Re: Locking server implementation: libsvn_repos or libsvn_fs

Ben Collins-Sussman wrote:

> It seems like most of the smaller issues have dropped out of this
> thread. Allow me to oversimplify.
>
> * The remaining 'big' argument for locks-in-fs is "it's more
> convenient for
> callers."
>
> * The remaining big argument for locks-in-repos is "it's easier to
> implement".

Change "easier" to "less work", and I agree.

> * Everyone wants to implement the lock-table (or tree) without using a
> database, regardless of whether we do locks-in-fs or locks-in-repos.

How did you reach that conclusion? IMHO if we put locks in the FS, it's
the back-end's choice where to put the locks. I can't imagine the BDB
back-end _not_ putting locks into a BDB table.

> * How 'svnadmin dump' deals with locks is also independent of where
> locks are implemented. Either we tell people to back up the locks
> data manually (like we do with hooks, authz file, etc), or we
> augment the dump format and make svnadmin grab the lock data,
> wherever it may be.

Yup.

-- Brane

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 30 09:23:43 2004