John, have you really, actually looked at the state of JEE these days? It's trivial to set up, trivial to develop in, and IMHO, easier to learn than many other things in Java.
Yes, the "old pre-JEE5 days were bad" but JEE 5,6,7 are a whole different animal. You can get it up and running in less than 10 minutes (Glassfish) and it's much easier than trying to cobble up the same environment from different pieces. Also, we developed a module (FlowLogix) that integrates most things JEE with Tapestry and Shiro. I have seen this again and again, my customers have the perception of "JEE bad/complicated" and spending at least 5x the time trying to cobble up an environment that tries (unsuccessfully) to replicate what JEE is doing. On Aug 27, 2013, at 3:49 AM, John wrote: > I'm reluctant to move my small/medium sized pure Tapestry project into a JEE > environment even though I am updating accross multiple databases and have > non-trivial needs for transactions. I don't see sufficient justifcation to > use JEE unless there is substantial need of EJBs. > > So far I have got away with using CommitAfter and some more manual coding to > handle the transactions. The main reason CommitAfter sucks IMO is because it > is coded on the interface and so breaks the rule about hiding implementation. > For that reason alone IMO it needs to be addressed. > > If I do decide to go into XA I will use Bitronix or similar over JEE and code > my transactions manually in the implementing classes, unless more demand for > EJB arises. > > John > > > ----- Original Message ----- > From: Taha Hafeez Siddiqi > To: Tapestry development > Sent: Sunday, August 25, 2013 1:17 AM > Subject: Re: Transactions and AfterCommit > > > I have used spring and JEE in the past and I don't think every project needs > them. The transaction support comes with a lot-n-lots of dependencies (at > least at that time it was the case :)) and some people don't like it. > > All we need is a support for @Transactional->Required /readonly. I think if > we support them, most of common requirements are met. > > regards > Taha > > On 25-Aug-2013, at 4:18 AM, Lenny Primak <[email protected]> wrote: > >> I would leave everything as is now. >> Tapestry should not try to implement or re-implement full transaction >> support. >> This has already been done with JEE or spring. If a user wants this support, >> they should just use what already exists out there. >> >> >> On Aug 24, 2013, at 3:18 PM, "Thiago H de Paula Figueiredo" >> <[email protected]> wrote: >> >>> On Sat, 24 Aug 2013 09:39:11 -0300, Taha Siddiqi <[email protected]> >>> wrote: >>> >>>> Hi everyone >>> >>> Hi! >>> >>>> There are two @CommitAfters and both work differently from each other. >>> >>> This is a problem >>> >>> Here's my suggestion: >>> >>> 1) Leave the @CommitAfter implementations the way they are now for >>> backward-compatibility reasons. >>> 2) Mark them as deprecated. >>> 3) Use EJB's @TransactionAttribute annotation instead of tapestry-hibernate >>> and tapestry-jpa defining different annotations. >>> 4) Implement the different transaction attribute types described in >>> http://docs.oracle.com/cd/B32110_01/web.1013/b28221/servtran002.htm. >>> >>> Question: use JTA? I don't know. >>> >>> -- >>> Thiago H. de Paula Figueiredo >>> >>> --------------------------------------------------------------------- >>> 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] >> > > > --------------------------------------------------------------------- > 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]
