Re: Fault in guile 2.9

[Stefan Israelsson Tampe]

Stack trace:

Backtrace:
In ice-9/boot-9.scm:
   838:4 19 (with-throw-handler _ _ _)
In system/base/compile.scm:
   59:11 18 (_)
  155:11 17 (_ #<closed: file 565113183230>)
  224:14 16 (read-and-compile _ #:from _ #:to _ #:env _ #:opts _)
   255:6 15 (compile _ #:from _ #:to _ #:env _ #:opts _)
  183:32 14 (compile-fold _ _ _ (#:to-file? #t #:warnings (# # # …)))
In language/cps/compile-bytecode.scm:
  683:12 13 (compile-bytecode _ #<directory (language prolog modul…> …)
  675:12 12 (lower-cps _ (#:to-file? #t #:warnings (format # # #)))
In language/cps/optimize.scm:
    84:0 11 (optimize-higher-order-cps _ _)
In language/cps/dce.scm:
  402:33 10 (eliminate-dead-code _)
  110:18  9 (compute-live-code #<intmap 0-462173>)
In language/cps/intmap.scm:
   519:5  8 (visit-branch #(#(#(#(#<cps (kfun () 0 329…> …) …) …) …) …)
   519:5  7 (visit-branch #(#(#(#<cps (kargs (arg) (23580…> …) …) …) …)
   519:5  6 (visit-branch #(#(#<cps (kargs (arg) (23580) (co…> …) …) …)
   519:5  5 (visit-branch #(#<cps (kargs (arg) (23825) (continu…> …) …)
In language/cps/type-checks.scm:
   41:15  4 (elide-type-checks #<intmap 0-462173> 33263 #<intmap 0-…>)
In language/cps/types.scm:
  1758:5  3 (worklist-fold* #<procedure visit-cont (label typev)> _ _)
 1899:26  2 (visit-cont 33278 #<intmap 33263+0-15>)
   381:4  1 (constant-type _)
In ice-9/boot-9.scm:
  749:25  0 (dispatch-exception 0 misc-error (#f "unhandled cons…" …))

ice-9/boot-9.scm:749:25: In procedure dispatch-exception:
unhandled constant #<procedure src (lam u)>

On Sat, Nov 10, 2018 at 12:25 AM Stefan Israelsson Tampe <address@hidden> wrote:

Hi,

I got this error,

what's this error message? unhandled constant #<procedure src (lam u)>

So a <const> with got a src meta informatoin in it's field somehow, a bug in guile's compiler 

for sure. This is for a real stress test, a compile of a huge file that used to take 10minutes to compile. So I ill not be able to track this down and make a bugreport But now you know, there 

exist a bug!

Regards

Stefan