Issue/23245 breakage

[som-snytt]

[skip community_build]
[skip docs]
[skip test_sbt]
Maybe there's no more skipping...

[hamzaremmal]

Maybe there's no more skipping...

There is no more skipping.

[som-snytt]

Added a bit of output for more stack trace. The test will always fail, to make re-running easier.

My first guess was NPE in clinit. My second guess was a bad "nullable field" getting nulled after lazy val init; perhaps the outer pointer itself, which was used only for the lazy init. But the outer is not a class param in the example; its rhs is just a static module. Maybe the trait init? but that is empty. Maybe a class isn't initialized? Maybe a missing inner class, except the module is top level.

I could not replicate locally (WSL) but interacting with the test rig made me want to contribute improvements, as I intended.

I see that the flaky test is not included in the group with -Ylegacy-lazy-vals, which was my fourth guess.

[som-snytt]

Woo-hoo!

Output from 'tests/run/i23245a' did not match check file. Actual output:
java.lang.NullPointerException: Cannot invoke "java.io.PrintStream.println(String)" because "java.lang.System.err" is null
	at logadapter.LogAdapter.info(api.scala:23)
	at Test$.main(test_2.scala:3)
	at Test.main(test_2.scala)
	at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
	at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.base/java.lang.reflect.Method.invoke(Method.java:569)
	at dotty.tools.vulpix.ChildJVMMain.runMain(ChildJVMMain.java:40)
	at dotty.tools.vulpix.ChildJVMMain.main(ChildJVMMain.java:47)

Oh wait, does that make any sense? I'll look at it later.