[FFmpeg-devel] [PATCH] RoQ video encoder, take 2

[Reimar Doeffinger]

Hello,
On Mon, May 28, 2007 at 04:25:14PM +0200, Michael Niedermayer wrote:
> On Mon, May 28, 2007 at 04:24:45PM +0200, Reimar Doeffinger wrote:
> > If anything the bug is in the C spec for giving it a name that promises
> > something that can't be done. Something like uint_native8_t would be
> 
> hmm, hmm
> it could be done
> 
> for example int_fast*_t could be defined as a type which is at least 8bit
> and which should be fast for the opertaions it is used in
> and that 2 int_fastX_t variables can differ in their type, which also
> would cause sizeof(int_fast*_t) to be invalid C

Then it wouldn't be any C types anymore, but more something like
-ffast-math...
I agree that sometimes it would be nice if you could tell the compiler
even for integer operations "I do not care about overflow cases", then
it would automatically use 32 or 64 bits for loop variables on AMD64
depending on whether it can use the variable directly for addressing or
not (which might even depend oncompiler options).
But asking the int_fast*_t types to do that is asking too much IMO ;-).

Greetings,
Reimar Doeffinger