From: Karl Ove Hufthammer (huftis@bigfoot.com)
Date: Wed Apr 24 2002 - 06:21:48 EDT
"Joaquin Cuenca Abela" <cuenca@pacaterie.u-psud.fr> wrote in
news:03da01c1eb61$aee36de0$231611ac@lou:
>> Selection should select the entire ligature.
>
> The problem here is that some ligatures look "mostly" as the
> original glyphs (occidental "ff" "fi" "ffi" "st", etc.)
> Selecting the entire ligature here is VERY surprising
> behaviour for the users.
OK. Here's a proposal. There are very few ligatures in the Latin
alphabet (only the ones you've mention are in actual use, I
believe), so we can hardcode a different/special behaviour for
these. Selecting an 'fi' ligature should select approx. half the
glyph. This will work adequately. There *will* be some overlap,
but that's not really a *big* problem. Note that there will also
be some overlap when selecting kerned characters (I certainly hope
Pango supports kerning!). E.g. in the string 'Ty', most of the 'y'
will be positioned under the crossline of the 'T', and selecting
the 'y' will also mark part of this crossline. Again, this is IMO
not a problem.
In other cases (Arabic &c.), we should select the entire glyph (or
do whatever other Arabic-aware WPs do).
-- Karl Ove Hufthammer
This archive was generated by hypermail 2.1.4 : Wed Apr 24 2002 - 06:24:21 EDT