I see.  Fixing the old impl (let's call this JDBCCacheLoader and the new one IndexingJDBCCacheLoader to prevent confusion) to persist the root will break running systems though, I fear.  

I do think this should be fixed in the JDBCCacheLoader though; perhaps a param like "persist.root = true" but by default it should NOT persist the root so it doesn't break existing systems.

Or will it even break existing systems?  Am I being unnecessarily careful here?
--
Manik Surtani

Lead, JBoss Cache
JBoss, a division of Red Hat

Email: manik@jboss.org
Telephone: +44 7786 702 706
MSN: manik@surtani.org
Yahoo/AIM/Skype: maniksurtani

On 1 Feb 2007, at 13:01, Mircea Markus wrote:

Hi,
New impl is 100% backward compatible - but I have to insert the root as it is missing from old impl.

Cheers,
Mircea

On 2/1/07, Manik Surtani <manik@jboss.org> wrote:
So without the root element you're not going to be able to persist any data stored directly in the root node (if anyone ever does this!)

Is the new impl in any way compatible with the existing JDBCCacheLoader anyway?  I thought their storage formats were different.


--
Manik Surtani

Lead, JBoss Cache
JBoss, a division of Red Hat

Telephone: +44 7786 702 706
Yahoo/AIM/Skype: maniksurtani



On 1 Feb 2007, at 03:49, Mircea Markus wrote:

In order to make new implementation 'backward' compatible (i.e. to work on existing data managed by JDBCacheLoader) I need root to be persisted - implementation heavily relies on that.
My thought is to check at startup whether it exists, and create it. Any cons/alternatives?

Cheers,
Mircea

On 2/1/07, Mircea Markus < mircea.markus@gmail.com> wrote:
Hi,

loader.put(any_fqn);
assertTrue( loader.exists(Fqn.ROOT)); //fails

The problem is that root element is not persisted.

Cheers,
Mircea


_______________________________________________
jbosscache-dev mailing list