# Lab — ch09-runtime

Chapter: Relurpic Internals — Inside Look, Chapter 9 (“Runtime, Sessions, and the Sidecar”). Pin: tag BETA-RC1-FALL2026; captured at commit 1bc57c62029641dcc919c04a8076593794246409 while the tag is pending.

What this lab is: queue a job, watch the runner drain it — with the sidecar built from source, since no release has ever carried it (decision D10). The runner self-submits a scheduled knowledge.refresh job via the spool, claims it, executes it, and completes it; the e2e suite then puts the whole jobs contract on record. No model backend, no sandbox.

# Command card

# Honesty notes

• D10 applies: the runner has never shipped in any release; the lab builds it from source. The seam working is not the release path working — Chapter 1 §1.2 and Chapter 9 §9.4 keep the two separate.
• The job.claimed line appears twice in the transcript (the second is the handler’s own completion log reusing the claim log type with message: "knowledge.refresh complete"); the job.completed record is the store’s terminal state.
• marked_stale: 0 is the empty-substrate fixture’s value; the claim is that the refresh handler reports what it marked, not the number.
• Correlate IDs and timestamps differ per run; compare event kinds, producers, and the claimed→completed ordering.
• On a workspace with an existing relurpify_cfg/, the runner reads it through userconfig like everything else; the lab’s scratch workspace runs on defaults.

# Verification

verify.sh ch09 (from the labs root) replays card items 1–4 and diffs the deterministic outputs (event-kind sequences, test verdicts, exit codes).