Re: Per-thread working directory and umask proposal Lassi Kortela 24 Apr 2020 08:43 UTC

> - Even all R7RS procedures so far behave consistently with per-thread
> working-directory, we still eventually need to explicitly get/set
> per-process working directory when we have subprocess srfi.

Not sure if I understand what you mean. If we want to give a different
working directory to the subprocess than the parent process, the `spawn`
procedure can take an explicit working-directory argument.

> - Not specifying working-directory is thread-specific or not, it can't
> be used portably for the code intended to be thread-safe.

True. SRFI 170 could require working-directory to be thread-specific in
all multi-threaded Scheme implementations.

> It's true
> that so many aspects of R7RS are vague about thread safety, but we don't
> need to add another ambiguity.

Excellent point.