From: Andrew Dunbar (hippietrail@yahoo.com)
Date: Fri Apr 26 2002 - 04:02:41 EDT
--- Tomas Frydrych <tomas@frydrych.uklinux.net>
wrote: >
> > > (1) Make glib 2.0 available on all our platforms
> >
> > Can you give us a rough idea what feautures or
> > specific functions from glib are required? We may
> be
> > able to split out what is needed. This would also
> > count toward "Draining the Swamp".
>
> Pango uses glib stuff throughout, types, lists,
> everything -- I think
> forking Pango is not a realistic option, and whether
> we could use
> just a reduced version of glib, and how reduced,
> could only be told
> by trying -- I think porting the whole thing might
> be easier. (They
> already have a makefile system for win32)
That means we'll have multiple implementations of
some basic concepts inside Abi which sounds a bit ugly
but I don't know how else to avoid it other than
embracing glib whole-heartedly and using its types
and lists ourselves - that also sounds ugly ):
> > > (4) Get FreeType to compile on all our
> platforms;
> > > this should not be
> > > a problem.
> >
> > I've been told it already compiles on all our
> > platforms.
>
> There is not a build system for QNX as far as I
> know, but the
> FreeType code is very clean, I think this would be a
> formality.
>
> > > (6) replace the current shaping engine with
> Pango.
> >
> > What about keeping both like we have done so far
> with
> > the bidi/non-bidi builds? It might take a while
> for
> > the Pango version to be useable for those not
> working
> > on the Pango-related stuff.
>
> The current shaping engine is inside the
> bidi-enabled build, and the
> Pango engine would go only inside the bidi-enabled
> branch, there
> is no point to use it in the non-bidi branch,
> especially if to start with
> it will only do Hebrew and Arabic. So while we are
> woking on it,
> there would be a non-bidi build available, but
> eventually the non-bidi
> code should be completely removed from the sources.
> I am not
> prepared to maintain two bidi builds though --
> sorry.
Does this mean we'll have to do some actual
experimentation with Pango to see if it's really going
to be viable. If it doesn't work straight away we'll
have no working build. Since we're going to be using
FreeType/OpenType either directly or indirectly (via
Pango). If we need to enhance Pango before it works
for us, what will happen in the meantime? I thought
going with dual FreeType/OpenType vs. Pango builds
might be the only way until Pango does what we want.
Since I think the layout and the rendering code is
going to change no matter what, keeping the current
bidi/non-bidi stuff isn't going to work anyway is it?
And if Pango gives us generic stuff such as line
breaks
and combining characters and selection then it will be
of use even for those not using Arabic/Hebrew.
Shaping
isn't the whole of Pango.
Andrew Dunbar.
> Tomas
=====
http://linguaphile.sourceforge.net http://www.abisource.com
__________________________________________________
Do You Yahoo!?
Everything you'll ever need on one web page
from News and Sport to Email and Music Charts
http://uk.my.yahoo.com
This archive was generated by hypermail 2.1.4 : Fri Apr 26 2002 - 04:04:13 EDT