* let-values, and-let* no variables
@ 2004-08-20 1:27 Kevin Ryde
2004-08-20 10:28 ` Marius Vollmer
0 siblings, 1 reply; 4+ messages in thread
From: Kevin Ryde @ 2004-08-20 1:27 UTC (permalink / raw)
I'm contemplating the way "begin" is used for the body of let-values
and and-let* with no bindings. It has the unhappy effect of allowing
what is visually an internal define to go out into the containing
environment, eg.
(use-modules (srfi srfi-11))
(let-values ()
(define x 123)
456)
(display x)
;; prints 123
The reference implementation of srfi-11 seems pretty clear that
"begin" is intended for let-values, and the formal semantics part of
srfi-2 suggests the same for and-let*. Dunno if that's only an
oversight regarding begin versus internal defines though.
I guess nobody would write an empty and-let* or let-values
deliberately, but it could arise out of a macro. I'm thinking to put
some words of caution in the manual, to make it clear the body isn't
the same as a plain "let" or "let*".
_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: let-values, and-let* no variables
2004-08-20 1:27 let-values, and-let* no variables Kevin Ryde
@ 2004-08-20 10:28 ` Marius Vollmer
2004-09-07 0:12 ` Kevin Ryde
0 siblings, 1 reply; 4+ messages in thread
From: Marius Vollmer @ 2004-08-20 10:28 UTC (permalink / raw)
Kevin Ryde <user42@zip.com.au> writes:
> I'm contemplating the way "begin" is used for the body of let-values
> and and-let* with no bindings. It has the unhappy effect of allowing
> what is visually an internal define to go out into the containing
> environment, eg.
Uhh, intuitively I'd say that this is wrong behavior. Maybe you
should ask on comp.lang.scheme.
> I guess nobody would write an empty and-let* or let-values
> deliberately, but it could arise out of a macro. I'm thinking to put
> some words of caution in the manual, to make it clear the body isn't
> the same as a plain "let" or "let*".
That would be good, even if we 'fix' our implementation to behave
differently.
_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: let-values, and-let* no variables
2004-08-20 10:28 ` Marius Vollmer
@ 2004-09-07 0:12 ` Kevin Ryde
2004-09-21 22:09 ` Marius Vollmer
0 siblings, 1 reply; 4+ messages in thread
From: Kevin Ryde @ 2004-09-07 0:12 UTC (permalink / raw)
Cc: guile-devel
[-- Attachment #1: Type: text/plain, Size: 593 bytes --]
Marius Vollmer <marius.vollmer@uni-dortmund.de> writes:
>
> Uhh, intuitively I'd say that this is wrong behavior.
One that's guile specific is let-optional and let-keywords, eg.
(use-modules (ice-9 optargs))
(let ()
(let-keywords '() #f ()
(define localvar 123))
localvar)
=> 123
which could be fixed fairly easily,
* optargs.scm (let-optional-template, let-keywords-template): Change
"(begin body)" to "(let () body)" for empty bindings, since the former
lets syntactically internal defines in body leak out to the
surrounding environment.
[-- Attachment #2: optargs.scm.begin.diff --]
[-- Type: text/plain, Size: 1013 bytes --]
--- optargs.scm.~1.20.~ 2003-04-07 08:04:58.000000000 +1000
+++ optargs.scm 2004-09-07 09:16:29.000000000 +1000
@@ -1,6 +1,6 @@
;;;; optargs.scm -- support for optional arguments
;;;;
-;;;; Copyright (C) 1997, 1998, 1999, 2001, 2002 Free Software Foundation, Inc.
+;;;; Copyright (C) 1997, 1998, 1999, 2001, 2002, 2004 Free Software Foundation, Inc.
;;;;
;;;; This library is free software; you can redistribute it and/or
;;;; modify it under the terms of the GNU Lesser General Public
@@ -125,7 +125,7 @@
(define (let-optional-template REST-ARG BINDINGS BODY let-type)
(if (null? BINDINGS)
- `(begin ,@BODY)
+ `(let () ,@BODY)
(let-o-k-template REST-ARG BINDINGS BODY let-type
(lambda (optional)
`(,(car optional)
@@ -140,7 +140,7 @@
(define (let-keywords-template REST-ARG ALLOW-OTHER-KEYS? BINDINGS BODY let-type)
(if (null? BINDINGS)
- `(begin ,@BODY)
+ `(let () ,@BODY)
(let* ((kb-list-gensym (gensym "kb:G"))
(bindfilter (lambda (key)
`(,(car key)
[-- Attachment #3: Type: text/plain, Size: 143 bytes --]
_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: let-values, and-let* no variables
2004-09-07 0:12 ` Kevin Ryde
@ 2004-09-21 22:09 ` Marius Vollmer
0 siblings, 0 replies; 4+ messages in thread
From: Marius Vollmer @ 2004-09-21 22:09 UTC (permalink / raw)
Kevin Ryde <user42@zip.com.au> writes:
> Marius Vollmer <marius.vollmer@uni-dortmund.de> writes:
>>
>> Uhh, intuitively I'd say that this is wrong behavior.
>
> One that's guile specific is let-optional and let-keywords, eg.
>
> (use-modules (ice-9 optargs))
> (let ()
> (let-keywords '() #f ()
> (define localvar 123))
> localvar)
>
> => 123
>
> which could be fixed fairly easily,
Yes, please do. Thanks!
--
GPG: D5D4E405 - 2F9B BCCC 8527 692A 04E3 331E FAF8 226A D5D4 E405
_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2004-09-21 22:09 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-08-20 1:27 let-values, and-let* no variables Kevin Ryde
2004-08-20 10:28 ` Marius Vollmer
2004-09-07 0:12 ` Kevin Ryde
2004-09-21 22:09 ` Marius Vollmer
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).