Interesting. According to that page, Kawa also has a per-thread working
directory. Per, do you feel strongly like Marc that a per-thread rather
than per-process CWD is the right way to go?
It's important to define exactly what CWD means. I can think of several different definitions:
- what (current-working-directory) and the like return
- the directory in which "foo" will be sought in (open-input-file "foo")
- the initial OS-level working directory of a newly created subprocess, regardless of the language the process is running
All of these can be controlled the same way, but that's not strictly necessary.
My inclination is that all three should be controlled by a thread-local parameter.