Bug 115789 - MathML Acid1 fails to render properly
Summary: MathML Acid1 fails to render properly
Alias: None
Product: WebKit
Classification: Unclassified
Component: MathML (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Nobody
URL: http://fred-wang.github.io/AcidTestsM...
Keywords: InRadar
Depends on: 44208 115786 115787 119038 122297
  Show dependency treegraph
Reported: 2013-05-08 03:10 PDT by Frédéric Wang (:fredw)
Modified: 2014-06-18 00:30 PDT (History)
5 users (show)

See Also:

Screenshot on nightly + patches for 115786, 124838, 115787 (80.12 KB, image/png)
2014-02-26 01:14 PST, Frédéric Wang (:fredw)
no flags Details
Screenshot Nightly WebKitGTK+ (26.58 KB, image/png)
2014-06-18 00:29 PDT, Frédéric Wang (:fredw)
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Frédéric Wang (:fredw) 2013-05-08 03:10:08 PDT
Webkit almost renders the MathML Acid1 properly, but a few requirements are not respected:


- the spacing of operators (especially invisible operators and commas) is not always good
- in test 3, "bang" and "whimper" <mi>'s should not be italic
- the stretching of parenthesis (e.g. in test 6) in is not always done for me (with Martin Robinson's patch), but that seems a random failure. I'm not sure what the problem is but that will probably be solved if we have size variants.
Comment 1 Frédéric Wang (:fredw) 2013-05-08 14:53:38 PDT
> but that seems a random failure.

I'm wondering if that has something to do with the reflow that happens when the Web fonts arrive.
Comment 2 Jon Lee 2013-11-26 18:37:30 PST
Comment 3 Frédéric Wang (:fredw) 2014-02-26 01:14:45 PST
Created attachment 225241 [details]
Screenshot on nightly + patches for 115786, 124838, 115787

We are getting close. However, in test 4 the fences do not stretch. I think this is because the height is too small and so we fallback to the base size:

>  // Sometimes we cannot stretch an operator properly, so in that case, we should just use the original size.
>  m_stretchyCharacter = m_isStretched ? findAcceptableStretchyCharacter(stretchedCharacter) : 0;
>  if (!m_stretchyCharacter)
>    m_isStretched = false;

So I think we will need the work of bug 122297 to be able to use size variants for small sizes (and we could do bug 119038 at the same time).
Comment 4 Frédéric Wang (:fredw) 2014-06-18 00:29:45 PDT
Created attachment 233291 [details]
Screenshot Nightly WebKitGTK+

Testing on WebKitGTK+, I no longer see the problem with stretchy parenthesis in test 4 and the radical sign looks slightly better (even with the current default non-OpenTypeMATH font).