The @QuarkusTest suite for casehub-life is fully green. 286 tests run, 0 fail, 0 skip — including all eight integration test classes that were blocked while engine#536 and engine#537 were open.

life#36 was on the backlog: LifeCaseResourceTest failing with a NoSuchMethodError on CaseHubRuntime.signal(UUID, String, Object). The issue was filed because the engine SNAPSHOT had changed the method signature. The obvious next step was to update the call site in LifeCaseService.

I brought Claude in to work through it. Before writing anything, I had Claude check the installed SNAPSHOT jar directly with javap. The method was there — same signature LifeCaseService was already using. We ran the test. It passed first time, no changes needed. The SNAPSHOT had moved again between when the issue was filed and when we picked it up.

This happens with SNAPSHOT dependencies. The artifact in your local Maven cache reflects the state at the last successful update, not the current one. Between filing an issue and picking it up, the upstream may already have changed. The right first move for a NoSuchMethodError against a SNAPSHOT is to check the jar before writing any fix code.

We cleaned up the aftermath: removed the “@QuarkusTest status (2026-06-18)” warning row from CLAUDE.md, and updated ARC42STORIES.MD’s improvement refs to mark both engine issues resolved. The status note had been accurate — it documented a real block — but once the upstream fixes landed it was just misleading noise waiting to ambush someone reading the file cold.


<
Previous Post
The Accessor That Was Lying
>
Next Post
Teaching an Agent What It Cannot Do