I had considered using the reporter interface more than once.
(So I could benefit from your work!)
But currently many things in svk depend on the mixed-revision root
that reflects the checkout copy. The operation done to them
are not just dir_delta, but also check_path, dir_entries, etc.
So in the end I probably have to implement a root class proxying
calls to individual revision_root. Additionally in that case
I could refactor what libsvn_fs does today, if by the point
of time we still only have bdb as fs backend.
Or perhaps this mixed-revision transaction-less root should be done
in svn, so new fs backends could just use it if they have
revision_root implemented...
On Mon, Feb 23, 2004 at 01:14:57PM -0500, Greg Hudson wrote:
> > The original post mentioned removal of dir_delta.
>
> Yeah, I think that was a pipe dream; I'm told svk uses it, for
> instance. It might be possible to factor the two bits of code together
> at some point, though.
Cheers,
CLK
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Feb 23 19:34:07 2004