On Wed, 11 Nov 2026, 繠䴧 via "Bug reports for GUILE, GNU's Ubiquitous Extension 
Language" <[email protected]> wrote:
> on guile 3.0.11, something is broken when integer?, throw, and floor or 
> ceiling are used together.
> just see my experiment below, you can try the commands if you have guix 
> installed for a clean guile 3.0.11 environment.
> I really could not figure out why the combination breaks something inferior.
> In short, executing `(lambda (x) (unless (integer? x) (throw 'bad)) 
> (floor x))` with guile 3.0.11 would fail, which is not mentioned by the 
> documentation, so it is a bug instead of feature.

Looks like a bug in the compiler:

Backtrace:
In ice-9/eval.scm:
    619:8 19 (_ #(#(#<directory (guile-user) 7f05026d8320>)))
In /home/old/softwares/guile/build/./meta/guild:
    72:17 18 (main _)
In srfi/srfi-1.scm:
    750:9 17 (for-each #<procedure 7f0502635870 at scripts/compile.…> …)
In scripts/compile.scm:
   290:27 16 (_ _)
In system/base/target.scm:
     80:6 15 (with-target _ _)
In system/base/compile.scm:
    188:6 14 (compile-file "foo.scm" #:output-file _ #:from _ #:to _ …)
     53:4 13 (call-with-output-file/atomic _ _ _)
In ice-9/boot-9.scm:
  1764:12 12 (with-exception-handler _ _ #:unwind? _ # _)
In system/base/compile.scm:
    69:11 11 (_)
   191:11 10 (_ #<closed: file 7f050262e620>)
   354:28  9 (read-and-compile _ #:from _ #:to _ #:env _ # _ # _ # _)
   267:44  8 (_ _ _)
   263:33  7 (_ #<intmap 0-19> #<module (#{ g459}#) 7f04f9b596e0>)
In language/cps/optimize.scm:
   134:12  6 (_ _ #<module (#{ g459}#) 7f04f9b596e0>)
    101:3  5 (optimize-higher-order-cps _ _)
In language/cps/type-fold.scm:
   1077:7  4 (type-fold _)
  1070:27  3 (fold-functions-in-renumbered-program #<procedure loca…> …)
   899:10  2 (local-type-fold _ 14 _)
     48:4  1 (materialize-constant _ _ _ _ _)
In ice-9/boot-9.scm:
  1685:22  0 (raise-exception _ #:continuable? _)

-- 
Olivier Dion



  • bug#82219: floo... Bug reports for GUILE, GNU's Ubiquitous Extension Language
    • bug#82219:... Olivier Dion

Reply via email to