Re: glibc and egcs....
On Wed, 27 Jan 1999, Paul Slootman wrote:
> Looks great. One problem with egcs, though:
>
> -------------------------------------
> __complex__ float wp[2048];
>
> int FFT() {
> int h;
> int zp;
> int k=0;
> int p=0;
> __complex__ float c[10];
>
> for(h = 10; h != 0; h--) {
> zp = 0;
> for(k=0; k < 10; k++) {
> c[k+p] = (c[k] - c[k+p])*wp[zp];
> }
> }
>
> return 0;
> }
> -------------------------------------
>
> This crashes every version egcs I have (all two of them :-) 1.03 and the
> current one). Perhaps you know who to report this to?
I think I can squeeze this one in. I'll even submit it for a testcase if
I can't get it fixed right off or appears to be a reoccurring problem.
> This comes from one of the not-yet compiled sound packages (I believe),
> after I reduced it to the minimal problem (so don't complain the
> function doesn't make any sense; it did before I hacked it; egcs
> crashed in the same way before that).
Hehehehe...I feel your pain. I'll take a look at the compiler messages
and submit from there. Thanks!
C
Reply to: