> > Indeed this looks primising and removes some of the caveats I
> > mentioned with threads in EJBs. So will the JSR 339 spec get
> > overhauled by replacing "normal" executors by managed ones?
> > Particularly the examples showing how to use executors should be
> > changed so people will understand that they should use managed ones
> in non-SE environments.
>
> We could at least include a reference to this spec. Note that the
> current spec does not show any code related to the creation of
> executors, it assumes they were created elsewhere. In any case, I will
> update that section to mention 236.
>
> -- Santiago
Actually page 47 of the public draft showed a code example which utilizes
Executors.newSingleThreadExecutor. As you remember, I complained that this
is invalid in an EJB scenario. So the new managed executor could be used
here instead.