Re: [css3-text] grapheme clusters across element boundary

coloring characters that participate in a grapheme cluster, ligature, etc.,
or produce context for surrounding characters (e.g., Arabic), is a
legitimate use case and a real world requirement, so it can't be simply
dismissed or discouraged; however, there may be constraints when a font
transition (family or face) is required across a boundary: that would be
sufficient to block ligature formation, and perhaps even contextual
substitution

On Tue, Jan 17, 2012 at 1:40 AM, Jonathan Kew <jonathan@jfkew.plus.com>wrote:

> On 16 Jan 2012, at 23:15, fantasai wrote:
> >
> > Added:
> >  | The rendering characteristics of a <i>character</i> divided by an
> >  | element boundary is undefined: it may be rendered as belonging to
> >  | either side of the boundary, or as some approximation of belonging
> >  | to both. Authors should avoid dividing grapheme clusters by element
> >  | boundaries.
> >
> > How's that?
>
> That appears to specifically _discourage_ examples such as colouring a
> diacritic differently from its base character, which is a use case that
> gets requested from time to time:
>
>  <div style="font:18px Times">
>    Careful authors use the spelling
>    ‘na&#x131;<span style="color:red">&#x308;</span>ve’
>    rather than
>    ‘na&#x131;<span style="color:red">&#x307;</span>ve’.
>  </div>
>
> Support for this is currently quite limited (the above renders nicely in
> Gecko on OS X, where the Times font does not use precomposed glyphs for
> these combinations, but the colour is lost on a platform where Times is
> replaced by Times New Roman, for example); still, it seems unfortunate to
> tell authors that they shouldn't be attempting to do this at all, rather
> than just warning them that the desired rendering may not (yet?) be widely
> implemented.
>
> JK
>
>
>