From 0b3974d7739029d133a6526773480e3698a19602 Mon Sep 17 00:00:00 2001 From: jochen Date: Thu, 19 Nov 2026 21:32:27 +0100 Subject: [PATCH] Answer A-490j: a holder takes a publish while its build runs when the reservation names no digest yet Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy --- 03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md | 8 ++++++++ 1 file changed, 8 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 e02f53a8..435a2088 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 @@ -2096,3 +2096,11 @@ the principle that settles it, cited as "design 47 G A-n". transport fault until it is answered and the indexer writes `call-unanswered` five minutes after its first unanswered try, as ADR 0668 has; the indexer holds no timer of its own, and its guard against timers stands. Each module's floor read runs on its own, so one unanswered read holds no other module's floor. +- **A-490j** (§369, §381, A-415, A-490; principles 3 and 7) A package holder takes a build module's one-part publish + made while it builds when the builder holds the name at that version as reserved and the reserved artifact names + that digest or names none yet: the builder records a build's digests only when its build ends (§369, A-415), so + "digest reserved" (§381) holds from that record on, and no verb or event is added to carry a digest earlier. A + second publish of the same name and version with another digest is refused as the forge holds the first (A-490); a + build whose recorded digest differs from what the holder holds at its version fails loudly, version taken, and is + never marked released. (Live 2026-11-19 21:09 and 21:19: every publish refused "reserved for another digest", the + reservation's digests empty until the build ends.) -- 2.54.0