The testcase for bug 1105268 fires a lot of warnings: WARNING: have unconstrained inline-size; this should only result from very large sizes, not attempts at intrinsic inline-size calculation: 'NS_UNCONSTRAINEDSIZE != computedISizeCBWM && NS_UNCONSTRAINEDSIZE != availISizeCBWM', file /Users/smontagu/mozwork/hgtree/mozilla/layout/generic/nsHTMLReflowState.cpp, line 2370 I was able to minimize the testcase to data:text/html,<div style="writing-mode: vertical-lr"> This is possibly regression from bug 1096244, but I haven't checked. The attached patch is just a copy of the fix from bug 1088547. I can't see a way to avoid duplicating the code, but maybe I'm missing something.

N.B. the patch fixes the minimized testcase, but not the original testcase from bug 1105268; investigating further.

(In reply to Simon Montagu :smontagu from comment #0) > This is possibly regression from bug 1096244, but I haven't checked. That would be bug 1096224

This fixes the original testcase (but not the minimized testcase, so we need both patches). If the comment above is true, setting UNCONSTRAINEDSIZE on this availSpace.BSize should have no effect, but waiting for a try run before requesting review (tryserver is CLOSED at time of writing).

(When I say "the comment above", I mean the comment in the first line of the diff: // The availSpace here is irrelevant to our needs - all we want // out if this setup is the block-start margin value which doesn't depend // on the childs available space.

Try run: https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=1368801bdc24

Comment on attachment 8529842 [details] [diff] [review] Patch 1 Review of attachment 8529842 [details] [diff] [review]: ----------------------------------------------------------------- LGTM

https://hg.mozilla.org/mozilla-central/rev/f3bdbfd4e117 https://hg.mozilla.org/mozilla-central/rev/1289066eccbe