Closed Bug 774689 Opened 13 years ago Closed 7 years ago

Intermittent 632781-verybig.html | image comparison (==), max difference: 255, number of differing pixels: 14400

Categories

(Core :: Graphics, defect, P3)

Product:
Core
Component:
Graphics
Platform:
x86
Windows XP
Type:
defect

Tracking

()

Status:
RESOLVED INCOMPLETE

People

(Reporter: emorley, Assigned: milan)

References

Details

(Keywords: intermittent-failure, Whiteboard: [leave open][marked random-if on Windows and OS X 10.10])

Attachments

(2 files)

Reftest log
14.38 KB, text/plain
Details
use a 100000x120 canvas
1.50 KB, patch
bas.schouten
: review+
Details | Diff | Splinter Review
Attached file Reftest log —
Rev3 WINNT 5.1 mozilla-inbound debug test reftest on 2012-07-16 17:45:53 PDT for push 7fe11a265ac5 slave: talos-r3-xp-009 https://tbpl.mozilla.org/php/getParsedLog.php?id=13589847&tree=Mozilla-Inbound { REFTEST TEST-START | file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/bugs/632781-verybig.html | 2826 / 8055 (35%) ++DOMWINDOW == 37 (083E17B0) [serial = 7790] [outer = 07838058] ??????????p: TWARNING: Invalid type to RecordMemoryUsedForSurfaceType!: file e:/builds/moz2_slave/m-in-w32-dbg/build/gfx/thebes/gfxASurface.cpp, line 616 WARNING: Invalid type to RecordMemoryUsedForSurfaceType!: file e:/builds/moz2_slave/m-in-w32-dbg/build/gfx/thebes/gfxASurface.cpp, line 616 JavaScript error: file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/bugs/632781-verybig.html, line 19: NS_ERROR_FAILURE: Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIDOMCanvasRenderingContext2D.fillRect] ??????????p: TWARNING: Invalid type to RecordMemoryUsedForSurfaceType!: file e:/builds/moz2_slave/m-in-w32-dbg/build/gfx/thebes/gfxASurface.cpp, line 616 WARNING: Invalid type to RecordMemoryUsedForSurfaceType!: file e:/builds/moz2_slave/m-in-w32-dbg/build/gfx/thebes/gfxASurface.cpp, line 616 REFTEST TEST-START | file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/bugs/632781-ref.html | 2826 / 8055 (35%) ++DOMWINDOW == 38 (2455E078) [serial = 7791] [outer = 07838058] REFTEST TEST-UNEXPECTED-FAIL | file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/bugs/632781-verybig.html | image comparison (==) REFTEST number of differing pixels: 14400 max difference: 255 }
Rev3 WINNT 5.1 mozilla-inbound debug test reftest on 2012-07-16 22:52:00 PDT for push 33bece9dee65 slave: talos-r3-xp-022 https://tbpl.mozilla.org/php/getParsedLog.php?id=13595032&tree=Mozilla-Inbound
Whiteboard: [orange]
Resolving WFM keyword:intermittent-failure bugs last modified >3 months ago, whose whiteboard contains none of: {random,disabled,marked,fuzzy,todo,fails,failing,annotated,time-bomb,leave open} There will inevitably be some false positives; for that (and the bugspam) I apologise. Filter on orangewfm.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → WORKSFORME
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
Summary: Intermittent 632781-verybig.html | image comparison (==) → Intermittent 632781-verybig.html | image comparison (==), max difference: 255, number of differing pixels: 14400
Ehsan, as test author could you take a look at this top #5 orange? :-)
Flags: needinfo?(ehsan)
Whiteboard: [disable-me 2013-06-05]
(In reply to Ed Morley [:edmorley UTC+1] from comment #132) > Ehsan, as test author could you take a look at this top #5 orange? :-) Sorry, I know about zero things about graphics. Forwarding to Joe to see who needs to look at it.
Flags: needinfo?(ehsan) → needinfo?(joe)
Flags: needinfo?(joe) → needinfo?(milan)
Too many failures & no action in the bug = marking as failing on WinXP: https://hg.mozilla.org/integration/mozilla-inbound/rev/189b50c8abdf
Whiteboard: [disable-me 2013-06-05]
Whiteboard: [marked fails-if WinXP]
I think you really wanted to mark it as random-if rather than fails-if; otherwise you're going to get an UNEXPECTED PASS orange on the "happy" occasions when it doesn't actually fail.
Ah whoops yeah random-if is what I was thinking of - thank you. remote: https://hg.mozilla.org/integration/mozilla-inbound/rev/742dbd5e76fe
Whiteboard: [marked fails-if WinXP] → [marked random-if on WinXP]
https://hg.mozilla.org/mozilla-central/rev/189b50c8abdf https://hg.mozilla.org/mozilla-central/rev/742dbd5e76fe
Whiteboard: [marked random-if on WinXP] → [leave open][marked random-if on WinXP]
To be fair, this not only tests the proper scrolling for large canvases, but also proper display for them. It's the display part that fails in here. It does point to a problem, to be sure.
Assignee: nobody → milan
Flags: needinfo?(milan)
Somewhere along the line someone switched from just WinXP to all of winWidget, but even with disabling it on 10.10 we'll still be running it on, um, Linux and OS X 10.6 for as long as we keep testing there.
Status: REOPENED → NEW
Whiteboard: [leave open][marked random-if on WinXP] → [leave open][marked random-if on Windows and OS X 10.10]
This test is typically the kind of reftest that we can't make work 100% of time because it relies on making a very big allocation for the canvas, and it is reasonable for the test runners to fail that allocation sometimes. Probably better as a crash test, or have something like "If the allocation fails, ignore the test" which is hard to do.
See Also: → 991566
This will test the same thing as the previous version of the test (namely forcing software rendering on accelerated canvas), while allocating substantially less memory and reduce the likelihood of running out of contiguous address space.
Attachment #8710001 - Flags: review?(bas)
Try push on the platforms where the test runs (although it is marked as random on windows and OSX) https://treeherder.mozilla.org/#/jobs?repo=try&revision=c8a3551d36c7
Blocks: 1180942
Attachment #8710001 - Flags: review?(bas) → review+
Bulk assigning P3 to all open intermittent bugs without a priority set in Firefox components per bug 1298978.
Priority: -- → P3
Mass-closing old bugs I filed that have not had recent activity/no longer affect me.
Status: NEW → RESOLVED
Closed: 12 years ago → 7 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: