PerryTS · jdalton · Aug 16, 2026
diff --git a/.github/workflows/node-compat-matrix.yml b/.github/workflows/node-compat-matrix.yml
@@ -1,6 +1,6 @@
 name: Node Compat Matrix Guard
 
-# Gates scripts/node_compat_matrix.mjs --check: a breadth sweep over every
+# Gates scripts/node_compat_matrix.mts --check: a breadth sweep over every
 # require("module").builtinModules entry, both import forms (M and node:M),
 # against a PINNED, SRI-verified Node oracle (external-tools.json tools.node,
 # currently 26.5.1). It compares Perry's export-SHAPE fingerprint per module
@@ -11,7 +11,7 @@ name: Node Compat Matrix Guard
 #
 # Its OWN job on purpose: the runner downloads the pinned Node dist tarball
 # (~40MB) and verifies it, which must not slow the main cargo-test job. The
-# node used by setup-node here only EXECUTES the .mjs; the oracle Node is the
+# node used by setup-node here only EXECUTES the .mts; the oracle Node is the
 # pinned dist the runner fetches + SRI-verifies itself.
 #
 # Decoupled from the (not-yet-enabled) merge queue like node-suite-guard.yml:
@@ -76,7 +76,7 @@ jobs:
         uses: actions/setup-node@v7
         with:
           # Single source of truth: .node-version at the repo root. This node
-          # only RUNS node_compat_matrix.mjs; the compat oracle is the pinned
+          # only RUNS node_compat_matrix.mts; the compat oracle is the pinned
           # Node the runner downloads + SRI-verifies from external-tools.json.
           node-version-file: .node-version
 
@@ -91,6 +91,6 @@ jobs:
           set -euo pipefail
           echo '### Node builtin compat matrix guard' >> "$GITHUB_STEP_SUMMARY"
           echo '```' >> "$GITHUB_STEP_SUMMARY"
-          node scripts/node_compat_matrix.mjs --check \
+          node scripts/node_compat_matrix.mts --check \
             | tee -a "$GITHUB_STEP_SUMMARY"
           echo '```' >> "$GITHUB_STEP_SUMMARY"
diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
@@ -294,7 +294,7 @@ jobs:
       # newly-soaked upstream releases as advisories.
       - name: Binding upstream pins (lock-step)
         if: ${{ !cancelled() }}
-        run: node scripts/binding_pins.mjs --check
+        run: node scripts/binding_pins.mts --check
 
       # GC write-barrier store-site inventory: every raw heap-slot store in
       # perry-codegen / perry-runtime / perry-stdlib must be barriered or

diff --git a/.gitignore b/.gitignore
@@ -6,7 +6,7 @@ __pycache__/
 *.py[cod]
 
 # Pinned Node oracle for the builtin-module compat matrix
-# (scripts/node_compat_matrix.mjs downloads + SRI-verifies + caches here).
+# (scripts/node_compat_matrix.mts downloads + SRI-verifies + caches here).
 .cache/node-pin/
 
 # Android Gradle: caches and build outputs are regenerable. Source under
@@ -192,7 +192,7 @@ private-*
 # mdBook output
 docs/book/
 
-# scripts/fp_fuzz.mjs failure dumps (regenerable, machine-local)
+# scripts/fp_fuzz.mts failure dumps (regenerable, machine-local)
 fp_fuzz_failures/
 
 test-files/.perry-cache/

diff --git a/CLAUDE.md b/CLAUDE.md
@@ -29,22 +29,22 @@ Three workflows are deliberately exempt and say so inline: `node-core-subset.yml
 - **Test/CI mechanics** — `#794` (per-category parity thresholds), `#796` (gap-suite output truncation + O(n²) `normalize_output`), `#812` (42-module behavioral matrix), `#806/#807/#808` (test harnesses for mixins / async context / ≥300-init scale).
 - **Skip-list audit** — `#797` covers `test-parity/known_failures.json` provenance (issue # + date per entry).
 
-### Node builtin compatibility matrix (`scripts/node_compat_matrix.mjs`)
+### Node builtin compatibility matrix (`scripts/node_compat_matrix.mts`)
 
 Breadth sweep over EVERY `require("module").builtinModules` entry, both import forms (`M` and `node:M`), against a **pinned, SRI-verified Node** (the "latest stable" oracle, pinned in `external-tools.json` `tools.node.version` — currently **26.5.1**, independent of the `.node-version` gap-suite oracle). It compares Perry's export-SHAPE fingerprint (sorted `name:typeof` over the module namespace + the default export's typeof) to the oracle's. This is the systematic version of the #812 "42-module behavioral matrix" — shape, not deep behavior (behavioral cases stay in the node-suite).
 
 ```bash
 # FAST LOOP — reach for this first when iterating on ONE builtin:
-node scripts/node_compat_matrix.mjs --module fs                      # one module, both forms
-node scripts/node_compat_matrix.mjs --module fs,path,crypto          # a few
-node scripts/node_compat_matrix.mjs --module fs --method readFileSync,promises  # only these exports
-node scripts/node_compat_matrix.mjs --only fs.readFileSync,path.join # combined mod.export form
+node scripts/node_compat_matrix.mts --module fs                      # one module, both forms
+node scripts/node_compat_matrix.mts --module fs,path,crypto          # a few
+node scripts/node_compat_matrix.mts --module fs --method readFileSync,promises  # only these exports
+node scripts/node_compat_matrix.mts --only fs.readFileSync,path.join # combined mod.export form
 # (the pinned Node download is skipped once cached under .cache/node-pin/)
 
 # FULL SWEEP + GATE:
-node scripts/node_compat_matrix.mjs                 # whole matrix + summary table
-node scripts/node_compat_matrix.mjs --check         # CI gate: exit 1 on regressions vs the baseline
-node scripts/node_compat_matrix.mjs --update-baseline   # rewrite test-parity/node-compat-matrix.baseline.json
+node scripts/node_compat_matrix.mts                 # whole matrix + summary table
+node scripts/node_compat_matrix.mts --check         # CI gate: exit 1 on regressions vs the baseline
+node scripts/node_compat_matrix.mts --update-baseline   # rewrite test-parity/node-compat-matrix.baseline.json
 ```
 
 A `--module` selector scopes `--check`/`--update-baseline` to just that slice (a single-module refresh never rewrites the whole baseline). A `--method`/`--only` subset is a print-only fast diagnostic (it narrows the fingerprint, so it is refused for `--check`/`--update-baseline`). **Bump the pinned Node** by editing `tools.node.version` in `external-tools.json` (add per-platform sha512 SRI), then `--update-baseline` and review the diff. Needs the release binary (`cargo build --release -p perry`). Full page: `docs/src/testing/node-compat-matrix.md`.

diff --git a/changelog.d/8241-scripts-mts.md b/changelog.d/8241-scripts-mts.md
@@ -0,0 +1 @@
+Converted the three remaining `scripts/**/*.mjs` files (`binding_pins`, `fp_fuzz`, `node_compat_matrix`) to TypeScript `.mts`, standardizing onto the same module + type-stripping convention the soak scripts already use. Declared `engines.node >= 22.18.0` in `package.json` (the floor for native `.mts` type stripping; the repo's `.node-version` pin of 26.5.1 is well above it). All invocation sites — CI workflows, docs, config JSON, in-script help text, Rust comments — updated to reference the new `.mts` paths.
diff --git a/crates/perry/src/commands/compile/well_known.rs b/crates/perry/src/commands/compile/well_known.rs
@@ -71,7 +71,7 @@ pub struct WellKnownBinding {
     /// Upstream provenance pin — which release of the npm package this
     /// wrapper ports, and when it was last reviewed against it. See
     /// `docs/src/native-libraries/upstream-pins.md` and the lock-step
-    /// gate in `scripts/binding_pins.mjs`. `None` for entries exempt
+    /// gate in `scripts/binding_pins.mts`. `None` for entries exempt
     /// from pinning (`node_builtin`, `alias_of`, and perry-owned
     /// packages).
     // Provenance metadata parsed from `well_known_bindings.toml` and consulted
@@ -133,7 +133,7 @@ fn binding_is_faithful(
 ///
 /// The **lock-step rule**: `ported_at` must equal `version`. Re-pinning
 /// an upstream release without re-reviewing the wrapper against the
-/// upstream diff reds the `binding_pins.mjs --check` gate until
+/// upstream diff reds the `binding_pins.mts --check` gate until
 /// `ported_at` advances with the review — an upstream release can never
 /// go silently stale, and a pin bump can never outrun its port.
 #[derive(Debug, Clone, PartialEq, Eq)]
@@ -317,7 +317,7 @@ fn parse_well_known_toml(raw: &str) -> Result<BTreeMap<String, WellKnownBinding>
                 let version = required("version")?;
                 let ported_at = required("ported-at")?;
                 // Parse-time lock-step backstop. The authoritative gate is
-                // `scripts/binding_pins.mjs --check` (CI); failing here too
+                // `scripts/binding_pins.mts --check` (CI); failing here too
                 // means a skewed pin can't even ship inside the binary.
                 if ported_at != version {
                     return Err(format!(
@@ -741,7 +741,7 @@ mod tests {
 
     /// The lock-step rule enforced at parse time: a pin bump
     /// (`version`) that outruns its review (`ported-at`) must refuse
-    /// to load — the authoritative CI gate is binding_pins.mjs
+    /// to load — the authoritative CI gate is binding_pins.mts
     /// --check, but a skewed pin must not even ship inside the binary.
     #[test]
     fn upstream_pin_rejects_lock_step_violation() {
@@ -793,7 +793,7 @@ mod tests {
         assert!(
             unpinned.is_empty(),
             "bindings without an [bindings.<name>.upstream] pin — provision one with \
-             `node scripts/binding_pins.mjs --set <name>`:\n  {}",
+             `node scripts/binding_pins.mts --set <name>`:\n  {}",
             unpinned.join("\n  ")
         );
     }

diff --git a/docs/po/de.po b/docs/po/de.po
@@ -52183,13 +52183,13 @@ msgstr "Korrektheits-Zahlen"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — generiert zufällig TS-Programme, die die sechs "
+"`scripts/fp_fuzz.mts` — generiert zufällig TS-Programme, die die sechs "
 "Muster ausüben, die am ehesten Per-Instruction-FMFs auslösen (Left-Fold-, "
 "Tree-Fold-, Right-Fold-Reduktionen; FMA-förmige Ketten; algebraische "
 "Identitäten wie `(a/b)*b`; Cancellation-Prädikate). Jedes Programm wird "

diff --git a/docs/po/es.po b/docs/po/es.po
@@ -52566,13 +52566,13 @@ msgstr "Números de corrección"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — genera aleatoriamente programas TS que ejercitan los"
+"`scripts/fp_fuzz.mts` — genera aleatoriamente programas TS que ejercitan los"
 " seis patrones con más probabilidad de activar FMFs por instrucción "
 "(reducciones left-fold, tree-fold, right-fold; cadenas con forma FMA; "
 "identidades algebraicas como `(a/b)*b`; predicados de cancelación). Cada "

diff --git a/docs/po/fr.po b/docs/po/fr.po
@@ -52693,13 +52693,13 @@ msgstr "Chiffres de correction"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — génère aléatoirement des programmes TS exercant les "
+"`scripts/fp_fuzz.mts` — génère aléatoirement des programmes TS exercant les "
 "six motifs les plus susceptibles de déclencher des FMF par instruction "
 "(réductions left-fold, tree-fold, right-fold ; chaînes de forme FMA ; "
 "identités algébriques comme `(a/b)*b` ; prédicats d'annulation). Chaque "

diff --git a/docs/po/id.po b/docs/po/id.po
@@ -51605,13 +51605,13 @@ msgstr "Angka kebenaran"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — secara acak menghasilkan program TS yang menjalankan"
+"`scripts/fp_fuzz.mts` — secara acak menghasilkan program TS yang menjalankan"
 " enam pola yang paling mungkin memicu FMF per-instruksi (reduksi left-fold, "
 "tree-fold, right-fold; rantai berbentuk FMA; identitas aljabar seperti "
 "`(a/b)*b`; predikat pembatalan). Setiap program dikompilasi dengan Node dan "

diff --git a/docs/po/it.po b/docs/po/it.po
@@ -52042,13 +52042,13 @@ msgstr "Numeri sulla correttezza"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — genera casualmente programmi TS che esercitano i sei"
+"`scripts/fp_fuzz.mts` — genera casualmente programmi TS che esercitano i sei"
 " pattern più suscettibili di attivare i FMF per istruzione (riduzioni left-"
 "fold, tree-fold, right-fold; catene con forma FMA; identità algebriche come "
 "`(a/b)*b`; predicati di cancellazione). Ogni programma viene compilato sia "

diff --git a/docs/po/ja.po b/docs/po/ja.po
@@ -49531,13 +49531,13 @@ msgstr "正確性の数字"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — 命令ごとの FMF を発生させやすい 6 "
+"`scripts/fp_fuzz.mts` — 命令ごとの FMF を発生させやすい 6 "
 "つのパターン（左折りたたみ、木折りたたみ、右折りたたみリダクション、FMA 形のチェーン、`(a/b)*b` "
 "のような代数恒等式、キャンセル述語）を行使する TS プログラムをランダムに生成します。各プログラムは Node と Perry "
 "の両方でコンパイルされ、stdout はバイト単位で diff されます。"

diff --git a/docs/po/ko.po b/docs/po/ko.po
@@ -49482,13 +49482,13 @@ msgstr "정확성 수치"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — 명령어별 FMF를 트리거할 가능성이 가장 높은 여섯 가지 패턴(왼쪽 폴드, 트리 폴드, 오른쪽"
+"`scripts/fp_fuzz.mts` — 명령어별 FMF를 트리거할 가능성이 가장 높은 여섯 가지 패턴(왼쪽 폴드, 트리 폴드, 오른쪽"
 " 폴드 리덕션; FMA 형태 체인; `(a/b)*b`와 같은 대수 항등식; 소거 술어)을 실행하는 TS 프로그램을 무작위로 생성합니다. "
 "각 프로그램은 Node와 Perry 양쪽에서 컴파일되며 stdout을 바이트 단위로 비교합니다."
 

diff --git a/docs/po/messages.pot b/docs/po/messages.pot
@@ -46721,7 +46721,7 @@ msgstr ""
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node "

diff --git a/docs/po/th.po b/docs/po/th.po
@@ -51127,13 +51127,13 @@ msgstr "ตัวเลขความถูกต้อง"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — สร้างโปรแกรม TS "
+"`scripts/fp_fuzz.mts` — สร้างโปรแกรม TS "
 "แบบสุ่มที่ทดสอบรูปแบบหกแบบที่มีแนวโน้มจะทำให้ per-instruction FMF สะดุด "
 "(การรวม left-fold, tree-fold, right-fold; chain รูปแบบ FMA; identity "
 "เชิงพีชคณิตเช่น `(a/b)*b`; predicate การยกเลิก) แต่ละโปรแกรมถูก compile "

diff --git a/docs/po/vi.po b/docs/po/vi.po
@@ -51541,13 +51541,13 @@ msgstr "Số liệu tính chính xác"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — tạo ngẫu nhiên các chương trình TS thực thi sáu mẫu "
+"`scripts/fp_fuzz.mts` — tạo ngẫu nhiên các chương trình TS thực thi sáu mẫu "
 "có khả năng kích hoạt FMF từng lệnh nhất (rút gọn left-fold, tree-fold, "
 "right-fold; chuỗi hình FMA; đẳng thức đại số như `(a/b)*b`; vị từ triệt "
 "tiêu). Mỗi chương trình được biên dịch bằng cả Node và Perry, và stdout được"

diff --git a/docs/po/zh-CN.po b/docs/po/zh-CN.po
@@ -49008,13 +49008,13 @@ msgstr "正确性数据"
 
 #: src/cli/fast-math.md:123
 msgid ""
-"`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the six "
+"`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the six "
 "patterns most likely to trip per-instruction FMFs (left-fold, tree-fold, "
 "right-fold reductions; FMA-shaped chains; algebraic identities like "
 "`(a/b)*b`; cancellation predicates). Each program is compiled with both Node"
 " and Perry, and stdout is diffed byte-for-byte."
 msgstr ""
-"`scripts/fp_fuzz.mjs` — 随机生成 TS 程序，使用最有可能触发按指令 FMF 的六种模式（左折叠、树折叠、右折叠归约；FMA "
+"`scripts/fp_fuzz.mts` — 随机生成 TS 程序，使用最有可能触发按指令 FMF 的六种模式（左折叠、树折叠、右折叠归约；FMA "
 "形状链；像 `(a/b)*b` 的代数恒等式；抵消谓词）。每个程序都用 Node 和 Perry 编译，并对 stdout 进行逐字节比对。"
 
 #: src/cli/fast-math.md:129

diff --git a/docs/src/cli/fast-math.md b/docs/src/cli/fast-math.md
@@ -120,7 +120,7 @@ giving up bit-exact parity.
 
 ## Correctness numbers
 
-`scripts/fp_fuzz.mjs` — randomly generates TS programs exercising the
+`scripts/fp_fuzz.mts` — randomly generates TS programs exercising the
 six patterns most likely to trip per-instruction FMFs (left-fold,
 tree-fold, right-fold reductions; FMA-shaped chains; algebraic
 identities like `(a/b)*b`; cancellation predicates). Each program is
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		Converted the three remaining `scripts/*/.mjs` files (`binding_pins`, `fp_fuzz`, `node_compat_matrix`) to TypeScript `.mts`, standardizing onto the same module + type-stripping convention the soak scripts already use. Declared `engines.node >= 22.18.0` in `package.json` (the floor for native `.mts` type stripping; the repo's `.node-version` pin of 26.5.1 is well above it). All invocation sites — CI workflows, docs, config JSON, in-script help text, Rust comments — updated to reference the new `.mts` paths.
Comment thread coderabbitai[bot] marked this conversation as resolved.