mirror of
https://github.com/logos-co/logos-logoscore-cli.git
synced 2026-08-27 10:51:15 +00:00
modules_state lands in modules/ alongside capability_module, in BOTH the dev and the portable bundle. Missing the portable one would have shipped it in dev builds and silently not in distributed ones -- a variant split that stays invisible until someone tests the wrong artifact. CORRECTS A COMMENT THAT WAS STOPPING THIS. The bundle note said "its doc-tests assert the exact module list, so anything extra there would change what it reports". The second half is true; the first half is not, and it is the half that reads as a prohibition. Every assertion in doctests/ is expect_contains: 107 of them, zero exact-match, one expect_not_contains about a REMOVED package. Adding a bundled module breaks none of them. The comment now says what is true -- the split keeps logoscore's default surface deliberate -- and states outright that the assertions are containment, so the next person does not have to re-derive it. logosctl-modules-bundle.test.yaml gets the same treatment: four modules rather than three, and a note at both assertion sites that containment is deliberate. On flake.lock: +8432 lines net, but the diff reads as ~65k/57k because this lock carries 14813 nodes and adding one input renumbers the _N-suffixed ones. The root input list differs by exactly one entry. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>