Copyright 1999, 2001, 2002, 2003 Free Software Foundation.
Contributed by the Spaces project, INRIA Lorraine.

This file is part of the MPFR Library.

The MPFR Library is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2.1 of the License, or (at your
option) any later version.

The MPFR Library is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
License for more details.

You should have received a copy of the GNU Lesser General Public License
along with the MPFR Library; see the file COPYING.LIB.  If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA.

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

Probably many bugs.

Known bugs:

* The overflows/underflows are not yet implemented in all functions.

* problem with tsqrt on ia64-hpux (problem is that variable a in check4 has
	a different value than in the call, may be a memory problem, since
	it disappears when we remove the mpfr_set_str instruction, and
	disappears without -O2)
mpfr_sqrt failed for a=4.15383748682786210282e+34, rnd_mode=GMP_RNDZ
expected 0.10010010111100001100100110000001011110111111001010000[00000000000]E53
got      0.10110101000001001111001100110011111110011101111001100[00000000000]E58

* texp fails on ia64-hpux: mpfr_exp_2 and mpfr_exp3 for prec=601

* acosh is wrong near EMAX (idem asinh):
   rounding mode GMP_RNDU:
      6096987078287.281250 ulp(s) for x=1.75062535692767808555e+308
      [mpfr: 7.097561817130842e2 libm: 7.10449328893644064919e+02]

Potential bugs:

* Possible integer overflows on some machines. Some types are used
  inconsistently.

* Possible infinite loop in some functions for particular cases: when
  the exact result is an exactly representable number or the middle of
  consecutive two such numbers.

* mpfr_set_d may give wrong results on some architectures.

* Error analysis in mpfr_sin is incorrect. A test has been added to
  show the bug. Check the other functions too...

* mpfr_hypot may fail for x very large, y very small and a very large
  target precision.