[Webkit-unassigned] [Bug 170272] Paint <mo> correctly.

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Jan 2 20:13:17 PST 2018


https://bugs.webkit.org/show_bug.cgi?id=170272

--- Comment #36 from Minsheng Liu <lambda at liu.ms> ---
Sorry that I cannot get to you sooner.

> - "<mo> is painted correctly" is vague.

Agreed.

> - Is <mo lspace="100px" rspace="100px"> a good case? It seems that it won't
> be affected by the old "centering" behavior?

Actually it is fine. The old behavior is add a center offset (lspace plus rspace divided by 2) to the existing lspace. The effect of that code to the test case I propose would be having a left space of 200px and right space of 0px. Though I do agree with you that we could add three different cases, where

left = right
left < right
left > right

> - While we are here, maybe test the RTL case too with <math dir="rtl">?

And double the three to six.

> - I would put some visible <mspace> elements before and after the <mo>
> operator, so that one can really see the operator spacing. Actually, you
> could use such <mspace> elements to indicate the reference 100px/200px
> widths.

My intention was to have them aligned so that a vertical comparison is possible. But you are right that the operator spacing would be hard to see in this case. I might change it to have a combination of both.

> - red/green is generally used to suggest failure/success, so the reference
> widths could have more neutral colors. See
> http://web-platform-tests.org/writing-tests/rendering.html

I was not aware of that : )


(In reply to Frédéric Wang (:fredw) [back 03/01/2018] from comment #35)
> Thanks for finishing the work on this Minsheng. Some quick comments:
> 
> - "<mo> is painted correctly" is vague. Can you provide more details in the
> bug summary and description? If I remember correctly the issue was that
> operator spacing was incorrect for horizontally stretched operators, because
> some centering was performed before? And also maybe because the logical
> width was incorrect too?
> 
> - Is <mo lspace="100px" rspace="100px"> a good case? It seems that it won't
> be affected by the old "centering" behavior? I guess you want to check a
> case where lspace != rspace e.g. <mo lspace="100px" and rspace="200px">?
> 
> - While we are here, maybe test the RTL case too with <math dir="rtl">?
> 
> - I would put some visible <mspace> elements before and after the <mo>
> operator, so that one can really see the operator spacing. Actually, you
> could use such <mspace> elements to indicate the reference 100px/200px
> widths.
> 
> - red/green is generally used to suggest failure/success, so the reference
> widths could have more neutral colors. See
> http://web-platform-tests.org/writing-tests/rendering.html

-- 
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/20180103/db5b433c/attachment.html>


More information about the webkit-unassigned mailing list