CLtL2 and the dpANS allow (but do not require) some standard CL types,
such as streams, readtables, and random-states, to be implemented via
DEFSTRUCT. Since EQUALP descends structures, does this mean that
EQUALP's behavior on such datatypes is implementation-dependent? In
other words, can the following return either true or false?
(equalp (make-broadcast-stream) (make-broadcast-stream))
(equalp (copy-readtable) (copy-readtable))
(equalp (make-random-state) (make-random-state))
--
Lawrence G. Mayka
AT&T Bell Laboratories
···@iexist.att.com
Standard disclaimer.
In article <·················@excalibur.ATT.COM> you write:
>CLtL2 and the dpANS allow (but do not require) some standard CL types,
>such as streams, readtables, and random-states, to be implemented via
>DEFSTRUCT. Since EQUALP descends structures, does this mean that
>EQUALP's behavior on such datatypes is implementation-dependent?
I don't think it ever came up at a meeting, but that seems like a logical
consequence of what we wrote. It's annoying, but it hardly seems like a
critical flaw. It might catch some users by surprise, though.
In the opinion of several X3J13 members, EQUALP is even more of an
abomination than EQUAL is. There were some serious suggestions to delete
them. But there's too much code that uses them, so we didn't dare.
--
Barry Margolin
System Manager, Thinking Machines Corp.
······@think.com {uunet,harvard}!think!barmar