Re: rendering and verification of glyphs

Ray Kiddy wrote:
> 
> I find this to be a problem generally and I am surprised to see this in 
> the CSS spec. Or am I misunderstanding something?
> 
> See: http://www.w3.org/Style/CSS/Test/CSS2.1/current/t1202-counter-09-b.htm
> 
> I am on a Mac running Mac OS X 10.4.9. With Firefox 2.0.0.4, if I 
> display this page, I see either two rows of question marks, or two rows 
> of glyphs that are made of squares with a 4-digit hex number in them, 
> which is obviously how FF is displaying unknown Unicode code points.
> 
> With a debug build of the Firefox trunk (to be FF 3), I see only the 
> hex-square glyphs.
> 
> With Opera, I see just squares.
> 
> In all these cases, the test would seem to be passing, as it says "The 
> following two lines should look the same:".
> 
> Should all of these tests be passing?
> 
> Is the CSS saying something about what we should be seeing here? Or is 
> it just saying that, whatever one sees, one should see two identical 
> rows of them?
> 
> The latter would make sense. The former opens up big questions. How can 
> a visual comparison within a web page, such as this, verify correct 
> glyph rendering? I am working on tests that involve image comparisons to 
> get around this limitation, but I am not sure that would be desirable here.
> 
> Can anyone speak to the intent of this and similar tests?

You're right, there's a problem here. You need to have the appropriate
fonts for any test results for these to be valid, and we should add that
to the test description. Can you make a list of tests with these problems?
You can add them to
   http://csswg.inkedblade.net/test/css2.1/issues#tests-assume-non-latin-font-availability
or just post them here.

~fantasai

Received on Tuesday, 3 July 2007 02:27:11 UTC