[Decision] service-automation: should a refusing end INSIDE a structured region propagate out and terminate the run, or stay a loud refusal at the region boundary? #18112

Raised as an open question by the domain:services dev that delivered #15788 (lane 2 of the #14945 ruling 2′, PR #18109), and filed here by the seat so the decision is not lost. ⛔ The dev does not file; it reports.

The question

The #14945 ruling says nothing about regions. An end node carrying outcome: 'refused' inside a loop body or a try_catch region is therefore undecided.

What shipped, and why it is an interim rather than an answer

Option A shipped: a refusing end inside a region is a loud failure at the region boundary — the same line where runRegion already refuses a durable pause.

The dev's reasoning, which this seat accepted:

• A refusal terminates the run, and a region body cannot end one.
• Left to propagate today, the signal unwinds into try_catch's own catch arm, which reads every throw as the try region FAILING. So the author's refusal would run the error path and the run would still record completed — i.e. the old silence with an extra step.
• ⛔ Nothing an author had is narrowed: an end inside a region was a no-op whatever its outcome, so this shape has never once been honoured.

⭐ It is the interim that 「prefer failing to falling back」 decides, and it forecloses a silent wrong answer. The loud message is the discoverable place a reader lands.

Option B — the real design question

Propagate the signal out of every container, so a refusal inside a loop or try region terminates the whole run. That needs:

1. try_catch (and a check of parallel / map) taught to rethrow the refusal rather than catch it — a refusal is a successful evaluation that says no, ⛔ not a failure, so a catch arm is the wrong receiver.
2. A decision about the in-flight siblings of a parallel gateway — ⚠️ the same scope ADR-0019 M1 explicitly left out for the durable pause. Whatever is decided here should be decided for both, or the two signals diverge inside the same containers.

⇒ This is a ruling-seat decision, ⛔ not a lane-2 implementation detail, which is why PR #18109 did not ride it in.

Precedent to weigh

The durable pause (FlowSuspendSignal) already refuses at the same boundary, and ADR-0019 M1 left the parallel-sibling question open there too. Option A keeps the two signals consistent; option B fixes both or deliberately splits them.

Related: #14945 · #15788 · PR #18109 · ADR-0019 M1

⛔ domain:*, type and priority are the triage seat's; this card is filed ungraded. The DECISION belongs to the seat that owns the #14945 ruling.

Generated by Claude Code