feat(codemode): native coercion parity in interpreter (#37608)

Author: rekram1-node

packages/codemode/interpreter-support.md

@@ -80,7 +80,7 @@ ultimate source of truth.
 - [x] Expression and block function bodies.
 - [x] User callbacks for the supported Array, Map, Set, URLSearchParams, sort, string-replacement, and `Array.from`
       mapper APIs, with one shared acceptance rule everywhere including promise reactions.
-- [x] `Boolean`, `Number`, `String`, `parseInt`, `parseFloat`, and URI helpers as callbacks.
+- [x] `Boolean`, `Number`, `String`, `parseInt`, `parseFloat`, `isFinite`, `isNaN`, and URI helpers as callbacks.
 - [x] Built-in method references as callbacks, such as `values.map(Math.abs)`, `records.map(JSON.stringify)`,
       `items.forEach(console.log)`, and `Promise.resolve(-1).then(Math.abs)`. Extra callback arguments a built-in
       does not consume are ignored, like JS; consumed arguments stay strictly validated (`Math.floor` still rejects a
@@ -210,9 +210,12 @@ ultimate source of truth.
 - [x] `localeCompare`; locale and options arguments are currently ignored.
 - [x] `toString`, `length`, numeric indexing, spread, and `for...of` by Unicode code point.
 - [x] Static `String.fromCharCode` and `String.fromCodePoint`.
-- [ ] Native argument coercion for supported String methods; for example, `includes(1)` and `slice("1")` currently
-      reject instead of coercing.
-- [ ] Native no-argument parity for `match()` and `search()`.
+- [x] Native argument coercion for supported String methods; for example, `includes(1)` and `slice("1")` coerce like
+      native JS, `split(undefined)` returns the whole string, and `includes`/`startsWith`/`endsWith` reject regular
+      expressions with a native-style `TypeError`. Opaque runtime references still reject as data errors, and
+      `repeat` still requires a finite non-negative count.
+- [x] Native no-argument parity for `match()`, `matchAll()`, and `search()`; all behave as an empty pattern. Present
+      arguments must still be a regular expression or string pattern.
 
 ## Numbers and Math
 
@@ -226,13 +229,16 @@ ultimate source of truth.
 - [x] Math methods: `random`, `max`, `min`, `abs`, `acos`, `acosh`, `asin`, `asinh`, `atan`, `atan2`, `atanh`,
       `floor`, `ceil`, `round`, `trunc`, `sign`, `sqrt`, `cbrt`, `pow`, `hypot`, `cos`, `cosh`, `sin`, `sinh`,
       `tan`, `tanh`, `log`, `log2`, `log10`, `log1p`, `exp`, `expm1`, `f16round`, `fround`, `clz32`, and `imul`.
-- [ ] Native zero-argument behavior for `Number()` and `String()`; they currently do not produce `0` and `""`.
-- [ ] `++` and `--` must use CodeMode numeric coercion and reject opaque runtime references; they currently call host
-      `Number(...)` directly.
-- [ ] Unknown static members must read as `undefined` for feature detection; some currently appear callable or throw
-      during property access.
+- [x] Native zero-argument behavior for `Number()` and `String()`: they produce `0` and `""`, while
+      `Number(undefined)` stays `NaN` and `String(undefined)` stays `"undefined"`.
+- [x] `++` and `--` use CodeMode numeric coercion (numeric strings increment, plain data objects become `NaN`, Dates
+      use their epoch time) and reject opaque runtime references as data errors.
+- [x] Unknown static members on global namespaces and on `Number`/`String`/the coercion functions read as `undefined`
+      for feature detection. Calling any undefined value reports a native-style `TypeError` naming the callee, for
+      example `Math.sumPrecise is not a function.` Blocked members (`constructor`, `__proto__`, ...) still throw,
+      and unknown `Promise` statics keep their descriptive error.
 - [ ] `Math.sumPrecise`.
-- [ ] Global coercing `isFinite` and `isNaN`.
+- [x] Global coercing `isFinite` and `isNaN`; opaque runtime references reject as data errors, like `Number(...)`.
 
 ## JSON and console
 

packages/codemode/src/interpreter/methods.ts

@@ -12,7 +12,7 @@ import {
   PromiseNamespace,
   UriFunction,
 } from "./model.js"
-import { rejectCircularInsertion, typeofValue } from "./references.js"
+import { containsOpaqueReference, rejectCircularInsertion, typeofValue } from "./references.js"
 import { isBlockedMember, type SafeObject } from "../tool-runtime.js"
 import {
   CodeModeDate,
@@ -137,21 +137,31 @@ export const invokeGlobalMethod = (ref: GlobalMethodReference, args: Array<unkno
   return invokeJsonMethod(ref.name, args, node)
 }
 
-const invokeStringMethod = (value: string, name: string, args: Array<unknown>, node: AstNode): unknown => {
-  const str = (index: number): string => {
-    const arg = args[index]
-    if (typeof arg !== "string")
-      throw new InterpreterRuntimeError(`String.${name} expects argument ${index + 1} to be a string.`, node)
-    return arg
-  }
-  const num = (index: number): number => {
-    const arg = args[index]
-    if (typeof arg !== "number")
-      throw new InterpreterRuntimeError(`String.${name} expects argument ${index + 1} to be a number.`, node)
-    return arg
+const requireDataArgument = (name: string, index: number, arg: unknown, node: AstNode): unknown => {
+  if (containsOpaqueReference(arg)) {
+    throw new InterpreterRuntimeError(
+      `String.${name} expects argument ${index + 1} to be a data value.`,
+      node,
+      "InvalidDataValue",
+    )
   }
+  return arg
+}
+
+const invokeStringMethod = (value: string, name: string, args: Array<unknown>, node: AstNode): unknown => {
+  // Coerce arguments like native JS; opaque runtime references still reject.
+  const str = (index: number): string => coerceToString(requireDataArgument(name, index, args[index], node))
+  const num = (index: number): number => coerceToNumber(requireDataArgument(name, index, args[index], node))
   const optNum = (index: number): number | undefined => (args[index] === undefined ? undefined : num(index))
   const optStr = (index: number): string | undefined => (args[index] === undefined ? undefined : str(index))
+  const rejectRegex = (): void => {
+    if (args[0] instanceof CodeModeRegExp) {
+      throw new InterpreterRuntimeError(
+        `String.${name} cannot take a regular expression; use regex.test(string) or String.search instead.`,
+        node,
+      ).as("TypeError")
+    }
+  }
 
   let result: unknown
   switch (name) {
@@ -187,8 +197,11 @@ const invokeStringMethod = (value: string, name: string, args: Array<unknown>, n
       break
     }
     case "split": {
-      if (args.length === 0) {
-        result = [value]
+      // Native: an undefined separator returns the whole string, not a split on "undefined",
+      // unless the limit truncates to zero.
+      if (args[0] === undefined) {
+        const requestedLimit = optNum(1)
+        result = requestedLimit !== undefined && requestedLimit >>> 0 === 0 ? [] : [value]
         break
       }
       if (args[0] instanceof CodeModeRegExp) {
@@ -203,12 +216,15 @@ const invokeStringMethod = (value: string, name: string, args: Array<unknown>, n
       result = value.slice(optNum(0), optNum(1))
       break
     case "includes":
+      rejectRegex()
       result = value.includes(str(0), optNum(1))
       break
     case "startsWith":
+      rejectRegex()
       result = value.startsWith(str(0), optNum(1))
       break
     case "endsWith":
+      rejectRegex()
       result = value.endsWith(str(0), optNum(1))
       break
     case "indexOf":
@@ -263,7 +279,7 @@ const invokeStringMethod = (value: string, name: string, args: Array<unknown>, n
     case "repeat": {
       const count = num(0)
       if (!Number.isFinite(count) || count < 0)
-        throw new InterpreterRuntimeError("String.repeat expects a finite non-negative count.", node)
+        throw new InterpreterRuntimeError("String.repeat expects a finite non-negative count.", node).as("RangeError")
       result = value.repeat(count)
       break
     }
@@ -301,6 +317,8 @@ const invokeStringMethod = (value: string, name: string, args: Array<unknown>, n
   return boundedData(result, `String.${name} result`)
 }
 
+export const arrayStatics = new Set(["isArray", "of", "from"])
+
 const invokeArrayStatic = (name: string, args: Array<unknown>, node: AstNode): unknown => {
   switch (name) {
     case "isArray":
@@ -400,11 +418,9 @@ const invokeStringReplacer = <R>(
     if (name === "replace") value.replace(pattern.regex, collect)
     else value.replaceAll(pattern.regex, collect)
   } else {
-    if (typeof pattern !== "string") {
-      throw new InterpreterRuntimeError(`String.${name} expects argument 1 to be a string.`, node)
-    }
-    if (name === "replace") value.replace(pattern, collect)
-    else value.replaceAll(pattern, collect)
+    const search = coerceToString(requireDataArgument(name, 0, pattern, node))
+    if (name === "replace") value.replace(search, collect)
+    else value.replaceAll(search, collect)
   }
 
   return Effect.gen(function* () {

packages/codemode/src/interpreter/model.ts

@@ -105,7 +105,7 @@ export class GlobalMethodReference {
 }
 
 export class CoercionFunction {
-  constructor(readonly name: "Number" | "String" | "Boolean" | "parseInt" | "parseFloat") {}
+  constructor(readonly name: "Number" | "String" | "Boolean" | "parseInt" | "parseFloat" | "isFinite" | "isNaN") {}
 }
 
 export class UriFunction {

packages/codemode/src/interpreter/runtime.ts

@@ -10,6 +10,7 @@ import {
   ErrorConstructorReference,
   GlobalMethodReference,
   GlobalNamespace,
+  type GlobalNamespaceName,
   getArray,
   getBoolean,
   getNode,
@@ -34,7 +35,7 @@ import {
   UriFunction,
 } from "./model.js"
 import { caughtErrorValue, constructErrorValue } from "./errors.js"
-import { type CallbackRunner, invokeArrayFrom, invokeGlobalMethod, invokeIntrinsic } from "./methods.js"
+import { arrayStatics, type CallbackRunner, invokeArrayFrom, invokeGlobalMethod, invokeIntrinsic } from "./methods.js"
 import {
   constructPromise,
   invokePromiseInstanceMethod,
@@ -46,17 +47,19 @@ import { containsOpaqueReference, isRuntimeReference, rejectCircularInsertion, t
 import { ScopeStack } from "./scope.js"
 import { arrayMethods, mapMethods, setMethods, spreadItems } from "../stdlib/collections.js"
 import { consoleMethods, formatConsoleMessage } from "../stdlib/console.js"
-import { dateMethods } from "../stdlib/date.js"
-import { mathConstants } from "../stdlib/math.js"
+import { dateMethods, dateStatics } from "../stdlib/date.js"
+import { jsonStatics } from "../stdlib/json.js"
+import { mathConstants, mathMethods } from "../stdlib/math.js"
 import { numberConstants, numberMethods, numberStatics } from "../stdlib/number.js"
-import { objectMethodsPreservingIdentity } from "../stdlib/object.js"
+import { objectMethodsPreservingIdentity, objectStatics } from "../stdlib/object.js"
 import { promiseStatics } from "../stdlib/promise.js"
 import { escapeRegexHint, regexpMethods, regexpProperties, regexFailureReason } from "../stdlib/regexp.js"
 import { stringMethods, stringStatics } from "../stdlib/string.js"
 import {
   urlMethods,
   urlProperties,
   urlSearchParamsMethods,
+  urlStatics,
   urlWritableProperties,
   invokeUriFunction,
   uriArgument,
@@ -83,6 +86,32 @@ import {
   CodeModeURLSearchParams,
 } from "../values.js"
 
+const globalStaticMembers: Partial<Record<GlobalNamespaceName, Set<string>>> = {
+  Object: objectStatics,
+  Math: mathMethods,
+  JSON: jsonStatics,
+  Array: arrayStatics,
+  console: consoleMethods,
+  Date: dateStatics,
+  URL: urlStatics,
+}
+
+const calleeDescription = (callee: AstNode): string => {
+  if (callee.type === "Identifier") return getString(callee, "name")
+  if (callee.type === "MemberExpression") {
+    const object = getNode(callee, "object")
+    const property = getNode(callee, "property")
+    const key =
+      callee.computed !== true && property.type === "Identifier"
+        ? getString(property, "name")
+        : property.type === "Literal" && typeof property.value === "string"
+          ? property.value
+          : undefined
+    if (object.type === "Identifier" && key !== undefined) return `${getString(object, "name")}.${key}`
+  }
+  return "The called value"
+}
+
 const instanceofValue = (lhs: unknown, rhs: unknown, node: AstNode): boolean => {
   if (rhs instanceof ErrorConstructorReference) {
     const brand = errorBrandName(lhs)
@@ -199,6 +228,8 @@ export class Interpreter<R> {
     globalScope.set("console", { mutable: false, value: new GlobalNamespace("console") })
     globalScope.set("parseInt", { mutable: false, value: new CoercionFunction("parseInt") })
     globalScope.set("parseFloat", { mutable: false, value: new CoercionFunction("parseFloat") })
+    globalScope.set("isFinite", { mutable: false, value: new CoercionFunction("isFinite") })
+    globalScope.set("isNaN", { mutable: false, value: new CoercionFunction("isNaN") })
     globalScope.set("Date", { mutable: false, value: new GlobalNamespace("Date") })
     globalScope.set("RegExp", { mutable: false, value: new GlobalNamespace("RegExp") })
     globalScope.set("Map", { mutable: false, value: new GlobalNamespace("Map") })
@@ -1454,10 +1485,23 @@ export class Interpreter<R> {
       throw new InterpreterRuntimeError(`Unsupported update operator '${operator}'.`, node)
     }
 
+    // CodeMode numeric coercion, not host Number(): null-prototype data objects would make
+    // the host throw during ToPrimitive, and opaque runtime references must reject clearly.
+    const operand = (current: unknown): number => {
+      if (containsOpaqueReference(current)) {
+        throw new InterpreterRuntimeError(
+          `'${operator}' requires a data value in CodeMode.`,
+          argument,
+          "InvalidDataValue",
+        )
+      }
+      return coerceToNumber(current)
+    }
+
     if (argument.type === "Identifier") {
       return Effect.sync(() => {
         const name = getString(argument, "name")
-        const current = Number(this.scopes.get(name, argument))
+        const current = operand(this.scopes.get(name, argument))
         const next = current + increment
         this.scopes.set(name, next, argument)
         return prefix ? next : current
@@ -1466,7 +1510,7 @@ export class Interpreter<R> {
 
     if (argument.type === "MemberExpression") {
       return this.modifyMember(argument, (current) => {
-        const value = Number(current)
+        const value = operand(current)
         const next = value + increment
         return Effect.succeed({ write: true, next, result: prefix ? next : value })
       })
@@ -1563,6 +1607,9 @@ export class Interpreter<R> {
         callable.settle(args[0])
         return undefined
       }
+      if (callable === undefined || callable === null) {
+        throw new InterpreterRuntimeError(`${calleeDescription(callee)} is not a function.`, callee).as("TypeError")
+      }
       throw new InterpreterRuntimeError("Only tools are callable in CodeMode.", callee)
     })
   }
@@ -1833,16 +1880,18 @@ export class Interpreter<R> {
       }
 
       if (objectValue instanceof GlobalNamespace) {
-        if (typeof key !== "string" || isBlockedMember(key)) {
-          throw new InterpreterRuntimeError(
-            `${objectValue.name}.${String(key)} is not available in CodeMode.`,
-            propertyNode,
-          )
+        if (typeof key === "string" && isBlockedMember(key)) {
+          throw new InterpreterRuntimeError(`${objectValue.name}.${key} is not available in CodeMode.`, propertyNode)
         }
+        if (typeof key !== "string") return new ComputedValue(undefined)
         if (objectValue.name === "Math" && mathConstants.has(key)) {
           return new ComputedValue((Math as unknown as Record<string, number>)[key])
         }
-        return new GlobalMethodReference(objectValue.name, key)
+        if (globalStaticMembers[objectValue.name]?.has(key)) {
+          return new GlobalMethodReference(objectValue.name, key)
+        }
+        // Unknown static members read as undefined so feature detection works like native JS.
+        return new ComputedValue(undefined)
       }
 
       if (typeof objectValue === "string") {
@@ -1858,12 +1907,21 @@ export class Interpreter<R> {
         return new ComputedValue(undefined)
       }
 
-      if (objectValue instanceof CoercionFunction && typeof key === "string" && !isBlockedMember(key)) {
+      if (objectValue instanceof CoercionFunction) {
+        if (typeof key === "string" && isBlockedMember(key)) {
+          throw new InterpreterRuntimeError(`${objectValue.name}.${key} is not available in CodeMode.`, propertyNode)
+        }
+        if (typeof key !== "string") return new ComputedValue(undefined)
         if (objectValue.name === "Number" && numberConstants.has(key)) {
           return new ComputedValue((Number as unknown as Record<string, number>)[key])
         }
-        if (objectValue.name === "Number" && numberStatics.has(key)) return new GlobalMethodReference("Number", key)
-        if (objectValue.name === "String" && stringStatics.has(key)) return new GlobalMethodReference("String", key)
+        if (objectValue.name === "Number" && numberStatics.has(key)) {
+          return new GlobalMethodReference("Number", key)
+        }
+        if (objectValue.name === "String" && stringStatics.has(key)) {
+          return new GlobalMethodReference("String", key)
+        }
+        return new ComputedValue(undefined)
       }
 
       if (objectValue instanceof CodeModeDate) {

packages/codemode/src/stdlib/date.ts

@@ -23,6 +23,8 @@ export const dateMethods = new Set([
   "getTimezoneOffset",
 ])
 
+export const dateStatics = new Set(["now", "parse", "UTC"])
+
 export const invokeDateStatic = (name: string, args: Array<unknown>, node: AstNode): number => {
   switch (name) {
     case "now":

packages/codemode/src/stdlib/json.ts

@@ -2,6 +2,8 @@ import { type AstNode, InterpreterRuntimeError, supportedSyntaxMessage } from ".
 import { typeofValue } from "../interpreter/references.js"
 import { copyIn, copyOut } from "../tool-runtime.js"
 
+export const jsonStatics = new Set(["parse", "stringify"])
+
 export const invokeJsonMethod = (name: string, args: Array<unknown>, node: AstNode): unknown => {
   switch (name) {
     case "stringify": {

packages/codemode/src/stdlib/object.ts

@@ -6,6 +6,8 @@ import { boundedData, coerceToString } from "./value.js"
 
 export const objectMethodsPreservingIdentity = new Set(["assign", "values", "entries", "fromEntries"])
 
+export const objectStatics = new Set(["keys", "values", "entries", "hasOwn", "is", "assign", "fromEntries"])
+
 export const invokeObjectMethod = (name: string, args: Array<unknown>, node: AstNode): unknown => {
   const requireObject = (): Record<string, unknown> => {
     const input = args[0]

packages/codemode/src/stdlib/regexp.ts

@@ -19,6 +19,8 @@ export const escapeRegexHint =
   'To match special characters like ( ) [ ] { } + * ? . literally, escape them with a backslash (e.g. "\\\\(") or test for them with String.includes instead.'
 
 export const toHostRegex = (arg: unknown, method: string, node: AstNode, extraFlags = ""): RegExp => {
+  // Native parity: an undefined pattern behaves as an empty pattern.
+  if (arg === undefined) return new RegExp("", extraFlags)
   if (arg instanceof CodeModeRegExp) return arg.regex
   if (typeof arg === "string") {
     try {