On Tuesday, August 30 at 01:57 PM, quoth Ingo van Lil:
>On 30 Aug 2005, Kyle Wheeler wrote:
>
>> Anyway - have at ye! Let me know about any problems you run into.
>
>Yep, haven't yet managed to compile it. What version of mpfr did you
>use? I have the most recent gmp version (4.1.4) installed and wcalc
>complains about several undeclared functions (e.g. mpfr_get_si,
>mpfr_equal_p and mpfr_mul_si). Plus, mpfr_urandomb returns void, so
>not ignoring the function result (as in scanner.l, lines 158 and 163)
>isn't allowed.

I ran into that too when trying to compile it on my OpenBSD machine
(only gmp is in ports, not mpfr, sadly). But, if you install mpfr 2.1.1,
from http://www.mpfr.org, it should work correctly.

>Oh, and help.c has to include limits.h, because MPFR_PREC_MAX maps to
>(ULONG_MAX>>1). No idea why mpfr.h doesn't include it itself.

Hmmm. This *should* be fixed by installing the latest MPFR, I think. In
the latest MPFR headers, MPFR_PREC_MAX maps to:

((mpfr_prec_t)((!(mpfr_prec_t)0)>>1))

Which is better and more cross-platform anyway -- not all Unixes
conveniently define ULONG_MAX for you (the same goes for UINT32_MAX,
which you'll notice I had to explicitly detect). Why "stdint.h" isn't
*standard*, I have no idea.

~Kyle
--
Once again the conservative sandwich-heavy portfolio pays off for the
hungry investor!
-- Zoidberg