[bv-dev] Validation of constructor constraints

[Emmanuel Bernard]

I agree with Michael and you. We would only come up with fairly arbitrary and unintuitive rules. Depending on the interception technology we even might end up with double validations.

On 30 juin 2012, at 11:11, Michael Nascimento wrote:

> There is method overriding, but no constructor overriding. Therefore,
> it seems perfectly fine for me.
> 
> Regards,
> Michael
> 
> On Sat, Jun 30, 2012 at 4:41 AM, Hardy Ferentschik <hardy at hibernate.org> wrote:
>> Hi everyone,
>> 
>> I hope you had some time to experiment with the functionality in Hibernate Validator 5.0.0.Alpha1.
>> One question Gunnar and I discussed when implementing the latest spec was constructor validation and constraints on super types.
>> 
>> Probably best to use an example:
>> 
>> 
>>        public class Foo {
>>                Foo(@Length(max = 10) String s) {
>>                }
>> 
>>                Foo(@Min(5) Integer i) {
>>                }
>> 
>>                public void foobar(@NotNull String s) {
>>                        // ...
>>                }
>>        }
>> 
>>        public class Bar extends Foo {
>>                public Bar(String s) {
>>                        super( s );
>>                }
>> 
>>                public void foobar(String s) {
>>                        // ...
>>                }
>>        }
>> 
>> In this case we have Bar extending Foo. What happens if Bar#foobar is called? In this case the @NotNull constraint from
>> Foo#foobar get applied. So far so good. The question is what happens with constructor validation?
>> 
>> In the example Bar has one constructor which defines no constraint. The matching super type constructor Foo(String) defines
>> an @Length constraint. Should this constraint be applied? In the example Bar(String) calls the corresponding string constructor
>> in the super type, but it might as well call the integer constructor. There is no telling from a Bean Validation point of view.
>> 
>> For this reason Hibernate Validator does not apply any super types constraints when validating constructors. We think that is
>> the right behavior, but wanted to run this by everyone. Either way it is probably worth to clarify the spec around this.
>> 
>> Thoughts?
>> 
>> --Hardy
>> _______________________________________________
>> beanvalidation-dev mailing list
>> beanvalidation-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/beanvalidation-dev
> 
> _______________________________________________
> beanvalidation-dev mailing list
> beanvalidation-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/beanvalidation-dev