[Webkit-unassigned] [Bug 180923] [Followup to 179682] Add comment and code style improvement to MUnderOver's layout functions

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=180923

--- Comment #10 from Frédéric Wang (:fredw) [back 03/01/2018] <fred.wang at free.fr> ---
Comment on attachment 329737
  --> https://bugs.webkit.org/attachment.cgi?id=329737
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=329737&action=review

> Source/WebCore/rendering/mathml/RenderMathMLUnderOver.cpp:93
> +    // stretched to cover the base rather than the text comment, the latter of which could be very wide.

I don't think it's appropriate to insert a (vague) example in the general description of the algorithm. If you really want examples, they should be in separate paragraph describing some use cases in my opinion. But I believe it's better to have use cases as non-regression tests so that people don't change the behavior so easily in the future ;-) In any case, the "even if" part of the 4th point is rather to handle this situation of attachment 328081:

<munder>
  <mtext>SHORT TEXT</mtext>
  <msub>
    <mo>...</mo>
    <mtext>VERY LONG TEXT</mtext>
  </msub>
</munder>

where the unstretched size of the <msub> is *already* large enough to cover the width of the base. Stretching the <mo> at the core to cover the base width makes the <msub> even wider. So one interpretation of the spec might be that we actually don't need to stretch the <mo> at all. The 4th item says we don't care and stretch it anyway.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20171221/75f2d294/attachment-0001.html>