Design 47 H2: H-A-468 second addendum, SeatQueue is the one derivation

This commit is contained in:
2026-11-22 11:35:19 +01:00
parent dd758d6e5f
commit e9192d7c7e
@@ -840,3 +840,8 @@ The lead's one-line answers to the questions H2's develop and review raise (impl
- **H-A-468 (addendum, 11:45)** The grants part grants each holder of a seat with an accept not pulled by kind the
subscribe of the bus's advisories CREATED and DELETED for that seat's shared worker, in the same change, so the
holder's bind and re-bind hold under scoped grants and not only while every account may do everything. P-failure-loud.
- **H-A-468 (second addendum, 11:50)** (3) is met without a new SDK: mesh-sdk's buscompose.SeatQueue is the one
derivation of a seat's queue and its worker; the grants part creates the worker from the configuration SeatQueue
returns, unchanged (no conversion of its own), and the builder's whole test makes it from SeatQueue the same way, so
the two cannot differ; a test in the grants part proves the created worker equals SeatQueue's. No SDK change tonight
(one would rebuild every importer again). P3, rules-are-weighed (said).