feat: add support for ES2025 duplicate named capturing groups

BREAKING CHANGE: add support for ES2025 duplicate named capturing groups

Author: ota-meshi

src/ast.ts

@@ -432,7 +432,7 @@ export interface Backreference extends NodeBase {
     type: "Backreference"
     parent: Alternative | Quantifier
     ref: number | string
-    resolved: CapturingGroup
+    resolved: CapturingGroup[]
 }
 
 /**

src/ecma-versions.ts

@@ -10,4 +10,5 @@ export type EcmaVersion =
     | 2022
     | 2023
     | 2024
-export const latestEcmaVersion = 2024
+    | 2025
+export const latestEcmaVersion = 2025

src/group-specifiers.ts

@@ -0,0 +1,104 @@
+/**
+ * Holds information for all GroupSpecifiers included in the pattern.
+ */
+export interface GroupSpecifiers {
+    /**
+     * @returns true if there are no GroupSpecifiers included in the pattern.
+     */
+    isEmpty: () => boolean
+    clear: () => void
+    /**
+     * Called when visiting the Alternative.
+     * For ES2025, manage nesting with new Alternative scopes.
+     */
+    enterAlternative: () => void
+    /**
+     * Called when leaving the Alternative.
+     */
+    leaveAlternative: () => void
+    /**
+     * Checks whether the given group name is within the pattern.
+     */
+    hasInPattern: (name: string) => boolean
+    /**
+     * Checks whether the given group name is within the current scope.
+     */
+    hasInScope: (name: string) => boolean
+    /**
+     * Adds the given group name to the current scope.
+     */
+    addToScope: (name: string) => void
+}
+
+export class GroupSpecifiersAsES2018 implements GroupSpecifiers {
+    private groupName = new Set<string>()
+
+    public clear(): void {
+        this.groupName.clear()
+    }
+
+    public isEmpty(): boolean {
+        return !this.groupName.size
+    }
+
+    public hasInPattern(name: string): boolean {
+        return this.groupName.has(name)
+    }
+
+    public hasInScope(name: string): boolean {
+        return this.hasInPattern(name)
+    }
+
+    public addToScope(name: string): void {
+        this.groupName.add(name)
+    }
+
+    // eslint-disable-next-line class-methods-use-this
+    public enterAlternative(): void {
+        // Prior to ES2025, it does not manage alternative scopes.
+    }
+
+    // eslint-disable-next-line class-methods-use-this
+    public leaveAlternative(): void {
+        // Prior to ES2025, it does not manage alternative scopes.
+    }
+}
+
+export class GroupSpecifiersAsES2025 implements GroupSpecifiers {
+    private groupNamesInAlternative = new Set<string>()
+    private upperGroupNamesStack: Set<string>[] = []
+
+    private groupNamesInPattern = new Set<string>()
+
+    public clear(): void {
+        this.groupNamesInAlternative.clear()
+        this.upperGroupNamesStack.length = 0
+        this.groupNamesInPattern.clear()
+    }
+
+    public isEmpty(): boolean {
+        return !this.groupNamesInPattern.size
+    }
+
+    public enterAlternative(): void {
+        this.upperGroupNamesStack.push(this.groupNamesInAlternative)
+        this.groupNamesInAlternative = new Set(this.groupNamesInAlternative)
+    }
+
+    public leaveAlternative(): void {
+        this.groupNamesInAlternative = this.upperGroupNamesStack.pop()!
+    }
+
+    public hasInPattern(name: string): boolean {
+        return this.groupNamesInPattern.has(name)
+    }
+
+    public hasInScope(name: string): boolean {
+        return this.groupNamesInAlternative.has(name)
+    }
+
+    public addToScope(name: string): void {
+        this.groupNamesInAlternative.add(name)
+        this.groupNamesInPattern.add(name)
+    }
+}

src/parser.ts

@@ -36,7 +36,6 @@ type AppendableNode =
 
 const DUMMY_PATTERN: Pattern = {} as Pattern
 const DUMMY_FLAGS: Flags = {} as Flags
-const DUMMY_CAPTURING_GROUP: CapturingGroup = {} as CapturingGroup
 
 function isClassSetOperand(
     node: UnicodeSetsCharacterClassElement,
@@ -148,12 +147,12 @@ class RegExpParserState {
 
         for (const reference of this._backreferences) {
             const ref = reference.ref
-            const group =
-                typeof ref === "number"
-                    ? this._capturingGroups[ref - 1]
-                    : this._capturingGroups.find((g) => g.name === ref)!
-            reference.resolved = group
-            group.references.push(reference)
+            for (const group of typeof ref === "number"
+                ? [this._capturingGroups[ref - 1]]
+                : this._capturingGroups.filter((g) => g.name === ref)) {
+                reference.resolved.push(group)
+                group.references.push(reference)
+            }
         }
     }
 
@@ -480,7 +479,7 @@ class RegExpParserState {
             end,
             raw: this.source.slice(start, end),
             ref,
-            resolved: DUMMY_CAPTURING_GROUP,
+            resolved: [],
         }
         parent.elements.push(node)
         this._backreferences.push(node)
@@ -747,14 +746,15 @@ export namespace RegExpParser {
         strict?: boolean
 
         /**
-         * ECMAScript version. Default is `2024`.
+         * ECMAScript version. Default is `2025`.
          * - `2015` added `u` and `y` flags.
          * - `2018` added `s` flag, Named Capturing Group, Lookbehind Assertion,
          *   and Unicode Property Escape.
          * - `2019`, `2020`, and `2021` added more valid Unicode Property Escapes.
          * - `2022` added `d` flag.
          * - `2023` added more valid Unicode Property Escapes.
          * - `2024` added `v` flag.
+         * - `2025` added duplicate named capturing groups.
          */
         ecmaVersion?: EcmaVersion
     }

src/validator.ts

@@ -1,5 +1,10 @@
 import type { EcmaVersion } from "./ecma-versions"
 import { latestEcmaVersion } from "./ecma-versions"
+import type { GroupSpecifiers } from "./group-specifiers"
+import {
+    GroupSpecifiersAsES2018,
+    GroupSpecifiersAsES2025,
+} from "./group-specifiers"
 import { Reader } from "./reader"
 import { newRegExpSyntaxError } from "./regexp-syntax-error"
 import {
@@ -231,14 +236,15 @@ export namespace RegExpValidator {
         strict?: boolean
 
         /**
-         * ECMAScript version. Default is `2024`.
+         * ECMAScript version. Default is `2025`.
          * - `2015` added `u` and `y` flags.
          * - `2018` added `s` flag, Named Capturing Group, Lookbehind Assertion,
          *   and Unicode Property Escape.
          * - `2019`, `2020`, and `2021` added more valid Unicode Property Escapes.
          * - `2022` added `d` flag.
          * - `2023` added more valid Unicode Property Escapes.
          * - `2024` added `v` flag.
+         * - `2025` added duplicate named capturing groups.
          */
         ecmaVersion?: EcmaVersion
 
@@ -631,7 +637,7 @@ export class RegExpValidator {
 
     private _numCapturingParens = 0
 
-    private _groupNames = new Set<string>()
+    private _groupSpecifiers: GroupSpecifiers
 
     private _backreferenceNames = new Set<string>()
 
@@ -643,6 +649,10 @@ export class RegExpValidator {
      */
     public constructor(options?: RegExpValidator.Options) {
         this._options = options ?? {}
+        this._groupSpecifiers =
+            this.ecmaVersion >= 2025
+                ? new GroupSpecifiersAsES2025()
+                : new GroupSpecifiersAsES2018()
     }
 
     /**
@@ -763,7 +773,7 @@ export class RegExpValidator {
         if (
             !this._nFlag &&
             this.ecmaVersion >= 2018 &&
-            this._groupNames.size > 0
+            !this._groupSpecifiers.isEmpty()
         ) {
             this._nFlag = true
             this.rewind(start)
@@ -1301,7 +1311,7 @@ export class RegExpValidator {
     private consumePattern(): void {
         const start = this.index
         this._numCapturingParens = this.countCapturingParens()
-        this._groupNames.clear()
+        this._groupSpecifiers.clear()
         this._backreferenceNames.clear()
 
         this.onPatternEnter(start)
@@ -1322,7 +1332,7 @@ export class RegExpValidator {
             this.raise(`Unexpected character '${c}'`)
         }
         for (const name of this._backreferenceNames) {
-            if (!this._groupNames.has(name)) {
+            if (!this._groupSpecifiers.hasInPattern(name)) {
                 this.raise("Invalid named capture referenced")
             }
         }
@@ -1380,7 +1390,9 @@ export class RegExpValidator {
 
         this.onDisjunctionEnter(start)
         do {
+            this._groupSpecifiers.enterAlternative()
             this.consumeAlternative(i++)
+            this._groupSpecifiers.leaveAlternative()
         } while (this.eat(VERTICAL_LINE))
 
         if (this.consumeQuantifier(true)) {
@@ -1846,8 +1858,8 @@ export class RegExpValidator {
     private consumeGroupSpecifier(): boolean {
         if (this.eat(QUESTION_MARK)) {
             if (this.eatGroupName()) {
-                if (!this._groupNames.has(this._lastStrValue)) {
-                    this._groupNames.add(this._lastStrValue)
+                if (!this._groupSpecifiers.hasInScope(this._lastStrValue)) {
+                    this._groupSpecifiers.addToScope(this._lastStrValue)
                     return true
                 }
                 this.raise("Duplicate capture group name")

test/fixtures/parser/literal/basic-valid-2015-u.json

@@ -1690,7 +1690,9 @@
                   "end": 6,
                   "raw": "\\1",
                   "ref": 1,
-                  "resolved": "♻️../0"
+                  "resolved": [
+                    "♻️../0"
+                  ]
                 }
               ]
             }
@@ -1741,7 +1743,9 @@
                   "end": 3,
                   "raw": "\\1",
                   "ref": 1,
-                  "resolved": "♻️../1"
+                  "resolved": [
+                    "♻️../1"
+                  ]
                 },
                 {
                   "type": "CapturingGroup",
@@ -2104,7 +2108,9 @@
                   "end": 34,
                   "raw": "\\10",
                   "ref": 10,
-                  "resolved": "♻️../9"
+                  "resolved": [
+                    "♻️../9"
+                  ]
                 }
               ]
             }
@@ -2465,7 +2471,9 @@
                   "end": 37,
                   "raw": "\\11",
                   "ref": 11,
-                  "resolved": "♻️../10"
+                  "resolved": [
+                    "♻️../10"
+                  ]
                 }
               ]
             }

test/fixtures/parser/literal/basic-valid-2015.json

@@ -3800,7 +3800,9 @@
                   "end": 6,
                   "raw": "\\1",
                   "ref": 1,
-                  "resolved": "♻️../0"
+                  "resolved": [
+                    "♻️../0"
+                  ]
                 }
               ]
             }
@@ -3851,7 +3853,9 @@
                   "end": 3,
                   "raw": "\\1",
                   "ref": 1,
-                  "resolved": "♻️../1"
+                  "resolved": [
+                    "♻️../1"
+                  ]
                 },
                 {
                   "type": "CapturingGroup",
@@ -4444,7 +4448,9 @@
                   "end": 34,
                   "raw": "\\10",
                   "ref": 10,
-                  "resolved": "♻️../9"
+                  "resolved": [
+                    "♻️../9"
+                  ]
                 }
               ]
             }
@@ -5135,7 +5141,9 @@
                   "end": 37,
                   "raw": "\\11",
                   "ref": 11,
-                  "resolved": "♻️../10"
+                  "resolved": [
+                    "♻️../10"
+                  ]
                 }
               ]
             }