I just receive a more complete answer, so here it goes:
To be compatible with a spec, an implementation
- Must pass the TCK
- Adhere to the compatibility requirements in the TCK user guide which is typically in chapter 2 of the user guide
- Fully implement all required assertions in the spec regardless of whether there is a current TCK test
The RI typically is used to indicate that the spec as written can be implemented.
HTH,
Marek
> On 19 Oct 2015, at 19:18, Marek Potociar <marek.potociar_at_oracle.com> wrote:
>
> Passing the TCK is what matters (which hopefully also means that you follow the words of the spec).
>
> If the TCK has a hole and does not cover some corner case, then it is advisable to stick to what the spec says and how RI behaves, but IIUC that does not impact the formal spec compliance.
>
> Marek.
>
>> On 17 Oct 2015, at 17:53, Markus KARG <markus_at_headcrashing.eu> wrote:
>>
>> There is a second question. What actually means "compliant to spec": Passing the TCK, following the words of the spec, or behaving like the RI?
>>
>> -----Original Message-----
>> From: Marek Potociar [mailto:marek.potociar_at_oracle.com]
>> Sent: Freitag, 16. Oktober 2015 20:37
>> To: jsr370-experts_at_jax-rs-spec.java.net
>> Subject: Re: [jax-rs-spec users] @Path with regular expression overrides other @Path's with the same URI
>>
>> Here’s what I found:
>>
>> From the Sun/Oracle Java EE side, we historically only add/remove tests to/from the TCK when an MR or JSR has been released for technology. IOW, we only add new tests when we do a new MR or full JSR. At that point, whoever wants to be compatible with new MR or JSR, must pass the new TCK tests.
>>
>> As such, you can be compatible with JAX-RS 2.0 but you may fail to be compatible with JAX-RS 2.1 if new tests are added to the TCK for JAX-RS 2.1 are covering some new areas that were not covered in the older TCK versions.
>>
>> HTH,
>> Marek
>>
>>> On 15 Oct 2015, at 14:23, Marek Potociar <marek.potociar_at_oracle.com> wrote:
>>>
>>> Not sure, need to check.
>>>
>>> Marek
>>>
>>>> On 15 Oct 2015, at 14:20, arjan tijms <arjan.tijms_at_gmail.com> wrote:
>>>>
>>>> Hi,
>>>>
>>>> On Thu, Oct 15, 2015 at 2:11 PM, Marek Potociar
>>>> <marek.potociar_at_oracle.com> wrote:
>>>>> If there is a hole in TCK coverage, it may get fixed later and then you would need to pass the fixed TCK to be compliant.
>>>>
>>>> That's only for a major new version of the spec, isn't it?
>>>>
>>>> If a TCK is updated for some spec, some version, then this not
>>>> automatically invalidates all existing certifications, does it?
>>>>
>>>> Kind regards,
>>>> Arjan Tijms
>>>
>>
>>
>