Fix false positives for symbols in regexp/prefer-quantifier rule (#25)

Author: ota-meshi

lib/rules/prefer-quantifier.ts

@@ -7,6 +7,7 @@ import {
     getRegexpRange,
     isDigit,
     isLetter,
+    isSymbol,
 } from "../utils"
 
 class CharBuffer {
@@ -73,14 +74,16 @@ class CharBuffer {
         if (this.elements.length < 2) {
             return true
         }
-        let charKind: "digit" | "letter" | null = null
+        let charKind: "digit" | "letter" | "symbol" | null = null
         for (const element of this.elements) {
             if (element.type === "Character") {
                 if (charKind == null) {
                     if (isDigit(element.value)) {
                         charKind = "digit"
                     } else if (isLetter(element.value)) {
                         charKind = "letter"
+                    } else if (isSymbol(element.value)) {
+                        charKind = "symbol"
                     } else {
                         return false
                     }
@@ -93,7 +96,9 @@ class CharBuffer {
             // It is valid when the same numbers are consecutive.
             charKind === "digit" ||
             // It is valid when the same letter character continues twice.
-            (charKind === "letter" && this.elements.length <= 2)
+            (charKind === "letter" && this.elements.length <= 2) ||
+            // It is valid when the same symbol character continues three times.
+            (charKind === "symbol" && this.elements.length <= 3)
         ) {
             return true
         }

lib/utils/unicode.ts

@@ -4,6 +4,14 @@ export const CP_VT = 11
 export const CP_FF = 12
 export const CP_CR = 13
 export const CP_SPACE = " ".codePointAt(0)!
+export const CP_BAN = "!".codePointAt(0)!
+export const CP_SLASH = "/".codePointAt(0)!
+export const CP_COLON = ":".codePointAt(0)!
+export const CP_AT = "@".codePointAt(0)!
+export const CP_OPENING_BRACE = "[".codePointAt(0)!
+export const CP_BACKTICK = "`".codePointAt(0)!
+export const CP_OPENING_BRACKET = "{".codePointAt(0)!
+export const CP_TILDE = "~".codePointAt(0)!
 export const CP_NEL = "\u0085".codePointAt(0)!
 export const CP_NBSP = "\u00a0".codePointAt(0)!
 export const CP_OGHAM_SPACE_MARK = "\u1680".codePointAt(0)!
@@ -91,6 +99,19 @@ export function isLetter(codePoint: number): boolean {
         isCodePointInRange(codePoint, CP_RANGE_CAPITAL_LETTER)
     )
 }
+/**
+ * Checks if the given code point is symbol.
+ * @param codePoint The code point to check
+ * @returns {boolean} `true` if the given code point is symbol.
+ */
+export function isSymbol(codePoint: number): boolean {
+    return (
+        isCodePointInRange(codePoint, [CP_BAN, CP_SLASH]) ||
+        isCodePointInRange(codePoint, [CP_COLON, CP_AT]) ||
+        isCodePointInRange(codePoint, [CP_OPENING_BRACE, CP_BACKTICK]) ||
+        isCodePointInRange(codePoint, [CP_OPENING_BRACKET, CP_TILDE])
+    )
+}
 
 /**
  * Checks if the given code point is space.

tests/lib/rules/prefer-quantifier.ts

@@ -9,7 +9,15 @@ const tester = new RuleTester({
 })
 
 tester.run("prefer-quantifier", rule as any, {
-    valid: [`/regexp/`, `/\\d_\\d/`, `/Google/`, `/2000/`],
+    valid: [
+        `/regexp/`,
+        `/\\d_\\d/`,
+        `/Google/`,
+        `/2000/`,
+        `/<!--/`,
+        `/\`\`\`/`,
+        `/---/`,
+    ],
     invalid: [
         {
             code: `/\\d\\d/`,
@@ -25,8 +33,8 @@ tester.run("prefer-quantifier", rule as any, {
             ],
         },
         {
-            code: `/__/`,
-            output: `/_{2}/`,
+            code: `/  /`,
+            output: `/ {2}/`,
             errors: [
                 {
                     message:
@@ -73,10 +81,10 @@ tester.run("prefer-quantifier", rule as any, {
             ],
         },
         {
-            code: String.raw`/__..\s\s\S\S\p{ASCII}\p{ASCII}/u`,
-            output: String.raw`/_{2}..\s{2}\S\S\p{ASCII}{2}/u`,
+            code: String.raw`/aaa..\s\s\S\S\p{ASCII}\p{ASCII}/u`,
+            output: String.raw`/a{3}..\s{2}\S\S\p{ASCII}{2}/u`,
             errors: [
-                'Unexpected consecutive same characters. Use "{2}" instead.',
+                'Unexpected consecutive same characters. Use "{3}" instead.',
                 'Unexpected consecutive same character sets. Use "{2}" instead.',
                 'Unexpected consecutive same character sets. Use "{2}" instead.',
                 'Unexpected consecutive same character sets. Use "{2}" instead.',