Re: CVS timestamps

[Mark Ferrell]

Matthew Berney writes:
> 
> I find it hard to believe that a tool that works across multiple timezones
> does not support a way to localize the timestamps?  Why not use the TZ
> environment variable?  This is pretty standard isn't it?  It would be nice
> to have the timestamps in the repository actually match our local time.

I would think that maintaining timestamps in the repository, and transmitting
that information via the protocol in UTC would be the simplest way to handle
it.  Allow the client to utilize the functionality of time.h, and or TZ if it
exists, to figure out the TZ for the client and adjust the visable timestamp
from there.  Apon a commit, simply reverse the process and hand the client the
UTC of the file based agains the same timezone information.

time.h on most systems should conform to one, if not all of the standards
found ... BSD4.3, ISO9899, SVID3, and POSIX

That way whenever you have a file it's timestamp is relative to your own clock,
but the repository itself is uneffected as everything is adjusted to UTC on the
inbound.

Admittedly, you will have a file checked out into germany that claims a
different timestamp then one checked out into california.  As the visable times
are relative to the files current location.  But it would hardly be relivant
when considering that they would all share the same UTC anyway.

-- 
Mark Ferrell <address@hidden>
Monta Vista Software Inc. <http://www.mvista.com>

pgpJe1H2F7dDL.pgp
Description: PGP signature