imo we should provide the api as well as the corresponding tests early (at
least as an information until we have bv.next).
regards,
gerhard
2013/7/26 Gunnar Morling <gunnar(a)hibernate.org>
Hi,
To give some background: the idea was to let integration frameworks decide
which executables should be validated to allow for consistency with their
own rules for interceptor enablement and request processing; e.g. JAX-RS
validates getter-style resource methods by default, or a DI framework might
have its own annotation for controlling validation or descriptor format for
registering interceptors.
Based on that, @ValidateOnExecution (and the settings in XML) is provided
as the *recommended* way of controlling executable validation but its not
obligatory: "This section expresses the behavior that integration with
interception frameworks should follow. Any deviation should be considered
with care as it will surprise Bean Validation users."
That being said, I agree that adding a facility for retrieving these
settings via the BV API makes sense to make life easier for those
integrators wishing to adhere to the BV behavior; integrators with deriving
behavior could still use this API as input for their evaluation.
--Gunnar
2013/7/26 Emmanuel Bernard <emmanuel(a)hibernate.org>
> I will let Hardy and Gunnar answer that one, this has been discussed a
> lot and some arguments they provided lead to delegate the rules
> implementations to the integration framework.
>
> On 25 juil. 2013, at 22:55, Gerhard Petracek <gerhard.petracek(a)gmail.com>
> wrote:
>
> > hi @ all,
> >
> > currently every framework which needs to integrate (bv based)
> method-validation has to implement the rules specified for
> @ValidateOnExecution itself.
> > there is no method in the api for delegating the evaluation to the
> bv-implementation (and everything passed to ExecutableValidator gets
> validated
> > independent of the xml based or annotation based config).
> >
> > it isn't just a higher effort for such frameworks, it (already) leads
> to inconsistencies across frameworks.
> > -> imo it should be one of the first topics for bv.next.
> >
> > regards,
> > gerhard
> > _______________________________________________
> > beanvalidation-dev mailing list
> > beanvalidation-dev(a)lists.jboss.org
> >
https://lists.jboss.org/mailman/listinfo/beanvalidation-dev
>
> _______________________________________________
> beanvalidation-dev mailing list
> beanvalidation-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/beanvalidation-dev
>
_______________________________________________
beanvalidation-dev mailing list
beanvalidation-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/beanvalidation-dev