* set_sj and set_uj 
  2018-31-07
I tried... and failed, to add mpfi_set_sj and mpfi_set_uj:
setting an interval from a huge interval. Pb with compilation: apparently I did not get how to test whether 
the corresponding types (intmax_t and uintmax_t) are known, or not, by the compiler.
Everything is ready: the code for these functions, some tests (actually, the same as for set_si and set_ui,
thus more cases, and especially cornercases, should be added) and the lines in the Makefile.am. Makefile.in
are already written. But it does not compile... so far.

* Fix predicates
From: "Sisyphus" <sisyphus1@optusnet.com.au>
To: "philippe theveny" <philippe.theveny@inrialpes.fr>,
        <Nathalie.Revol@ens-lyon.fr>
Subject: mpfi-1.4: Some surprises wrt nans
Date: Mon, 1 Feb 2010 21:26:01 +1100

Hi,

Here's the demo prog:

###############################
#include <stdio.h>
#include <mpfi.h>
#include <mpfi_io.h>

int main(void) {
mpfi_t x;

mpfi_init(x);
mpfi_out_str(stdout, 10, 0, x);

printf("\n%d %d %d %d %d %d\n",
        mpfi_is_pos(x),
        mpfi_is_strictly_pos(x),
        mpfi_is_nonneg(x),
        mpfi_is_neg(x),
        mpfi_is_strictly_neg(x),
        mpfi_is_nonpos(x));

return 0;
}

###############################

It shows that mpfi_is_pos(), mpfi_is_strictly_pos(), mpfi_is_nonneg(), 
mpfi_is_neg(), mpfi_is_strictly_neg() and mpfi_is_nonpos() all return true 
for nans.

I expected that they all would return false .... though there might be some 
room for debate on that as regards nonpos and nonneg.

(Btw, there's no need for any patches - I can wait for the next release. 
Main purpose of this post is just to make sure you're aware of this 
behaviour.)

Cheers,
Rob 
------
