Hi!
> |> Rickard posted its solution and I implemented it;
> |> Rickard restored the throw new RemoteException("Reentrant call")
> |(and so I
> |> did) that is now remarked (around line 185 in
EntityInstanceInterceptor).
> |> This led to the break of my 11-th test, 'cause when I remarked
> |it locktest
> |> run fine.
> |> This part of code is entered because the tx associated with the
> |context is
> |> null.
> |
> |Yes, I saw that reentrancy was not checked. I remember vaguely some
> |discussion on that topic, and apparently it was removed. Why?
> |
> |In any case, is it back now? (i.e. reentrant calls on non-reentrant beans
> |throws remoteexception?)
>
>
> rickard,
>
> the reentrancy test is wrong on entity.
>
> If threads come in the bean with the same transaction, it is perfectly
valid
> to go on invocation it is not a "reentrant" call from the instance through
> its EJBObject (you know the "many threads per TX thing ;-)
EJBObject calls, true. Reentrant calls in one tx on an Entity marked as
non-reentrant should fail however. I could not see in code that this was
dealt with.
/Rickard
