I have experienced this too, about a week ago.

Is there any way - in a generic interface - to recursively call the generic interface w/o it calling the same method implementation? For example:

#:methods gen:foobar [(define (foo x) (foo (x-child x)))] This fails because even though foo-child returns the right thing, which is a gen:foobar, it’s not calling the foo implementation for the right type as opposed to the same foo it was just in?

Figured out a work-around

Any reason why command line racket and DrRacket (both built from source) would have different behavior with respect to loading libraries?

I seem to have an issue with the order of directories searched for a library where DrRacket works fine, but command line racket fails.

I would guess, based on previous experience with a similar issue with IntelliJ, that CLI and DrRacket end up running with two different sets of environment variables.

How would that happen?

Environment configuration that happens in rcfiles doesn’t necessarily apply to apps that are launched from a graphical file explorer. I think that behavior is OS-dependent.

Yes, that’s definitely it. PATH and DYLD_LIBRARY_PATH are totally different for the two.

This is what define/generic is for https://docs.racket-lang.org/reference/struct-generics.html#%28form._%28%28lib._racket%2Fgeneric..rkt%29._define%2Fgeneric%29%29