On Fri, Mar 1, 2024 at 2:47 PM Lassi Kortela <xxxxxx@lassi.io> wrote:
>
> According to POSIX, time_t (the type for an integer number of seconds)
> can be either signed or unsigned depending on the implementation.
>
> That adds more ambiguity, but perhaps not to this SRFI. We can keep
> saying that the seconds field is signed, and if the OS won't return
> negative seconds, that's fine.

Yes, but that inconsistency hurts the brain.

Maybe make the representation of times before the epoch undefined?

What do you think?