Re: Locking server implementation: libsvn_repos or libsvn_fs

Greg Hudson <ghudson@MIT.EDU> writes:

> > > * The data currently managed by libsvn_fs is dumped and loaded, so
> > > adding locks would be breaking precedent.
> > >
> > The changes table isn't...
>
> The changes table isn't data; it's an index. (And one could argue that
> since it's used to determine what to dump, it is dumped.)

Yah. Subversion's repository dump mechanism is *not* a database
replication mechanism. It's a versioned history transport mechanism.
We don't dump stuff that is reconstructed by the load (like the
changes table). We don't dump stuff that isn't connected to a
finalized revision (like transactions and the not-yet-committed node
dangling off of them). We don't dump orphaned representations or
strings (if there are any). And so on.

Locks, also, are not a piece of versioned history. They exist outside
that history, just like <insert auth mechanism here>, just like hook
scripts and other processes that affect that history but aren't
actually part of it.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Oct 29 17:56:07 2004