On Thursday 19 October 2006 16:32, Jason Thorpe wrote:
> On Oct 19, 2006, at 3:10 AM, Havard Eidnes wrote:

> > sysconf() returns long, which isn't neccessarily
> > assignment-compatible with size_t, as evidenced by sh5, so add an
> > explicit cast.
>
> The sh5 port is wrong.  It should be unsigned long in both the LP64
> and the ILP32 case.  E.g. sparc/sparc64 does this.

ISTR having to do this in order to play nice with the original SuperH 
version of gcc used to bring-up the sh5 port. There's a good chance the 
hack (in sh5/ansi.h) can now be removed.

Cheers, Steve