Re: [CSS21] contradiction on computed value of 'float'

On 28/10/2011 04:25, Sylvain Galineau wrote:
>
> [L. David Baron:]
>>
>> http://www.w3.org/TR/CSS21/visuren.html#propdef-float says that the
>> computed value of 'float' is as specified.
>>
>> http://www.w3.org/TR/CSS21/visuren.html#dis-pos-flo , step 2, says that if
>> 'position' is 'absolute' or 'fixed', then the computed value of 'float' is
>> forced to 'none'.
>>
>> These statements contradict each other.  I believe the second one is what
>> was intended, and we missed this when we introduced the "Computed value:"
>> lines.
>>
>> A testcase is attached.  Gecko and Opera agree with my assessment; WebKit
>> does not.
>>
> IE9 and IE10 also do not.

The forcing of 'float' to 'none' has been around since CSS2, whereas the 
computed value lines were introduced in 2002 by the looks of it.  The 
one for the 'float' property does indeed seem to have been written 
incorrectly.  (Note how the 'display' property is handled better.)

[The notion of specified/computed/used value, both formally and 
informally, seems to have varied during the lifetime of CSS21, and we 
could ask whether it was originally intended that the _used_ value of 
float be the one forced to 'none'.  However, given that the modern spec 
works heavily with computed value I don't think that we could revert to 
that without a lot of undesirable editorial work.]

Let's note this problem for errata.

Cheers,
Anton Prowse
http://dev.moonhenge.net