Made progress in porting it to CDI 1.2

- Also used JSONP 1.0 (instead of 1.1)
- Used Producers instead of beans for each @Claim @Inject (as described
here https://github.com/eclipse/microprofile-jwt-auth/issues/32)

Still a bit hacky and not all producers created, but my demo app with
injection of JsonWebToken, JsonString Claim and String Claim are already
working.

Rudy



On 23 April 2018 at 16:41, Romain Manni-Bucau <[email protected]> wrote:

> This is what does the extension, the not CDI 1 features used are only
> configurators to override the @Claim model (this one is only supported for
> OWB >= 2.0.5 even in the spec since more - at least 2.0.0) + to create
> beans (this one is easy to solve adding a custom Bean+PassivationCapable
> impl)
>
>
> Romain Manni-Bucau
> @rmannibucau <https://twitter.com/rmannibucau> |  Blog
> <https://rmannibucau.metawerx.net/> | Old Blog
> <http://rmannibucau.wordpress.com> | Github
> <https://github.com/rmannibucau> | LinkedIn
> <https://www.linkedin.com/in/rmannibucau> | Book
> <https://www.packtpub.com/application-development/java-ee-8-high-performance>
>
> 2018-04-23 16:38 GMT+02:00 John D. Ament <[email protected]>:
>
>> I can look at the code later but what I had to do before is capture all
>> of the claim injection points and provide specific producers for each.
>>
>> On Mon, Apr 23, 2018, 10:35 AM Mark Struberg <[email protected]> wrote:
>>
>>> Qualifiers are per CDI spec not AnnotatedTypes.
>>> So if we rely on this then it's not spec compliant anyway.
>>>
>>> LieGrue,
>>> strub
>>>
>>> > Am 23.04.2018 um 14:30 schrieb Romain Manni-Bucau <
>>> [email protected]>:
>>> >
>>> > the extension modifies @Claim to remove @NonBinding. This requires the
>>> impl to support to read qualifiers as AnnotatedType and only OWB 2.0.5
>>> supports it in OWB series ATM
>>> >
>>> >
>>> > Romain Manni-Bucau
>>> > @rmannibucau |  Blog | Old Blog | Github | LinkedIn | Book
>>> >
>>> > 2018-04-23 14:18 GMT+02:00 John D. Ament <[email protected]>:
>>> > Whats the qualifier issue you're referring to?
>>> >
>>> > On Mon, Apr 23, 2018, 8:05 AM Romain Manni-Bucau <
>>> [email protected]> wrote:
>>> > Same here, I just doubt we have an owb impl supporting the qualifier
>>> model change today so we can stay on OWB 2.0.5 or need to backport it to
>>> 1.x as well (which can likely be the case as well but can need to be done
>>> in parallel).
>>> >
>>> >
>>> > Romain Manni-Bucau
>>> > @rmannibucau |  Blog | Old Blog | Github | LinkedIn | Book
>>> >
>>> > 2018-04-23 13:17 GMT+02:00 John D. Ament <[email protected]>:
>>> > If you already have a PR submitted even better.  We should accept it.
>>> >
>>> > On Mon, Apr 23, 2018, 7:07 AM Rudy De Busscher <[email protected]>
>>> wrote:
>>> > Not that hard, except maybe for the NonBinding thing which is removed
>>> from @Claim.
>>> >
>>> > All the rest was done in 20 minutes or so.
>>> >
>>> > On 23 April 2018 at 13:03, Jean-Louis MONTEIRO <[email protected]>
>>> wrote:
>>> > Overall same view here.
>>> > How hard is it to make it 1.2 compliant?
>>> >
>>> >
>>> > Le lun. 23 avr. 2018 à 12:25, John D. Ament <[email protected]> a
>>> écrit :
>>> > MP has made it very clear they don't care about portable libraries,
>>> and only care about the vendor provided solutions.  The requirement is that
>>> vendors provide a CDI 1.2 runtime to use.  Liberty provides a way to switch
>>> between them (1.2, 2.0).  I think Swarm may have moved to 2.0; not sure.
>>> >
>>> > I think Safeguard also compiles against CDI 2.0, but I don't think I'm
>>> using any 2.0 features in it so it may run properly against 1.2.
>>> >
>>> > Personally, if we have a user who wants it for 1.2, and the effort is
>>> minimal we should appease that user to help build out the community.
>>> >
>>> > John
>>> >
>>> >
>>> > On Mon, Apr 23, 2018 at 2:17 AM Romain Manni-Bucau <
>>> [email protected]> wrote:
>>> > Hi guys,
>>> >
>>> > current codebase uses cdi 2.0 which means it can be used on tomee,
>>> meecrowave,  openwebbeans etc...
>>> >
>>> > Rudy opened https://issues.apache.org/jira/browse/GERONIMO-6604 to
>>> move it to cdi 1.2 - BTW "Microprofile depends on CDI 1.2, so using 2.0 is
>>> wrong." is wrong since some years you can always use a version *>=* of the
>>> minimum requirement for spec impls.
>>> > Technically I don't see a strong need to do it but I'd like to get
>>> your feeling about it to know what we do of the issue.
>>> >
>>> >
>>> > Romain Manni-Bucau
>>> > @rmannibucau |  Blog | Old Blog | Github | LinkedIn | Book
>>> >
>>> >
>>> >
>>>
>>>
>

Reply via email to