Hi Daniel, I meant the way that a new transaction is acquired after the commit, in the 5.0.14, it always call: transaction = session.beginTransaction() to get a new transaction. I don't know if this really fix to all my problems, but i do have some bugs that i can't understand, example, I have dispatcher, page and service, dispatcher create a persisted log object, page created a persisted parent record while service created a child record, this was working until i found out that i need to commit in the dispatcher in order to persist the object, now everything wrong, Referential Integrity exceptions pop out, removing the cp30 works, putting it back failed, until I put a commit also in the page class before calling the service method to make it work, and I don't know why I have to put a commit in the page while page and service transactions are actually the same one, this trial and error approach took several days, a workaround is found, but not so sure why. upgraded to 5.0.14, and I removed the commit in the page, everything works, at least I can explain this time that I don't need a manual commit in the page before calling service method as eventually it will be commited due to the @CommitAfter in the page class. I guess, if you use hibernate in a normal way, say, purely in the page, or you do not put referential integrity in the table level, 5.0.13 might work, otherwise it's really very frustrating to trace those problems without really knowing what the problem is.
I believe this problem was there in the early versions, remember I have a post http://www.nabble.com/T5%3A-how-to-use-Hibernate-properly--to15580722.html#a15580722 How to use Hibernate properly before? rule no. 1 in that post is:1. Don't commit, let Tapestry-Hibernate take care of that. I think that's actually due to this bug, it really kind of disappointing when you have no control over the transaction, that actually means you have no control over the integrity of the data. if you use tapestry-hibernate heavily and this bug is there, really a problem, taking time to debug, actually you are just trying to find a workaround:( glad 5.0.14 finally solved this problem and it has to be released asap. Angelo Daniel Jue wrote: > > Do you mean the "bug" of not committing automatically? If so, there > have been posts about that one, because it was actually a fix. > > > On Fri, Aug 1, 2008 at 9:37 PM, Angelo Chen <[EMAIL PROTECTED]> > wrote: >> >> I strongly recommend to release this 5.0.14-SNAPSHOT as soon as possible, >> 5.0.13 has a serious bug in its commit method with hibernate, it >> generates >> unpredictable results making debugging very difficult, switching to >> 5.0.14 >> solves the problem, I'd say quite impossible to proceed in 5.0.13 if you >> have some hibernate code. >> >> >> >> Peter Stavrinides wrote: >>> >>> Hi >>> >>> I am just curious to know if the 14 snapshot is due for release anytime >>> soon? >>> >>> Thanks, >>> Peter >>> >>> >>> --------------------------------------------------------------------- >>> To unsubscribe, e-mail: [EMAIL PROTECTED] >>> For additional commands, e-mail: [EMAIL PROTECTED] >>> >>> >>> >> >> -- >> View this message in context: >> http://www.nabble.com/Tapestry-5.0.14-SNAPSHOT-tp18770439p18785186.html >> Sent from the Tapestry - User mailing list archive at Nabble.com. >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [EMAIL PROTECTED] >> For additional commands, e-mail: [EMAIL PROTECTED] >> >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > > -- View this message in context: http://www.nabble.com/Tapestry-5.0.14-SNAPSHOT-tp18770439p18795792.html Sent from the Tapestry - User mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
