Kartikaya Gupta (email:kats@mozilla.staktrace.com) Assignee
	Description • 8 years ago

The test_bug982141 (added in bug 961289) was one of the first APZ mochitests and has withstood the ravages of time remarkably well. However, in the move from B2G to desktop (bug 1251608), I neglected to properly take into account the effect of non-overlay scrollbars. In particular, the hard-coded "50" in the test should really subtract off the scrollbar width/height when there are non-overlay scrollbars. Or more simply, it should just use the clientWidth/clientHeight of the subframe which already excludes those values (test_scroll_subframe_scrollbar.html uses that to good effect).

This bug was masked by the fact that the displayport gets aligned to tiles (or 128 on non-tiling platforms), and so even if there are scrollbars present the displayport is large enough to cover the required 50 pixels. However in bug 1216924 I remove this alignment when displayport suppression is occurring, and that exposes this bug in the test.

	Botond Ballo [:botond]
	Comment 2 • 8 years ago

Comment on attachment 8729838 [details] [diff] [review]
Patch

Review of attachment 8729838 [details] [diff] [review]:
-----------------------------------------------------------------

Good catch, thanks for fixing this!

	Pulsebot
	Comment 3 • 8 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/48c0d3800c35

	Carsten Book [:Tomcat]
	Comment 4 • 8 years ago
bugherder

https://hg.mozilla.org/mozilla-central/rev/48c0d3800c35