feat: require-unicode-regexp add suggestions (#17007)

* feat: `require-unicode-regexp` add suggestions

* Review fixups: invalid patterns; sequence expressions

* I promise I know how CJS modules work

* Handled non-string literals

* Update lib/rules/require-unicode-regexp.js

Co-authored-by: Milos Djermanovic <milos.djermanovic@gmail.com>

* Don't concatenate + 'u'

* Test the new cases

* Touch up regular-expressions.js templating

* Add myself as author to regular-expressions.js

---------

Co-authored-by: Milos Djermanovic <milos.djermanovic@gmail.com>

lib/rules/no-misleading-character-class.js

@@ -4,16 +4,15 @@
 "use strict";
 
 const { CALL, CONSTRUCT, ReferenceTracker, getStringIfConstant } = require("@eslint-community/eslint-utils");
-const { RegExpValidator, RegExpParser, visitRegExpAST } = require("@eslint-community/regexpp");
+const { RegExpParser, visitRegExpAST } = require("@eslint-community/regexpp");
 const { isCombiningCharacter, isEmojiModifier, isRegionalIndicatorSymbol, isSurrogatePair } = require("./utils/unicode");
 const astUtils = require("./utils/ast-utils.js");
+const { isValidWithUnicodeFlag } = require("./utils/regular-expressions");
 
 //------------------------------------------------------------------------------
 // Helpers
 //------------------------------------------------------------------------------
 
-const REGEXPP_LATEST_ECMA_VERSION = 2022;
-
 /**
  * Iterate character sequences of a given nodes.
  *
@@ -185,38 +184,10 @@ module.exports = {
             }
         }
 
-        /**
-         * Checks if the given regular expression pattern would be valid with the `u` flag.
-         * @param {string} pattern The regular expression pattern to verify.
-         * @returns {boolean} `true` if the pattern would be valid with the `u` flag.
-         * `false` if the pattern would be invalid with the `u` flag or the configured
-         * ecmaVersion doesn't support the `u` flag.
-         */
-        function isValidWithUnicodeFlag(pattern) {
-            const { ecmaVersion } = context.languageOptions;
-
-            // ecmaVersion <= 5 doesn't support the 'u' flag
-            if (ecmaVersion <= 5) {
-                return false;
-            }
-
-            const validator = new RegExpValidator({
-                ecmaVersion: Math.min(ecmaVersion, REGEXPP_LATEST_ECMA_VERSION)
-            });
-
-            try {
-                validator.validatePattern(pattern, void 0, void 0, /* uFlag = */ true);
-            } catch {
-                return false;
-            }
-
-            return true;
-        }
-
         return {
             "Literal[regex]"(node) {
                 verify(node, node.regex.pattern, node.regex.flags, fixer => {
-                    if (!isValidWithUnicodeFlag(node.regex.pattern)) {
+                    if (!isValidWithUnicodeFlag(context.languageOptions.ecmaVersion, node.regex.pattern)) {
                         return null;
                     }
 
@@ -242,7 +213,7 @@ module.exports = {
                     if (typeof pattern === "string") {
                         verify(refNode, pattern, flags || "", fixer => {
 
-                            if (!isValidWithUnicodeFlag(pattern)) {
+                            if (!isValidWithUnicodeFlag(context.languageOptions.ecmaVersion, pattern)) {
                                 return null;
                             }
 

lib/rules/prefer-regex-literals.js

@@ -13,13 +13,12 @@ const astUtils = require("./utils/ast-utils");
 const { CALL, CONSTRUCT, ReferenceTracker, findVariable } = require("@eslint-community/eslint-utils");
 const { RegExpValidator, visitRegExpAST, RegExpParser } = require("@eslint-community/regexpp");
 const { canTokensBeAdjacent } = require("./utils/ast-utils");
+const { REGEXPP_LATEST_ECMA_VERSION } = require("./utils/regular-expressions");
 
 //------------------------------------------------------------------------------
 // Helpers
 //------------------------------------------------------------------------------
 
-const REGEXPP_LATEST_ECMA_VERSION = 2022;
-
 /**
  * Determines whether the given node is a string literal.
  * @param {ASTNode} node Node to check.

lib/rules/require-unicode-regexp.js

@@ -15,6 +15,8 @@ const {
     ReferenceTracker,
     getStringIfConstant
 } = require("@eslint-community/eslint-utils");
+const astUtils = require("./utils/ast-utils.js");
+const { isValidWithUnicodeFlag } = require("./utils/regular-expressions");
 
 //------------------------------------------------------------------------------
 // Rule Definition
@@ -31,7 +33,10 @@ module.exports = {
             url: "https://eslint.org/docs/rules/require-unicode-regexp"
         },
 
+        hasSuggestions: true,
+
         messages: {
+            addUFlag: "Add the 'u' flag.",
             requireUFlag: "Use the 'u' flag."
         },
 
@@ -47,7 +52,20 @@ module.exports = {
                 const flags = node.regex.flags || "";
 
                 if (!flags.includes("u")) {
-                    context.report({ node, messageId: "requireUFlag" });
+                    context.report({
+                        messageId: "requireUFlag",
+                        node,
+                        suggest: isValidWithUnicodeFlag(context.languageOptions.ecmaVersion, node.regex.pattern)
+                            ? [
+                                {
+                                    fix(fixer) {
+                                        return fixer.insertTextAfter(node, "u");
+                                    },
+                                    messageId: "addUFlag"
+                                }
+                            ]
+                            : null
+                    });
                 }
             },
 
@@ -59,11 +77,46 @@ module.exports = {
                 };
 
                 for (const { node: refNode } of tracker.iterateGlobalReferences(trackMap)) {
-                    const flagsNode = refNode.arguments[1];
+                    const [patternNode, flagsNode] = refNode.arguments;
+                    const pattern = getStringIfConstant(patternNode, scope);
                     const flags = getStringIfConstant(flagsNode, scope);
 
                     if (!flagsNode || (typeof flags === "string" && !flags.includes("u"))) {
-                        context.report({ node: refNode, messageId: "requireUFlag" });
+                        context.report({
+                            messageId: "requireUFlag",
+                            node: refNode,
+                            suggest: typeof pattern === "string" && isValidWithUnicodeFlag(context.languageOptions.ecmaVersion, pattern)
+                                ? [
+                                    {
+                                        fix(fixer) {
+                                            if (flagsNode) {
+                                                if ((flagsNode.type === "Literal" && typeof flagsNode.value === "string") || flagsNode.type === "TemplateLiteral") {
+                                                    const flagsNodeText = sourceCode.getText(flagsNode);
+
+                                                    return fixer.replaceText(flagsNode, [
+                                                        flagsNodeText.slice(0, flagsNodeText.length - 1),
+                                                        flagsNodeText.slice(flagsNodeText.length - 1)
+                                                    ].join("u"));
+                                                }
+
+                                                // We intentionally don't suggest concatenating + "u" to non-literals
+                                                return null;
+                                            }
+
+                                            const penultimateToken = sourceCode.getLastToken(refNode, { skip: 1 }); // skip closing parenthesis
+
+                                            return fixer.insertTextAfter(
+                                                penultimateToken,
+                                                astUtils.isCommaToken(penultimateToken)
+                                                    ? ' "u",'
+                                                    : ', "u"'
+                                            );
+                                        },
+                                        messageId: "addUFlag"
+                                    }
+                                ]
+                                : null
+                        });
                     }
                 }
             }

lib/rules/utils/regular-expressions.js

@@ -0,0 +1,42 @@
+/**
+ * @fileoverview Common utils for regular expressions.
+ * @author Josh Goldberg
+ * @author Toru Nagashima
+ */
+
+"use strict";
+
+const { RegExpValidator } = require("@eslint-community/regexpp");
+
+const REGEXPP_LATEST_ECMA_VERSION = 2022;
+
+/**
+ * Checks if the given regular expression pattern would be valid with the `u` flag.
+ * @param {number} ecmaVersion ECMAScript version to parse in.
+ * @param {string} pattern The regular expression pattern to verify.
+ * @returns {boolean} `true` if the pattern would be valid with the `u` flag.
+ * `false` if the pattern would be invalid with the `u` flag or the configured
+ * ecmaVersion doesn't support the `u` flag.
+ */
+function isValidWithUnicodeFlag(ecmaVersion, pattern) {
+    if (ecmaVersion <= 5) { // ecmaVersion <= 5 doesn't support the 'u' flag
+        return false;
+    }
+
+    const validator = new RegExpValidator({
+        ecmaVersion: Math.min(ecmaVersion, REGEXPP_LATEST_ECMA_VERSION)
+    });
+
+    try {
+        validator.validatePattern(pattern, void 0, void 0, /* uFlag = */ true);
+    } catch {
+        return false;
+    }
+
+    return true;
+}
+
+module.exports = {
+    isValidWithUnicodeFlag,
+    REGEXPP_LATEST_ECMA_VERSION
+};