STR: (1) Load https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/reftests/forms/input/range/range-orient-horizontal.html on a linux debug build (2) click the range. We trip the following assertion: http://searchfox.org/mozilla-central/rev/02a56df6474a97cf84d94bbcfaa126979970905d/layout/forms/nsRangeFrame.cpp#565 Looks like the width is 840 but the height is 0.
Looks like we end up taking this path: http://searchfox.org/mozilla-central/rev/02a56df6474a97cf84d94bbcfaa126979970905d/widget/gtk/gtk3drawing.cpp#2557 Which stores 0 in thumb_height. Not sure if this is a bug in gecko's expectations or in the gtk glue.
I'd guess this is a bug in gecko's expectations. The |style| passed through to gtk_style_context_get includes the orientation of the control. I'd guess it leaves/sets the dimension perpendicular to the thumb's direction of travel as/to zero. This code doesn't care about that dimension anyway so we should change the assertion to only assert about the dimension that matters to us.
Created attachment 8832268 [details] [diff] [review] patch
Assignee: nobody → jwatt
Attachment #8832268 - Flags: review?(mstange)
Attachment #8832268 - Flags: review?(mstange) → review+
Pushed by email@example.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/9f130bfec2d5 Fix crash in nsRangeFrame::GetValueAtEventPoint on linux. r=mstange
Status: NEW → RESOLVED
Last Resolved: 2 years ago
status-firefox54: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla54
You need to log in before you can comment on or make changes to this bug.