test-finalizers may break in interpreted mode when importing posix

[sjamaan]

I've seen the final assertion in tests/test-finalizers.scm fail under CHICKEN 5 when you add (import (chicken posix)) to the top.

It fails because n has value 0 at the end instead of the expected 1, which means the bump procedure (which is registered as a finalizer on the string constant) wasn't run.

I haven't been able to reproduce this consistently, so this may also affect CHICKEN 4.

[felix winkelmann]

I can not reproduce this. I tried changing the nursery size, but nothing happens.

If someone can, try replacing the final constant string with something dynamically generated (say, (make-string 100)). Does the error still occur?

[sjamaan]

Let's ignore this for now. It might've been accidentally fixed or not, but if we can't reproduce it it makes no sense to keep this ticket open. We may re-open it later. I've added a comment to the test so we can find the ticket again.