From 33d26479854f7d73dd1a7f9b9b126379b85a7e9b Mon Sep 17 00:00:00 2001 From: jochen Date: Sat, 21 Nov 2026 19:15:38 +0100 Subject: [PATCH] Design 47 H: H-A-294, a test that needs tooling its step lacks says NOT RUN, and the need is declared --- 03-DESIGN/01-to-be/47-milestones/H-lead-answers.md | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/03-DESIGN/01-to-be/47-milestones/H-lead-answers.md b/03-DESIGN/01-to-be/47-milestones/H-lead-answers.md index d08c42a7..785e6b36 100644 --- a/03-DESIGN/01-to-be/47-milestones/H-lead-answers.md +++ b/03-DESIGN/01-to-be/47-milestones/H-lead-answers.md @@ -1700,3 +1700,11 @@ Where two preparers disagreed, the chosen one is named. its info), never asking a second page and never refusing. And H-A-291 (1) is read per series: the newest reservation and the newest release are found per module AND series (MAJOR.MINOR), so a standing 0.2 never retires a 0.1.x build. P-failure-loud (an answer that says what it left out), H-A-292. +- **H-A-294 (a test that needs tooling its step lacks says NOT RUN, and the need is declared, 19:15)** Met live: the + forge module's whole-seat test (added 20 Nov, never passed in a pipeline) needs the module's npm packages installed + and fails every Go integration step, which fetches the tree without them; it sank every pipeline carrying the + forge. In H1 the test checks its needs first (node, npm, the installed SDK package) and, when one is missing, ends + NOT RUN with the reason and where it was proven (the build module's own precedent), never a failure; a NOT RUN is + said in the step's summary. Owed, H2: the Go integration level receives a module's npm tree installed by the Node + build module behind a declared need (check-needs node-modules), so the test runs on the mesh. Also owed: a read tool + for a run's test output (node-build-go.output answers nothing through the mesh MCP server). P-failure-loud, P3, H-A-267.