Update quarkus-http, create test coverage for Undertow session context events

[manovotn]

The idea is to enable @SessionScoped beans to observe its own @Initialized and @BeforeDestroyed events (which they should be able to according to the specification).

The fix itself is in quarkus-http, see quarkusio/quarkus-http#174

Fixes #46363

[quarkus-bot]

Status for workflow Quarkus CI

This is the status report for running Quarkus CI on commit d188f5c.

✅ 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.

[Ladicek]

LGTM, just wondering if we maybe wanna assert that session-scoped beans cannot observe @Destroyed(SessionScoped.class)?

[mkouba]

LGTM, just wondering if we maybe wanna assert that session-scoped beans cannot observe @Destroyed(SessionScoped.class)?

Except that it's not exactly defined what happens if an observer method is invoked and the relevant context is not active 🤔...

[Ladicek]

Except that it's not exactly defined what happens if an observer method is invoked and the relevant context is not active 🤔...

You mean when the event is fired? I'm thinking the observer method should not be called at all. But if it's more complex than what I imagine, then let's leave it :-)

[manovotn]

LGTM, just wondering if we maybe wanna assert that session-scoped beans cannot observe @Destroyed(SessionScoped.class)?

I am not sure we test this for any other scope but TBF it doesn't make sense to me - we always deactivate context and then fire the event. And that is the only sensible order of actions to me so I see such test as completely redundant...but maybe I am missing something 🤷

You mean when the event is fired? I'm thinking the observer method should not be called at all.

I agree, and that is what Weld does too.

[mkouba]

LGTM, just wondering if we maybe wanna assert that session-scoped beans cannot observe @Destroyed(SessionScoped.class)?

I am not sure we test this for any other scope but TBF it doesn't make sense to me - we always deactivate context and then fire the event. And that is the only sensible order of actions to me so I see such test as completely redundant...but maybe I am missing something 🤷

You mean when the event is fired? I'm thinking the observer method should not be called at all.

I agree, and that is what Weld does too.

And ArC does the same since 3.0.0.Final.

[manovotn]

@mkouba do you then agree to keep this PR as is or should I change something?
I am not sure how to interpret your comment :)

[mkouba]

@mkouba do you then agree to keep this PR as is or should I change something? I am not sure how to interpret your comment :)

I do agree.