Rand, 2026-09-23: "we should be using sc-lint-boundary" (upstream), not the in-tree copy.
State today:
just lint runs the vendored crates/sc-lint-boundary via cargo run; it is the only parser that accepts our boundaries/**/*.toml.- Released sc-lint 0.4.0 / 0.5.0 reject
public.trait (54 files). sc-lint develop (e2477c38857537c8bcd34cf3873ee9726ef8d12a, 0.6.0 unreleased) accepts trait but rejects no_in_repo_implementation (boundaries/atm-core/atm-protocol.toml:39); further unknown fields may follow.
- Found because the sc-publish kit's Release Preflight ran
sc-lint lint sc-boundary as a gate (sc-publish#67 regression, removed by sc-publish#115). Publication no longer depends on this.
Work: diff the vendored crate against upstream sc-lint-boundary, upstream the inventory fields we rely on (or drop them), get an sc-lint release that parses our inventory, switch .just/lint_sc_boundary.py to the released binary, delete crates/sc-lint-boundary (and crates/sc-lint-directives if it is the same story). Post-1.6.1.
🤖 Generated with Claude Code
Rand, 2026-09-23: "we should be using sc-lint-boundary" (upstream), not the in-tree copy.
State today:
just lintruns the vendoredcrates/sc-lint-boundaryviacargo run; it is the only parser that accepts ourboundaries/**/*.toml.public.trait(54 files). sc-lint develop (e2477c38857537c8bcd34cf3873ee9726ef8d12a, 0.6.0 unreleased) acceptstraitbut rejectsno_in_repo_implementation(boundaries/atm-core/atm-protocol.toml:39); further unknown fields may follow.sc-lint lint sc-boundaryas a gate (sc-publish#67 regression, removed by sc-publish#115). Publication no longer depends on this.Work: diff the vendored crate against upstream sc-lint-boundary, upstream the inventory fields we rely on (or drop them), get an sc-lint release that parses our inventory, switch
.just/lint_sc_boundary.pyto the released binary, deletecrates/sc-lint-boundary(andcrates/sc-lint-directivesif it is the same story). Post-1.6.1.🤖 Generated with Claude Code