Re: [Geiser-users] Re-evaluation fails when embedding Geiser into Racket program

[Mikhail Maluyk]

Hello,

This is my theory: in your example for hello.rkt, you had the form
(require geiser/server-start "world.rkt"): that's probably loading
world.rkt before geiser, and once you compile a module without setting
compile-enforce-module-constants to #t, it remains unmodifiable forever
(even if you change compile-enforce-module-constants to #f afterwards,
as geiser does).

You theory was just proved. Error occurred because compile-enforce-module-constants was not set correctly. Setting it at a top level fixed it. Another interesting thing is that setting this flag in module directly didn't fix the issue, but that has nothing to do with geiser (probably just matter of initialization order). 

Thanks for your help, and software!

--
Regards,
Mikhail