On Tue, Mar 3, 2009 at 7:28 AM, stan <stanb AT panix DOT com> wrote:
>> The major addition on the client side is the Application API.  While
>> we could probably patch a minimal client to support client-side
>> configuration files, such a minimal client will never support the
>> Application API.  But that's OK!
>
> Could you give me a quick description of this?

It's a client-side API that allows Amanda to back up anything you can
write a script for - ZFS partitions, PostgreSQL databases, ORACLE
databases, and so on.  It's far more featureful and flexible than the
old system of writing a wrapper around gnutar.  There's a related
Pre/Post-script API, too, which allows you to run scripts of your
choosing before or after a dump.

> Sorry, I am not completly current on source code control systesm (I'm still
> stuck with SCCS in places)> So let me get an understanding. Current is in
> git, 0.0 -? some version is in CVS, some version to some other version is
> in Subversion?

I don't know when CVS begins, but it contains versions up through
2.5.1p3.  SVN was initially set up with a 'cvs export' and 'svn
commit' rather than the more sensible cvs2svn, so the Subversion
repository doesn't contain any of the history that's in the CVS
repository.  Subversion continues to be authoritative to this day.

The active developers are using git in parallel with Subversion (git
is a *lot* more flexible and makes development a lot easier), so
you'll see exactly the same source code checking out 'trunk' from the
Subversion repo on http://sourceforge.net/projects/amanda and checking
out 'master' from http://github.com/nikolasco/amanda/.

Git has the interesting ability to "splice" history together, so at
some point I would like to import all of the history from CVS into
git, and "splice" that onto the history already imported from
Subversion, such that we have all of the history available in a single
repository.

Dustin

-- 
Storage Software Engineer
http://www.zmanda.com