On 1/5/06, Tim Janik <[EMAIL PROTECTED]> wrote: > On Thu, 5 Jan 2006, Mike Emmel wrote: > > > On 1/5/06, Tim Janik <[EMAIL PROTECTED]> wrote: > >> On Thu, 5 Jan 2006, ANDREW PAPROCKI, BLOOMBERG/ 731 LEXIN wrote: > >> > >>> I disagree with this completely. We compile all of our code on > >>> Solaris/AIX with > >>> SunPRO cc & IBM xlc without c99 extensions enabled for our own reasons. > >> > >> please tell us what kind of reasons those are. considering and > >> re-evaluating them > >> is exactly what this thread is for ;) > >> > > > > Well if issues like this are going to come up I did some investigation > > on using libffi for signal dispatch and my conclusion was it would be > > significantly faster then the current dispatch system. > > Just food for thought. > > it wouldn't, i evaluated it for emission when i wrote the > marshaller generation code for GSignal. > Really ? Can you explain a bit more. I saw that it could reduce dispatch cost by about 50% at least since your looking up the function and directly pushing the args.
> > Mike > > --- > ciaoTJ > _______________________________________________ gtk-devel-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/gtk-devel-list
