The catalog half is read through mesh-delivery's source as well as mesh-catalog's verbs, and a runner with an unread
seat list admitted only the latter. The bootstrapped copy is mesh-delivery main 21eaa55b's contract file, byte for
byte; the tests' seat list and the captured beside copy take the same file.
Builds: design 47 G lead answers A-362, A-353
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
Main reverted G's half merge (#189, reverted by 8afc3eb in #191) and then took the reader release's S1-S5, which
copied G's code. G's form holds where they differ (process_bus.go, seats.go, state.go and their tests); the
release's kept names stay so main's callers keep building: SeatViewPrefix beside SeatReadPrefix, and QueueInvoke's
empty strings on not-ok (with its test TestQueueInvoke). The tree differs from G's head f30dee54 only by those and
the release's two tests (objects cap option, seat states and stream kinds).
Builds: design 47 G (the milestone's sdk part, as at f30dee54)
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
mesh/check being carried to 8afc3ebdf371 on 93e0a7b50d90 | from 27da949140c6
The group feat/47-G-join11 merged 3 of 17 members on 2026-11-13; G's mesh-sdk alone on main breaks 25 trunk
modules. The operator chose to put the order right: a reader release first, then G whole. This restores main
to 8fbba810 exactly.
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
HoldersIn is the one holder function a part gives seatlist.Part.Holders: the modules holding a seat on
any node, from the seat-holders view as the part last heard it (a views follower's or a watched state's
values), so a callers entry `seat:<seat>` is resolved with no call per checked call.
Builds: design 47 G A-295, A-210
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
A tool's callers list named modules ("node-tools", "mesh-healers"). A-295: an entry `seat:<seat>`
allows a module caller whose module holds that seat, as a holder function the part is given answers
it from what it last heard of the assigner's seat-holders view, so no call is made per check.
VerbCallers.AllowedWith and seatlist.RefusalWith/ToolsWith take the function; Allowed, Refusal and
Tools keep their form (no seat entry allows anyone without one); a seatlist.Part is given it by
Holders.
Builds: design 47 G A-210, A-295
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
The definition of a seat on the list a part follows, read through the part's reader at each call: none until the list
is read, none for a seat the list does not hold. Every holder passes it instead of writing its own.
Builds: design 47 G A-278, A-186, A-179
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
Four delivery parts each wrote the same wiring of the seat-list reader
(start from the bootstrap, give the catalogue what it holds, read again
when events go live, two tools over a reader set later). seatlist.Part
is that wiring, its tools named by the part's prefix and its again
answered to the callers the part gives (ADR 0636 §7), so W4-W8 delete
their copies.
Builds: design 47 G §204, §246
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
mesh/check checking 2f156e7f3050 on 5b63c67ce852 | novox/mesh-sdk#184 in group fix/1966-seat-list-unread
A node-engine started in the five seconds its holder re-served read no seat
list and nothing asked again: the reader read only on the seat's next event.
An ask the bus delivered to no one is now asked again within busretry's bound;
an answer, a refusal or a timeout never is (issue 1966).
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ