[Webkit-unassigned] [Bug 66295] [Meta] Support a reftest which follows CSS test format guidelines defined by CSS working group

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=66295





--- Comment #8 from Ryosuke Niwa <rniwa at webkit.org>  2011-11-02 22:32:44 PST ---
(In reply to comment #7)
> The code in single_test_runner runs different code paths based on whether it's a ref test or not, and that branch is done before we run the test. So we'd either have to open the test file ourselves and figure out that it's a reftest (making things somewhat fragile because we'd have to duplicate the logic in DRT and incurring a perf hit), or restructure the code so that we get that info handed back from DRT (as Ojan suggests in comment #3, which *I think* is true, but I'm not 100% positive).

I've looked into this just now but it appears like we'd have to make quite few changes to each port's DRT. It's doable but I'd rather avoid it if possible.

But I'm also strongly against special-casing directories. So I guess it all depends on the perf hit. Also, we might want to try http://lxml.de/html5parser.html since it implements HTML5 parsing algorithm.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.