"Glenn A. Thompson" <gthompson@cdr.net> writes:
> Ouch! My brain can't handle that.
>
> What series of events causes this? Is it branches that have nodes
> that have branches? I'm playing with svnadmin lsrevs. This should
> help to get my brain wrapped around this stuff a little better. If
> I know what commands I issued and have the state of the FS I think
> can follow along in the code easier. But this definitely creates
> some problems for a SQL DB. Yikes!
The cause is pretty straightforward, and is (I think) directly the
result of our overloading our node-ids with the responsibility of
reporting ancestry as well as just being UIDs. I won't go into the
algorithm now (you can find it in various comments in the libsvn_fs/
code...in fact, read 'structure' in that directory for a grand
overview of the filesystem), but Tutt's new schema should make all the
icky go away.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat May 4 17:12:36 2002