Quarkus REST - reuse CDI request context if it exists

[manovotn]

Related to: #40307

Allows reuse of existing CDI req. context if present and monitors if req. context was activated from within AbstractResteasyReactiveContext. Only performs cleanup if the context was initialized here.

[manovotn]

I'll turn this into a regular PR to get full test coverage.

@michalvavrik feel free to add the tests you mentioned straight into this PR.

[michalvavrik]

Hello @manovotn , thank you for prompt debugging, feedback and fix. I have prepared reproducer inside OpenAPI IT module (because that is where Quarkus REST and Reactive Routes already are, no better place I could find).

• 125453e

I cannot push it to your PR, please just pickup the commit, refactor it to your liking and use it.

[mkouba]

I do agree with the idea (and it's consistent with what other extensions are doing; e.g. reactive routes) but I don't feel qualified to comment on the implementation 🤷 .

[manovotn]

Hello @manovotn , thank you for prompt debugging, feedback and fix. I have prepared reproducer inside OpenAPI IT module (because that is where Quarkus REST and Reactive Routes already are, no better place I could find).

* [125453e](https://github.com/quarkusio/quarkus/commit/125453ecd23905b5da2664c82b987289f0a1f0af)

I cannot push it to your PR, please just pickup the commit, refactor it to your liking and use it.

Ah, I didn't realize.
Cherry picked the commit into this PR so we should be good once CI gives us green light 👍

[manovotn]

Failures are related. Those are SR CP tests asserting destruction of beans in a no-propagation scenario that now doesn't happen correctly.
I will debug tomorrow but my prime suspect is that I made the destruction logic conditional whereas before we did it at all times which seemed weird but this might have been the reason.

[quarkus-bot]

Status for workflow Quarkus CI

This is the status report for running Quarkus CI on commit 864be71.

✅ The latest workflow run for the pull request has completed successfully.

It should be safe to merge provided you have a look at the other checks in the summary.

You can consult the Develocity build scans.

[michalvavrik]

@manovotn please change PR desc so that it closes #40307, thanks.

[manovotn]

@manovotn please change PR desc so that it closes #40307, thanks.

It is already linked to that issue and will autoclose it on merge (see the Development section of the PR)

[gsmet]

My understanding is that this should be backported as far back as 3.8 as the original report is that 3.8.4 introduces the issue?

[gsmet]

I added the backport but please I'd like someone to confirm :).

[geoand]

Yeah, that makes sense