Created attachment 63660 [details] Test case In the attached test case, there is too much whitespace in the rows. Firefox's rendering is much tighter. I think that some of the stretching in RenderMathMLRow::layout() might be overly aggressive.
(In reply to comment #0) > Created an attachment (id=63660) [details] > Test case > > In the attached test case, there is too much whitespace in the rows. Firefox's rendering is much tighter. I think that some of the stretching in RenderMathMLRow::layout() might be overly aggressive. Yes, there is still a lot of work to do about the layout of MathML elements in general and specially for mrow. Layout and robustness vs zooming are the two things we plan to focus our attention on in the coming weeks/months.
Something else strange is going on here. The operator height is being set to the container size but then then alignment is set to near bottom of the box. This causes the whole thing to stretch in strange ways. While it is coming from row layout, it isn't operator stretching. Instead, it comes from attempting to align the operator. Either way, this is very broken.
Created attachment 70887 [details] Reduced Example of Problem This attached example demonstrates the problem as well (although less dramatic). The operators are extending the vertical height because the operator placement in the block and their baseline is wrong. This causes the overall row to have a larger space on top than it should. The bug is in RenderMathMLOperator and its relationship to the row.
Created attachment 70979 [details] Patch that removes the extra space above non-stretched operators
Comment on attachment 70979 [details] Patch that removes the extra space above non-stretched operators Nice
Comment on attachment 70979 [details] Patch that removes the extra space above non-stretched operators Clearing flags on attachment: 70979 Committed r69926: <http://trac.webkit.org/changeset/69926>
All reviewed patches have been landed. Closing bug.