Closed Bug 1164274 Opened 8 years ago Closed 8 years ago

Intermittent test_wheel_scroll.html | We should not have scrolled to the bottom of the scrollframe - got false, expected true

Categories

(Core :: Panning and Zooming, defect)

40 Branch
defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla41
Tracking Status
firefox39 --- unaffected
firefox40 --- unaffected
firefox41 --- fixed
firefox-esr31 --- unaffected
firefox-esr38 --- unaffected

People

(Reporter: KWierso, Assigned: kats)

References

(Depends on 1 open bug)

Details

(Keywords: intermittent-failure, Whiteboard: gfx-noted)

Attachments

(1 file)

12:28:57 INFO - 91 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 92 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 93 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 94 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 95 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 96 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 97 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 98 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 99 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 100 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | Wheel event successfully synthesized
12:28:57 INFO - 101 INFO TEST-PASS | gfx/layers/apz/test/test_wheel_scroll.html | We should have scrolled down somewhat
12:28:57 INFO - 102 INFO TEST-UNEXPECTED-FAIL | gfx/layers/apz/test/test_wheel_scroll.html | We should not have scrolled to the bottom of the scrollframe - got false, expected true
12:28:57 INFO - 103 INFO TEST-UNEXPECTED-FAIL | gfx/layers/apz/test/test_wheel_scroll.html | The rotation should have been adjusted - got false, expected true
12:28:57 INFO - MEMORY STAT vsize after test: 731107328
12:28:57 INFO - MEMORY STAT vsizeMaxContiguous not supported in this build configuration.
12:28:57 INFO - MEMORY STAT residentFast after test: 149639168
12:28:57 INFO - MEMORY STAT heapAllocated after test: 58481032
12:28:57 INFO - 104 INFO TEST-OK | gfx/layers/apz/test/test_wheel_scroll.html | took 1716ms
12:28:58 INFO - 105 INFO TEST-START | image/test/mochitest/test_ImageContentLoaded.html
kats, this is pretty frequent on Mulet. Can you please take a look?
Component: Graphics: Layers → Panning and Zooming
Flags: needinfo?(bugmail.mozilla)
Will do
Assignee: nobody → bugmail.mozilla
Flags: needinfo?(bugmail.mozilla)
Given a local build of Mulet, how do I run mochitests on it? |mach mochitest| doesn't seem to do it and the help options for that command also don't seem to say anything about it.
Flags: needinfo?(ahalberstadt)
(In reply to Kartikaya Gupta (email:kats@mozilla.com) from comment #55)
> Given a local build of Mulet, how do I run mochitests on it? |mach
> mochitest| doesn't seem to do it and the help options for that command also
> don't seem to say anything about it.

At the moment |mach mochitest| is desktop Firefox only. I have a patch in the works that will make |mach mochitest| work with all platforms as well as deprecate all the other mochitest-<foo> commands. Until then, try using |mach mochitest-plain|.
Flags: needinfo?(ahalberstadt)
Thanks! I can run the test with that. It looks like like a regular Linux build though, so I'm not sure there's much value in running Mulet-Linux mochitests in automation.

Anyway, I'll try to repro the intermittent failure locally and if I can't I might just disable the test on Mulet.
I think I managed to identify what's going on here. Since APZ is disabled by default, we're going through the main-thread wheel handling code. Although smooth scrolling is disabled by the test, the scroll is still "async" (using the main-thread async code). When it gets to [1] sometimes the mAsyncScroll from the previous wheel event still hasn't been dealt with. So the new wheel event just accumulates onto it, replacing the pending instant scroll with a new instant scroll of the accumulated amount. If this happens enough times in a row we can jump right past the scrollbox that is supposed to catch the wheel events. In a sense this is already a nondeterministic version of what mstange suggested we do at bug 1166871 comment 4.

[1] http://mxr.mozilla.org/mozilla-central/source/layout/generic/nsGfxScrollFrame.cpp?rev=31459ec4ef93#2079
I'm not sure why the test is intermittent-failing on Mulet only instead of everywhere since all desktop platforms should have this problem, I think. Maybe just timing. Nonetheless I'm happy to disable it on Mulet for now until we figure out what we're doing with the wheel event targeting over.
Attachment #8609412 - Flags: review?(mstange)
Comment on attachment 8609412 [details] [diff] [review]
Disable test on mulet

Sure.
Attachment #8609412 - Flags: review?(mstange) → review+
Whiteboard: gfx-noted
Actually I'm just going to close this for now. If we change behaviour in bug 1168182 we'll likely need to revisit this test anyway since the proposed behaviour there will break this test.
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.