Message-ID: <737978.33107.qm@web25501.mail.ukl.yahoo.com>
Date: Tue, 15 Jun 2010 06:57:45 +0000 (GMT)
From: Marco Atzeri <marco_atzeri AT yahoo DOT it>
Subject: Re: bug in NINT() in gfortran
To: cygwin AT cygwin DOT com
In-Reply-To: <4C139E54.9000807@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Precedence: bulk
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com

--- Sab 12/6/10, Dave Korn  ha scritto:

> On 12/06/2010 15:13, Marco Atzeri
> wrote:
> > --- Sab 12/6/10, Dave Korn=A0 ha scritto:
>=20
> >> yesterday, and my investigations so far suggest
> that it's actually a
> >> problem in the underlying maths functions in
> newlib.=A0=20
>=20
> > it seems a fault coming from newlib lround
>=20
> =A0 Yes, that's what I did in my "investigations so
> far".=A0 I've now tracked it
> down to undefined behaviour (a right-shift by 32 bits) in
> newlib; a patch will
> follow shortly.
>=20
> =A0 =A0 cheers,
> =A0 =A0 =A0 DaveK
>=20
>=20

I see.=20

Testing similar functions I found that we have=20=20
"undefined reference to `_llround'"
"undefined reference to `_llroundf'"

It seems that we are incorrectly looking in
newlib/libm/machine/spu

while the functionalities exist in
newlib/libm/common/s_llround.c
newlib/libm/common/sf_llround.c

llrint and llrintf have not such problem.

Marco
=20


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple