Thanks, Marina.
Regarding 3). The new wording is much better. Using the Local or Remote
annotation *with an empty value* on an interface of the bean class would
provide more clarity, but maybe this is/can be assumed since a value
should not be provided?
A few more minor comments -
4) Section 5.4.6, pg131 - The compatibility note indicates that the MDC is
available via dependency injection, but makes no mention that the context
could also be looked up via java:comp/EJBContext.
5) Section 11.5.1.1 pg259 - bullet 3, "Stateles annotation" (missing an
s)
6) pg356 - the formatting of the the doc in the header in the xsd is a bit
off
-Jeremy
From: Marina Vatkina <marina.vatkina_at_oracle.com>
To: jsr345-experts_at_ejb-spec.java.net,
Cc: Jeremy Bauer/Rochester/IBM_at_IBMUS
Date: 01/07/2013 09:43 PM
Subject: [ejb-spec users] [jsr345-experts] Re: Re: Public Draft doc
candidates uploaded to the project area for review
Hi Jeremy,
Thanks for reviewing the spec.
On 1/7/13 12:50 PM, Jeremy Bauer wrote:
Hi Marina,
No comments on the optional doc. Had a few comments on core pd4 thus far.
I'm still reading through the core updates so I may have a more comments.
If so, I'll get them to you by tomorrow AM.
1) The text in 4.6.1 was updated to state that an SFSB can be removed when
CDI context ends. It would also be useful to update the Figure 5 (Life
Cycle of a Stateful Session Bean Instance).
Good point. There is not much space available, but I can add "or CDI
context destroyed" in-between "Remove method/or timeout" lines.
2) The footnote [27] on p90 is the only use of the term "failover" in the
spec. I assume it means what we think it means, but it seems out of place
to use the terminology without introducing it. I would say either the
footnote should be removed, or a section should be introduced to explain
the meaning of the term (and state that it's not required, and that
configuration is beyond the scope of the spec), and then the footnote
should say something like "If the container supports failover, disabling
passivation of a stateful session bean might prevent that functionality
from working.". Also, the current text also has a typo: "that is
failover" -> "that if failover".
Thanks. This was a request from an internal Java EE architecture review.
I'll follow up on it.
3) In section 4.7, I think would be clearer with the following
clarifications :
The bean class must implement the interface or the interface must be
designated as a local or remote business interface of the bean by means of
the Local or Remote annotation or in the deployment descriptor. The
following rules apply:
The following interfaces are excluded when determining whether the bean
class has business interfaces: java.io.Serializable;
java.io.Externalizable; any of the interfaces defined by the javax.ejb
package.
If a bean class has more than one interface—excluding the interfaces
listed above—and at least one interface of the bean class is explicitly
designated as a business interface of the bean by means of the Local or
Remote annotation *with a non-empty value* on the bean class or *the Local
or Remote annotations with an empty value on an* interface *on implements
clause of the bean class* or in the deployment descriptor, or if the bean
exposes a no-interface view, all business interfaces must be explicitly
designated as such.
How about this for the 2nd bullet:
If a bean class has more than one interface—excluding the interfaces
listed above—*all* business interfaces must be explicitly designated as
such if any of the following is true:
* the bean exposes a no-interface view
* *any* interface of the bean class is explicitly designated as a
business interface of the bean by either of the following means:
* the Local or Remote annotation with a non-empty value on the bean
class
* the Local or Remote annotation on an interface of the bean
class
* in the deployment descriptor
thanks,
-marina
Otherwise:
If the bean class is annotated with the Remote annotation *with an empty
value*, all such interfaces are assumed to be remote business interfaces
of the bean.
If the bean class is annotated with the Local annotation *with an empty
value*, or if the bean class is annotated with neither the Local nor the
Remote annotation, all such interfaces are assumed to be local business
interfaces of the bean.
Even with that, the second bullet is quite dense, and it's not entirely
clear whether class-level @Local/_at_Remote with non-empty value (or a
business-local/remote in XML) should prevent the implements clause from
being searched or not, or whether or not empty class-level @Local should
override or conflict with interface-level @Local/_at_Remote.
-Jeremy
From: Marina Vatkina <marina.vatkina_at_oracle.com>
To: jsr345-experts_at_ejb-spec.java.net,
Date: 01/07/2013 02:13 PM
Subject: [ejb-spec users] [jsr345-experts] Re: Public Draft doc
candidates uploaded to the project area for review
Experts,
If I don't hear otherwise, I'll push the specs to the JCP by the COB
tomorrow.
-marina
On 12/28/12 6:47 PM, Marina Vatkina wrote:
> Experts,
>
> Please review ASAP so that I can submit the docs to the JCP on-time.
> I'll upload the javadoc when I figure out why it's not being built in
> my workspace.
>
> *Core document:
>
http://java.net/projects/ejb-spec/downloads/download/ejb-3_2-core-pd-4.pdf*
>
> Major changes since pd-3 (the complete list is in A.5, the change-bars
> reflect changes since EDR):
> - Removed restrictions on javax.ejb.Timer and javax.ejb.TimerHandle
> access (and removed restriction on cancelling timers returned by
> getAllTimers method)
> - EJBContainer class now implements the java.lang.AutoCloseable
interface
> - Added references to the CDI spec and rules on constructors and
> removal of the CDI-enabled beans
> - Removed requirements for the container provider to publish the
> java:comp/uniqueMDBName name (it will be a new API in the Connector
spec)
> - Updated XML Schema definition to the latest version
> - Restored references to entity beans in Ch 10, 11, and 12 as
> separating the rules added an unnecessary complexity (see changes to
> the Optional Doc). Added footnotes that component contract and client
> view of entity beans are described in the EJB Optional Features
document.
>
> *Optional document:
>
http://java.net/projects/ejb-spec/downloads/download/ejb-3_2-optional-prd-candidate.pdf*
>
> Major changes since EDR (the complete list is in A.3, the change-bars
> reflect changes since EDR):
> - Replaced section numbers with the section titles when referencing
> EJB Core Contracts and Requirements document.
> - Removed text from Ch 9 and 10 that was exactly the same as in the
> EJB Core Contracts and Requirements document.
> - Removed most of the text from Ch 11 as it was an unnecessary
> duplication of the corresponding chapter in the EJB Core Contracts and
> Requirements document. Left only section “Mapping Objects Returned by
> Entity Bean Finder Methods to IDL”
> - Removed completely chapters “Enterprise Bean Environment” and
> “Security Management” as it was an unnecessary duplication of the
> corresponding chapters in the EJB Core Contracts and Requirements
> document.
>
> Have a Happy New Year!
> -marina