[campaign] PMPL-1.0 → MPL-2.0 SPDX licence header sweep (~12,000 files)

[hyperpolymath]

Finding

A 2026-05-30 estate-wide grep shows 12,223 files still carry `SPDX-License-Identifier: PMPL-1.0-or-later` (predecessor estate licence) vs 55,245 files on the target `SPDX-License-Identifier: MPL-2.0`. `MPL-1.0` is already at zero — that conversion completed.

Per estate language policy

Per [[feedback_estate_lang_policy_2026_05_25]] memory:

MPL-1.0/PMPL-1.0 → MPL-2.0

PMPL-1.0 is the predecessor estate licence; MPL-2.0 is the target.

Top hotspots by repo

Repo	PMPL-1.0 files	Notes
iser-tools	1,517	Likely all estate-authored
.claude (worktrees?)	1,310	Probably ephemeral worktree shadows
v-ecosystem	1,159	Vendored upstream V-lang — exempt
rescript-ecosystem	882	Vendored upstream ReScript — exempt
.github (workflows)	845	Mostly workflow files
asdf-augmenters	820	Estate-authored
asdf-plugin-collection	706	Estate-authored
docs (various)	495	Documentation
verification-ecosystem	453	Estate-authored
a2ml	325	Estate-authored
k9-svc	293	Estate-authored
affinescript-ecosystem	272	Estate-authored
...	...	...

Standards repo itself: 2,382 PMPL-1.0 vs 350 MPL-2.0 — the canonical-policy repo is itself ~87% on the predecessor licence.

Scope considerations

• Vendored exclusions: v-ecosystem, rescript-ecosystem, hyperpolymath-archive/*, linguist, repos-monorepo — these are upstream forks and should keep their original headers per upstream-fidelity policy.
• Realistic remaining scope: ~8,000 files across ~40 estate-authored repos.
• Bulk-replace tractability: Mechanical sed substitution; per-PR scope = per-repo would be ~40 PRs.

Recommended approach

Phase 1 — standards itself

Convert standards' own files (2,032 estimated after carve-outs). Single large PR or 3-4 smaller PRs split by directory (workflows/, docs/, scripts/, lib/). Establishes the pattern.

Phase 2 — top-10 hotspots

Per-repo PRs for iser-tools, asdf-augmenters, asdf-plugin-collection, verification-ecosystem, a2ml, k9-svc, affinescript-ecosystem, container-stack, etc. Mechanical sed -i 's|PMPL-1.0-or-later|MPL-2.0|' plus header-only MPL-1.0 variants.

Phase 3 — long-tail

Subagent fan-out across remaining repos, paced for rate-limit. ~6-8 hours total.

Out of scope

• Code-level licence changes (just SPDX headers)
• Upstream-vendored content
• Files where PMPL is INTENTIONAL (none currently identified)

Owner decision needed

• (a) Approve Phase 1 (standards itself) now
• (b) Defer — PMPL-1.0-or-later is contractually equivalent to MPL-2.0 in practice; cosmetic
• (c) Trigger full sweep — would consume ~6-8 hours of CI/Actions budget

Refs [[feedback_estate_lang_policy_2026_05_25]] (policy source)

[hyperpolymath]

Status sweep 2026-05-30 — refresh + reality check.

Count refresh

Re-ran inventory against current estate:

Metric	At filing	2026-05-30 (now)	Δ
PMPL-1.0 files	~12,223	~21,108	+8,885
MPL-2.0 files	55,245	(not re-measured)	n/a
MPL-1.0 (word-boundary, excl. PMPL substrings)	0	0 estate-authored	0

The PMPL-1.0 count grew, not shrank — likely from the in-flight CLAUDE.md docs flip (65-PR background script per memory) propagating fresh PMPL-1.0 stamps before this sweep approves. The MPL-1.0 conversion claim in the body remains accurate (only matches are license/PMPL-1.0.txt filename hits, not estate-authored content).

Progress since filing

PR	Title	Status
standards#249	chore(spdx): migrate 25 workflow SPDX headers PMPL-1.0-or-later → MPL-2.0	MERGED 2026-05-28
standards#201	feat(governance): add licence-consistency CI check	MERGED 2026-05-27
standards#196	docs(audits): estate-wide licence-debt audit 2026-05-26	MERGED 2026-05-27

The "Phase 1 — standards itself" candidate (2,032 files at filing → ~3,822 now) has had only the 25-file workflow precursor merged. Owner decision (a/b/c) still pending.

Owner-decision options (re-stated for clarity)

• (a) Approve Phase 1 (standards itself) now — pattern-establishing, ~3,800-file PR or 3-4 split PRs.
• (b) Defer — PMPL-1.0-or-later is contractually equivalent to MPL-2.0; cosmetic-only.
• (c) Trigger full sweep — ~21,000 files, ~6-8 hours of Actions budget, ~40 PRs.

This is a genuinely-open campaign awaiting owner decision, not a stale-claim closure candidate. Keeping open. No per-phase sub-issues exist yet — they'd be filed once option (a) or (c) is approved.

Cross-ref: also note [[feedback_echidna_license_docs_mpl_intentional.md]] memory — echidna's CLAUDE.md+FAQ.md AGPL reconciliation PR #149 was reverted 2026-05-30, so echidna docs intentionally stay MPL-2.0 despite AGPL LICENSE; do not sweep that repo's SPDX headers unprompted.