Followup from bug 1240799. I was able to get the test case to pass by adapting Rule 15d, App. G, TeXbook (used for bar case) to the stack case Differences from the bar case: actualRuleThickness is set to 0 minClearanceNum == minClearanceDen == minClearance/2 Patch to come tomorrow.
Interestingly, the axis alignment is now also incorrect on WebKit after https://bugs.webkit.org/show_bug.cgi?id=155714. We'll have to figure out whether it's a bug in implementations, in the implementation documents or if it's just the expected TeX / OpenType MATH behavior (and so the test should be fixed): https://github.com/MathML/MathMLinHTML5-tests/issues/2
Created attachment 8771610 [details] [diff] [review] patch Patch. I won't flag it for review until there is a decision over the correct behaviour.
(In reply to James Kitchener (:jkitch) from comment #2) > Created attachment 8771610 [details] [diff] [review] > patch > > Patch. > > I won't flag it for review until there is a decision over the correct > behaviour. Wrong patch? (this seems to be for bug 1274796)
Created attachment 8771611 [details] [diff] [review] patch Sigh. Correct patch this time.
Attachment #8771610 - Attachment is obsolete: true
Thanks James. So as I understand the MATH table, nothing ensures that the stack gap is split symmetrically around the axis height, even if that sounds natural and makes things more consistent with the fraction case. I guess we would have to ask to Murray Sargent but for now I disabled the tests for stacks: https://github.com/MathML/MathMLinHTML5-tests/commit/630c20db3337b68ec137153ebef20099bcf7049b
Priority: -- → P5
Assignee: jkitch.bug → nobody
Status: ASSIGNED → NEW
You need to log in before you can comment on or make changes to this bug.