You are probably releasing something else then an exception :)
But I can't reproduce the problem (though I'm seeing another one -
deploy tries to run java2db). I'll reassign 303 to Tom and add
my comments, and I'll file a new one for java2db.
thanks,
-marina
Gordon Yorke wrote:
> Issue 303 appears to be a different issue. On EntityManager.close() command we are attempting to release the database exception and this is resulting in an exception.
> --Gordon
>
> -----Original Message-----
> From: Marina Vatkina [mailto:Marina.Vatkina_at_Sun.COM]
> Sent: Sunday, February 26, 2006 4:27 PM
> To: Gordon Yorke
> Cc: Krogh Peter W; Ware Tom G; CMP-EJBTeam_at_Sun.COM
> Subject: Re: [Fwd: Re: changes post-beta]
>
>
> OK. Thanks. Is issue #303 the same or a different problem?
> I'm just installing the latest build now.
>
> thanks,
> -marina
>
> Gordon Yorke wrote:
>
>>Yes, on both issues. Changes to resolve issue 300 were checked in by me on friday. Those changes exposed another issue which I will patch and checkin.
>>Thanks,
>>--Gordon
>>
>>-----Original Message-----
>>From: Marina Vatkina [mailto:Marina.Vatkina_at_Sun.COM]
>>Sent: Sunday, February 26, 2006 3:21 PM
>>To: Yorke Gordon J
>>Subject: [Fwd: Re: changes post-beta]
>>
>>
>>Gordon,
>>
>>Are you aware of this?
>>
>>thanks,
>>-marina
>>
>>-------- Original Message --------
>>From: Shelly (Donna) McGowan <Shelly.McGowan_at_Sun.COM>
>>Subject: Re: changes post-beta
>>To: Scott Oaks <Scott.Oaks_at_Sun.COM>
>>CC: CMP-EJBTeam_at_Sun.COM
>>
>>
>>Scott,
>>
>>I believe the behavior your seeing is related to entity-persistence
>>issue 300 which was just fixed yesterday. See:
>>
>>https://glassfish.dev.java.net/issues/show_bug.cgi?id=300
>>
>>With that fix in todays build, we see IllegalStateException if find is
>>invoked outside TX context.
>>
>>Shelly
>>
>>
>>Scott Oaks wrote On 02/25/06 16:30,:
>>
>>
>>>I've attached a sample program which works quite well in glassfish beta.
>>>In recent builds (at least 37 and 38; I didn't check the ones bewteen
>>>32f and 37), it no longer is fuctionally correct: the application runs,
>>>but no data is ever persisted to the database.
>>>
>>>What changed in the interim that caused this code (and, as it turns out,
>>>the performace test for Oracle) to break?
>>>
>>>If you need to run the code to see, compile it with
>>>ant ear
>>>
>>>Deploy it to an appserver that has a connection pool named jdbc/SPECDB
>>>and a schema with one table:
>>>CREATE TABLE c_customer ( c_id integer not null);
>>>
>>>Visit the app like this:
>>>http://host:port/Test/TestServlet?action=3.0&id=1
>>>
>>>That should insert a row into c_customer with id of 1 (so change the id
>>>on subsequent tests). Which is what happens in b32. In b37+, the servlet
>>>executes but nothing is ever written to the database.
>>>
>>>Did I do something wrong (of course) or is something fundamentally
>>>broken?
>>>
>>>-Scott
>>
>>
>>
>>
>>
>
>
>