#9 emits rollout-error for events that are not errors:
tailResumeRolloutFile → switchTo: a successful committed-tail switch to a detected fork.isResumeForkCandidate: an expected rejection of a same-cwd sibling agent's rollout.tailNewRolloutFile: a multi-rollout startup ambiguity notice.
Routing operational diagnostics through rollout-error makes that channel noisy and conflates "the fix worked / behaved as designed" with "something failed."
Proposal: add a dedicated non-error event to CodexHeadlessEvents (e.g. rollout-diagnostic with a small typed payload, or specific rollout-tail-switched / rollout-candidate-rejected events) and route these notices through it. Update downstream consumers and the agent-code IPC bridge accordingly so they surface as diagnostics, not errors.
Refs #9.
#9emitsrollout-errorfor events that are not errors:tailResumeRolloutFile→switchTo: a successful committed-tail switch to a detected fork.isResumeForkCandidate: an expected rejection of a same-cwd sibling agent's rollout.tailNewRolloutFile: a multi-rollout startup ambiguity notice.Routing operational diagnostics through
rollout-errormakes that channel noisy and conflates "the fix worked / behaved as designed" with "something failed."Proposal: add a dedicated non-error event to
CodexHeadlessEvents(e.g.rollout-diagnosticwith a small typed payload, or specificrollout-tail-switched/rollout-candidate-rejectedevents) and route these notices through it. Update downstream consumers and the agent-code IPC bridge accordingly so they surface as diagnostics, not errors.Refs #9.