RE: border-*-width-036/047 (etc.?) make invalid assumptions about rounding

On Thursday, October 14, 2010 10:08 AM Gérard Talbot wrote:
> L. David Baron wrote:
> > There are a bunch of border-related tests that fail in Gecko because
> > the tests make invalid assumptions about rounding behavior.  Gecko has
> > rounding behavior that has served us very well on the Web:  we have
> > different rounding behavior for border-widths (where pixel-width
> > consistency is important) than for most other dimensions.  (This
> > different behavior is shared by outline widths and hopefully soon by
> > line heights.)
> >
> > The following tests fail in Gecko, despite Gecko's behavior being
> > valid according to the spec:
> >
> > border-bottom-width-036
> > border-bottom-width-047
> > border-left-width-036
> > border-left-width-047
> > border-right-width-036
> > border-right-width-047
> > border-top-width-036
> > border-top-width-047
> 
> 
> David,
> 
> In
> 
> http://lists.w3.org/Archives/Public/public-css-testsuite/2010Aug/0059.html
> 
> I reported these problems and proposed replacements for each of them.
> They are still available:
> 
> http://www.gtalbot.org/BrowserBugsSection/css21testsuite/border-
> bottom-width-036.htm
> 
> http://www.gtalbot.org/BrowserBugsSection/css21testsuite/border-top-
> width-036.htm
> 
> http://www.gtalbot.org/BrowserBugsSection/css21testsuite/border-
> bottom-width-047.htm
> 
> http://www.gtalbot.org/BrowserBugsSection/css21testsuite/border-left-
> width-047.htm
> 
> http://www.gtalbot.org/BrowserBugsSection/css21testsuite/border-right-
> width-047.htm
> 
> http://www.gtalbot.org/BrowserBugsSection/css21testsuite/border-top-
> width-047.htm
> 
> 
> I also reported these again in
> 
> http://lists.w3.org/Archives/Public/public-css-testsuite/2010Sep/0098.html
> 
> and then again (along with others in) in
> 
> http://lists.w3.org/Archives/Public/public-css-testsuite/2010Sep/0161.html
> 
> -------------------
> 
> Missing in your list are:
> 
> http://test.csswg.org/suites/css2.1/20101001/html4/padding-top-050.htm
> 
> and
> 
> http://test.csswg.org/suites/css2.1/20101001/html4/padding-top-051.htm
> 
> You have to view these padding-top-05? in Opera 10.63.
> 

I have updated all the border-*-036 and 047 cases. I would like one or both of you to review them before I work on a similar fix for padding, etc...

Please take a look and let me know. Also don't forget to remove the cases from the http://wiki.csswg.org/test/css2.1/invalid page if the updates are acceptable.

--
Thanks,
Arron Eicholz

Received on Thursday, 27 January 2011 19:38:42 UTC