deps: update to cjs-module-lexer@1.2.1

PR-URL: #38450 Backport-PR-URL: #39422 Reviewed-By: Bradley Farias <bradley.meck@gmail.com> Reviewed-By: Antoine du Hamel <duhamelantoine1995@gmail.com> Reviewed-By: Jan Krems <jan.krems@gmail.com> Reviewed-By: Rich Trott <rtrott@gmail.com>
richardlau · Jul 29, 2021 · Mar 6, 2021 · Apr 25, 2021 · May 3, 2021 · Jul 3, 2021
commit b3c698a5d8357152504a100768c2c68ea25b5c5d
diff --git a/deps/cjs-module-lexer/CHANGELOG.md b/deps/cjs-module-lexer/CHANGELOG.md
@@ -1,3 +1,10 @@
+1.2.1
+- Support Unicode escapes in strings (https://github.com/guybedford/cjs-module-lexer/pull/55)
+- Filter export strings to valid surrogate pairs (https://github.com/guybedford/cjs-module-lexer/pull/56)
+
+1.2.0
+- Support for non-identifier exports (https://github.com/guybedford/cjs-module-lexer/pull/54, @nicolo-ribaudo)
+
 1.1.1
 - Better support for Babel reexport getter function forms (https://github.com/guybedford/cjs-module-lexer/issues/50)
 - Support Babel interopRequireWildcard reexports patterns (https://github.com/guybedford/cjs-module-lexer/issues/52)

diff --git a/deps/cjs-module-lexer/README.md b/deps/cjs-module-lexer/README.md
@@ -70,29 +70,27 @@ IDENTIFIER: As defined by ECMA-262, without support for identifier `\` escapes,
 
 STRING_LITERAL: A `"` or `'` bounded ECMA-262 string literal.
 
-IDENTIFIER_STRING: ( `"` IDENTIFIER `"` | `'` IDENTIFIER `'` )
-
 MODULE_EXPORTS: `module` `.` `exports`
 
 EXPORTS_IDENTIFIER: MODULE_EXPORTS_IDENTIFIER | `exports`
 
 EXPORTS_DOT_ASSIGN: EXPORTS_IDENTIFIER `.` IDENTIFIER `=`
 
-EXPORTS_LITERAL_COMPUTED_ASSIGN: EXPORTS_IDENTIFIER `[` IDENTIFIER_STRING `]` `=`
+EXPORTS_LITERAL_COMPUTED_ASSIGN: EXPORTS_IDENTIFIER `[` STRING_LITERAL `]` `=`
 
-EXPORTS_LITERAL_PROP: (IDENTIFIER  (`:` IDENTIFIER)?) | (IDENTIFIER_STRING `:` IDENTIFIER)
+EXPORTS_LITERAL_PROP: (IDENTIFIER  (`:` IDENTIFIER)?) | (STRING_LITERAL `:` IDENTIFIER)
 
 EXPORTS_SPREAD: `...` (IDENTIFIER | REQUIRE)
 
 EXPORTS_MEMBER: EXPORTS_DOT_ASSIGN | EXPORTS_LITERAL_COMPUTED_ASSIGN
 
-EXPORTS_DEFINE: `Object` `.` `defineProperty `(` EXPORTS_IDENFITIER `,` IDENTIFIER_STRING
+EXPORTS_DEFINE: `Object` `.` `defineProperty `(` EXPORTS_IDENFITIER `,` STRING_LITERAL
 
 EXPORTS_DEFINE_VALUE: EXPORTS_DEFINE `, {`
   (`enumerable: true,`)?
   (
     `value:` |
-    `get` (`: function` IDENTIFIER? )?  `() {` return IDENTIFIER (`.` IDENTIFIER | `[` IDENTIFIER_STRING `]`)? `;`? `}` `,`?
+    `get` (`: function` IDENTIFIER? )?  `() {` return IDENTIFIER (`.` IDENTIFIER | `[` STRING_LITERAL `]`)? `;`? `}` `,`?
   )
   `})`
 
@@ -127,8 +125,8 @@ EXPORT_STAR_LIB: `Object.keys(` IDENTIFIER$1 `).forEach(function (` IDENTIFIER$2
 Spacing between tokens is taken to be any ECMA-262 whitespace, ECMA-262 block comment or ECMA-262 line comment.
 
 * The returned export names are taken to be the combination of:
-  1. All `IDENTIFIER` and `IDENTIFIER_STRING` slots for `EXPORTS_MEMBER` and `EXPORTS_LITERAL` matches.
-  2. The first `IDENTIFIER_STRING` slot for all `EXPORTS_DEFINE_VALUE` matches where that same string is not an `EXPORTS_DEFINE` match that is not also an `EXPORTS_DEFINE_VALUE` match.
+  1. All `IDENTIFIER` and `STRING_LITERAL` slots for `EXPORTS_MEMBER` and `EXPORTS_LITERAL` matches.
+  2. The first `STRING_LITERAL` slot for all `EXPORTS_DEFINE_VALUE` matches where that same string is not an `EXPORTS_DEFINE` match that is not also an `EXPORTS_DEFINE_VALUE` match.
 * The reexport specifiers are taken to be the combination of:
   1. The `REQUIRE` matches of the last matched of either `MODULE_EXPORTS_ASSIGN` or `EXPORTS_LITERAL`.
   2. All _top-level_ `EXPORT_STAR` `REQUIRE` matches and `EXPORTS_ASSIGN` matches whose `IDENTIFIER` also matches the first `IDENTIFIER` in `EXPORT_STAR_LIB`.

diff --git a/deps/cjs-module-lexer/dist/lexer.js b/deps/cjs-module-lexer/dist/lexer.js
diff --git a/deps/cjs-module-lexer/dist/lexer.mjs b/deps/cjs-module-lexer/dist/lexer.mjs
diff --git a/deps/cjs-module-lexer/lexer.js b/deps/cjs-module-lexer/lexer.js
@@ -37,8 +37,6 @@ const Import = 0;
 const ExportAssign = 1;
 const ExportStar = 2;
 
-const strictReserved = new Set(['implements', 'interface', 'let', 'package', 'private', 'protected', 'public', 'static', 'yield', 'enum']);
-
 function parseCJS (source, name = '@') {
   resetState();
   try {
@@ -49,14 +47,39 @@ function parseCJS (source, name = '@') {
     e.loc = pos;
     throw e;
   }
-  const result = { exports: [..._exports].filter(expt => !unsafeGetters.has(expt)), reexports: [...reexports] };
+  const result = { exports: [..._exports].filter(expt => expt !== undefined && !unsafeGetters.has(expt)), reexports: [...reexports].filter(reexpt => reexpt !== undefined) };
   resetState();
   return result;
 }
 
-function addExport (name) {
-  if (!strictReserved.has(name))
-    _exports.add(name);
+function decode (str) {
+  if (str[0] === '"' || str[0] === '\'') {
+    try {
+      const decoded = (0, eval)(str);
+      // Filter to exclude non-matching UTF-16 surrogate strings
+      for (let i = 0; i < decoded.length; i++) {
+        const surrogatePrefix = decoded.charCodeAt(i) & 0xFC00;
+        if (surrogatePrefix < 0xD800) {
+          // Not a surrogate
+          continue;
+        }
+        else if (surrogatePrefix === 0xD800) {
+          // Validate surrogate pair
+          if ((decoded.charCodeAt(++i) & 0xFC00) !== 0xDC00)
+            return;
+        }
+        else {
+          // Out-of-range surrogate code (above 0xD800)
+          return;
+        }
+      }
+      return decoded;
+    }
+    catch {}
+  }
+  else {
+    return str;
+  }
 }
 
 function parseSource (cjsSource) {
@@ -173,10 +196,8 @@ function parseSource (cjsSource) {
         // TODO: <!-- XML comment support
         break;
       case 39/*'*/:
-        singleQuoteString();
-        break;
       case 34/*"*/:
-        doubleQuoteString();
+        stringLiteral(ch);
         break;
       case 47/*/*/: {
         const next_ch = source.charCodeAt(pos + 1);
@@ -329,11 +350,9 @@ function tryParseObjectDefineOrKeys (keys) {
         pos++;
         ch = commentWhitespace();
         if (ch !== 39/*'*/ && ch !== 34/*"*/) break;
-        let quot = ch;
-        const exportPos = ++pos;
-        if (!identifier() || source.charCodeAt(pos) !== quot) break;
-        expt = source.slice(exportPos, pos);
-        pos++;
+        const exportPos = pos;
+        stringLiteral(ch);
+        expt = source.slice(exportPos, ++pos);
         ch = commentWhitespace();
         if (ch !== 44/*,*/) break;
         pos++;
@@ -360,7 +379,7 @@ function tryParseObjectDefineOrKeys (keys) {
           pos += 5;
           ch = commentWhitespace();
           if (ch !== 58/*:*/) break;
-          addExport(expt);
+          _exports.add(decode(expt));
           pos = revertPos;
           return;
         }
@@ -403,8 +422,7 @@ function tryParseObjectDefineOrKeys (keys) {
           else if (ch === 91/*[*/) {
             pos++;
             ch = commentWhitespace();
-            if (ch === 39/*'*/) singleQuoteString();
-            else if (ch === 34/*"*/) doubleQuoteString();
+            if (ch === 39/*'*/ || ch === 34/*"*/) stringLiteral(ch);
             else break;
             pos++;
             ch = commentWhitespace();
@@ -427,13 +445,13 @@ function tryParseObjectDefineOrKeys (keys) {
           pos++;
           ch = commentWhitespace();
           if (ch !== 41/*)*/) break;
-          addExport(expt);
+          _exports.add(decode(expt));
           return;
         }
         break;
       }
       if (expt) {
-        unsafeGetters.add(expt);
+        unsafeGetters.add(decode(expt));
       }
     }
     else if (keys && ch === 107/*k*/ && source.startsWith('eys', pos + 1)) {
@@ -801,7 +819,7 @@ function tryParseObjectDefineOrKeys (keys) {
 
         const starExportSpecifier = starExportMap[id];
         if (starExportSpecifier) {
-          reexports.add(starExportSpecifier);
+          reexports.add(decode(starExportSpecifier));
           pos = revertPos;
           return;
         }
@@ -863,7 +881,7 @@ function tryParseExportsDotAssign (assign) {
         const endPos = pos;
         ch = commentWhitespace();
         if (ch === 61/*=*/) {
-          addExport(source.slice(startPos, endPos));
+          _exports.add(decode(source.slice(startPos, endPos)));
           return;
         }
       }
@@ -874,19 +892,15 @@ function tryParseExportsDotAssign (assign) {
       pos++;
       ch = commentWhitespace();
       if (ch === 39/*'*/ || ch === 34/*"*/) {
-        pos++;
         const startPos = pos;
-        if (identifier() && source.charCodeAt(pos) === ch) {
-          const endPos = pos++;
-          ch = commentWhitespace();
-          if (ch !== 93/*]*/)
-            break;
-          pos++;
-          ch = commentWhitespace();
-          if (ch !== 61/*=*/)
-            break;
-          addExport(source.slice(startPos, endPos));
-        }
+        stringLiteral(ch);
+        const endPos = ++pos;
+        ch = commentWhitespace();
+        if (ch !== 93/*]*/) break;
+        pos++;
+        ch = commentWhitespace();
+        if (ch !== 61/*=*/) break;
+        _exports.add(decode(source.slice(startPos, endPos)));
       }
       break;
     }
@@ -921,39 +935,21 @@ function tryParseRequire (requireType) {
     if (ch === 40/*(*/) {
       pos++;
       ch = commentWhitespace();
-      const reexportStart = pos + 1;
-      if (ch === 39/*'*/) {
-        singleQuoteString();
-        const reexportEnd = pos++;
-        ch = commentWhitespace();
-        if (ch === 41/*)*/) {
-          switch (requireType) {
-            case ExportAssign:
-              reexports.add(source.slice(reexportStart, reexportEnd));
-              return true;
-            case ExportStar:
-              reexports.add(source.slice(reexportStart, reexportEnd));
-              return true;
-            default:
-              lastStarExportSpecifier = source.slice(reexportStart, reexportEnd);
-              return true;
-          }
-        }
-      }
-      else if (ch === 34/*"*/) {
-        doubleQuoteString();
-        const reexportEnd = pos++;
+      const reexportStart = pos;
+      if (ch === 39/*'*/ || ch === 34/*"*/) {
+        stringLiteral(ch);
+        const reexportEnd = ++pos;
         ch = commentWhitespace();
         if (ch === 41/*)*/) {
           switch (requireType) {
             case ExportAssign:
-              reexports.add(source.slice(reexportStart, reexportEnd));
+              reexports.add(decode(source.slice(reexportStart, reexportEnd)));
               return true;
             case ExportStar:
-              reexports.add(source.slice(reexportStart, reexportEnd));
+              reexports.add(decode(source.slice(reexportStart, reexportEnd)));
               return true;
             default:
-              lastStarExportSpecifier = source.slice(reexportStart, reexportEnd);
+              lastStarExportSpecifier = decode(source.slice(reexportStart, reexportEnd));
               return true;
           }
         }
@@ -982,7 +978,7 @@ function tryParseLiteralExports () {
         }
         ch = source.charCodeAt(pos);
       }
-      addExport(source.slice(startPos, endPos));
+      _exports.add(decode(source.slice(startPos, endPos)));
     }
     else if (ch === 46/*.*/ && source.startsWith('..', pos + 1)) {
       pos += 3;
@@ -996,21 +992,20 @@ function tryParseLiteralExports () {
       ch = commentWhitespace();
     }
     else if (ch === 39/*'*/ || ch === 34/*"*/) {
-      const startPos = ++pos;
-      if (identifier() && source.charCodeAt(pos) === ch) {
-        const endPos = pos++;
+      const startPos = pos;
+      stringLiteral(ch);
+      const endPos = ++pos;
+      ch = commentWhitespace();
+      if (ch === 58/*:*/) {
+        pos++;
         ch = commentWhitespace();
-        if (ch === 58/*:*/) {
-          pos++;
-          ch = commentWhitespace();
-          // nothing more complex than identifier expressions for now
-          if (!identifier()) {
-            pos = revertPos;
-            return;
-          }
-          ch = source.charCodeAt(pos);
-          addExport(source.slice(startPos, endPos));
+        // nothing more complex than identifier expressions for now
+        if (!identifier()) {
+          pos = revertPos;
+          return;
         }
+        ch = source.charCodeAt(pos);
+        _exports.add(decode(source.slice(startPos, endPos)));
       }
     }
     else {
@@ -1248,26 +1243,10 @@ function lineComment () {
   }
 }
 
-function singleQuoteString () {
-  while (pos++ < end) {
-    let ch = source.charCodeAt(pos);
-    if (ch === 39/*'*/)
-      return;
-    if (ch === 92/*\*/) {
-      ch = source.charCodeAt(++pos);
-      if (ch === 13/*\r*/ && source.charCodeAt(pos + 1) === 10/*\n*/)
-        pos++;
-    }
-    else if (isBr(ch))
-      break;
-  }
-  throw new Error('Unterminated string.');
-}
-
-function doubleQuoteString () {
+function stringLiteral (quote) {
   while (pos++ < end) {
     let ch = source.charCodeAt(pos);
-    if (ch === 34/*"*/)
+    if (ch === quote)
       return;
     if (ch === 92/*\*/) {
       ch = source.charCodeAt(++pos);

diff --git a/deps/cjs-module-lexer/package.json b/deps/cjs-module-lexer/package.json
@@ -1,6 +1,6 @@
 {
   "name": "cjs-module-lexer",
-  "version": "1.1.1",
+  "version": "1.2.1",
   "description": "Lexes CommonJS modules, returning their named exports metadata",
   "main": "lexer.js",
   "exports": {

diff --git a/doc/api/esm.md b/doc/api/esm.md
@@ -1284,7 +1284,7 @@ success!
 [`Uint8Array`]: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Uint8Array
 [dynamic instantiate hook]: #esm_code_dynamicinstantiate_code_hook
 [`util.TextDecoder`]: util.md#util_class_util_textdecoder
-[cjs-module-lexer]: https://github.com/guybedford/cjs-module-lexer/tree/1.1.1
+[cjs-module-lexer]: https://github.com/guybedford/cjs-module-lexer/tree/1.2.1
 [special scheme]: https://url.spec.whatwg.org/#special-scheme
 [the official standard format]: https://tc39.github.io/ecma262/#sec-modules
 [transpiler loader example]: #esm_transpiler_loader

diff --git a/test/fixtures/es-modules/cjs-exports.mjs b/test/fixtures/es-modules/cjs-exports.mjs
@@ -3,11 +3,14 @@ import { strictEqual, deepEqual } from 'assert';
 import m, { π } from './exports-cases.js';
 import * as ns from './exports-cases.js';
 
-deepEqual(Object.keys(ns), ['default', 'isObject', 'z', 'π']);
+deepEqual(Object.keys(ns), ['?invalid', 'default', 'invalid identifier', 'isObject', 'package', 'z', 'π', '\u{d83c}\u{df10}']);
 strictEqual(π, 'yes');
 strictEqual(typeof m.isObject, 'undefined');
 strictEqual(m.π, 'yes');
 strictEqual(m.z, 'yes');
+strictEqual(m.package, 10);
+strictEqual(m['invalid identifier'], 'yes');
+strictEqual(m['?invalid'], 'yes');
 
 import m2, { __esModule as __esModule2, name as name2 } from './exports-cases2.js';
 import * as ns2 from './exports-cases2.js';

diff --git a/test/fixtures/es-modules/exports-cases.js b/test/fixtures/es-modules/exports-cases.js
@@ -1,7 +1,9 @@
 if (global.maybe)
   module.exports = require('../is-object');
-exports['invalid identifier'] = 'no';
-module.exports['?invalid'] = 'no';
+exports['invalid identifier'] = 'yes';
+module.exports['?invalid'] = 'yes';
 module.exports['π'] = 'yes';
-exports.package  = 10; // reserved word -> not used
+exports['\u{D83C}'] = 'no';
+exports['\u{D83C}\u{DF10}'] = 'yes';
+exports.package  = 10; // reserved word
 Object.defineProperty(exports, 'z', { value: 'yes' });