From gwyn@arl.army.mil  Fri Feb 27 01:46:29 2004
Received: from soiree.arl.army.mil (soiree.arl.army.mil [128.63.39.181])
	by dkuug.dk (8.12.10/8.9.2) with SMTP id i1R0kP1S085622
	for <embedded-c@dkuug.dk>; Fri, 27 Feb 2004 01:46:27 +0100 (CET)
	(envelope-from gwyn@arl.army.mil)
Date:     Thu, 26 Feb 2004 19:37:56 EST
From: Doug Gwyn (CISD) <gwyn@arl.army.mil>
To: John Hauser <jhauser@jhauser.us>
cc: Embedded-C <embedded-c@dkuug.dk>
Subject:  Re:  (embedded-c.195) Repeat: US-40
Message-ID:  <200402261937.aa12609@soiree.arl.army.mil>
X-Spam-Score: 0 () 

The proposal does sound like it is along the right lines, particularly
in backing out the other changes to the standard that led to US-40 in
the first place.  I think I would change "minus" to "less" or "without",
though.

I'm not clear about why an object declared with an address-space qualifer
does not have the address-space qualifier in its effective type.  I can
see that you don't want to impress the address-space quality upon a
malloc()ed object, although I don't see why anybody would try to do that
anyway.  Is the real problem that you want to be able to copy between
address-space qualified objects and ordinary objects without needing a
cast?  If so, that might be handled more simply in the semantics for
address-space qualification.