[campaign #239] STEP 6 — PLUGIN-HOST BINDINGS: cross-ref AS top-50 roadmap

[hyperpolymath]

Part of #239. Cross-references project_affinescript_bindings_top50_roadmap.

Goal

Coordinate with the AS bindings top-50 roadmap so that the host APIs needed by plugin repos land BEFORE step 7 starts.

Plugin-host APIs needed

Host	Repo(s)	API surface needed
WordPress (wp-admin, blocks)	wordpress-tools	wp.blocks.*, wp.element.*, websocket-server
Firefox/Thunderbird WebExt	universal-chat-extractor, double-track-browser, thunderbird-template-reloaded	browser.tabs.*, browser.runtime.*, MailExtensions
Next.js / React	polyglot-i18n, lcb-website?	React component model, Next.js Pages Router

Acceptance

• Each host API gets a bindings PR in affinescript repo OR an explicit "won't bind" rationale
• This issue closes when ALL plugin-blocked repos in step 7 have at least one path forward

[hyperpolymath]

Co-design questions to resolve before execution

Marking this issue as Co-design owner on the project board because the strategic calls below need operator judgement before Claude can sequence work. Each question is being worked through with a Plus/Minus/Interesting analysis (de Bono PMI) in the operator session log; resolutions will land here as edits.

Q1: Which plugin hosts to bind at all?

Host	Estate users	Compat question
WordPress	wordpress-tools (43 files)	WP-PHP doesn't load WASM. Plugin entry point is PHP. AS can reach Gutenberg blocks JS, not server-side. Permanent ceiling?
WebExtensions	universal-chat-extractor (7), double-track-browser (6)	Workable — JS-loaded, but ~200+ fn surface
Thunderbird MailExt	thunderbird-template-reloaded (7)	Subset of WebExt + ~30 mail-specific fns
React/Next.js	polyglot-i18n (2)	Tiny estate footprint, huge framework surface

Q2: Surface scope per chosen host

Full surface bind (multi-month, future-proof) vs scoped to actual estate usage (fast, brittle for downstream).

Q3: Async ABI maturity

AS issue #103 (async-extern ABI) — precondition, or ship sync-first with Thenable/Async wrapping?

Q4: Manifest V3 vs V2

WebExt MV3 only / MV2 only / both?

Q5: Top-50 roadmap integration

Fold into the AS bindings top-50 as new tiers, keep as parallel track (current framing), or drop in favour of broader-impact bindings?

Q6: Prioritization order

If "bind some but not all" — which first? WebExt (3 estate repos), MailExt (smallest surface), or neither?

Q7: Plugin compatibility test bench

Per-host test fixtures + CI infra, or manual smoke tests at port time?

---

Resolution sequence: Q1 → Q6 → Q5 → Q2 → Q3 → Q4 → Q7 (scope before tactics).

[hyperpolymath]

Re-verified 2026-08-26 — closing: the premise is gone

This issue exists to provide AffineScript bindings so the plugin-host TypeScript in five repos could be ported. There is no TypeScript left to port.

Measured from full recursive trees, truncated: false on all five:

repo	.ts files	what it actually is now
wordpress-tools	0	138 .php, Rust, PowerShell
universal-chat-extractor	0	11 .idr
thunderbird-template-reloaded	0	9 .idr, 8 .affine
double-track-browser	0	27 .affine, 11 .rs
polyglot-i18n	0	79 .js, 31 .affine

#246 — the step this was meant to unblock — inventories 43+7+7+6+2 = 65 TypeScript files. The live count is 0.

Two further checks:

• The named technical blocker, affinescript#103 (async-extern ABI), is CLOSED.
• Nothing was ever bound: affinescript returns 0 path hits for wordpress / gutenberg / webext / browser.tabs / thunderbird / mailext, and docs/bindings-roadmap.adoc carries no plugin-host entry in any of its five tiers. The cross-reference this issue asks for was never made — and the consumers it would have served no longer exist.

One thing worth separating out

I did not establish whether that TypeScript was migrated or simply deleted. Given the estate's documented pattern of scoring absence as completion, that distinction may deserve its own issue. It does not change this verdict either way: the stated need — bindings to unblock a TS→AffineScript port — is gone.

Closing. Reopen if the migrate-vs-delete question turns out to matter.