In the real world, I don't think you can depend on timestamps to be
monotonically increasing. People do change their system clocks
sometimes, and it's not always the safe way that keeps time from going
backwards. The well-orderderd-ness of the revision history should NOT
depend on timestamps.
-David Mankin
On 3 May 2002, Ben Collins-Sussman wrote:
> Well, that's kind of the point. Once we have a real SQL table or
> somesuch in our filesystem back end, then the revision "names" don't
> become so important. All revisions have datestamp properties attached
> to them. As long as we are always able to sort our revisions table by
> date, we'll have a well-ordered time machine. The revision names
> become irrelevant.
>
> Alan Langford <jal@ambitonline.com> writes:
>
> > Pick a random integer when the repository is created and use that as
> > an exclusive or mask on all the revision numbers!
> >
> > Then of course you need an "is a later revision" query that takes two
> > encoded revnums and returns a boolean.
> >
> > At 2002/05/03 11:51 -0500, Ben Collins-Sussman wrote:
> > >But beyond that: our code uses a custom 'svn_revnum_t' type, which we
> > >can define to be whatever we want in the future. We're thinking,
> > >actually, of switching all our revision "numbers" to arbitrary hex
> > >strings soon. Just to stop people from thinking that it means
> > >anything important about their project. :-)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 3 20:24:55 2002