RE: Internet Explorer Implementation Report

John Daggett and I discussed this offline and found that IE9 had a regression in the beta build that caused us to fail these test cases. IE passes them in the builds leading up to the Beta release, and we pass them in the current internal builds. In order to make sure I do the right thing, I'm going to start a new thread on the list, because it may pertain to more than just IE and I don't want it to be missed...

-John

> -----Original Message-----
> From: public-css-testsuite-request@w3.org [mailto:public-css-testsuite-
> request@w3.org] On Behalf Of John Jansen
> Sent: Tuesday, October 26, 2010 11:09 PM
> To: John Daggett
> Cc: public-css-testsuite@w3.org
> Subject: RE: Internet Explorer Implementation Report
> 
> 
> 
> > -----Original Message-----
> > From: John Daggett [mailto:jdaggett@mozilla.com]
> > Sent: Tuesday, October 26, 2010 8:49 PM
> > To: John Jansen
> > Cc: public-css-testsuite@w3.org
> > Subject: Re: Internet Explorer Implementation Report
> >
> > John Jansen wrote:
> >
> > > > Hi John,
> > > >
> > > > > Apologies for not sending this announcement to the public list
> sooner.
> > > > > Our Implementation is up on the site[1] based on our publicly
> > > > > available Beta.
> > > > >
> > > > > (HTML and xHTML tests):
> > > > >
> > > > > Pass: 18960 — 97.40%
> > > > >
> > > > > Fail: 383 — 1.97%
> > > > >
> > > > > Invalid: 60 — 0.31%
> > > > >
> > > > > Skip: 63 — 0.32%
> > > > >
> > > > > As with others, we expect to continue to improve our results
> > > > > over time and will continue to update our IR as the suite changes as
> well.
> > > >
> > > > Was this done with a publicly available IE9 beta or with an
> > > > unreleased version?
> > > > IE claims passing all font-family-name-* tests which I can't reproduce.
> > > >
> > >
> > > It was with the publicly released beta that shipped on Sept 15th.
> > > Did you
> > remember to install the ahem font?
> >
> > The font-family-name-* tests are my tests, they use the CSSTest fonts
> > not Ahem. Among those tests is an explicit check for font family name
> > matching using GDI which the latest IE9 beta appears to still use for
> > font family lookups.  GDI allows matching on names other than the font
> > family and this won't work across platforms (ditto for Postscript name
> > lookups in Webkit).  If the tested version of IE9 uses the
> > FindFamilyName method of a IDWriteFontCollection system font object,
> > then everything will match correctly.
> >
> > Here's an example of what I'm talking about, IE matches "Arial Bold"
> > and Webkit/Mac matches "ArialMT-Bold" as *family* names:
> >
> >   http://people.mozilla.org/~jdaggett/tests/arialvariations.html

> >
> > I constructed explicit checks for these which is why I'm puzzled how
> > all these tests passed.  Maybe they were run without the CSSTest fonts
> installed?
> >
> 
> [John Jansen]
> Ah, apologies about the ahem font, I was looking at the 'font-family', not the
> 'font-family-name' tests. Let's take this offline, because when I look at
> arialvariations.html, it looks to me like IE passes, but I want to make sure I
> understand this and I don't want to paste screen shots here :-).
> 
> I'll email you the screen shot I'm seeing, so we can discuss.
> 
> -John
> 
> > Cheers,
> >
> > John

Received on Wednesday, 27 October 2010 18:45:22 UTC