Re: CSS3 Text and UAX14

On 2/20/2007 3:27 AM, fantasai wrote:
> Anne van Kesteren wrote:
> >
> > On Tue, 20 Feb 2007 11:22:23 +0100, fantasai
> > <> wrote:
> >> Your argument has convinced me that CSS3 Text should be normatively
> >> requiring the correct implementation of UAX14's normative line 
> breaking
> >> classes.
> >
> >
> Yep. But there's no reason the behavior specified in UAX14 for mandatory
> breaks (CR, LF, NEL, etc) shouldn't be required. For CSS3 Text, the
> behavior specified for BK, CR, LF, CM, NL classes can, I think, be safely
> required in all cases. The behavior for WJ, ZW, and GL should be required
> in normal text wrapping.
with rule 12 modified as in the modifications proposed for 5.0.1, i.e to 
break between, SP, BA, HY on one side and GL on the other, as needed
to support Portuguese and Polish which use sequences of <SHY, NBHY>
(the draft document that explains all of that adjustment will be up 
> I don't think we care about SG either way.
SG only applies to broken encoding, either unpaired surrogates in UTF-16,
or surrogate code points mistakenly present in UTF-32 or UTF-8. There's
no need to deviate from Unicode's treatment of these, as by definition there
can't be an overriding requirement that's based on established convention.

The final normative LB class in UAX#14 is SP (which contains only U+0020).
See previous message). Let's see whether we can settle the issues around
> References to any other parts of UAX14 should be informative only.

Because we allow tailoring in UAX#14, CSS could describe certain widely used
tailorings (e.g. the Korean 'space based" line breaking) as well as further
allow UA level tailoring.

Therefore, it should be possible to get what you want, even with a normative
reference to UAX#14. (in light of the discussion in my earlier message).
> (The required behavior of SP is imho, not clearly defined. I don't 
> want to
> import normative text that may or may not contradict the CSS spec 
> itself.)
Let's settle the issue around SP. I need more detail, for that one.
See other message.

> ~fantasai

Received on Tuesday, 20 February 2007 22:22:27 UTC