<html>
<head>
<base href="https://bugs.webkit.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_RESOLVED bz_closed"
title="RESOLVED DUPLICATE - imported/mathml-in-html5/mathml/presentation-markup/scripts/subsup-parameters-1.html is a flaky timeout"
href="https://bugs.webkit.org/show_bug.cgi?id=158993#c7">Comment # 7</a>
on <a class="bz_bug_link
bz_status_RESOLVED bz_closed"
title="RESOLVED DUPLICATE - imported/mathml-in-html5/mathml/presentation-markup/scripts/subsup-parameters-1.html is a flaky timeout"
href="https://bugs.webkit.org/show_bug.cgi?id=158993">bug 158993</a>
from <span class="vcard"><a class="email" href="mailto:fred.wang@free.fr" title="Frédéric Wang (:fredw) <fred.wang@free.fr>"> <span class="fn">Frédéric Wang (:fredw)</span></a>
</span></b>
<pre>(In reply to <a href="show_bug.cgi?id=158993#c6">comment #6</a>)
<span class="quote">> (In reply to <a href="show_bug.cgi?id=158993#c5">comment #5</a>)
> > (In reply to <a href="show_bug.cgi?id=158993#c4">comment #4</a>)
> > > But it's an imported test, so we can't really modify it.
> >
> > Indeed, the ideal would be not to change it.
>
> Ideally, if it is getting imported once, we should be able to reimport the
> tests again and again. Tests should be able to improve, like source code.
>
> Is there a git repository, like for W3C wpt/css test suites?
> If so, WebKit changes should be contributed to that repo.
>
> For WPT tests, when I need to change a test, I add a comment like: //WEBKIT
> change:... to the WebKit test.
>
> I then upstream the changes to W3C repo.
> When I reimport the tests from W3C repo (once a while), the only remaining
> change is the removal of the // WEBKIT comment.</span >
Yes, the repo is at <a href="https://github.com/MathML/MathMLinHTML5-tests">https://github.com/MathML/MathMLinHTML5-tests</a> for now and indeed I agree with you that it is great if browser vendors contribute there when there are mistakes in the test. The change discussed here is actually a workaround for font loading bug in WebKit (<a class="bz_bug_link
bz_status_NEW "
title="NEW - [Font Loading] The callback passed to document.fonts.ready should always be called"
href="show_bug.cgi?id=158884">bug 158884</a>). So my point is that it is fine to do a small hack so that we can run the actual MathML test but ideally we should not deviate too much from upstream and even less contribute back such hacks ;-). The best remains to fix <a class="bz_bug_link
bz_status_NEW "
title="NEW - [Font Loading] The callback passed to document.fonts.ready should always be called"
href="show_bug.cgi?id=158884">bug 158884</a>, so that the test passes when we open <a href="http://tests.mathml-association.org/mathml/presentation-markup/scripts/subsup-parameters-1.html">http://tests.mathml-association.org/mathml/presentation-markup/scripts/subsup-parameters-1.html</a> in WebKit :-)
BTW, this test also has another change to disable the SuperscriptShiftUpCramped part until <a class="bz_bug_link
bz_status_NEW "
title="NEW - Use SuperscriptShiftUpCramped parameter from the OpenType MATH table"
href="show_bug.cgi?id=156401">bug 156401</a> is fixed.
Finally, the path to the fonts has also been modified to be executed... That also seems to be what is done in some tests for LayoutTests/imported/w3c/canvas when loading the CanvasTest.ttf font...</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the assignee for the bug.</li>
</ul>
</body>
</html>