lib, url: add a windows option to path parsing

PR-URL: #52509 Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: Moshe Atlow <moshe@atlow.co.il> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Zeyu "Alex" Yang <himself65@outlook.com> Reviewed-By: LiviaMedeiros <livia@cirno.name>
nodejs · May 3, 2024 · 722fe64 · 722fe64
1 parent 4d5ef4f
commit 722fe64
Show file tree

Hide file tree

Showing 6 changed files with 267 additions and 216 deletions.
diff --git a/doc/api/esm.md b/doc/api/esm.md
@@ -1157,7 +1157,7 @@ resolution for ESM specifiers is [commonjs-extension-resolution-loader][].
 [`package.json`]: packages.md#nodejs-packagejson-field-definitions
 [`path.dirname()`]: path.md#pathdirnamepath
 [`process.dlopen`]: process.md#processdlopenmodule-filename-flags
-[`url.fileURLToPath()`]: url.md#urlfileurltopathurl
+[`url.fileURLToPath()`]: url.md#urlfileurltopathurl-options
 [cjs-module-lexer]: https://github.com/nodejs/cjs-module-lexer/tree/1.2.2
 [commonjs-extension-resolution-loader]: https://github.com/nodejs/loaders-test/tree/main/commonjs-extension-resolution-loader
 [custom https loader]: module.md#import-from-https
@@ -1166,4 +1166,4 @@ resolution for ESM specifiers is [commonjs-extension-resolution-loader][].
 [special scheme]: https://url.spec.whatwg.org/#special-scheme
 [status code]: process.md#exit-codes
 [the official standard format]: https://tc39.github.io/ecma262/#sec-modules
-[url.pathToFileURL]: url.md#urlpathtofileurlpath
+[url.pathToFileURL]: url.md#urlpathtofileurlpath-options
diff --git a/doc/api/url.md b/doc/api/url.md
@@ -1137,13 +1137,23 @@ console.log(url.domainToUnicode('xn--iñvalid.com'));
 // Prints an empty string
 ```
 
-### `url.fileURLToPath(url)`
+### `url.fileURLToPath(url[, options])`
 
 <!-- YAML
 added: v10.12.0
+changes:
+  - version: REPLACEME
+    pr-url: https://github.com/nodejs/node/pull/52509
+    description: The `options` argument can now be used to
+                 determine how to parse the `path` argument.
 -->
 
 * `url` {URL | string} The file URL string or URL object to convert to a path.
+* `options` {Object}
+  * `windows` {boolean|undefined} `true` if the `path` should be
+    return as a windows filepath, `false` for posix, and
+    `undefined` for the system default.
+    **Default:** `undefined`.
 * Returns: {string} The fully-resolved platform-specific Node.js file path.
 
 This function ensures the correct decodings of percent-encoded characters as
@@ -1237,13 +1247,23 @@ console.log(url.format(myURL, { fragment: false, unicode: true, auth: false }));
 // Prints 'https://測試/?abc'
 ```
 
-### `url.pathToFileURL(path)`
+### `url.pathToFileURL(path[, options])`
 
 <!-- YAML
 added: v10.12.0
+changes:
+  - version: REPLACEME
+    pr-url: https://github.com/nodejs/node/pull/52509
+    description: The `options` argument can now be used to
+                 determine how to return the `path` value.
 -->
 
 * `path` {string} The path to convert to a File URL.
+* `options` {Object}
+  * `windows` {boolean|undefined} `true` if the `path` should be
+    treated as a windows filepath, `false` for posix, and
+    `undefined` for the system default.
+    **Default:** `undefined`.
 * Returns: {URL} The file URL object.
 
 This function ensures that `path` is resolved absolutely, and that the URL

diff --git a/lib/internal/url.js b/lib/internal/url.js
@@ -44,6 +44,7 @@ const {
   getConstructorOf,
   removeColors,
   kEnumerableProperty,
+  kEmptyObject,
   SideEffectFreeRegExpPrototypeSymbolReplace,
 } = require('internal/util');
 
@@ -1452,14 +1453,15 @@ function getPathFromURLPosix(url) {
   return decodeURIComponent(pathname);
 }
 
-function fileURLToPath(path) {
+function fileURLToPath(path, options = kEmptyObject) {
+  const windows = options?.windows;
   if (typeof path === 'string')
     path = new URL(path);
   else if (!isURL(path))
     throw new ERR_INVALID_ARG_TYPE('path', ['string', 'URL'], path);
   if (path.protocol !== 'file:')
     throw new ERR_INVALID_URL_SCHEME('file');
-  return isWindows ? getPathFromURLWin32(path) : getPathFromURLPosix(path);
+  return (windows ?? isWindows) ? getPathFromURLWin32(path) : getPathFromURLPosix(path);
 }
 
 // The following characters are percent-encoded when converting from file path
@@ -1481,11 +1483,12 @@ const tabRegEx = /\t/g;
 const questionRegex = /\?/g;
 const hashRegex = /#/g;
 
-function encodePathChars(filepath) {
+function encodePathChars(filepath, options = kEmptyObject) {
+  const windows = options?.windows;
   if (StringPrototypeIndexOf(filepath, '%') !== -1)
     filepath = RegExpPrototypeSymbolReplace(percentRegEx, filepath, '%25');
   // In posix, backslash is a valid character in paths:
-  if (!isWindows && StringPrototypeIndexOf(filepath, '\\') !== -1)
+  if (!(windows ?? isWindows) && StringPrototypeIndexOf(filepath, '\\') !== -1)
     filepath = RegExpPrototypeSymbolReplace(backslashRegEx, filepath, '%5C');
   if (StringPrototypeIndexOf(filepath, '\n') !== -1)
     filepath = RegExpPrototypeSymbolReplace(newlineRegEx, filepath, '%0A');
@@ -1496,8 +1499,9 @@ function encodePathChars(filepath) {
   return filepath;
 }
 
-function pathToFileURL(filepath) {
-  if (isWindows && StringPrototypeStartsWith(filepath, '\\\\')) {
+function pathToFileURL(filepath, options = kEmptyObject) {
+  const windows = options?.windows;
+  if ((windows ?? isWindows) && StringPrototypeStartsWith(filepath, '\\\\')) {
     const outURL = new URL('file://');
     // UNC path format: \\server\share\resource
     const hostnameEndIndex = StringPrototypeIndexOf(filepath, '\\', 2);
@@ -1518,20 +1522,22 @@ function pathToFileURL(filepath) {
     const hostname = StringPrototypeSlice(filepath, 2, hostnameEndIndex);
     outURL.hostname = domainToASCII(hostname);
     outURL.pathname = encodePathChars(
-      RegExpPrototypeSymbolReplace(backslashRegEx, StringPrototypeSlice(filepath, hostnameEndIndex), '/'));
+      RegExpPrototypeSymbolReplace(backslashRegEx, StringPrototypeSlice(filepath, hostnameEndIndex), '/'),
+      { windows },
+    );
     return outURL;
   }
-  let resolved = path.resolve(filepath);
+  let resolved = (windows ?? isWindows) ? path.win32.resolve(filepath) : path.posix.resolve(filepath);
   // path.resolve strips trailing slashes so we must add them back
   const filePathLast = StringPrototypeCharCodeAt(filepath,
                                                  filepath.length - 1);
   if ((filePathLast === CHAR_FORWARD_SLASH ||
-       (isWindows && filePathLast === CHAR_BACKWARD_SLASH)) &&
+       ((windows ?? isWindows) && filePathLast === CHAR_BACKWARD_SLASH)) &&
       resolved[resolved.length - 1] !== path.sep)
     resolved += '/';
 
   // Call encodePathChars first to avoid encoding % again for ? and #.
-  resolved = encodePathChars(resolved);
+  resolved = encodePathChars(resolved, { windows });
 
   // Question and hash character should be included in pathname.
   // Therefore, encoding is required to eliminate parsing them in different states.

diff --git a/lib/url.js b/lib/url.js
@@ -1020,10 +1020,10 @@ Url.prototype.parseHost = function parseHost() {
 // When used internally, we are not obligated to associate TypeError with
 // this function, so non-strings can be rejected by underlying implementation.
 // Public API has to validate input and throw appropriate error.
-function pathToFileURL(path) {
+function pathToFileURL(path, options) {
   validateString(path, 'path');
 
-  return _pathToFileURL(path);
+  return _pathToFileURL(path, options);
 }
 
 module.exports = {

diff --git a/test/parallel/test-url-fileurltopath.js b/test/parallel/test-url-fileurltopath.js
@@ -49,106 +49,120 @@ assert.throws(() => url.fileURLToPath('https://a/b/c'), {
   }
 }
 
-{
-  let testCases;
-  if (isWindows) {
-    testCases = [
-      // Lowercase ascii alpha
-      { path: 'C:\\foo', fileURL: 'file:///C:/foo' },
-      // Uppercase ascii alpha
-      { path: 'C:\\FOO', fileURL: 'file:///C:/FOO' },
-      // dir
-      { path: 'C:\\dir\\foo', fileURL: 'file:///C:/dir/foo' },
-      // trailing separator
-      { path: 'C:\\dir\\', fileURL: 'file:///C:/dir/' },
-      // dot
-      { path: 'C:\\foo.mjs', fileURL: 'file:///C:/foo.mjs' },
-      // space
-      { path: 'C:\\foo bar', fileURL: 'file:///C:/foo%20bar' },
-      // question mark
-      { path: 'C:\\foo?bar', fileURL: 'file:///C:/foo%3Fbar' },
-      // number sign
-      { path: 'C:\\foo#bar', fileURL: 'file:///C:/foo%23bar' },
-      // ampersand
-      { path: 'C:\\foo&bar', fileURL: 'file:///C:/foo&bar' },
-      // equals
-      { path: 'C:\\foo=bar', fileURL: 'file:///C:/foo=bar' },
-      // colon
-      { path: 'C:\\foo:bar', fileURL: 'file:///C:/foo:bar' },
-      // semicolon
-      { path: 'C:\\foo;bar', fileURL: 'file:///C:/foo;bar' },
-      // percent
-      { path: 'C:\\foo%bar', fileURL: 'file:///C:/foo%25bar' },
-      // backslash
-      { path: 'C:\\foo\\bar', fileURL: 'file:///C:/foo/bar' },
-      // backspace
-      { path: 'C:\\foo\bbar', fileURL: 'file:///C:/foo%08bar' },
-      // tab
-      { path: 'C:\\foo\tbar', fileURL: 'file:///C:/foo%09bar' },
-      // newline
-      { path: 'C:\\foo\nbar', fileURL: 'file:///C:/foo%0Abar' },
-      // carriage return
-      { path: 'C:\\foo\rbar', fileURL: 'file:///C:/foo%0Dbar' },
-      // latin1
-      { path: 'C:\\fóóbàr', fileURL: 'file:///C:/f%C3%B3%C3%B3b%C3%A0r' },
-      // Euro sign (BMP code point)
-      { path: 'C:\\€', fileURL: 'file:///C:/%E2%82%AC' },
-      // Rocket emoji (non-BMP code point)
-      { path: 'C:\\🚀', fileURL: 'file:///C:/%F0%9F%9A%80' },
-      // UNC path (see https://docs.microsoft.com/en-us/archive/blogs/ie/file-uris-in-windows)
-      { path: '\\\\nas\\My Docs\\File.doc', fileURL: 'file://nas/My%20Docs/File.doc' },
-    ];
-  } else {
-    testCases = [
-      // Lowercase ascii alpha
-      { path: '/foo', fileURL: 'file:///foo' },
-      // Uppercase ascii alpha
-      { path: '/FOO', fileURL: 'file:///FOO' },
-      // dir
-      { path: '/dir/foo', fileURL: 'file:///dir/foo' },
-      // trailing separator
-      { path: '/dir/', fileURL: 'file:///dir/' },
-      // dot
-      { path: '/foo.mjs', fileURL: 'file:///foo.mjs' },
-      // space
-      { path: '/foo bar', fileURL: 'file:///foo%20bar' },
-      // question mark
-      { path: '/foo?bar', fileURL: 'file:///foo%3Fbar' },
-      // number sign
-      { path: '/foo#bar', fileURL: 'file:///foo%23bar' },
-      // ampersand
-      { path: '/foo&bar', fileURL: 'file:///foo&bar' },
-      // equals
-      { path: '/foo=bar', fileURL: 'file:///foo=bar' },
-      // colon
-      { path: '/foo:bar', fileURL: 'file:///foo:bar' },
-      // semicolon
-      { path: '/foo;bar', fileURL: 'file:///foo;bar' },
-      // percent
-      { path: '/foo%bar', fileURL: 'file:///foo%25bar' },
-      // backslash
-      { path: '/foo\\bar', fileURL: 'file:///foo%5Cbar' },
-      // backspace
-      { path: '/foo\bbar', fileURL: 'file:///foo%08bar' },
-      // tab
-      { path: '/foo\tbar', fileURL: 'file:///foo%09bar' },
-      // newline
-      { path: '/foo\nbar', fileURL: 'file:///foo%0Abar' },
-      // carriage return
-      { path: '/foo\rbar', fileURL: 'file:///foo%0Dbar' },
-      // latin1
-      { path: '/fóóbàr', fileURL: 'file:///f%C3%B3%C3%B3b%C3%A0r' },
-      // Euro sign (BMP code point)
-      { path: '/€', fileURL: 'file:///%E2%82%AC' },
-      // Rocket emoji (non-BMP code point)
-      { path: '/🚀', fileURL: 'file:///%F0%9F%9A%80' },
-    ];
-  }
+const windowsTestCases = [
+  // Lowercase ascii alpha
+  { path: 'C:\\foo', fileURL: 'file:///C:/foo' },
+  // Uppercase ascii alpha
+  { path: 'C:\\FOO', fileURL: 'file:///C:/FOO' },
+  // dir
+  { path: 'C:\\dir\\foo', fileURL: 'file:///C:/dir/foo' },
+  // trailing separator
+  { path: 'C:\\dir\\', fileURL: 'file:///C:/dir/' },
+  // dot
+  { path: 'C:\\foo.mjs', fileURL: 'file:///C:/foo.mjs' },
+  // space
+  { path: 'C:\\foo bar', fileURL: 'file:///C:/foo%20bar' },
+  // question mark
+  { path: 'C:\\foo?bar', fileURL: 'file:///C:/foo%3Fbar' },
+  // number sign
+  { path: 'C:\\foo#bar', fileURL: 'file:///C:/foo%23bar' },
+  // ampersand
+  { path: 'C:\\foo&bar', fileURL: 'file:///C:/foo&bar' },
+  // equals
+  { path: 'C:\\foo=bar', fileURL: 'file:///C:/foo=bar' },
+  // colon
+  { path: 'C:\\foo:bar', fileURL: 'file:///C:/foo:bar' },
+  // semicolon
+  { path: 'C:\\foo;bar', fileURL: 'file:///C:/foo;bar' },
+  // percent
+  { path: 'C:\\foo%bar', fileURL: 'file:///C:/foo%25bar' },
+  // backslash
+  { path: 'C:\\foo\\bar', fileURL: 'file:///C:/foo/bar' },
+  // backspace
+  { path: 'C:\\foo\bbar', fileURL: 'file:///C:/foo%08bar' },
+  // tab
+  { path: 'C:\\foo\tbar', fileURL: 'file:///C:/foo%09bar' },
+  // newline
+  { path: 'C:\\foo\nbar', fileURL: 'file:///C:/foo%0Abar' },
+  // carriage return
+  { path: 'C:\\foo\rbar', fileURL: 'file:///C:/foo%0Dbar' },
+  // latin1
+  { path: 'C:\\fóóbàr', fileURL: 'file:///C:/f%C3%B3%C3%B3b%C3%A0r' },
+  // Euro sign (BMP code point)
+  { path: 'C:\\€', fileURL: 'file:///C:/%E2%82%AC' },
+  // Rocket emoji (non-BMP code point)
+  { path: 'C:\\🚀', fileURL: 'file:///C:/%F0%9F%9A%80' },
+  // UNC path (see https://docs.microsoft.com/en-us/archive/blogs/ie/file-uris-in-windows)
+  { path: '\\\\nas\\My Docs\\File.doc', fileURL: 'file://nas/My%20Docs/File.doc' },
+];
+const posixTestCases = [
+  // Lowercase ascii alpha
+  { path: '/foo', fileURL: 'file:///foo' },
+  // Uppercase ascii alpha
+  { path: '/FOO', fileURL: 'file:///FOO' },
+  // dir
+  { path: '/dir/foo', fileURL: 'file:///dir/foo' },
+  // trailing separator
+  { path: '/dir/', fileURL: 'file:///dir/' },
+  // dot
+  { path: '/foo.mjs', fileURL: 'file:///foo.mjs' },
+  // space
+  { path: '/foo bar', fileURL: 'file:///foo%20bar' },
+  // question mark
+  { path: '/foo?bar', fileURL: 'file:///foo%3Fbar' },
+  // number sign
+  { path: '/foo#bar', fileURL: 'file:///foo%23bar' },
+  // ampersand
+  { path: '/foo&bar', fileURL: 'file:///foo&bar' },
+  // equals
+  { path: '/foo=bar', fileURL: 'file:///foo=bar' },
+  // colon
+  { path: '/foo:bar', fileURL: 'file:///foo:bar' },
+  // semicolon
+  { path: '/foo;bar', fileURL: 'file:///foo;bar' },
+  // percent
+  { path: '/foo%bar', fileURL: 'file:///foo%25bar' },
+  // backslash
+  { path: '/foo\\bar', fileURL: 'file:///foo%5Cbar' },
+  // backspace
+  { path: '/foo\bbar', fileURL: 'file:///foo%08bar' },
+  // tab
+  { path: '/foo\tbar', fileURL: 'file:///foo%09bar' },
+  // newline
+  { path: '/foo\nbar', fileURL: 'file:///foo%0Abar' },
+  // carriage return
+  { path: '/foo\rbar', fileURL: 'file:///foo%0Dbar' },
+  // latin1
+  { path: '/fóóbàr', fileURL: 'file:///f%C3%B3%C3%B3b%C3%A0r' },
+  // Euro sign (BMP code point)
+  { path: '/€', fileURL: 'file:///%E2%82%AC' },
+  // Rocket emoji (non-BMP code point)
+  { path: '/🚀', fileURL: 'file:///%F0%9F%9A%80' },
+];
 
-  for (const { path, fileURL } of testCases) {
-    const fromString = url.fileURLToPath(fileURL);
-    assert.strictEqual(fromString, path);
-    const fromURL = url.fileURLToPath(new URL(fileURL));
-    assert.strictEqual(fromURL, path);
-  }
+for (const { path, fileURL } of windowsTestCases) {
+  const fromString = url.fileURLToPath(fileURL, { windows: true });
+  assert.strictEqual(fromString, path);
+  const fromURL = url.fileURLToPath(new URL(fileURL), { windows: true });
+  assert.strictEqual(fromURL, path);
+}
+
+for (const { path, fileURL } of posixTestCases) {
+  const fromString = url.fileURLToPath(fileURL, { windows: false });
+  assert.strictEqual(fromString, path);
+  const fromURL = url.fileURLToPath(new URL(fileURL), { windows: false });
+  assert.strictEqual(fromURL, path);
+}
+
+const defaultTestCases = isWindows ? windowsTestCases : posixTestCases;
+
+// Test when `options` is null
+const whenNullActual = url.fileURLToPath(new URL(defaultTestCases[0].fileURL), null);
+assert.strictEqual(whenNullActual, defaultTestCases[0].path);
+
+for (const { path, fileURL } of defaultTestCases) {
+  const fromString = url.fileURLToPath(fileURL);
+  assert.strictEqual(fromString, path);
+  const fromURL = url.fileURLToPath(new URL(fileURL));
+  assert.strictEqual(fromURL, path);
 }