Mail Archives: djgpp/1994/10/27/01:22:54
In message <9410270204 DOT AA06397 AT cfa0 DOT harvard DOT edu DOT HARVARD DOT EDU>, Bob Babcock write
s:
> I wrote:
> > and print each rgb triplet. On either of the systems I've tried it on it
> > dumps in the readpal routine if I optimize it.
>
> Your readpal routine uses _go32_dpmi_simulate_int() without initializing ss
> and sp. Unless they both happen to be zero, they are used as-is. So, some
> random area of memory is being used as stack space. Turning on optimization
[...]
> registers which you really care about. The docs don't mention it, but some
> recent messages indicated that the value of the flags register could also
> cause trouble if, say, the single-step bit happened to be set.
I had several answers that said this same thing... Thanks dj and er uh,
yeah I nuked the other name. Any way I thought that adding either of
regs.x.ss = regs.x.sp = reags.x.flags = 0;
memset(®s, (unsigned char)0, sizeof(_go32_dpmi_registers));
would fix it... It didn't. Anything else that I did aor am doing wrong?
-Rob
- Raw text -