Re: someone could port libgc
There are some potential problems with your patch. It doesn't look to me
like it would really work. I think it is a bit premature to send patches
like this upstream when you have not been able to adequately test them.
The type of signal handlers on the Hurd is BSD-style:
void (*handler) (int sig, int code, struct sigcontext *scp)
For SIGSEGV, the code will be the fault address.
There is no need to reset signal handlers every time on the Hurd (or on
linux with glibc).
As to the stack bounds, the problem here is that it's not a fixed range of
addresses. The value you got (0x101a000) will not necessarily be the same
with other executables, or in theory from run to run, it's just not
something you can rely on statically.