W3C home > Mailing lists > Public > public-css-testsuite@w3.org > December 2010

Re: line-height-014 is invalid

From: L. David Baron <dbaron@dbaron.org>
Date: Thu, 2 Dec 2010 10:31:19 -0800
To: Arron Eicholz <Arron.Eicholz@microsoft.com>
Cc: "public-css-testsuite@w3.org" <public-css-testsuite@w3.org>
Message-ID: <20101202183119.GA24518@pickering.dbaron.org>
On Thursday 2010-12-02 17:35 +0000, Arron Eicholz wrote:
> On Wednesday, December 01, 2010 7:45 PM: L. David Baron wrote:
> > On Wednesday 2010-12-01 19:43 -0800, L. David Baron wrote:
> > > On Thursday 2010-12-02 02:27 +0000, Arron Eicholz wrote:
> > > > On Friday, October 15, 2010 5:04 PM, L. David Baron wrote:
> > > > > These tests:
> > > > > http://test.csswg.org/suites/css2.1/20101001/html4/line-height-014
> > > > > .htm
> > > > > http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-01
> > > > > 4.xht are invalid because of rounding issues.
> > > > >
> > > > > They use a font-size of 16px and a line-height of 1.3333px, and
> > > > > assert that the top of the font in the first line and the bottom
> > > > > of the font in the next line are separated by 17px.  However,
> > > > > since CSS does not define rounding behavior, they could reasonably be
> > separated by 18px.
> > > > >
> > > > > I'd recommend testing rounder values instead, or loosening the pass
> > criteria.
> > > >
> > > > Fixed, changed the font value to 20px.
> > >
> > > While that does happen to fix it in one case (Gecko on Linux with my
> > > default font size), I don't think it's sufficient since the
> > > line-height is still 1.3333px.
> > 
> > Also, when the test does pass, the pass condition is rather unclear since the
> > boxes are touching (and therefore look like a single box, not two).
> 
> Fixed the spacing issue with the 2 boxes.
> 
> The 1pt value can't be changed however since this test is
> specifically testing the boundary conditions of minimum value plus
> one. This will mean that the value will always be a fraction of a
> pixel and the user agent needs to properly account for that in a
> consistent/predictable way.

There are plenty of consistent and predictable ways that don't match
your test assertion.  If you specifically want to test the 1pt
value, you need to find another way to do it.

-David

-- 
L. David Baron                                 http://dbaron.org/
Mozilla Corporation                       http://www.mozilla.com/
Received on Thursday, 2 December 2010 18:31:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 2 December 2010 18:31:58 GMT