nothing detects "another roar run/roar build may currently be in-flight in the same active session" when a bare roar register publishes it (e.g. terminal A running a long job, terminal B running bare roar register publishes a prior completed job in the same session with no warning that another run is still active). I looked for a cheap existing signal to surface in the new prompt and didn't find one: job rows are only written to the DB after a job completes (ExecutionJobRecorder.record() calls record_job() only once both duration and exit code are known — nothing is written at job start), and Job.status's "running" state (roar/core/models/job.py) is defined in the schema but nothing in the codebase ever sets it — it's effectively vestigial.
nothing detects "another
roar run/roar buildmay currently be in-flight in the same active session" when a bareroar registerpublishes it (e.g. terminal A running a long job, terminal B running bareroar registerpublishes a prior completed job in the same session with no warning that another run is still active). I looked for a cheap existing signal to surface in the new prompt and didn't find one: job rows are only written to the DB after a job completes (ExecutionJobRecorder.record()callsrecord_job()only once both duration and exit code are known — nothing is written at job start), andJob.status's"running"state (roar/core/models/job.py) is defined in the schema but nothing in the codebase ever sets it — it's effectively vestigial.