Would it be feasible to raise a warning or error in cases where mapping annotations are given on an element not matching the default access type for that entity and @Access has not been given explicitly either?

The most common case of this issue which I saw several times happen is where @Id is given at the field level but some other mapping annotations are given on property getters

for those not super-deep into the details of access type retrieval.

A warning might help them.

+1 this is a very common mistake for beginners. But I'm with Hardy in not trying to shoehorn such support in our current metadata handling, but rather to keep it in mind when moving it all to Jandex. (how to remember that?) On 12 January 2016 at 11:49, Hardy Ferentschik <hardy(a)hibernate.org&gt; wrote: > Hi, > >> Would it be feasible to raise a warning or error in cases where >> mapping annotations are given on an element not matching the default >> access type for that entity and @Access has not been given explicitly >> either? > > Sure, but I think it would need some doing. AFAIR the code determines the access type > and from then on just reflects on either fields or methods. Maybe as a performance > optimisation as well. Not sure. I agree having some warning for the case you describe > is a good idea, but don't expect this to be a one lines in the code. > > That said, things would be easier with Jandex. There it would be easier to determine > inconsistent configuration. I seem to vaguely recall that I might even have done though > when working in the metamodel branch. > > Personally I'd probably address this in conjunction with a re-write of the mapping code > and switch to Jandex. I think that is still somewhere on the horizon. > > If not, it is for sure feasible, just take quite some effort. > >> The most common case of this issue which I saw several times happen is >> where @Id is given at the field level but some other mapping >> annotations are given on property getters > > Right, that's a usual mistake. > >> for those not super-deep into the details of access type retrieval. > > I don't think you need to be super-deep into ORM/JPA to understand that. > There are warnings and notes about this everywhere. > >> A warning might help them. > > For sure, but as said, I'd tackle this as part of a bigger refactoring. > > --Hardy > > > _______________________________________________ > hibernate-dev mailing list > hibernate-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/hibernate-dev _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev

There you go: https://hibernate.atlassian.net/browse/HHH-10440 2016-01-12 16:08 GMT+01:00 Steve Ebersole <steve(a)hibernate.org&gt;: > Create the issue and use the 'jandex-binding' version in Jira > > On Tue, Jan 12, 2016 at 5:57 AM Sanne Grinovero <sanne(a)hibernate.org&gt; wrote: > >> +1 this is a very common mistake for beginners. >> >> But I'm with Hardy in not trying to shoehorn such support in our >> current metadata handling, but rather to keep it in mind when moving >> it all to Jandex. (how to remember that?) >> >> On 12 January 2016 at 11:49, Hardy Ferentschik <hardy(a)hibernate.org&gt; >> wrote: >> > Hi, >> > >> >> Would it be feasible to raise a warning or error in cases where >> >> mapping annotations are given on an element not matching the default >> >> access type for that entity and @Access has not been given explicitly >> >> either? >> > >> > Sure, but I think it would need some doing. AFAIR the code determines >> the access type >> > and from then on just reflects on either fields or methods. Maybe as a >> performance >> > optimisation as well. Not sure. I agree having some warning for the case >> you describe >> > is a good idea, but don't expect this to be a one lines in the code. >> > >> > That said, things would be easier with Jandex. There it would be easier >> to determine >> > inconsistent configuration. I seem to vaguely recall that I might even >> have done though >> > when working in the metamodel branch. >> > >> > Personally I'd probably address this in conjunction with a re-write of >> the mapping code >> > and switch to Jandex. I think that is still somewhere on the horizon. >> > >> > If not, it is for sure feasible, just take quite some effort. >> > >> >> The most common case of this issue which I saw several times happen is >> >> where @Id is given at the field level but some other mapping >> >> annotations are given on property getters >> > >> > Right, that's a usual mistake. >> > >> >> for those not super-deep into the details of access type retrieval. >> > >> > I don't think you need to be super-deep into ORM/JPA to understand that. >> > There are warnings and notes about this everywhere. >> > >> >> A warning might help them. >> > >> > For sure, but as said, I'd tackle this as part of a bigger refactoring. >> > >> > --Hardy >> > >> > >> > _______________________________________________ >> > hibernate-dev mailing list >> > hibernate-dev(a)lists.jboss.org >> > https://lists.jboss.org/mailman/listinfo/hibernate-dev >> _______________________________________________ >> hibernate-dev mailing list >> hibernate-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/hibernate-dev >> > _______________________________________________ > hibernate-dev mailing list > hibernate-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/hibernate-dev