>
> > GL text drawing has been recently implemented in 3 steps on Mac OS:
> > 1) draw text to a bitmap using standard OS text drawing routines
> > 2) convert this bitmap into a GL texture
> > 3) map this texture to the GL scene
>
> When was this implemented?
This came with the cocoa-ization of FLTK (r7027).

>
> As of Sunday evening, testing on my OSX 10.4.11 PPC box, GL text
> rendering was not working.

Thanks for pointing that. I had not tested the ppc architecture.
I have now put back the old algo (with its drawbacks) for the ppc
architecture. Cube and fractals now display text as needed
also with ppc.

>
> Also, STR #2342 refers, which I raised on Saturday.

I believe I will close this one now.

>
>
> > Because the pile height has no obvious universal adequate value,
> > I have added 2 mac-specific functions:
> > void gl_texture_pile_height(int max);
> > int gl_texture_pile_height();
> > to set and get the texture pile height.
>
> Not keen on that - how can the programmer know what is a sensible thing
> to set these too?
> Can we do something at runtime to allocate RAM dynamically based on
> <something> e.g. installed RAM or such like...?
>
>

This is an annoying aspect of this solution to improve the speed
of GL text rendering. The rule is simple: set the pile height
to the number of simultaneously displayed strings on a GL scene.
This rule is mentionned in the doc.
The situation is that we --have to-- do some texture pre-computation
or else GL becomes annoyingly slow.
_______________________________________________
fltk-dev mailing list
[email protected]
http://lists.easysw.com/mailman/listinfo/fltk-dev

Reply via email to