From d4f02e4bf1b9ae4e1fc8f2bc4e4851ae3c36a127 Mon Sep 17 00:00:00 2001
From: Yosuke Ota <otameshiyo23@gmail.com>
Date: Wed, 26 Jul 2023 20:52:17 +0900
Subject: [PATCH] feat: `no-control-regex` support `v` flag (#17405)

* feat: `no-control-regex` support v flag

* test: add testcases

* fix: false positives for `uv` flags
---
 lib/rules/no-control-regex.js       | 17 +++++++++++++++--
 tests/lib/rules/no-control-regex.js | 19 ++++++++++++++++++-
 2 files changed, 33 insertions(+), 3 deletions(-)

diff --git a/lib/rules/no-control-regex.js b/lib/rules/no-control-regex.js
index 086747f3746..dc412fcabd5 100644
--- a/lib/rules/no-control-regex.js
+++ b/lib/rules/no-control-regex.js
@@ -14,6 +14,16 @@ const collector = new (class {
     }
 
     onPatternEnter() {
+
+        /*
+         * `RegExpValidator` may parse the pattern twice in one `validatePattern`.
+         * So `this._controlChars` should be cleared here as well.
+         *
+         * For example, the `/(?<a>\x1f)/` regex will parse the pattern twice.
+         * This is based on the content described in Annex B.
+         * If the regex contains a `GroupName` and the `u` flag is not used, `ParseText` will be called twice.
+         * See https://tc39.es/ecma262/2023/multipage/additional-ecmascript-features-for-web-browsers.html#sec-parsepattern-annexb
+         */
         this._controlChars = [];
     }
 
@@ -32,10 +42,13 @@ const collector = new (class {
 
     collectControlChars(regexpStr, flags) {
         const uFlag = typeof flags === "string" && flags.includes("u");
+        const vFlag = typeof flags === "string" && flags.includes("v");
+
+        this._controlChars = [];
+        this._source = regexpStr;
 
         try {
-            this._source = regexpStr;
-            this._validator.validatePattern(regexpStr, void 0, void 0, uFlag); // Call onCharacter hook
+            this._validator.validatePattern(regexpStr, void 0, void 0, { unicode: uFlag, unicodeSets: vFlag }); // Call onCharacter hook
         } catch {
 
             // Ignore syntax errors in RegExp.
diff --git a/tests/lib/rules/no-control-regex.js b/tests/lib/rules/no-control-regex.js
index 14abfbce450..3bfc87bace1 100644
--- a/tests/lib/rules/no-control-regex.js
+++ b/tests/lib/rules/no-control-regex.js
@@ -33,7 +33,10 @@ ruleTester.run("no-control-regex", rule, {
         String.raw`new RegExp("\\u{20}", "u")`,
         String.raw`new RegExp("\\u{1F}")`,
         String.raw`new RegExp("\\u{1F}", "g")`,
-        String.raw`new RegExp("\\u{1F}", flags)` // when flags are unknown, this rule assumes there's no `u` flag
+        String.raw`new RegExp("\\u{1F}", flags)`, // when flags are unknown, this rule assumes there's no `u` flag
+        String.raw`new RegExp("[\\q{\\u{20}}]", "v")`,
+        { code: String.raw`/[\u{20}--B]/v`, parserOptions: { ecmaVersion: 2024 } }
+
     ],
     invalid: [
         { code: String.raw`var regex = /\x1f/`, errors: [{ messageId: "unexpected", data: { controlChars: "\\x1f" }, type: "Literal" }] },
@@ -85,6 +88,20 @@ ruleTester.run("no-control-regex", rule, {
         {
             code: String.raw`new RegExp("\\u{1F}", "gui")`,
             errors: [{ messageId: "unexpected", data: { controlChars: "\\x1f" }, type: "Literal" }]
+        },
+        {
+            code: String.raw`new RegExp("[\\q{\\u{1F}}]", "v")`,
+            errors: [{ messageId: "unexpected", data: { controlChars: "\\x1f" }, type: "Literal" }]
+        },
+        {
+            code: String.raw`/[\u{1F}--B]/v`,
+            parserOptions: { ecmaVersion: 2024 },
+            errors: [{ messageId: "unexpected", data: { controlChars: "\\x1f" }, type: "Literal" }]
+        },
+        {
+            code: String.raw`/\x11/; RegExp("foo", "uv");`,
+            parserOptions: { ecmaVersion: 2024 },
+            errors: [{ messageId: "unexpected", data: { controlChars: "\\x11" }, type: "Literal", column: 1 }]
         }
     ]
 });