Re: [css3-text] incorrect order of text operations

On 08/15/2013 07:16 AM, John Daggett wrote:
>
> The CSS3 Text spec has a normative appendix listing the ordering of text operations:
> http://dev.w3.org/csswg/css-text/#appendix-f-text-processing-order-of-oper
>
> # 1. white space processing part I (pre-wrapping)
> # 2. text combination [CSS3-WRITING-MODES]
> # 3. text transformation
> # 4. text wrapping while applying per line:
> #    1. indentation
> #    2. bidirectional reordering [CSS21] / [CSS3-WRITING-MODES]
> #    3. white space processing part II
> #    4. text orientation [CSS3-WRITING-MODES]
> #    5. ‘letter-spacing’ and ‘word-spacing’
> #    6. font/glyph selection and positioning [CSS21] / [CSS3-FONTS]
> #    7. hanging punctuation
> # 5. justification (which may affect glyph selection and/or text
> #    wrapping, looping back into that step)
> # 6. text alignment
>
> There are a couple problems here:
>
> 1. The handling of 'text-combine-horizontal' takes place with
>     'text-orientation', it's the point at which a text run is sliced
>     into individual spans that are either horizontal or vertical in
>     nature. It definitely happens *after* 'text-transform'.

Ok, swapped order of #2 and #3, and shifted "text orientation" to
right after that item. Let me know if it's good:
   http://dev.w3.org/csswg/css-text/#order

> 2. Text alignment and justification should be grouped in the same
>     step.  I think the steps for handling both of these are intertwined
>     and it would be simpler to just group them together rather than
>     making the handling of one normatively precede the other.

While they're both specified in the same set of author controls,
justification and alignment are really two independent operations.
So I'd rather leave it that way in the list.

(If a UA wants to do them together somehow, that's fine; the UA
can do anything it wants as long as the results are the same.)

~fantasai

Received on Tuesday, 20 August 2013 18:47:47 UTC