From owner-freebsd-hackers  Fri Jun 23 12:21:24 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id MAA22848
          for hackers-outgoing; Fri, 23 Jun 1995 12:21:24 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id MAA22840
          for <hackers@freebsd.org>; Fri, 23 Jun 1995 12:21:00 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id FAA21290; Sat, 24 Jun 1995 05:18:33 +1000
Date: Sat, 24 Jun 1995 05:18:33 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199506231918.FAA21290@godzilla.zeta.org.au>
To: hackers@freebsd.org, luigi@labinfo.iet.unipi.it
Subject: Re: Any experience on tcl7.4/tk4.0 ?
Sender: hackers-owner@freebsd.org
Precedence: bulk

>I have tried to build "GroupKit" with tcl 7.4 and tk4.0b4, but with not
>much success. tcl 7.4 as distributed fails some tests with a FP
>exception. The documentation suggests a fix which has to do with the
>rounding mode and disable the fpmask. With this fix tcl completes the
>tests, but then tk fails several tests (here again there appear to be
>errors in the FP calculations: a lot of results are off by 1 WRT the
>expected values).

The FreeBSD port of tcl 7.3 has diffs for the fpmask and the rounding
mode.  These diffs apparently still apply (FreeBSD certainly hasn't
changed).  It also has diffs for configuring strtod().  I'm not sure
exactly what these do.  I don't know of any bugs in the FreeBSD
strtod().  Use of a toy portable strtod() could easily cause small FP
errors.

Bruce