On 8 May 2013, at 13:37, Pedro Ruivo wrote: > Hi guys, > > Is there any reason to enable the recovery when we pick from > TransactionConfigurationBuilder? > > public RecoveryConfigurationBuilder recovery() { > recovery.enable(); > return recovery; > } the method documents that as the the behaviour, can't remember the logic behind it. > > This will enable the Recovery in configurations like this: > > <namedCache name="noRecovery"> > <transaction (...) transactionMode="NON_TRANSACTIONAL"> > <recovery enabled="false" recoveryInfoCacheName="someName"/> > </transaction> > </namedCache> > > leading the RecoveryConfigTest to fail. This should result in a cache with recovery disabled so it is a bug. Please file a JIRA.

> > ps. I detected this when working on > https://issues.jboss.org/browse/ISPN-3063. > > Cheers, > Pedro > _______________________________________________ > infinispan-dev mailing list > infinispan-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/infinispan-dev Cheers,

On 8 May 2013, at 13:52, Pedro Ruivo wrote: >> This should result in a cache with recovery disabled so it is a bug. Please file a JIRA. > > Can I do the fastest solution and remove the recovery.enable()? -1. That would mean breaking public API. as that method would behave differently. > > or should I be safe and change the parser to invoke > TransactionConfigurationBuilder.recovery() only once before parsing the > attributed? +1. Or set the value as parsed explicitly. > Should I also change the parser for 5.1 and 5.2? please do.

Cheers,