* bug#14537: defect in guile with syntax transformation
@ 2013-06-02 8:14 Dr. M. Luedde
2013-06-02 21:48 ` Ludovic Courtès
0 siblings, 1 reply; 4+ messages in thread
From: Dr. M. Luedde @ 2013-06-02 8:14 UTC (permalink / raw)
To: 14537; +Cc: mirko.luedde
;; I think the below code exhibits a defect with guile 1.8.7.
;;
;; The `defrec' syntax transformer (defined below) does not behave
;; correctly within an imbedding `let' expression.
;;
;; The system is CYGWIN_NT-6.1-WOW64 WDFN00305859A 1.7.18(0.263/5/3)
;; 2013-04-19 10:39 i686 Cygwin
;;
;; Guile runs in GNU bash, version 4.1.10(4)-release (i686-pc-cygwin)
;; enable `define-syntax' functionality
(use-modules (ice-9 syncase))
;; OK
(let ( (n 99) )
(begin
(define
ev.OK? (lambda (x) (if (zero? x) #t
(od.OK? (- x 1)))))
(define
od.OK? (lambda (x) (if (zero? x) #f
(ev.OK? (- x 1))))))
(od.OK? 99))
;; OK, evaluates to #t
(define-syntax defrec
;; TYPE: syntax transformer.
;;
;; REQUIRES: (defrec . X) requires that X := ((S_1 V_1) ... (S_N
;; V_N)). The S_I are mutually distinct symbols. The V_J may contain
;; references to the S_I. On evaluation of any V_J the value bound
;; to a referenced S_I must not be needed.
;;
;; RETURNS: The value of the expression (begin (define S_1 #f)
;; ... (define S_N #f) (set! S_1 V_1) ... (set! S_N V_N)).
(syntax-rules ()
((_ (S V) ...)
(begin (define S #f) ...
(set! S V) ...))))
;; OK
(defrec
(ev.broken? (lambda (x) (if (zero? x) #t
(od.broken? (- x 1)))))
(od.broken? (lambda (x) (if (zero? x) #f
(ev.broken? (- x 1))))))
(od.broken? 99)
;; OK, evaluates to #t
(let ( (n 99) )
(defrec
(ev.broken? (lambda (x) (if (zero? x) #t
(od.broken? (- x 1)))))
(od.broken? (lambda (x) (if (zero? x) #f
(ev.broken? (- x 1))))))
(od.broken? n))
;; Produces an error message:
;; In standard input:
;; 14: 0* (let* ((n 99)) (defrec (ev.broken? #) (od.broken? #)) (od.broken? n))
;; standard input:14:1: In procedure memoization in expression (let* (#) (defrec # #) ...):
;; standard input:14:1: In file "standard input", line 14: Mixed definitions and expressions in (defrec (ev.broken? (lambda (x) (if (zero? x) #t (od.broken? (- x 1))))) (od.broken? (lambda (x) (if (zero? x) #f (ev.broken? (- x 1)))))).
;; Backtrace:
;; ABORT: (syntax-error)
--
Dr. M. Luedde, Rheinblick 26A, D-69226 Nussloch, Germany
+49-175-2779708, Skype: MirkoLuedde
^ permalink raw reply [flat|nested] 4+ messages in thread
* bug#14537: defect in guile with syntax transformation
2013-06-02 8:14 bug#14537: defect in guile with syntax transformation Dr. M. Luedde
@ 2013-06-02 21:48 ` Ludovic Courtès
2013-06-03 22:27 ` Dr. M. Luedde
0 siblings, 1 reply; 4+ messages in thread
From: Ludovic Courtès @ 2013-06-02 21:48 UTC (permalink / raw)
To: Dr. M. Luedde; +Cc: 14537, mirko.luedde
Hello,
Thanks for your report.
"Dr. M. Luedde" <mirko_luedde@yahoo.de> skribis:
> ;; I think the below code exhibits a defect with guile 1.8.7.
Would it be an option for you to try this code in Guile 2.0.x?
Guile 1.8.7 is very old and its support for hygienic macros (among other
things) was leaving a lot to be desired.
Ludo’.
^ permalink raw reply [flat|nested] 4+ messages in thread
* bug#14537: defect in guile with syntax transformation
2013-06-02 21:48 ` Ludovic Courtès
@ 2013-06-03 22:27 ` Dr. M. Luedde
2013-06-03 22:38 ` Ludovic Courtès
0 siblings, 1 reply; 4+ messages in thread
From: Dr. M. Luedde @ 2013-06-03 22:27 UTC (permalink / raw)
To: Ludovic Courtès; +Cc: 14537@debbugs.gnu.org, mirko.luedde@sap.com
Hello Ludo,
alright, I compiled guile 2.0.6 on my cygwin system.
Given the amount and content of the warnings, it came as a surprise to me that the compilation apparently was successul and the resulting guile did process my testcase without the previously exhibited errors.
I consider this defect report closed.
Thanks and best, Mirko
P.S.
This "...I compiled guile 2.0.6 on my cygwin system..." and this "... Guile 1.8.7 is very old" might give the cygwin maintainers a hint.
--
Dr. M. Luedde, Rheinblick 26A, D-69226 Nussloch, Germany
+49-175-2779708, Skype: MirkoLuedde
----- Ursprüngliche Message -----
> Von: Ludovic Courtès <ludo@gnu.org>
> An: Dr. M. Luedde <mirko_luedde@yahoo.de>
> CC: 14537@debbugs.gnu.org; mirko.luedde@sap.com
> Gesendet: 23:48 Sonntag, 2.Juni 2013
> Betreff: Re: bug#14537: defect in guile with syntax transformation
>
> Hello,
>
> Thanks for your report.
>
> "Dr. M. Luedde" <mirko_luedde@yahoo.de> skribis:
>
>> ;; I think the below code exhibits a defect with guile 1.8.7.
>
> Would it be an option for you to try this code in Guile 2.0.x?
>
> Guile 1.8.7 is very old and its support for hygienic macros (among other
> things) was leaving a lot to be desired.
>
> Ludo’.
>
^ permalink raw reply [flat|nested] 4+ messages in thread
* bug#14537: defect in guile with syntax transformation
2013-06-03 22:27 ` Dr. M. Luedde
@ 2013-06-03 22:38 ` Ludovic Courtès
0 siblings, 0 replies; 4+ messages in thread
From: Ludovic Courtès @ 2013-06-03 22:38 UTC (permalink / raw)
To: Dr. M. Luedde; +Cc: 14537@debbugs.gnu.org, mirko.luedde@sap.com
Hello,
"Dr. M. Luedde" <mirko_luedde@yahoo.de> skribis:
> alright, I compiled guile 2.0.6 on my cygwin system.
>
> Given the amount and content of the warnings, it came as a surprise to me that the compilation apparently was successul and the resulting guile did process my testcase without the previously exhibited errors.
>
> I consider this defect report closed.
OK, thanks!
I know my answer may have sounded a bit frustrating, but it’s been more
than two years since 2.0 was released. It’s really a huge step forward
compared to Guile 1.8, and we hope people can appreciate it and find it
a good incentive to upgrade.
Ludo’.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2013-06-03 22:38 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-06-02 8:14 bug#14537: defect in guile with syntax transformation Dr. M. Luedde
2013-06-02 21:48 ` Ludovic Courtès
2013-06-03 22:27 ` Dr. M. Luedde
2013-06-03 22:38 ` Ludovic Courtès
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).