- From: Gérard Talbot <css21testsuite@gtalbot.org>
- Date: Sun, 31 Jan 2010 09:13:11 -0800
- To: "public-css-testsuite@w3.org" <public-css-testsuite@w3.org>
- Cc: "Arron Eicholz" <Arron.Eicholz@microsoft.com>
Hello all, Some Microsoft tests assume that border: 0.1in solid green; can be rendered without issues or problems by user agents. The problem is that, with a resolution of the device set to 96 CSS pixels per inch, 0.1in computes to 9.6px and here - some user agent will just truncate the decimal part: e.g. Firefox 3.6 and Safari 4.0.4 use 9px - some user agent will round the value to the closest integer: e.g. Opera 10.10 build 1893 uses 10px - some user agent, like IE 8, will apparently use precisely 1 tenth of an inch ... albeit I don't see how this precision in rendering such length can be verified, confirmed. I checked the CSS 2.1 spec and I see no normative restraint or requirement on how fraction of a pixel is supposed to be handled for border (or other properties). Since any/all user agents have implementation limits and various capability limitations, I think those tests should be redone to avoid these. One example where this happens: http://www.w3.org/Style/CSS/Test/CSS2.1/20100127/html4/margin-006.htm A machine strictly and formally checking such test would "see" thin red lines with Firefox 3.6, Opera 10.10, Safari 4.0.4 and so would human testers too. There are, I'm sure, other tests relying on fractional parts; there may be many tests like that. regards, Gérard -- Contributions to the CSS 2.1 test suite: http://www.gtalbot.org/BrowserBugsSection/css21testsuite/ CSS 2.1 test suite (alpha 1; January 27th 2010): http://www.w3.org/Style/CSS/Test/CSS2.1/current/html4/ CSS 2.1 test suite contributors: http://test.csswg.org/source/contributors/
Received on Sunday, 31 January 2010 17:13:45 UTC