Re: [css-gcpm] order of text replace vs. whitespace processing

Also sprach Ludger Buenger:

 > http://www.w3.org/TR/css3-gcpm/#character
 > <http://www.w3.org/TR/css3-gcpm/#character> sais regarding the
 > text-replace property:
 > 
 > "This property is evaluated after the 'content' property, and
 > before 'text-transform'."
 > 
 > However (if I am not mistaken) it is undefined, whether
 > text-replace is to be applied before or after whitespace
 > processing.
 > 
 > Currently I expect that the sentence above implicitly assumes that
 > whitespace processing happens after text-transform but
 > 
 > CSS2.1 never defines when text-transform is applied with respect to
 > whitespace processing.
 > 
 > In 2.1 it did not matter - text-transform does not modify the
 > content in a way that it affects whitespace processing so applying
 > whitespace processing before text transformation does not change
 > the result.
 > 
 > However with text-replace now being defined to happen before
 > text-transform should't it be defined that/whether text replacement
 > also occurs before whitespace processing?

I'm open for proposals to change the text on this issue. I am,
however, worried about imposing constraints on implementations that
would otherwise not be there. 

When you say 'whitespace' processing, do you refer to the
'white-space' property?

I believe implementation keep the white-space around until just before
the presentation -- if they threw it out at parse time they wouldn't
be able to support the 'white-space' property.

As such, I would suggest specifying that 'text-replace' is evaluated
before 'white-space'.

What do implementors think?

-h&kon
              Håkon Wium Lie                          CTO °þe®ª
howcome@opera.com                  http://people.opera.com/howcome

Received on Tuesday, 31 March 2009 14:43:50 UTC