Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 14 Aug 2016 16:03:51 -0700
From:      Steve Kargl <sgk@troutmask.apl.washington.edu>
To:        freebsd-ports@freebsd.org
Subject:   Problems with out libgcc_s.so in base
Message-ID:  <20160814230351.GA10587@troutmask.apl.washington.edu>

next in thread | raw e-mail | index | archive | help

> The reason ports gcc now has this requirment on 4.6 or better is
> fortran standard says we have to support quad floating point math.
> e.g. /usr/local/lib/gccXX/libquadmath.so 

Diane,

Can you please stop with the dis-information?  No Fortran standard
has ever required quad floating point math. Ever.  The Fortran standard
requires REAL and DOUBLE PRECISION with the requirement that the
DP type must occupy exact twice the storage units of REAL.  Now,
the Fortran standard permits a Fortran processor to support more
than REAL and DP.  On IA32 and X86_64 hardware, gfortran happens
to support REAL(4), REAL(8), REAL(10), and REAL(16) (note I'm
using the kind type parameters of gfortran here.  The values for
the kind type parameter are not specified by a Fortran standard).
On sparc64, gfortran supports REAL(4), REaL(8), and REAL(16).

FreeBSD-ports could avoid libquadmath issues by building gcc
without it.  See --without-libquadmath or --without-quadmath (I
don't remember the config option because it would be questionable
to neuter one of gfortran's strength).

Freebsd-ports could also use a wrapper:
% cat ~/bin/gfc7
#! /bin/sh
DIR=`id -P sgk | sed 's/\:/\ /g' | awk '{print $9}'`
export DIR

LD_LIBRARY_PATH=$DIR/work/7/lib
export LD_LIBRARY_PATH

LD_RUN_PATH=$DIR/work/7/lib
export LD_RUN_PATH

$DIR/work/7/bin/gfortran -fno-backtrace $@

>
> We *could* lie in our libgcc_s and claim to be 4.6 which would
> simply mean libgfortran would not complain and simply load the
> missing libquadmath.
>
> If we updated the claimed GCC version in our libgcc_s.so to claim
> GCC_4.6, we really also should provide a libquadmath subsitute.
>

No offense intended, but given the complete lack of work on libm
(see FreeBSD-numerics mailinglist) after I was forced to return
my commit bit, I doubt that there is anyone within the FreeBSD
community who could replace libquadmath.

-- 
Steve



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20160814230351.GA10587>