From: Joaquin Cuenca Abela (cuenca@pacaterie.u-psud.fr)
Date: Wed Apr 24 2002 - 03:28:52 EDT
----- Original Message -----
From: "Andrew Dunbar" <hippietrail@yahoo.com>
To: <abiword-dev@abisource.com>
Sent: Tuesday, April 23, 2002 6:15 PM
Subject: Re: undo and combining characters
> --- Karl Ove Hufthammer <huftis@bigfoot.com> wrote: >
> "Tomas Frydrych" <tomas@frydrych.uklinux.net> wrote
> > in
> > news:3CC58433.16110.B1F416@localhost:
> >
> > > I think in the case of the Arabic ligature, these
> > have to be
> > > treated as two characters, i.e., pressing
> > backspace after the
> > > second one leaves you with the first one. This
> > case is not a
> > > real issue, because internally the ligature is
> > stored as two
> > > separate characters, ligature is just a way of
> > displaying in
> > > them in a way that looks better,
> >
> > But how does selection works? Are the glyphs
> > 'decomposed' to allow
> > selection (which causes a reflow), and religated
> > when you move the
> > selection?
>
> 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.
Other ligatures look radically different from the original glyphs (arabic
have some of these). To take an occidental analogy, put yourself in the
epoc before the ligature "&" (from the latin "et". It's more visible using
the roman form of "&") become a character itself.
Now, should abiword try to select the 't' of "&"? (hell, what 't'? The
original glyph it's absolutely invisible in the ligated glyph!)
So I think that for some ligatures individual selection of the characters
that build the ligature makes sense, and for others no (it's only me, or all
that unicode stuff is starting to sound *REALLY* hard?)
Cheers,
-- Joaquin Cuenca Abela cuenca@pacaterie.u-psud.fr
This archive was generated by hypermail 2.1.4 : Wed Apr 24 2002 - 03:26:33 EDT