[campaign] TypeScript → AffineScript estate migration (UMBRELLA)

Campaign overview

Estate-wide migration of TypeScript to AffineScript, parallel to the completed V→Zig sweep (2026-05-28) but at very different scale.

State: IN-FLIGHT. Multi-session campaign, ~50-80 PRs across ~38 repos.

Scope safety: only hyperpolymath/ repos that are non-fork (owner-only + shared-with-son). Upstream forks (rescript, servers, repos-monorepo, linguist) are carve-outs.

4-layer architecture

LAYER 1 - POLICY & ENFORCEMENT COMPLETE

hypatia#375 :typescript_detected gains path_allow_prefixes (7 cls)
hypatia#378 +linguist, +/deps/, +/vscode/ (9 cls)
standards#235 + docs+ci: mirror hypatia#378 TS carve-out extension (linguist + /deps/ + /vscode/) #238 + reposystem#82 docs + CI gate mirror
standards#247 ROADMAP campaign section
hypatia tests: 11 cases across 9 carve-out classes

LAYER 2 - STDLIB BINDING CAPACITY EXPANDING

Deno.affine: +6 externs via affinescript#504 = statIsFile / statIsDirectory / bytesLength / bytesByteAt / bytesAsciiSlice / importMetaUrl
json.affine v0.3 (Zig FFI parse), Vscode.affine (55 fns), VscodeLanguageClient.affine
Gaps tracked: [campaign #239] STEP 4-A — STDLIB FILL II: raw binary I/O (Bytes constructor + LE getters/setters) #326 binary I/O (Bytes new/fill/LE getters+setters), [campaign #239] STEP 4-B — STDLIB FILL III: randomness + time bindings #327 random + perf.now, recursive walk done (chore(deps): bump the actions group with 4 updates #445), Set / general Map, TOML, WebExt / WP-admin / Zotero APIs

LAYER 3 - PHYSICAL FILE PORTS (~285 files / 38 repos)

STEP 1 phronesis 54L LANDED
STEP 2 4 ports LANDED (panic-attack#82 + session-sentinel#25 + tropical-resource-typing#15 + nafa-app#23)
STEP 3 affinescript#504 armed (6 gaps in one PR)
STEP 4 sub-shape filed: 4-A binary ([campaign #239] STEP 4-A — STDLIB FILL II: raw binary I/O (Bytes constructor + LE getters/setters) #326) + 4-B random ([campaign #239] STEP 4-B — STDLIB FILL III: randomness + time bindings #327) + 4-C per-repo + 4-D cross-campaign

LAYER 4 - PARALLEL DEPENDENCY: AS bindings top-50 roadmap

Plugin-host APIs needed before plugin-repo ports can land

Step issues

Dependency graph

step 1 LANDED -> step 2 LANDED -> step 3 (~ done) -> step 4-A (#326) + step 4-B (#327) -> step 4-C (#243) -> step 4-D -> step 5

step 6 (host bindings, parallel) -> step 7 (plugin repos)

Cross-references

GitHub Projects board: https://github.com/users/hyperpolymath/projects/35 (Hyperpolymath Master Roadmap)
Memory tracker: project_estate_ts_to_affinescript_2026_05_28.md
2026-05-31 session: session_2026_05_31_ts_to_as_step3_closure.md
AS bindings parallel: project_affinescript_bindings_top50_roadmap.md
V-lang precedent (completed 2026-05-28): project_estate_vlang_remediation_2026_05_28.md
Layer-1 PRs: hypatia#375, hypatia#378, standards#235, standards#238, standards#247, reposystem#82

Scale summary

LANDED so far: standards#235, docs+ci: mirror hypatia#378 TS carve-out extension (linguist + /deps/ + /vscode/) #238, docs(ROADMAP): add TS→AffineScript estate migration campaign section #247, reposystem#82, hypatia#375 + hypatia#378, phronesis#19, panic-attack#82, session-sentinel#25, tropical-resource-typing#15, nafa-app#23 + affinescript STEP-3 chore(deps): bump the actions group with 4 updates #445 + secret-scanner: add canonical .gitleaks.toml allowlist (unblocker for re-pinning after #500) #504 (armed).
Realistic PR count remaining: ~50-70 across 38 repos
Sessions estimate: 7-13+ (vs V-lang's 1 session for 16 deletion PRs)
Critical path: STEP 3 ~ done -> STEP 4-A/4-B unlock STEP 4-C -> STEP 5 -> STEPs 6/7 already in parallel

docs(ROADMAP): add TS→AffineScript estate migration campaign section (#…

feat(vscode): AffineScript port of extension (TS→AS campaign step 1 /…

feat(stdlib): Deno scripting surface — walk/args/exit/iso/regex/stder…

refactor(scripts): call exit() directly in main() — drop wrapper-defe…

feat(scripts): AffineScript port of check-ts-allowlist (STEP 2 seed, …

feat(scripts): AffineScript port of check-ts-allowlist (TS->AS campai…

feat(scripts+ci): swap governance TS-allowlist check to AffineScript-…

fix(scripts): make check-ts-allowlist.affine compile cleanly (follow-…

STEP 4-A/4-B PRs now armed (2026-05-31 PM)

sub-issue	PR	state
#326 STEP 4-A binary I/O	affinescript#507	auto-merge SQUASH armed; 24 CI queued
#327 STEP 4-B random + perf.now	affinescript#509	auto-merge SQUASH armed; 24 CI queued
#242 STEP 3 stdlib fill (predecessor)	affinescript#504	auto-merge SQUASH armed; 24 CI queued

3 sibling PRs queued — file sections are disjoint enough that natural merge order is #504 → #507 → #509 with mechanical rebases (no symbol-level conflicts). After all three land, STEP 4-C (#243 per-repo ports) becomes the critical path — first realistic ports: raze-tui's 6 binary-ABI test files (unblocked by #507), bofig's property/bench files (unblocked by #509), bofig's unit/e2e/aspect files (no extra binding needed).

port(ts->affine): unit/evidence_graph_test → AffineScript (STEP 4-C S…

feat(stdlib): close 6 STEP-3 Deno-ESM gaps surfaced by TS→AS ports (R…

Status update — AS stdlib codegen landed; campaign waypoint shift (2026-06-01) 📌

For the campaign umbrella tracking here:

What landed (sourced):

affinescript#522 (2026-06-01) — Sqlite.affine codegen foundation; first stdlib extern with full Deno-ESM lowering + tests end-to-end.
affinescript#504 (2026-06-01) — 6 Deno-ESM stdlib externs (stat helpers, byte ops, importMetaUrl) + wildcard-let codegen fix; closes 6 of 10 STEP-3 gaps in standards#242.
affinescript#463 (2026-05-30) — non-ASCII Unicode escape emit in Deno-ESM.
affinescript#421 — stdlib/json v0.3 via hpm-json-rsr Zig FFI.

What this means for the campaign:

Stdlib extern → Deno-ESM lowering pattern is repeatable and tested (lib/codegen_deno.ml + tests/codegen-deno/*).
STEP 3 / STEP 4 / megaport STEP 8 prerequisites partially closed (stdlib-extern side).
Not yet sufficient for user-module codegen / hand-port → runtime replacement on its own; user-module emit + remaining stdlib primitives (affinescript#160 Http.fetch / proof-progress(typed-wasm): spec tracker stale — TW1 partial work landed, TW4 effectively discharged, TW3 ReScript doc-drift #162 Dict/Map / etc.) remain the keystone gaps.

Action: No new mass-port being kicked off from this post — surfacing the status so campaign waypoints stay accurate. Dependent sub-tickets can re-evaluate parking once they verify their specific blockers are in the "now-closed" set above.

🤖 Cross-estate AS-status sweep on 2026-06-01.

🔨 Owner ruling 2026-08-25 — AffineScript governs. This campaign stays LIVE.

This issue was blocked on a genuine contradiction between two canonical documents:

LANGUAGE-POLICY.adoc §1.2 (since 2026-07-29): "TypeScript is permitted under Bun … the former 'use ReScript instead' rule is retired"
.claude/CLAUDE.md (the file agents read first): bans TypeScript in favour of AffineScript

.claude/CLAUDE.md flagged the conflict against itself on 2026-08-07 and deliberately left it for an owner ruling rather than picking a side.

The owner has now ruled: AffineScript governs. LANGUAGE-POLICY.adoc §1.2 was the error.

What this means concretely


Runtime	Bun, tier 1, unchanged — the §1 ordering was never in dispute
Language	AffineScript for new application code; TypeScript only where AffineScript cannot reach (the narrow, transitional MCP-glue / runtime-API carve-out JavaScript already holds)
ReScript	ban uncontested; migration destination is AffineScript

The withdrawn text ran the runtime and language questions together. Separating them makes both documents coherent without touching Bun's position.

Why I am recording this here

I triaged this issue as a close-as-stale candidate on the strength of §1.2, and put the question to the owner before acting. Had I not asked, I would have closed a live campaign. Recording the ruling in-thread so the next reader — human or agent — does not repeat that reasoning from the same evidence.

One correction to the record

§1.2 justified retiring the rule by pointing at ts-blocker.yml being structurally incapable of failing (20 of 20 runs green; never once fired). That measurement is correct and stands. But it is evidence about the gate, not about the rule — a broken enforcement mechanism says nothing about whether the policy it failed to enforce was right. The remedy for a fake gate is to repair it and verify it both ways, not to withdraw the policy.

Correction PR: #632

policy: rule the TypeScript/AffineScript conflict — AffineScript gove…

Re-measured 2026-08-27 — 95 files in 9 repos, and 72 of those are deliberate

	claimed	measured
`.ts`/`.tsx` remaining	~285	95
repos	38	9

No node_modules / vendor / deps hits at all — the census is clean.

72 of the 95 are proven's published polyglot bindings — bindings/typescript/src 36, bindings/deno/src 28, bindings/assemblyscript/src 7. That is a deliberate consumer surface, not migration debt.

Excluding those: 23 files in 8 repos, largest metadatastician/stapeln 7 and metadatastician/svalinn 5 — both now in the other org.

Step status

Closed: #240, #241, #242, #245, #326, #327 — and #243, #244, #246 closed today on measured zero counts.

⚠️ Those closures stand (nothing left to port) but see #664: some tiers reached zero by deletion, and this campaign's metric cannot tell that apart from completion.

Recommend: this umbrella is effectively complete at 23 genuine files. Close it or rescope to those 23.

policy: the VSCode TypeScript carve-out's unblock condition is alread…

docs(roadmap): point the Projects board at metadatastician/projects/2

docs(roadmap): point the Projects board at metadatastician/projects/2 (…

Navigation Menu

Repository navigation

[campaign] TypeScript → AffineScript estate migration (UMBRELLA) #239

Campaign overview

4-layer architecture

Step issues

Dependency graph

Cross-references

Scale summary

6 remaining items

🔨 Owner ruling 2026-08-25 — AffineScript governs. This campaign stays LIVE.

What this means concretely

Why I am recording this here

One correction to the record

Re-measured 2026-08-27 — 95 files in 9 repos, and 72 of those are deliberate

Step status

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

[campaign] TypeScript → AffineScript estate migration (UMBRELLA) #239

Description

Campaign overview

4-layer architecture

Step issues

Dependency graph

Cross-references

Scale summary

Activity

6 remaining items

hyperpolymath commented on May 31, 2026

hyperpolymath commented on Jun 1, 2026

hyperpolymath commented on Aug 25, 2026

🔨 Owner ruling 2026-08-25 — AffineScript governs. This campaign stays LIVE.

What this means concretely

Why I am recording this here

One correction to the record

hyperpolymath commented on Aug 27, 2026

Re-measured 2026-08-27 — 95 files in 9 repos, and 72 of those are deliberate

Step status

Metadata

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Issue actions