1) I did not get this email from Marina. I am not on the persistence glassfish mailing list so if anybody wants me to get mail they need to send it to my email address.
2) The answer to Marina's question is that I have said nothing whatsoever about the semantics of specifying or not specifying the @Inheritance annotation, only that in its current form the spec has not defined it to be optional.
3) I don't personally have any issues with it being optional, and I expect that most vendors, the RI included, will likely support it being optional. We have been given instructions that the spec should not change, though, unless there is a critical issue that needs fixing. In my view this is not critical and likely not a reason to re-open the spec.
-Mike
> -----Original Message-----
> From: Sanjeeb.Sahoo_at_Sun.COM [mailto:Sanjeeb.Sahoo_at_Sun.COM]On Behalf Of
> Sanjeeb Kumar Sahoo
> Sent: Tuesday, April 04, 2006 2:48 AM
> To: persistence_at_glassfish.dev.java.net; Linda.Demichiel_at_Sun.COM; Mike
> Keith
> Subject: Re: [Fwd: [Issue 486] Problem with abstract @Entity
> and @Id in
> b41]
>
>
> Mike, Linda,
>
> Has a decision been made on this? I am hoping the spec to be
> reworded so
> that @Inheritance becomes optional.
>
> Thanks,
> Sahoo
>
> Marina Vatkina wrote:
> > Mike,
> >
> > Are you saying that Java inheritance does not apply to Java
> Persistence
> > without an extra annotation?
> >
> > thanks,
> > -marina
> >
> > Mike Keith wrote:
> >> Actually that isn't true. We don't say that Lob must be
> specified but
> >> we have been assuming that it must also be present if a
> column is a lob.
> >> Same for Embedded. I don't think we even say anything
> about requiring
> >> Embeddable (except for the EmbeddedId case), but we do. Same for
> >> Inheritance.
> >>
> >> Note that I am not saying that some of things could not be
> >> defaulted, because most of them can and should be, just like
> >> Inheritance could be defaulted. I am just saying that we didn't
> >> specify it to be defaultable.
> >>
> >>
> >>> -----Original Message-----
> >>> From: Linda DeMichiel [mailto:Linda.Demichiel_at_Sun.COM]
> >>> Sent: Wednesday, March 29, 2006 8:51 PM
> >>> To: Mike Keith
> >>> Cc: persistence_at_glassfish.dev.java.net; Sanjeeb Kumar Sahoo; Krogh
> >>> Peter
> >>> W
> >>> Subject: Re: [Fwd: [Issue 486] Problem with abstract
> @Entity and @Id in
> >>> b41]
> >>>
> >>>
> >>> I see. However, we state in the spec when things are required
> >>> to be specified.
> >>> We don't require that @Inheritance be specified, and otherwise
> >>> always treat normal Java language inheritance as being operative,
> >>> so I think this interpretation is counterintuitive, although I
> >>> can see why confusion might arise.
> >>>
> >>> Linda
> >>>
> >>>
> >>> Mike Keith wrote:
> >>>
> >>>> Actually, I can see exactly why there is confusion. I have always
> >>>> assumed that the @Inheritance annotation is present and in
> >>>
> >>> the spec we
> >>>
> >>>> do not use the wording that we use in the other sections
> that assume
> >>>> a default when the annotation is absent.
> >>>>
> >>>> For example, in every section where the annotation may
> be defaulted
> >>>> if it is absent we say:
> >>>>
> >>>> "If no <InsertAnnotationHere> annotation is specified, ..."
> >>>> "If the <InsertAnnotationHere> annotation is not specified..."
> >>>> "If the <InsertAnnotationHere> annotation is missing..."
> >>>>
> >>>> We say this for:
> >>>>
> >>>> @Table - "If no Table annotation is specified..."
> >>>> @SecondaryTable - "If no SecondaryTable annotation is
> specified..."
> >>>> @Column - "If no Column annotation is specified..."
> >>>> @JoinColumn - "If no JoinColumn annotation is specified..."
> >>>> @JoinTable - "If the JoinTable annotation is missing..."
> >>>>
> >>>> And for the two inheritance annotations that *are* optional:
> >>>>
> >>>> @DiscriminatorColumn - "If the DiscriminatorColumn
> >>>
> >>> annotation is missing..."
> >>>
> >>>> @DiscriminatorValue - "If the DiscriminatorValue annotation
> >>>
> >>> is not specified..."
> >>>
> >>>> But in the Inheritance section, we define the enum, and
> >>>
> >>> then talk about
> >>>
> >>>> the *type* (the thing in the annotation element) being
> >>>
> >>> specified, not the
> >>>
> >>>> annotation:
> >>>>
> >>>> "public enum InheritanceType
> >>>> { SINGLE_TABLE, JOINED, TABLE_PER_CLASS };
> >>>>
> >>>> Support for the TABLE_PER_CLASS mapping strategy is
> >>>
> >>> optional in this release.
> >>>
> >>>> If no inheritance type is specified..."
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>> -----Original Message-----
> >>>>> From: Linda DeMichiel [mailto:Linda.Demichiel_at_Sun.COM]
> >>>>> Sent: Wednesday, March 29, 2006 5:20 PM
> >>>>> To: persistence_at_glassfish.dev.java.net
> >>>>> Cc: Mike Keith
> >>>>> Subject: Re: [Fwd: [Issue 486] Problem with abstract @Entity
> >>>>> and @Id in
> >>>>> b41]
> >>>>>
> >>>>>
> >>>>> I don't see how this confusion could arise from the spec.
> >>>>>
> >>>>> Inheritance applies as a result of the Java language "extends"
> >>>>> clause. You don't have to specify @Inheritance for it to
> >>>>> apply.
> >>>>>
> >>>>> (The @Inheritance annotation is not required to be specified
> >>>>> for inheritance. See section 9.1.29, which states that if no
> >>>>> annotation
> >>>>> is present, the single table per class hierarchy inheritance
> >>>>> strategy applies.)
> >>>>>
> >>>>> Linda
> >>>>>
> >>>>>
> >>>>>
> >>>>> Sanjeeb Kumar Sahoo wrote:
> >>>>>
> >>>>>
> >>>>>
> >>>>>> Hi Mike,
> >>>>>>
> >>>>>> Would you mind having a look at this issue? The latest
> >>>>>
> >>>>> update suggest
> >>>>>
> >>>>>
> >>>>>> that user must specify @Inheritance in a class to indicate
> >>>
> >>> that that
> >>>
> >>>>>> class is the root of the inheritance hierarchy. Why is this
> >>>>>
> >>>>> the case?
> >>>>>
> >>>>>
> >>>>>> Why can't root be identified by the persistence
> provider runtime by
> >>>>>> walking up the tree?
> >>>>>>
> >>>>>> Thanks,
> >>>>>> Sahoo
> >>>>>>
> >>>>>>
> >>>>>
> >>>>> --------------------------------------------------------------
> >>>>> ----------
> >>>>>
> >>>>>
> >>>>>> Subject:
> >>>>>> [Issue 486] Problem with abstract @Entity and @Id in b41
> >>>>>> From:
> >>>>>> pkrogh_at_dev.java.net
> >>>>>> Date:
> >>>>>> Wed, 29 Mar 2006 15:48:07 +0000
> >>>>>> To:
> >>>>>> issues_at_glassfish.dev.java.net
> >>>>>>
> >>>>>> To:
> >>>>>> issues_at_glassfish.dev.java.net
> >>>>>>
> >>>>>>
> >>>>>> https://glassfish.dev.java.net/issues/show_bug.cgi?id=486
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> User pkrogh changed the following:
> >>>>>>
> >>>>>> What |Old value |New value
> >>>>>>
> >>>>>
> >>>>> ==============================================================
> >>>>> ==================
> >>>>>
> >>>>>
> >>>>>> Issue type|DEFECT |ENHANCEMENT
> >>>>>>
> >>>>>
> >>>>> --------------------------------------------------------------
> >>>>> ------------------
> >>>>>
> >>>>>
> >>>>>> Priority|P3 |P4
> >>>>>>
> >>>>>
> >>>>> --------------------------------------------------------------
> >>>>> ------------------
> >>>>>
> >>>>>
> >>>>>>
> >>>>>>
> >>>>>> ------- Additional comments from pkrogh_at_dev.java.net Wed
> >>>>>
> >>>>> Mar 29 15:48:07 +0000 2006 -------
> >>>>>
> >>>>>
> >>>>>> This bug seems to be about the provider determining that
> >>>>>
> >>>>> inheritance is implied
> >>>>>
> >>>>>
> >>>>>> without specifying the @Inheritance annotation. This is
> >>>>>
> >>>>> not defined in the
> >>>>>
> >>>>>
> >>>>>> spec, and I believe that it was not the intention of
> the spec. The
> >>>>>> @Inheritance tag defines that Inheritance is needed.
> >>>>>>
> >>>>>> It seems like a nice usability feature, but it would
> >>>>>
> >>>>> potentially require the
> >>>>>
> >>>>>
> >>>>>> provider to walk the whole hierarchy instead of just
> >>>>>
> >>>>> looking for @Inheritance.
> >>>>>
> >>>>>
> >>>>> ------------------------------------------------------------
> >>>
> >>> ---------
> >>>
> >>>>>> To unsubscribe, e-mail:
> issues-unsubscribe_at_glassfish.dev.java.net
> >>>>>> For additional commands, e-mail:
> issues-help_at_glassfish.dev.java.net
> >>>>>>
> >>>>>
> >>
>
>