Created attachment 182201 [details] Testcase See the testcase, where I use the CSS rule p { text-indent: 2em; } There are large gaps inside the formulas in Chrome 24. This seems similar to this old Gecko's bug: https://bugzilla.mozilla.org/show_bug.cgi?id=454085 and could be fixed the same way. Interestingly, if you uncomment the <script> and let MathJax process the <math>, then the page renders correctly.
Fixing this bug should just be a matter of adding "text-indent: 0" rules in Source/WebCore/css/mathml.css
So it seems that the bug appears with the latest LaTeXML version (with new CSS style). See for instance my blog: http://www.maths-informatique-jeux.com/blog/frederic/?post/2013/04/04/Suslin%E2%80%99s-Problem
Created attachment 200640 [details] Patch V1
Can someone please review the patch?
(In reply to comment #4) > Can someone please review the patch? This patch doesn't change any test results or include any new tests?
(In reply to comment #5) > This patch doesn't change any test results or include any new tests? This only resets the text-indent value on the <math> element. So this only has an effect on pages that modify text-indent and contain a <math> element. I thought it was a trivial change so didn't need tests. Apparently, everything is green above.
(In reply to comment #6) > (In reply to comment #5) > > This patch doesn't change any test results or include any new tests? > > This only resets the text-indent value on the <math> element. So this only has an effect on pages that modify text-indent and contain a <math> element. I thought it was a trivial change so didn't need tests. Apparently, everything is green above.
(In reply to comment # > (In reply to comment # > This patch doesn't change any test results or include any new tests? > > This only resets the text-indent value on the <math> element. So this only has an effect on pages that modify text-indent and contain a <math> element. I thought it was a trivial change so didn't need tests. Apparently, everything is green above. I think it could be useful for this to have a test so that we don't regress here. It seems that a reference test could be fairly trivial in this case actually. In general, changes to WebCore need tests and I'd say doubly so for changes to layout.
Created attachment 201981 [details] Patch V2
Comment on attachment 201981 [details] Patch V2 Clearing flags on attachment: 201981 Committed r150264: <http://trac.webkit.org/changeset/150264>
All reviewed patches have been landed. Closing bug.