[AFS3-std] Re: A call for consensus on draft-deason-afs3-type-time-02

[Russ Allbery]

Andrew Deason <adeason@sinenomine.net> writes:
> Russ Allbery <rra@stanford.edu> wrote:

>> Jeffrey has a good point, though: we lose representability of dates
>> that can currently be handled with CIFS.

> Then we just make the absolute timestamps signed. It just seems better
> to me to start from an epoch that's a bit more well-defined (or at
> least, more easily well-defined; we can always define 1 Jan 1600 as X
> seconds before 1 Jan 1970, but that seems strangely indirect).

I guess I'm ambivalent between signed timestamps and the Windows epoch.
Both of them are going to be at least somewhat unusual for typical UNIX
code.  My guess is that it will be easier to explain the Windows epoch
than it will be to explain signed arithmetic on times, but I could be
wrong.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>