users@jax-rs-spec.java.net

[jax-rs-spec users] [jsr339-experts] Re: Re: Re: Bring back BeanValidation

From: Bill Burke <bburke_at_redhat.com>
Date: Mon, 22 Oct 2012 09:47:37 -0400

On 10/22/2012 5:07 AM, Marek Potociar wrote:
>
> On Oct 19, 2012, at 7:33 PM, Bill Burke <bburke_at_redhat.com> wrote:
>
>>
>>
>> On 10/19/2012 11:50 AM, Marek Potociar wrote:
>>>
>>> On Oct 19, 2012, at 4:30 PM, Bill Burke <bburke_at_redhat.com> wrote:
>>>
>>>>
>>>>
>>>> On 10/19/2012 9:29 AM, Emmanuel Bernard wrote:
>>>>> On Fri 2012-10-19 8:40, Bill Burke wrote:
>>>>>>
>>>>>>
>>>>>> On 10/19/2012 8:32 AM, Emmanuel Bernard wrote:
>>>>>>> Bill, I am not in the jsr399 EG list so please forward my email if the
>>>>>>> list is not open to non registered persons. I tried to join with my
>>>>>>> Java.net account emmanuelbernard but I am no authorized.
>>>>>>>
>>>>>>> On Fri 2012-10-19 7:52, Bill Burke wrote:
>>>>>>>> BV 1.0 exists and is final, so your argument is invalid. The
>>>>>>>> proposed integration is simple, only for Java EE environments, and
>>>>>>>> is not BV version dependent.
>>>>>>>
>>>>>>> JAX-RS validation will require method validation which is a feature
>>>>>>> coming in BV 1.1.
>>>>>>>
>>>>>>
>>>>>> Ok then, I have no counterargument to Marke...Sorry Emmanuel.
>>>>>> JAX-RS 1.1 came out same time as Java EE 6 to address integration
>>>>>> issues. Marek seems to want to do the same thing. You ok with that
>>>>>> Emmanuel?
>>>>>
>>>>> Are you telling me that:
>>>>>
>>>>> - JAX-RS 2.0 comes out very early next year
>>>>> - JAX-RS 2.1 or 2.0.MR comes out before Java EE 7 is finalized and Java
>>>>> EE 7 when coming out will for sure include JAX-RS 2.0MR
>>>
>>> NO. It is not likely that JAX-RS 2.0MR will be out before Java EE 7, unless of course Java EE 7 does not slip for another 6 moths, which is unlikely. That is not the plan.
>>>
>>
>> Explain to me why we cannot reference BV 1.1 in JAx-RS 2.0 then? The suggested added wording for JAX-RS + BV integration is so general and high-level, IMO, you don't need a finished BV spec. Or, in other words, there shouldn't be anything added to BV that would hinder the additional wording.
>
> I believe I explained myself earlier. Again, any vague formulations should not get into the spec. We either can depend on BV in specific version minimum version that has all the facilities to satisfy the proposed validation mechanism or we cannot. In this case we cannot. Specifically, we can do the method parameter validation without BV 1.1.
>
>>
>> IMO, an integrated platform is just as important as an independent spec.
>
> +1. That's why we tried very hard to contact BV folks (for 2 months!) in the summer to find out what is their schedule. We tried convince them to release early and notified them about the risk of not delivering early. We also slipped our schedule as much as we could and postponed the JAX-RS PR submission for over a month while waiting for their decision in hope that BV will commit to an earlier release. Unfortunately, that didn't happen. So I refuse any statements that would indicate that we would not care about the integrated Java EE platform. Try to convince BV spec lead to release in the beginning of Feb and we can still re-introduce BV support.
>

Then, IMO, a 2.0MR should be done to coincide with EE7 release. The
thing is, the suggested addition to JAX-RS spec will still be the same
before and after BV 1.1 is released.

Bill

-- 
Bill Burke
JBoss, a division of Red Hat
http://bill.burkecentral.com