attr-2.html with Windows APZ enabled → "Intermittent" failure in reftest-no-accel scroll-rounding.html,iframe-scrolling-attr-1.html,iframe-scrolling-attr-2.html,image-scrolling-zoom-1.html with Windows APZ enabled
Blocks: 1154459
No longer depends on: 1154459
this bug stopped occurring about 2 weeks ago, did this get fixed magically?
"with Windows APZ enabled" - which was true for one nightly (plus some slop around merging it around and merging the pref flip around after that).
should we close this then?
I would like to leave it open until we know it's fixed with APZ enabled. We can change the decorations if it's turning up in some search that you don't want it anymore (e.g. remove the intermittent-failure keyword).
Whiteboard: [gfx-noted]
Note that comment 114 and 115 should be split into a separate bug; this bug is only for when APZ is enabled. I don't know if the bug summary should be changed so that it doesn't show up as a match for failures in TH.
Flags: needinfo?(cbook)
Flags: needinfo?(cbook)
The failures in this bug seem to be largely on:
- Win7 opt
- Win7 pgo
- Win8 opt
- Win8 pgo

I did a new try push at https://treeherder.mozilla.org/#/jobs?repo=try&revision=af08ad81505b with APZ enabled and retriggered the Win8 opt reftest-no-accel a bunch of times and it's green. I'm inclined to think this is WFM now, but I'll do a PGO push as well just to make sure.
The PGO push didn't really work: https://treeherder.mozilla.org/#/jobs?repo=try&revision=5aa7423d3f0f

On the assumption that "if I can't reproduce this on try, it doesn't count" I'm closing it as WFM.
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.