[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Issue: HASH-TABLE-PACKAGE-GENERATORS (version 3)
- To: Moon@STONY-BROOK.SCRC.Symbolics.COM
- Subject: Issue: HASH-TABLE-PACKAGE-GENERATORS (version 3)
- From: Jon L White <jonl@lucid.com>
- Date: Mon, 17 Oct 88 16:45:13 PDT
- Cc: cl-cleanup@sail.stanford.edu
- In-reply-to: David A. Moon's message of Tue, 11 Oct 88 13:05 EDT <19881011170501.9.MOON@EUPHRATES.SCRC.Symbolics.COM>
re: I still think the with-package-iterator syntax isn't right. The way
you have it now, there is a required argument <package> which by
default is ignored!
The only time the <package> argument is ignored is when the other arguments
imply DO-ALL-SYMBOLS. MLY's original suggestion (to me, privately) was to
have the <package> argument be optional; that would work too, and make for a
bit nicer syntax in the one case of DO-ALL-SYMBOLS. Richard [MLY] actually
suggested dropping the DO-ALL-SYMBOLS case, but I don't think we can do that.
However, there is a constraint that isn't easy to express in the existing
CL language: namely, that the <package> argument is required if any of the
other keyword arguments are supplied as non-nil. One way out of this bind
is to let the &optional <package> argument default to the symbol *PACKAGE*;
this way, there would be no problem of syntax consistency, but only that of
remembering that, in the DO-ALL-SYMBOLS case, passing the optional argument
doesn't do anything.
-- JonL --