From e671d865d3ee7c41ae1a5a6fa702cc47f3b0f846 Mon Sep 17 00:00:00 2001 From: jochen Date: Tue, 17 Nov 2026 13:08:08 +0100 Subject: [PATCH 1/2] Answer A-470u: a reader that leaves or is replaced has its wait ended --- 03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md | 3 +++ 1 file changed, 3 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 09171f8d..ab61f5fc 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 @@ -1804,3 +1804,6 @@ the principle that settles it, cited as "design 47 G A-n". - **A-470t** (G-24 review r1, runner; A-470q; principle 7) The runner publishes its step events with no message id, so the bus's deduplication drops none of its ends said again: A-470q's "the same holds for the runner" is met as built, and no id is added; test 109 proves it by reading the ends said again from the stream through a durable consumer. +- **A-470u** (G-24 review r1, assigner; A-469y; principle 3) A reader that leaves a run not yet taken, or whose place a + rerun takes, has its `wait-held` ended at once by `wait-ended` with how `superseded`; no wait is left held for a reader + the run no longer names. From aa52fa5345e3c9cd944c88764d8cd8bdb1b3d493 Mon Sep 17 00:00:00 2001 From: jochen Date: Tue, 17 Nov 2026 13:08:28 +0100 Subject: [PATCH 2/2] Answer A-470v: the search blames every refused module, not only the first --- 03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md | 4 ++++ 1 file changed, 4 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 ab61f5fc..d2b42236 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 @@ -1807,3 +1807,7 @@ the principle that settles it, cited as "design 47 G A-n". - **A-470u** (G-24 review r1, assigner; A-469y; principle 3) A reader that leaves a run not yet taken, or whose place a rerun takes, has its `wait-held` ended at once by `wait-ended` with how `superseded`; no wait is left held for a reader the run no longer names. +- **A-470v** (G-24 review r2, composer; §721, §722, A-470c; principle 3) A hold that changes the problem but leaves the + machine not composing keeps that module held and blamed, and the search goes on against the retry's problem within + the same `blame_tries`, until the machine composes or no hold changes it: every module the candidate's composition + refuses is blamed, never only the first.