serious inline math pb i386

[N8TM@aol.com]

In a message dated 7/7/98 4:01:00 PM Pacific Daylight Time, nbecker@fred.net
writes:

> segfault is in pow
>  (double __x, double __y) from __math.h, but the stack is apparantly
>  corrupted


I doubt that egcs-bugs is the place to mention these things, but I've avoided
using the inline-math templates from glibc because of the sprinkling of errors
of this type.  The culprits are likely to be the instructions which take 2
arguments from the stack or push 2 results on the stack.  Compare the way it's
done there with other people's versions.  pow() is a bit too much to do
reliably on one mouthful of asm() macros, in my opinion.