feat: add support v flag (#82)

Author: ota-meshi

README.md

@@ -79,15 +79,15 @@ Parse a regular expression literal.
 - **Return:**
     - The AST of the regular expression.
 
-#### parser.parsePattern(source, start?, end?, uFlag?)
+#### parser.parsePattern(source, start?, end?, flags?)
 
 Parse a regular expression pattern.
 
 - **Parameters:**
     - `source` (`string`) The source code to parse. E.g. `"abc"`.
     - `start?` (`number`) The start index in the source code. Default is `0`.
     - `end?` (`number`) The end index in the source code. Default is `source.length`.
-    - `uFlag?` (`boolean`) The flag to enable Unicode mode.
+    - `flags?` (`{ unicode?: boolean, unicodeSets?: boolean }`) The flags to enable Unicode mode, and Unicode Set mode.
 - **Return:**
     - The AST of the regular expression pattern.
 
@@ -118,15 +118,15 @@ Validate a regular expression literal.
     - `start?` (`number`) The start index in the source code. Default is `0`.
     - `end?` (`number`) The end index in the source code. Default is `source.length`.
 
-#### validator.validatePattern(source, start, end, uFlag)
+#### validator.validatePattern(source, start, end, flags)
 
 Validate a regular expression pattern.
 
 - **Parameters:**
     - `source` (`string`) The source code to validate.
     - `start?` (`number`) The start index in the source code. Default is `0`.
     - `end?` (`number`) The end index in the source code. Default is `source.length`.
-    - `uFlag?` (`boolean`) The flag to enable Unicode mode.
+    - `flags?` (`{ unicode?: boolean, unicodeSets?: boolean }`) The flags to enable Unicode mode, and Unicode Set mode.
 
 #### validator.validateFlags(source, start, end)
 
@@ -172,6 +172,6 @@ Please use GitHub's Issues/PRs.
 - `npm run watch` runs tests with `--watch` option.
 
 [`AST.Node`]: src/ast.ts#L4
-[`RegExpParser.Options`]: src/parser.ts#L539
-[`RegExpValidator.Options`]: src/validator.ts#L127
-[`RegExpVisitor.Handlers`]: src/visitor.ts#L204
+[`RegExpParser.Options`]: src/parser.ts#L743
+[`RegExpValidator.Options`]: src/validator.ts#L220
+[`RegExpVisitor.Handlers`]: src/visitor.ts#L291

scripts/update-fixtures.ts

@@ -47,11 +47,16 @@ for (const filename of Object.keys(Visitor.fixturesData)) {
             onCharacterClassEnter: enter,
             onCharacterClassRangeEnter: enter,
             onCharacterSetEnter: enter,
+            onClassIntersectionEnter: enter,
+            onClassStringDisjunctionEnter: enter,
+            onClassSubtractionEnter: enter,
+            onExpressionCharacterClassEnter: enter,
             onFlagsEnter: enter,
             onGroupEnter: enter,
             onPatternEnter: enter,
             onQuantifierEnter: enter,
             onRegExpLiteralEnter: enter,
+            onStringAlternativeEnter: enter,
             onAlternativeLeave: leave,
             onAssertionLeave: leave,
             onBackreferenceLeave: leave,
@@ -60,11 +65,16 @@ for (const filename of Object.keys(Visitor.fixturesData)) {
             onCharacterClassLeave: leave,
             onCharacterClassRangeLeave: leave,
             onCharacterSetLeave: leave,
+            onClassIntersectionLeave: leave,
+            onClassStringDisjunctionLeave: leave,
+            onClassSubtractionLeave: leave,
+            onExpressionCharacterClassLeave: leave,
             onFlagsLeave: leave,
             onGroupLeave: leave,
             onPatternLeave: leave,
             onQuantifierLeave: leave,
             onRegExpLiteralLeave: leave,
+            onStringAlternativeLeave: leave,
         })
 
         fixture.patterns[pattern] = history

src/ast.ts

@@ -11,11 +11,16 @@ export type BranchNode =
     | CapturingGroup
     | CharacterClass
     | CharacterClassRange
+    | ClassIntersection
+    | ClassStringDisjunction
+    | ClassSubtraction
+    | ExpressionCharacterClass
     | Group
     | LookaroundAssertion
     | Pattern
     | Quantifier
     | RegExpLiteral
+    | StringAlternative
 
 /**
  * The type which includes all leaf nodes.
@@ -41,17 +46,29 @@ export type QuantifiableElement =
     | Character
     | CharacterClass
     | CharacterSet
+    | ExpressionCharacterClass
     | Group
     | LookaheadAssertion
 
 /**
  * The type which includes all character class atom nodes.
  */
 export type CharacterClassElement =
+    | ClassRangesCharacterClassElement
+    | UnicodeSetsCharacterClassElement
+export type ClassRangesCharacterClassElement =
     | Character
     | CharacterClassRange
     | EscapeCharacterSet
     | UnicodePropertyCharacterSet
+export type UnicodeSetsCharacterClassElement =
+    | Character
+    | CharacterClassRange
+    | ClassStringDisjunction
+    | EscapeCharacterSet
+    | ExpressionCharacterClass
+    | UnicodePropertyCharacterSet
+    | UnicodeSetsCharacterClass
 
 /**
  * The type which defines common properties for all node types.
@@ -166,12 +183,35 @@ export interface Quantifier extends NodeBase {
  * The character class.
  * E.g. `[ab]`, `[^ab]`
  */
-export interface CharacterClass extends NodeBase {
+export type CharacterClass =
+    | ClassRangesCharacterClass
+    | UnicodeSetsCharacterClass
+interface BaseCharacterClass extends NodeBase {
     type: "CharacterClass"
-    parent: Alternative | Quantifier
+    parent:
+        | Alternative
+        | ExpressionCharacterClass
+        | Quantifier
+        | UnicodeSetsCharacterClass
+    unicodeSets: boolean
     negate: boolean
     elements: CharacterClassElement[]
 }
+export interface ClassRangesCharacterClass extends BaseCharacterClass {
+    parent: Alternative | Quantifier
+    unicodeSets: false
+    elements: ClassRangesCharacterClassElement[]
+}
+/** UnicodeSetsCharacterClass is the CharacterClass when in Unicode sets mode. So it may contain strings. */
+export interface UnicodeSetsCharacterClass extends BaseCharacterClass {
+    parent:
+        | Alternative
+        | ExpressionCharacterClass
+        | Quantifier
+        | UnicodeSetsCharacterClass
+    unicodeSets: true
+    elements: UnicodeSetsCharacterClassElement[]
+}
 
 /**
  * The character class.
@@ -239,7 +279,12 @@ export interface AnyCharacterSet extends NodeBase {
  */
 export interface EscapeCharacterSet extends NodeBase {
     type: "CharacterSet"
-    parent: Alternative | CharacterClass | Quantifier
+    parent:
+        | Alternative
+        | CharacterClass
+        | ClassIntersection
+        | ClassSubtraction
+        | Quantifier
     kind: "digit" | "space" | "word"
     negate: boolean
 }
@@ -248,14 +293,98 @@ export interface EscapeCharacterSet extends NodeBase {
  * The unicode property escape.
  * E.g. `\p{ASCII}`, `\P{ASCII}`, `\p{Script=Hiragana}`
  */
-export interface UnicodePropertyCharacterSet extends NodeBase {
+export type UnicodePropertyCharacterSet =
+    | CharacterUnicodePropertyCharacterSet
+    | StringsUnicodePropertyCharacterSet
+interface BaseUnicodePropertyCharacterSet extends NodeBase {
     type: "CharacterSet"
-    parent: Alternative | CharacterClass | Quantifier
+    parent:
+        | Alternative
+        | CharacterClass
+        | ClassIntersection
+        | ClassSubtraction
+        | Quantifier
     kind: "property"
+    strings: boolean
     key: string
     value: string | null
     negate: boolean
 }
+export interface CharacterUnicodePropertyCharacterSet
+    extends BaseUnicodePropertyCharacterSet {
+    strings: false
+    value: string | null
+    negate: boolean
+}
+/** StringsUnicodePropertyCharacterSet is Unicode property escape with property of strings. */
+export interface StringsUnicodePropertyCharacterSet
+    extends BaseUnicodePropertyCharacterSet {
+    strings: true
+    value: null
+    negate: false
+}
+
+/**
+ * The expression character class.
+ * E.g. `[a--b]`, `[a&&b]`,`[^a--b]`, `[^a&&b]`
+ */
+export interface ExpressionCharacterClass extends NodeBase {
+    type: "ExpressionCharacterClass"
+    parent:
+        | Alternative
+        | ExpressionCharacterClass
+        | Quantifier
+        | UnicodeSetsCharacterClass
+    negate: boolean
+    expression: ClassIntersection | ClassSubtraction
+}
+
+export type ClassSetOperand =
+    | Character
+    | ClassStringDisjunction
+    | EscapeCharacterSet
+    | ExpressionCharacterClass
+    | UnicodePropertyCharacterSet
+    | UnicodeSetsCharacterClass
+
+/**
+ * The character class intersection.
+ * E.g. `a&&b`
+ */
+export interface ClassIntersection extends NodeBase {
+    type: "ClassIntersection"
+    parent: ClassIntersection | ExpressionCharacterClass
+    left: ClassIntersection | ClassSetOperand
+    right: ClassSetOperand
+}
+
+/**
+ * The character class subtraction.
+ * E.g. `a--b`
+ */
+export interface ClassSubtraction extends NodeBase {
+    type: "ClassSubtraction"
+    parent: ClassSubtraction | ExpressionCharacterClass
+    left: ClassSetOperand | ClassSubtraction
+    right: ClassSetOperand
+}
+
+/**
+ * The character class string disjunction.
+ * E.g. `\q{a|b}`
+ */
+export interface ClassStringDisjunction extends NodeBase {
+    type: "ClassStringDisjunction"
+    parent: ClassIntersection | ClassSubtraction | UnicodeSetsCharacterClass
+    alternatives: StringAlternative[]
+}
+
+/** StringAlternative is only used for `\q{alt}`({@link ClassStringDisjunction}). */
+export interface StringAlternative extends NodeBase {
+    type: "StringAlternative"
+    parent: ClassStringDisjunction
+    elements: Character[]
+}
 
 /**
  * The character.
@@ -264,7 +393,14 @@ export interface UnicodePropertyCharacterSet extends NodeBase {
  */
 export interface Character extends NodeBase {
     type: "Character"
-    parent: Alternative | CharacterClass | CharacterClassRange | Quantifier
+    parent:
+        | Alternative
+        | CharacterClass
+        | CharacterClassRange
+        | ClassIntersection
+        | ClassSubtraction
+        | Quantifier
+        | StringAlternative
     value: number // a code point.
 }
 
@@ -292,4 +428,5 @@ export interface Flags extends NodeBase {
     multiline: boolean
     sticky: boolean
     unicode: boolean
+    unicodeSets: boolean
 }

src/ecma-versions.ts

@@ -9,3 +9,5 @@ export type EcmaVersion =
     | 2021
     | 2022
     | 2023
+    | 2024
+export const latestEcmaVersion = 2024