Re: floats-137 invalid

On 18/10/10 21:38, L. David Baron wrote:
> http://test.csswg.org/suites/css2.1/20101001/html4/floats-137.htm
> http://test.csswg.org/suites/css2.1/20101001/xhtml1/floats-137.xht
> requires implementations that pass to have a bug in which they do
> not place one of the floats correctly (even though there is exactly
> enough room next to the text prior to it to place it).
>
> (Hixie -- also, cc:ing you on this report, but I haven't cc:ed you
> on a bunch of prior reports of errors in your tests.  If you want me
> to gather up the list of such reports, let me know.)

(Hixie doesn't really maintain any of the tests he wrote — they fall 
into a general bin of unmaintained by anyone specific tests, and 
normally end up being done by fantasai (or occasionally myself).)

> I suggest that the text be fixed by changing the reference diagram
> from its current state:
>
> ############
> #AA1       #
> #AA        #
> #222222222 #
> #BCDEEF3   #
> #BC EE444  #
> # C55555555#
> # C66666   #
> # C        #
> #777777777 #
> #          #
> ############
>
> into:
>
> ############
> #AA1       #
> #AA        #
> #B222222222#
> #BCDEEF3   #
> # C EE444  #
> # C55555555#
> # C66666   #
> # C        #
> #777777777 #
> #          #
> ############
>
> The difference is that float "B" has moved up one row, and the text
> "2" has moved right one column.

Hmm, I'll try and take a look at that on Thursday (I don't really want 
to be editing tests without knowing what I'm doing, for obvious 
reasons). Equally, I'll fix body v. BODY, and the leading new line in 
the pre element then.

However, in general, it seems a good idea to change the test to a 
reftest, yet as it is now the reference differs from the test by way of 
AA artifacts. Any suggestion on how to create a reference that is identical?

-- 
Geoffrey Sneddon — Opera Software
<http://gsnedders.com>
<http://opera.com>

Received on Tuesday, 19 October 2010 22:45:46 UTC