Closed Bug 547342 Opened 10 years ago Closed 9 years ago
random orange in /tests/content/xul/templates/tests/test
135 INFO Running /tests/content/xul/templates/tests/test_bug441785.xul... ++DOMWINDOW == 25 (07A6CC00) [serial = 32] [outer = 074B5100] 136 ERROR TEST-UNEXPECTED-FAIL | /tests/content/xul/templates/tests/test_bug441785.xul | first tree row count datasource 1 - got 0, expected 10 137 INFO TEST-PASS | /tests/content/xul/templates/tests/test_bug441785.xul | second tree row count datasource 1 At the time this one popped up, bz said on irc that it is a problem with the test itself - i think.
WINNT 5.2 mozilla-central debug test mochitests-2/5 [testfailed] Started 15:46, finished 16:11 http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266623214.1266624657.22644.gz
Component: Layout → DOM
QA Contact: layout → general
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266621503.1266622074.14271.gz WINNT 5.2 mozilla-central opt test mochitests-2/5 [testfailed] Started 15:18, finished 15:28
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266616129.1266617148.28745.gz WINNT 5.2 mozilla-central opt test mochitests-2/5 [testfailed] Started 13:48, finished 14:06
Component: DOM → XP Toolkit/Widgets: XUL
QA Contact: general → xptoolkit.xul
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266633151.1266633684.16416.gz http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266632331.1266632919.14078.gz http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266625159.1266625726.26058.gz http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266623541.1266624121.21007.gz http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266632216.1266633760.16571.gz This is perma-orange on Windows.
OS: Linux → Windows 7
This works around the issue, but the real problem, imo, is that scrollToRow needs to flush. Otherwise mView can be null when someone calls it, in which case it will silently fail. Or someone might have set an explit view, in which case it can read random numbers from mPageLength (which the constructor does not initialize)... We should probably set mPageLength to 0 in the constructor, and have more box object methods here flushing layout. I'm just not sure which ones.
Oh, I pushed that as http://hg.mozilla.org/mozilla-central/rev/0264f8178ac9
Attachment #427906 - Flags: review?(roc) → review+
Comment on attachment 427906 [details] [diff] [review] Temporary fix I pushed Look ok for now. Flushing before calling looks ok too.
Attachment #427906 - Flags: review?(enndeakin) → review+
Attachment #463078 - Flags: review? → review?(enndeakin)
Whiteboard: [orange] → [need review][orange]
Whiteboard: [need review][orange] → [need checkin][orange]
Summary: random orange in content/xul/templates/tests/bug441785.xul → random orange in /tests/content/xul/templates/tests/test_bug441785.xul
Hmm. Looks like the patch I just checked in didn't work. Investigating.
Yeah, ok. That patch was wrong, because it fails to flush layout when we have a cached frame. Pushed fix for that. So overall, pushed: http://hg.mozilla.org/mozilla-central/rev/dd0c8936a36d http://hg.mozilla.org/mozilla-central/rev/3f91606bd115
Status: ASSIGNED → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Whiteboard: [need checkin][orange] → [orange]
Target Milestone: --- → mozilla2.0b4
Moving to Core:XUL per https://bugzilla.mozilla.org/show_bug.cgi?id=1455336
Component: XP Toolkit/Widgets: XUL → XUL
You need to log in before you can comment on or make changes to this bug.