From 1e309472dac1577eaeadb73a7489812ee9e60ec4 Mon Sep 17 00:00:00 2001 From: jochen Date: Thu, 19 Nov 2026 15:21:32 +0100 Subject: [PATCH] Answer A-489: an ask a build module ends itself is acknowledged once its end is written, never terminated Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy --- 03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md b/03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md index 7db4c9d4..7041cf5e 100644 --- a/03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md +++ b/03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md @@ -2026,3 +2026,10 @@ the principle that settles it, cited as "design 47 G A-n". A-488a holds for every build module whose runtime follows the seat list (`mesh-build-go`, `mesh-build-node`, `mesh-build-image`), and each also invokes the verb that reader calls, `seat:mesh-catalog.seat-definitions`, so its manifest names all the reader uses. (From the review of A-488 and A-488a, round 1.) +- **A-489** (principles 3 and 2) An ask a build module's puller ends itself, refused unread, refused for its asker, ended + poison or out of its deliveries, is acknowledged once its end is written, never terminated: on a work-queue stream a + terminated message is never delivered again yet stays, and the seat's queue then shows it waiting and the builder waits + on it as put, a stall nobody is told of. The written end is the one record of what became of it. A give-back (the + module cannot judge or run it now) still leaves it unacknowledged for the bus to give again, and a give-back is no + delivery that counts toward poison. (Live 2026-11-19: 56 messages held on `SEAT_NODE_BUILD_NODE`, ack floor 0, the + sdk's build waiting on three of them since 13:24.)