H. Peter Anvin wrote:
> Ed Beroset wrote:
>> H. Peter Anvin wrote:
>>> No leak, looks more like an overrun, especially since compiling with
>>> LOGALLOC moved the bug into the RAA layer, which looks like it, too,
>>> could use tossing out and restarting ... it has a bloody ldiv() in it...
>> What's wrong with ldiv()? That's both standard (C89) and efficient.
> Because using a divide when a shift will do is stupid at best.
OK, I thought you were saying that there was something intrinsically
wrong with using ldiv(). The real problem is that it's been misapplied
here. I just looked at the context, and you're right -- that code