[Gcl-devel] Re: TMP_ALLOC in mpn_mul_n

[Kevin Ryde]

Camm Maguire <address@hidden> writes:
>
> OK, then gcl will either have to step around, *or* rely on its
> traversal of the C stack to find pointers and relocate them.

I think you might be better off trying to make something with normal C
malloc semantics.

Perhaps we can work towards friendlier interaction with a lisp style
gc, but for now the requirements of the mp_set_memory_functions really
would want to be observed.

Once outside any gmp function it should be ok to move stuff, knowing
the mpz_t _mp_d pointer will be the only thing needing to be changed.
Of course that isn't documented but it's got to be a lot safer than
fiddling about in the C stack.

> I see no evidence of mpz_{set,clr}bit changes -- can you elaborate
> further?

That might have been followed up privately.  I'll see if I've still
got the mail.  It was the same issue, blocks moving in a realloc.