512405 - layout/base/crashtests/265986-1.html triggers NS_UNCONSTRAINEDSIZE assertions

Status: NEW

(Core :: Layout, defect)

Description

[Jesse Ruderman]

Split from bug 387080: layout/base/crashtests/265986-1.html triggers these assertions.

###!!! ASSERTION: computed width should always be computed: 'computedSize.width != NS_INTRINSICSIZE', file layout/xul/base/src/nsBoxFrame.cpp, line 708

###!!! ASSERTION: no unconstrained widths should be present anymore: 'NS_UNCONSTRAINEDSIZE != aReflowState.ComputedWidth()', file layout/generic/nsBlockReflowState.cpp, line 114

###!!! ASSERTION: should no longer be using unconstrained widths: 'aWidth != NS_UNCONSTRAINEDSIZE', file layout/generic/nsLineLayout.cpp, line 175

###!!! ASSERTION: shouldn't have unconstrained widths anymore: 'NS_UNCONSTRAINEDSIZE != aReflowState.availableWidth', file layout/generic/nsLineLayout.cpp, line 1099

###!!! ASSERTION: shouldn't have unconstrained widths anymore: 'psd->mRightEdge != NS_UNCONSTRAINEDSIZE', file layout/generic/nsLineLayout.cpp, line 2407

###!!! ASSERTION: shouldn't have unconstrained widths anymore: 'psd->mRightEdge != NS_UNCONSTRAINEDSIZE', file layout/generic/nsLineLayout.cpp, line 779

###!!! ASSERTION: shouldn't use unconstrained widths anymore: '(mFrameType == NS_CSS_FRAME_TYPE_INLINE && !frame->IsFrameOfType(nsIFrame::eReplaced)) || frame->GetType() == nsGkAtoms::textFrame || mComputedWidth != NS_UNCONSTRAINEDSIZE', file layout/generic/nsHTMLReflowState.cpp, line 287

###!!! ASSERTION: shouldn't use unconstrained widths anymore: 'availableWidth != NS_UNCONSTRAINEDSIZE', file layout/generic/nsHTMLReflowState.cpp, line 266

###!!! ASSERTION: unconstrained widths no longer supported: 'aContainingBlockWidth != NS_UNCONSTRAINEDSIZE', file layout/base/nsLayoutUtils.cpp, line 2077

Comment 1

[Jesse Ruderman]

The manifest has been annotated:

asserts(10) asserts-if(MOZ_WIDGET_TOOLKIT=="windows",8) load 265986-1.html

I'll update the manifest to also point at this bug report.

Comment 2

[(no longer active)]

(In reply to comment #1)
> The manifest has been annotated:
> 
> asserts(10) asserts-if(MOZ_WIDGET_TOOLKIT=="windows",8) load 265986-1.html
> 
> I'll update the manifest to also point at this bug report.

FWIW, on Mac OS X 10.6, I consistently get 8 assertions as well.  Can we just mark it as asserts(8-10)?

Comment 3

[Jesse Ruderman]

Sure, you can change it to asserts(8-10).

Comment 4

[(no longer active)]

Attachment: adjust the assertion count

Comment 5

[(no longer active)]

Attachment: adjust the assertion count (for check-in)

Comment 6

[[e-mail]]

Comment on attachment 468523 [details] [diff] [review]
adjust the assertion count (for check-in)

http://hg.mozilla.org/mozilla-central/rev/a2947419d601

Comment 7

[(no longer active)]

This is not actually fixed yet.

Comment 8

[Christian Holler (:decoder)]

I've been hitting an AddressSanitizer error on Linux, just on the crashtest mentioned here. The other bug is bug 800390, not sure if this is possibly related.