From owner-freebsd-hackers  Thu Dec 29 09:47:42 1994
Return-Path: hackers-owner
Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id JAA28529 for hackers-outgoing; Thu, 29 Dec 1994 09:47:42 -0800
Received: from seagull.rtd.com (root@Seagull.rtd.com [198.102.68.2]) by freefall.cdrom.com (8.6.9/8.6.6) with ESMTP id JAA28523 for <freebsd-hackers@freefall.cdrom.com>; Thu, 29 Dec 1994 09:47:41 -0800
Received: (from dgy@localhost) by seagull.rtd.com (8.6.9/8.6.9.1) id KAA22290; Thu, 29 Dec 1994 10:47:14 -0700
From: Don Yuniskis <dgy@seagull.rtd.com>
Message-Id: <199412291747.KAA22290@seagull.rtd.com>
Subject: Re: 1.1.5.1R libm
To: bde@zeta.org.au (Bruce Evans)
Date: Thu, 29 Dec 1994 10:47:12 -0700 (MST)
Cc: freebsd-hackers@freefall.cdrom.com (FreeBSD hackers)
In-Reply-To: <199412271620.DAA13423@godzilla.zeta.org.au> from "Bruce Evans" at Dec 28, 94 03:20:37 am
X-Mailer: ELM [version 2.4 PL23]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 740       
Sender: hackers-owner@freebsd.org
Precedence: bulk

> >while porting a few packages.  Specifically, it appears that gamma()
> >is discontinuous.
> 
> Well, it is discontinuous for some x <= 0 (negative integers?).  Are
> you aware that gamma() really is gamma() in libm, but it used to be
> the log of the real gamma().  It is the log of the real gamma() in

This appears to have been the problem. The (old) port of gnuplot
had been configured to use gamma() instead of lgamma() for the
GAMMA function.  I guess it could be a bit confusing... ;-)

Anyway, if built as originally ported, the "prob" demo would make
these problems very visible (discontinuities in PDF's which depend
on GAMMA(), CDF's that don't approach 1.0, missing graphs, etc.)

Someone might wish to check these on 2.0R...