Re: [RC6] white-space-processing-056: false positives

Le Sam 11 août 2012 18:19, L. David Baron a écrit :
> On Thursday 2012-07-26 22:27 -0400, "Gérard Talbot" wrote:
>> Hello,
>>
>> [RC6]
>> http://test.csswg.org/suites/css2.1/20110323/html4/white-space-processing-056.htm
>>
>> [nightly-unstable]
>> http://test.csswg.org/suites/css2.1/nightly-unstable/html4/white-space-processing-056.htm
>>
>> As far as I can see this, Firefox 4.0.1 under XP SP3, Firefox 14.0.1
>> under Linux KDE 4.8.4, Firefox 17.0.1a (nightly build 20120726) under
>> XP
>> SP3 and Safari 5.1.7 under XP SP3 fail such test.
>>
>> The submitted implementation reports suggests otherwise.
>>
>> I think the pass/fail conditions sentence is written in a way that is
>> not so clear. What's a box? If an irregular polygon can be a "box",
>> then
>> it's a pass..
>
> What behavior do you see?
>

David,

I have modified the test 1-2 hours ago.

Attached is screenshot of
http://test.csswg.org/source/approved/css2.1/src/text/white-space-processing-056.xht
with Firefox 14.0.1 under Linux KDE 4.9.0

white-space-processing-056-Firefox14.0.1LinuxKDE4.9.0.png , 600px by
300px, 21918 bytes


> I don't think the Ahem font provides a glyph for ideographic space,

I never considered such possibility; I would not know.

> so it seems that you'd get a glyph from a different font -- and I
> could imagine the results being substantially different depending on
> which font, and whether that font has an ideographic space that's
> exactly 1em wide.

Well, then I can not explain some browser passing and failing while
declaring the same browser default font.

For this test:
http://test.csswg.org/source/approved/css2.1/src/text/white-space-processing-056.xht

I declare DejaVu Serif as default font in all browsers I use under KDE
4.9.0; I declare Times New Roman as default font in all browsers under
Windows XP Pro SP3.

Opera 12.0.1 passes this test under Linux KDE 4.9.0 and Windows XP Pro
SP3.  Safari 5.1.7 under Windows XP Pro SP3 fails this test.
Chrome 21.0.1180.75 under Linux KDE 4.9.0 passes this test.

>
> (I see a 1px gap in the middle of what "should be" a single box, but
> nothing out of alignment.)

That's a 1px vertical gap, right?

Over here, I get a 4px (or so) horizontal offset under Linux KDE for the
beforementioned browsers; such horizontal offset seems bigger, wider
under Windows XP Pro SP3. I can attach (or upload) more screenshots if
needed.

Gérard
-- 
Contributions to the CSS 2.1 test suite:
http://www.gtalbot.org/BrowserBugsSection/css21testsuite/

CSS 2.1 Test suite RC6, March 23rd 2011:
http://test.csswg.org/suites/css2.1/20110323/html4/toc.html

CSS 2.1 test suite harness:
http://test.csswg.org/harness/

Contributing to to CSS 2.1 test suite:
http://www.gtalbot.org/BrowserBugsSection/css21testsuite/web-authors-contributions-css21-testsuite.html