Re: HTML Testing Task Force Conf Call Agenda 9/21/2010

Kris Krueger wrote:
> For #1 - yes lossy
> 
> A better test would not have a 'tolerance' for passing and failing.
> I see Firefox, Safari and IE9 all fail, though they get fail because the tolerance is too high.

Firefox apparently fails on toDataURL.jpeg.quality.basic because of 
https://bugzilla.mozilla.org/show_bug.cgi?id=564388 (the quality 
parameter isn't supported and it conflicts with a proprietary extension 
to toDataURL). Chromium on Linux passes the test for me. Haven't tested 
Safari. IE9 apparently returns (0,11,255,255) instead of (0,0,255,255), 
so increasing the tolerance to e.g. 16 or 32 would let it pass the test. 
(The purpose of the pixel value tests was just to verify that the JPEG 
encoder was doing something vaguely sane, and not returning red or black 
or whatever, so the tolerances can be high.)

-- 
Philip Taylor
pjt47@cam.ac.uk

Received on Tuesday, 21 September 2010 14:51:17 UTC