RE: Deficiency in Subversion - restore breaks working copies

On Mon, 18 Aug 2003, Wadsworth, Eric (Contractor) wrote:

> > Just wondering out loud: what if a wc A is upto revision 200, and the
> > repository is restored from backup to revision 100, THEN other
> > developers pump up the version number to >200. Does wc A go
> > completely wild now?
> This is one reason why I said, "And I don't think you can trust revision
> numbers to always be good indicators of state." There would need to be some
> additional information saved. Can anyone say "can 'o worms"? Can it even be
> done, I wonder?

(msg a little bit re-edited, please don't top-post)

I don't think Subversion wants to create a perfect solution, this
moves it too much out of the original "just beat CVS" goal. That said,
a simple solution which at least detects this kinds of mishaps would
be good, imho.

What about storing a UUID with each revision, and force client and
server to communicate using revision number and UUID? UUID can be used
to check for the errors I outlined above, while revision numbers are
for communicating with the user.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Aug 18 19:20:27 2003