Re: Re: Qt for GNOME?

[Emmanuele Bassi <ebassi gmail com>]

On Wed, 2008-07-16 at 17:45 +0200, Sascha Peilicke wrote:

> > > I'm not into all the details relating pango and cairo implementation but
> > > at least the latter has a weakness, it has only a double-precision
> > > rendering path, no single-precision and no int-based, which is bad for
> > > embedded devices which have limited capabilities (no floating point unit,
> > > ...).
> >
> > wrong. cairo has an internal fixed point path and floating point API,
> > and has had it for quite a while now. embedded devices have been using
> > cairo since release 1.2.0 or 1.4.0 - and now we are at 1.6.x.
> >
> > ciao,
> >  Emmanuele.
> 
> My apologies, didn't know that. Why is it not an external interface?

because the number of devices using softfloat is going to diminish in
the near future; because there's no sufficient proof that it would be a
good thing; because it would be quite out of place in the current API
scheme; because you can really do all the computations in fixed point,
using the precision you like, and then convert to floating point before
using cairo.

ciao,
 Emmanuele.

-- 
Emmanuele Bassi,
W: http://www.emmanuelebassi.net
B: http://log.emmanuelebassi.net