Design 47 H: H-A-328, the runner asks the build seat whether a package is published

This commit is contained in:
2026-11-21 21:18:03 +01:00
parent fc31df351a
commit 2226f10086
@@ -1974,3 +1974,11 @@ Where two preparers disagreed, the chosen one is named.
H-A-294: a test's setup bound is read from the step's own bound, not a value in the test; a whole-module test that
needs the bus's time under load says NOT RUN with the reason rather than failing; both fixed in their modules
before the next release that carries them. P-failure-loud, test-budget.
- **H-A-328 (the runner asks the build seat whether a package is published, 21:18)** Amends H-A-326 (1): the runner's
record knows only the publishes it heard, and a package published before the pipeline existed is in no record of
it; the builder owns publishes, so the seat mesh-builds gains a read verb published {module, artifact,
fingerprint} answering {published, version, digest, at} from the builder's reservations and the registries'
record (H-A-274), and the runner asks it once when it writes a mover's wait on a package, placing the mover when
the answer is published; the match is by the package's build fingerprint; the publish event stays the second
path. The group: the contracts repository (the seat's definition) and the delivery repository (the builder serves
it, the runner asks it). P-NATS-native (a call where an answer is needed; the owner answers), P7, ADR 0596.