From xemacs-m  Tue Feb 25 15:48:12 1997
Received: from cdc.noaa.gov (manager.Colorado.EDU [128.138.218.210])
	by xemacs.org (8.8.5/8.8.5) with SMTP id PAA08344
	for <xemacs-beta@xemacs.org>; Tue, 25 Feb 1997 15:48:11 -0600 (CST)
Received: from suomi by cdc.noaa.gov (SMI-8.6/SMI-SVR4)
	id OAA22854; Tue, 25 Feb 1997 14:48:05 -0700
Received: by suomi (SMI-8.6) id OAA08717; Tue, 25 Feb 1997 14:48:05 -0700
Sender: mdb@cdc.noaa.gov
To: xemacs-beta@xemacs.org
Subject: Re: message-make-message-id problems, anybody?
References: <199702251433.PAA18387@midnight.eng.ecf.teradyne.com> <rxsbu98fys3.fsf@midnight.ecf.teradyne.com> <m2pvxo7evy.fsf@altair.xemacs.org>
Organization: CIRES, University of Colorado
X-Attribution: mb
Mime-Version: 1.0 (generated by tm-edit 7.105)
Content-Type: text/plain; charset=US-ASCII
From: Mark Borges <mdb@cdc.noaa.gov>
Date: 25 Feb 1997 14:48:03 -0700
In-Reply-To: Steven L Baur's message of 25 Feb 1997 13:30:09 -0800
Message-ID: <vkzpwszhf0.fsf@cdc.noaa.gov>
Lines: 42
X-Mailer: Gnus v5.4.15/XEmacs 20.1

>> On 25 Feb 1997 13:30:09 -0800,
>> Steven L Baur(sb) wrote:
sb> Adrian Aichner writes:
>>>>>>> "Adrian" == Adrian Aichner <aichner@ecf.teradyne.com> writes:
>> Hello All,

>> That's what it boils down to:

>> What is (% 8 36) ?

>> Well that depends:

>> xemacs-20.1-b1 compiled '-g -xO5' says 8

sb> This is correct.

>> xemacs-20.1-b1 compiled '-g -fast' says 36

sb> This is not correct.

sb> It sounds like you have found a compiler optimization bug.  The "%"
sb> function is in src/data.c:

I wouldn't classify it more as a pitfall than as a bug, if it is what
I think.

If this is a SUNWspro compiler (which is consistent with the -xO5
flag; sorry I missed the earlier part of the thread), then `-fast'
must be used with caution. It is a convenience switch, which turns on
other options, one of which is `-dalign'. If the other libraries
you're linking against were not compiled with this (-dalign) flag, the
results are unpredictable.

I ran into this when compiling some code that links in the Numerical
Algorithm Group (NAg) mathematical libraries. The NAg libs were
compiled with `-dalign' but the default compile flag does not include
it, and I was seeing random bus errors.

Martin?

-- 
  -mb-

