TK> I would think about this in a different way. In my point of view, this
TK> is a code generation bug in the compiler (gcc?). In a perfect world,
TK> any piece of source code should compile and work perfectly even with
TK> the most exotic optimization flags turned on.
Not necessarily. C9X allows a compiler to make a number of
assumptions about aliasing (about whether two distinct pointers point
at the same data in memory) which older C standards do not allow.
X does a lot of pointer aliasing, and it wouldn't surprise me at all
if it violated some of the restrictions on aliasing in C9X. Recent
versions of gcc do enable -fstrict-aliasing at higher optimisation
levels.
Juliusz
_______________________________________________
Xpert mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/xpert
