`time' macro's "running stats for" can be misleading

[Derick Eddington]

`time' makes it look like (format A ~s B 123) was the timed expression, because `print-stats' uses "~a" for its format string and `time-macro' quotes the s-expression syntax. JTMI, if "~s" was used instead, the "running stats for" message for (time-it "Who Message" whatever) would look different -- maybe that's a reason to think of a solution other than just using "~s".

Ikarus Scheme version 0.0.3+ (revision 1372, build 2008-02-08)
Copyright (c) 2006-2008 Abdulaziz Ghuloum

> (time (format "A ~s B" 123))
running stats for (format A ~s B 123):
    no collections
    0 ms elapsed cpu time, including 0 ms collecting
    0 ms elapsed real time, including 0 ms collecting
    1504 bytes allocated
"A 123 B"
>

[Abdulaziz Ghuloum (aghuloum)]

Fixed in 1429. Thanks. You now get:

> (time (format "A ~s B" 123))
running stats for (format "A ~s B" 123):
    no collections
    0 ms elapsed cpu time, including 0 ms collecting
    0 ms elapsed real time, including 0 ms collecting
    1504 bytes allocated
"A 123 B"

[Abdulaziz Ghuloum (aghuloum)]

BTW, the solution was not to change the time-it procedure, but to change the time macro to construct the correct string:

--- scheme/psyntax.expander.ss 2008-03-16 01:00:16 +0000
+++ scheme/psyntax.expander.ss 2008-03-25 01:49:42 +0000
@@ -1301,7 +1301,11 @@
     (lambda (stx)
       (syntax-match stx ()
         ((_ expr)
- (bless `(time-it ',expr (lambda () ,expr)))))))
+ (let ([str
+ (let-values ([(p e) (open-string-output-port)])
+ (write (syntax->datum expr) p)
+ (e))])
+ (bless `(time-it ,str (lambda () ,expr))))))))