Bump RESTEasy to 4.5.8.Final

[ronsigal]

This is #12232 but with RESTEasy version 4.5.8.Final. I don't have permission to edit #12232 directly.

Fixes: #10133

[ronsigal]

"Initial JDK 11 Build' failed because RESTEasy 4.5.8.Final hasn't made it to https://repo.maven.apache.org/maven2.

[ronsigal]

RESTEasy 4.5.8.Final is now on https://repo.maven.apache.org/maven2. Could someone restart the tests?

[kenfinnigan]

done

[gastaldi]

That's awesome, but can you squash all these commits and add a Co-authored-by in the single commit? It seems to me that the upgrade only works when all these commits are applied, hence the request 😉

[ronsigal]

Hi @gastaldi,

re: "squash all these commits and add a Co-authored-by in the single commit?"

Done.

[gastaldi]

@ronsigal excellent, but I thought the Co-Authored-By would mention @asoldano instead ?

[ronsigal]

re: "would mention @asoldano instead"

Ah, I thought maybe there was some standard requiring a Co-authored-by in any case. I doubt that @asoldano would be offended, but, sure, I can update it. ;-)

[ronsigal]

Done.

[gastaldi]

@ronsigal thank you, yeah, it's just to give him credit for his commits ;)

[ronsigal]

re: "it's just to give him credit for his commits"

Cool. Maybe I can leverage this into a larger bonus. ;-)

[kenfinnigan]

Is it also worth replacing https://github.com/quarkusio/quarkus/blob/master/extensions/rest-client/runtime/src/main/java/io/quarkus/restclient/runtime/RestClientRecorder.java#L64-L69 with the static call to ResteasyClientBuilderImpl that was added to RESTEasy?

[ronsigal]

Hi @kenfinnigan,

re: "Is it also worth replacing"

Sorry, I'm not following ... . https://github.com/quarkusio/quarkus/blob/master/extensions/rest-client/runtime/src/main/java/io/quarkus/restclient/runtime/RestClientRecorder.java#L64-L69 doesn't reference ResteasyClientBuilderImpl. What am I missing?

[kenfinnigan]

Sorry @ronsigal

Those lines were added because we didn't have the ability to set the static, so I think they can be removed and replaced with a call to the static, similar to calling the RestClientBuilderImpl static to set the provider factory

[ronsigal]

Oh, man, I'm still not getting it. I blame 2020 because nothing else makes sense either. :(

[kenfinnigan]

No worries.

Instead of the lines I highlighted that peek into the RESTEasy provider factory and set it, just call ResteasyClientBuilderImpl.setProviderFactory(clientProviderFactory);

[ronsigal]

Mmmm, but ResteasyProviderFactory.instance and RestClientBuilderImpl.PROVIDER_FACTORY are two different things ... . I guess I still don't know the resteasy-client-microprofile code as well as I should.

[kenfinnigan]

They most definitely are different, but hijacking the RESTEasy provider factory for everything was the only way we could hack it so that the ResteasyClientBuilderImpl ended up with the provider factory we wanted it to.

With the static, we don't need those shenanigans and we can just set the static.

Sound ok?

[ronsigal]

Ahhhhhh. Yes, ok. ;-)

Would you like me to add that change to this PR?

[asoldano]

Given this is passing CI, I would deal with what @kenfinnigan is asking for (enhancements relying on the additions in 4.5.8.Final) in another PR.

[ronsigal]

@kenfinnigan , I've created issue #12325 "Remove unnecesary code from RestClientRecorder" and pull request #12326 to remove that bit of code.

[gsmet]

@kenfinnigan any chance you could address the static thing? You're probably the one with the best understanding of this.

[kenfinnigan]

@ronsigal has provided a PR in #12326