ASSERTION: Returning bogus 0-sized margin, because this margin depends on layout & isn't cached!: 'Error', file .../mozilla/layout/generic/nsIFrame.cpp:1617
Categories
(Calendar :: Calendar Frontend, defect)
Tracking
(Not tracked)
People
(Reporter: ishikawa, Unassigned)
Details
Attachments
(1 file)
I test my local patches for C-C TB using DEBUG version of TB.
Since the day before, I suddenly got tons of the following errors during mochitest.
More than 4K lines.
###!!! ASSERTION: Returning bogus 0-sized margin, because this margin depends on layout & isn't cached!: 'Error', file /NEW-SSD/NREF-COMM-CENTRAL/mozilla/layout/generic/nsIFrame.cpp:1617
It is possible that I did not run calendar tests before. (I may have excluded the tests due to lengthy timeouts experienced during some tests there.)
The number of Assertions for each test where the assertion occurs is as follows.
34 comm/calendar/test/browser/browser_dragEventItem.js
1602 comm/calendar/test/browser/browser_eventDisplay_multiWeekView.js
31 comm/calendar/test/browser/browser_eventUndoRedo.js
34 comm/calendar/test/browser/browser_import.js
140 comm/calendar/test/browser/browser_tabs.js
221 comm/calendar/test/browser/contextMenu/browser_edit.js
42 comm/calendar/test/browser/eventDialog/browser_attachMenu.js
34 comm/calendar/test/browser/invitations/browser_imipBar.js
175 comm/calendar/test/browser/recurrence/browser_annual.js
233 comm/calendar/test/browser/recurrence/browser_biweekly.js
56 comm/calendar/test/browser/recurrence/browser_daily.js
485 comm/calendar/test/browser/recurrence/browser_lastDayOfMonth.js
28 comm/calendar/test/browser/recurrence/browser_weeklyUntil.js
28 comm/calendar/test/browser/recurrence/browser_weeklyWithException.js
175 comm/calendar/test/browser/recurrence/browser_annual.js
233 comm/calendar/test/browser/recurrence/browser_biweekly.js
56 comm/calendar/test/browser/recurrence/browser_daily.js
485 comm/calendar/test/browser/recurrence/browser_lastDayOfMonth.js
28 comm/calendar/test/browser/recurrence/browser_weeklyUntil.js
28 comm/calendar/test/browser/recurrence/browser_weeklyWithException.js
34 comm/calendar/test/browser/views/browser_monthView.js
35 comm/calendar/test/browser/views/browser_propertyChanges.js
I am attaching the log from the last test above,
comm/calendar/test/browser/views/browser_propertyChanges.js
for your reference.
There are debug output done my local patches. And some line numbers may have been shifted due the local patches.
The assertion is here:
https://searchfox.org/mozilla-central/source/layout/generic/nsIFrame.cpp#1616
if (!StyleMargin()->GetMargin(margin)) {
// If we get here, our caller probably shouldn't be calling us...
NS_ERROR(
"Returning bogus 0-sized margin, because this margin "
"depends on layout & isn't cached!");
}
}
Given the comment, " // If we get here, our caller probably shouldn't be calling us...",
I think T-B and/or calendar is doing something wrong.
| Reporter | ||
Comment 1•3 years ago
|
||
I hasten to add that other people's test (debug version test) sees the same asserts on the try-comm-central.
For example,
Geoff's job
https://treeherder.mozilla.org/jobs?repo=try-comm-central&revision=37ddd6c4575ca98c4652892d4228b2edf795c07f
the log:
https://treeherder.mozilla.org/logviewer?job_id=404440652&repo=try-comm-central&lineNumber=185565
| Reporter | ||
Comment 2•3 years ago
|
||
This warning probably accounts for the sudden log file size increase for mochitest.
Locally it accounts for the increase of about 50MB, which is huge.
(I am dumping stracktrace from MOZ_ASSERT() to learn any hint it may give us regarding the cause of assertion failure.)
The latest assert is triggered 2376 times during mochitest run.
I check the size of the text related to the first assert (with traceback), and it is 22930.
Multiplied together, the result is 54481680, about 54.5 MB, which seems to explain the log size increase.
Comment 3•3 years ago
|
||
See bug 1813765
| Reporter | ||
Comment 4•3 years ago
|
||
(In reply to Magnus Melin [:mkmelin] from comment #3)
See bug 1813765
Now that the patch from the mentioned bug landed(?), I don't see the errors any more.
So I am closing this as WORKSFORME.
Description
•