Strengthen tests for @QuarkusMain

[holly-cummins]

Our test coverage for @QuarkusMain is all in the picocli-native project. Although it covers more complex scenarios relating to picocli commands, it doesn't cover the specific scenario in https://quarkus.io/guides/command-mode-reference#mocking. Since it's documented, we want it to keep working. #28997

I've slightly updated the wording in the docs, since I needed a bit of help to understand the intent of the example test, and I've added an integration test which exercises the scenario of using a test profile to adjust injected beans.

[holly-cummins]

My initial assumption is that the ClasspathInGraalITCase failure is unrelated, since I didn't touch that test, but it could be a port conflict or some other cleanup. I'll rebase to see if that changes things.

[github-actions]

🙈 The PR is closed and the preview is expired.

[holly-cummins]

The failure's pretty reproducible, so I assume there's some cross-talk:

 Error:  io.quarkus.it.extension.ClasspathInGraalITCase.testRuntimeInitMainResourceNoDuplicate -- Time elapsed: 0.011 s <<< ERROR!
java.lang.RuntimeException: java.lang.RuntimeException: Unable to successfully launch process '2780'. Exit code is: '0'.
	at io.quarkus.test.junit.QuarkusIntegrationTestExtension.throwBootFailureException(QuarkusIntegrationTestExtension.java:373)
	at io.quarkus.test.junit.QuarkusIntegrationTestExtension.beforeEach(QuarkusIntegrationTestExtension.java:115)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1511)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1511)
Caused by: java.lang.RuntimeException: Unable to successfully launch process '2780'. Exit code is: '0'.
	at io.quarkus.test.common.LauncherUtil.ensureProcessIsAlive(LauncherUtil.java:119)
	at io.quarkus.test.common.LauncherUtil.waitForCapturedListeningData(LauncherUtil.java:82)
	at io.quarkus.test.common.DefaultNativeImageLauncher.start(DefaultNativeImageLauncher.java:110)
	at io.quarkus.test.junit.IntegrationTestUtil.startLauncher(IntegrationTestUtil.java:157)
	at io.quarkus.test.junit.QuarkusIntegrationTestExtension.doProcessStart(QuarkusIntegrationTestExtension.java:301)
	at io.quarkus.test.junit.QuarkusIntegrationTestExtension.ensureStarted(QuarkusIntegrationTestExtension.java:168)
	at io.quarkus.test.junit.QuarkusIntegrationTestExtension.beforeAll(QuarkusIntegrationTestExtension.java:128)
	... 1 more

What's awesome is that I can reproduce it locally, and I can also reproduce it locally if I revert this changeset. Not sure how to debug that one.

[holly-cummins]

@Ladicek very kindly debugged this for me, since the problem test has always failed for me locally. He pointed out that I was mixing a @QuarkusMain and @QuarkusIntegrationTest in the same module. The @QuarkusIntegrationTest runs the Quarkus application, and since there's @QuarkusMain, that's what gets executed, and it terminates early, causing unhappiness for the @QuarkusIntegrationTest.

One solution would be to split the tests out into their own modules but I was reluctant to go super-granular with the modules, so I've applied the second fix @Ladicek suggested, which is just to take QuarkusMain off the default path by adding a qualifier.

That fixes CI, so I will squash.

[quarkus-bot]

Status for workflow Quarkus CI

This is the status report for running Quarkus CI on commit 6ef81f9.

✅ 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.

Warning

There are other workflow runs running, you probably need to wait for their status before merging.

You can consult the Develocity build scans.

[quarkus-bot]

Status for workflow Quarkus Documentation CI

This is the status report for running Quarkus Documentation CI on commit 6ef81f9.

✅ 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.

[gsmet]

Thanks @holly-cummins and @Ladicek !

[holly-cummins]

Good news! This test is already failing!

I mean, it's bad news that it's failing, but it's good news because it's failing on my classloading-rewrite branch, which means the coverage gap was real, and a regression got caught before making it to the field.