[gambit-list] [ANN] Gambit-C v4.4.3 released

Valeriya Pudova val1969 at neotech.lv
Fri Apr 24 06:49:27 EDT 2009


Hello good people,

I found that sometime gsi (or gsc) does not report the error location
refering information. It make hard job to find location of error.

After few tests I found simple code which display this issue.

> (eval '(define (foo) (display (/ 1 0)))) (foo)
*** ERROR IN foo -- Divide by zero,
(/ 1 0)

Not (eval) makes the problem, because this  variant works

> (eval '(define (foo)  (/ 1 0))) (foo)
*** ERROR IN "test2.scm"@1.33 -- Divide by zero
(/ 1 0)

And not (display):

> (define (foo)  (display (/ 1 0))) (foo)
*** ERROR IN foo, "test2.scm"@1.25 -- Divide by zero
(/ 1 0)

Could anybody explain this?

N.B. I know that with scheme sometime hard to see where code was
located before eval. The code may be generated by macros.



More information about the Gambit-list mailing list