Cesar Crusius <[EMAIL PROTECTED]> writes:
> why? is it better to have to implement the same thing over every
> single program that wants to put boxes modified internally in a
> generic way? right now, i think ggiPutBox and ggiGetBox are functions
> with very little practical use.
correct me if i'm wrong, but apart from displays with 32 bit pixels,
your routine would slow down a putbox operation considerably.
i'd take it you would want to implement this routine (briefly, example
for 8bpp) as:
uint8* dest = my_8bpp_array;
uint32* src = my_pixel_array;
for ( i = 0; i < width*height; i++ )
{
*dest++ = *src++;
}
most people wouldn't want such a slowdown. they would just render
into 8bpp arrays internally and ggiPutBox() that to the screen
directly.
what kind of software is it you are writing? if you care a bit about
speed, i don't see how you could come up with an idea like this. if
you don't, then i'd suggest to put this in an extension to ggi, but
please keep it out of libggi; it is, as you quoted Lee (his post
hasn't arrived here yet), ``bloat''.
most people know exactly what the layout of their display target is to
save on cpu and memory usage.
again, i'd like to stress that reading books might work as an
eye-opener here. perhaps you haven't considered the problems that
arise with indexed graphics modes or with black and white modes. you
can try all sorts of fancy generalizations that just don't work. so
they may have a nice place in an extension library, but i would never
consider it to be a useful part of the core functionality as found in
libggi.
perhaps i misunderstand you completely though, correct me if i went
wrong with this post; i hope this helped :/
--
Tijs van Bakel, <[EMAIL PROTECTED]>
