OK, I said something before, but I didn't have a concrete solution to
your problem. I'm going to commit the bugfix portion of the patch
today. In the future, plese post separate patches for bug fixes, and
new features.
As for the feature, I'm +0, because I don't like the implementation (not
the idea):
Will it really be necessary to notify more than one cache finalizer
object that an object is leaving the cache? You really just want to
clean up the object, it just needs one object that knows how to do that.
Chris, what do you think about changing the vector of listeners to a
single private CacheFinalizer object, and then changing the addListener
to setFinalizer. And get rid of the loop in the finalize method, and
just call the method on the one finalizer (if it's not null).
-Nissim
Chris Kimpton wrote:
>
> Hi,
>
> > I think you already gave some sort of description, but could you be more
> > specific and verbose about what the patch is for?
>
> There were several mails on the list. The main points are;
>
> Fixed the clearCache routine - it was never removing stale objects due to a bug in
>the use of the hashtable.
>
> Added a method to CachedObject to allow its expiry time to be reset (it just resets
>the created time to now).
>
> Added a facility for people to register listeners to be notified when an object is
>being removed from the cache.
>
> I need this because I am using a custom db implementation (which I am not currently
>looking at integrating with Turbine...) and I want to be put my db connection into
>the cache - but to be notified when it is expired - so that I can close it and free
>its resources.
>
> An example of usage of this is as follows (I hope the indenting gets sent this
>time...)
>
> // START OF SAMPLE CODE
>
> Database db = null;
> GlobalCacheService gs = null;
> try
> {
> // Look for the item in the cache.
> // If it doesn't exist or the item is stale,
> // the cache will throw an exception.
>
> gs = (GlobalCacheService)TurbineServices.getInstance()
> .getService(GlobalCacheService.SERVICE_NAME);
>
> CachedObject obj = gs.getObject(CACHE_NAME_FOR_DB);
> // got a db object - reset its expiry date - indicating its been used recently
> db = (Database)obj.getContents();
> obj.resetExpires();
>
> } catch(ObjectExpiredException gone)
> {
> db = org.melati.poem.odmg.ODMGFactory.getNewDatabase();
> db.open("vgdb",Database.OPEN_READ_WRITE);
> CachedObject obj = new CachedObject(db);
> // add this anonymous listener to close the db when its removed from the cache
> obj.addListener(new CachedObjectListener() {
> public void cachedObjectFinalizer(CachedObject cobj) {
> Log.note("in cachedObjectFinalizer for "+cobj);
> try {
> Database cdb = (Database)cobj.getContents();
> cdb.close();
> } catch (ODMGException exc) {
> Log.error(exc);
> }
> }
> });
> gs.addObject(CACHE_NAME_FOR_DB, obj);
> }
> return db;
>
> // END OF SAMPLE CODE
>
> > Also, I personally
> > won't tend to look at any patches that aren't generated as unidiffs
> (i.e. `cvs diff -u`). Thanks Chris.
>
> I will submit a unidiff patch tonight (the code is at home, if only this firewall
>didn't limit us to port 80 only...).
>
> Regards,
> Chris
> ---
> "surely it is madness to accept life as it is and not as it could be"
> ______________________________________________
> FREE Personalized Email at Mail.com
> Sign up at http://www.mail.com/?sr=signup
>
> ------------------------------------------------------------
> To subscribe: [EMAIL PROTECTED]
> To unsubscribe: [EMAIL PROTECTED]
> Search: <http://www.mail-archive.com/turbine%40list.working-dogs.com/>
> Problems?: [EMAIL PROTECTED]
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Search: <http://www.mail-archive.com/turbine%40list.working-dogs.com/>
Problems?: [EMAIL PROTECTED]
