Floating point exception on 0.15.13, alpha NetBSD 1.6.1

Matthias Andree matthias.andree at gmx.de
Sun Jan 11 23:45:38 CET 2004


On Sun, 11 Jan 2004, Paul Ackersviller wrote:

> On Sun, Jan 11, 2004 at 10:00:30AM -0500, Clint Adams wrote:
> > Try building with gcc -mieee.
> 
> Thanks, that did the trick.
> 
> Is there somewhere this should be documented, perhaps added to the FAQ,
> or put in a README file?

I'd like to make it a standard build option though and I wonder if it is
safe to feed the compiler the first of the "-mieee",
"-ieee_with_no_inexact" options that it accepts.

I've changed CVS configure.ac to just try these options and ignore them
when the compiler rejects them. The change should propagate to the next
release, unless trouble will be reported.

If you want to try the stable version (which is basically 0.15.13.1 with
just this change, make sure you have current automake >= 1.6 and
autoconf >= 2.53, grab the bogofilter-0_15_13 branch from CVS, like
this:

cvs -d:pserver:anonymous at cvs.sourceforge.net:/cvsroot/bogofilter login

cvs -z3 -d:pserver:anonymous at cvs.sourceforge.net:/cvsroot/bogofilter \
  co -r bogofilter-0_15_13 bogofilter

To build, just "cd bogofilter ; autoreconf -i" then configure and make
as usual (you only need this autoreconf after initial checkout "co" or
when the build fails for strange reasons and re-running configure
doesn't fix it).

To pull new versions, just "cd bogofilter" and "cvs update".

Note that anonymous CVS access and ViewCVS may lag behind a bit,
bribe SourceForge to fix that...
<URL:http://cvs.sourceforge.net/viewcvs.py/bogofilter/bogofilter/configure.ac?only_with_tag=bogofilter-0_15_13>

HTH,
Matthias




More information about the Bogofilter mailing list