fix(v2): fix encoding of markdown image/file inline file-loaders (#4736)

* try to reproduce windows edge case due to file encoding

* mdx loader => required file paths should be escaped

* revert bad change

* try to fix posix path issues

* try to fix posix path issues

* attempt to fix the file-loader edge cases with non-ascii chars

* Add more example image edge-cases

packages/docusaurus-utils/src/__tests__/escapePath.test.ts

@@ -0,0 +1,25 @@
+/**
+ * Copyright (c) Facebook, Inc. and its affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+import {escapePath} from '../escapePath';
+
+describe('escapePath', () => {
+  test('escapePath works', () => {
+    const asserts: Record<string, string> = {
+      'c:/aaaa\\bbbb': 'c:/aaaa\\\\bbbb',
+      'c:\\aaaa\\bbbb\\★': 'c:\\\\aaaa\\\\bbbb\\\\★',
+      '\\\\?\\c:\\aaaa\\bbbb': '\\\\\\\\?\\\\c:\\\\aaaa\\\\bbbb',
+      'c:\\aaaa\\bbbb': 'c:\\\\aaaa\\\\bbbb',
+      'foo\\bar': 'foo\\\\bar',
+      'foo\\bar/lol': 'foo\\\\bar/lol',
+      'website\\docs/**/*.{md,mdx}': 'website\\\\docs/**/*.{md,mdx}',
+    };
+    Object.keys(asserts).forEach((file) => {
+      expect(escapePath(file)).toBe(asserts[file]);
+    });
+  });
+});

packages/docusaurus-utils/src/__tests__/index.test.ts

@@ -38,7 +38,7 @@ import {sum} from 'lodash';
 
 describe('load utils', () => {
   test('aliasedSitePath', () => {
-    const asserts = {
+    const asserts: Record<string, string> = {
       'user/website/docs/asd.md': '@site/docs/asd.md',
       'user/website/versioned_docs/foo/bar.md':
         '@site/versioned_docs/foo/bar.md',
@@ -51,23 +51,8 @@ describe('load utils', () => {
     });
   });
 
-  test('posixPath', () => {
-    const asserts = {
-      'c:/aaaa\\bbbb': 'c:/aaaa/bbbb',
-      'c:\\aaaa\\bbbb\\★': 'c:\\aaaa\\bbbb\\★',
-      '\\\\?\\c:\\aaaa\\bbbb': '\\\\?\\c:\\aaaa\\bbbb',
-      'c:\\aaaa\\bbbb': 'c:/aaaa/bbbb',
-      'foo\\bar': 'foo/bar',
-      'foo\\bar/lol': 'foo/bar/lol',
-      'website\\docs/**/*.{md,mdx}': 'website/docs/**/*.{md,mdx}',
-    };
-    Object.keys(asserts).forEach((file) => {
-      expect(posixPath(file)).toBe(asserts[file]);
-    });
-  });
-
   test('genComponentName', () => {
-    const asserts = {
+    const asserts: Record<string, string> = {
       '/': 'index',
       '/foo-bar': 'FooBar096',
       '/foo/bar': 'FooBar1Df',
@@ -99,7 +84,7 @@ describe('load utils', () => {
   });
 
   test('docuHash', () => {
-    const asserts = {
+    const asserts: Record<string, string> = {
       '': '-d41',
       '/': 'index',
       '/foo-bar': 'foo-bar-096',
@@ -115,7 +100,7 @@ describe('load utils', () => {
   });
 
   test('fileToPath', () => {
-    const asserts = {
+    const asserts: Record<string, string> = {
       'index.md': '/',
       'hello/index.md': '/hello/',
       'foo.md': '/foo',
@@ -166,7 +151,7 @@ describe('load utils', () => {
   });
 
   test('genChunkName', () => {
-    const firstAssert = {
+    const firstAssert: Record<string, string> = {
       '/docs/adding-blog': 'docs-adding-blog-062',
       '/docs/versioning': 'docs-versioning-8a8',
       '/': 'index',
@@ -186,7 +171,7 @@ describe('load utils', () => {
     );
 
     // Even with same preferred name, still different chunk name for different path
-    const secondAssert = {
+    const secondAssert: Record<string, string> = {
       '/blog/1': 'blog-85-f-089',
       '/blog/2': 'blog-353-489',
     };
@@ -195,7 +180,7 @@ describe('load utils', () => {
     });
 
     // Only generate short unique id
-    const thirdAssert = {
+    const thirdAssert: Record<string, string> = {
       a: '0cc175b9',
       b: '92eb5ffe',
       c: '4a8a08f0',

packages/docusaurus-utils/src/__tests__/posixPath.test.ts

@@ -0,0 +1,25 @@
+/**
+ * Copyright (c) Facebook, Inc. and its affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+import {posixPath} from '../posixPath';
+
+describe('posixPath', () => {
+  test('posixPath works', () => {
+    const asserts: Record<string, string> = {
+      'c:/aaaa\\bbbb': 'c:/aaaa/bbbb',
+      'c:\\aaaa\\bbbb\\★': 'c:\\aaaa\\bbbb\\★',
+      '\\\\?\\c:\\aaaa\\bbbb': '\\\\?\\c:\\aaaa\\bbbb',
+      'c:\\aaaa\\bbbb': 'c:/aaaa/bbbb',
+      'foo\\bar': 'foo/bar',
+      'foo\\bar/lol': 'foo/bar/lol',
+      'website\\docs/**/*.{md,mdx}': 'website/docs/**/*.{md,mdx}',
+    };
+    Object.keys(asserts).forEach((file) => {
+      expect(posixPath(file)).toBe(asserts[file]);
+    });
+  });
+});

packages/docusaurus-utils/src/escapePath.ts

@@ -0,0 +1,23 @@
+/**
+ * Copyright (c) Facebook, Inc. and its affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+/**
+ * When you have a path like C:\X\Y
+ * It is not safe to use directly when generating code
+ * For example, this would fail due to unescaped \: `<img src={require('${filePath}')} />`
+ * But this would work: `<img src={require('${escapePath(filePath)}')} />`
+ *
+ * Workaround for issue in posixPath, maybe we won't need it anymore soon?
+ * https://github.com/facebook/docusaurus/issues/4730#issuecomment-833530370
+ * https://github.com/sindresorhus/slash/pull/16#issuecomment-833528479
+ */
+export function escapePath(str: string): string {
+  const escaped = JSON.stringify(str);
+
+  // Remove the " around the json string;
+  return escaped.substring(1, escaped.length - 1);
+}

packages/docusaurus-utils/src/index.ts

@@ -21,9 +21,14 @@ import {
 // @ts-expect-error: no typedefs :s
 import resolvePathnameUnsafe from 'resolve-pathname';
 
+import {posixPath as posixPathImport} from './posixPath';
+
+export const posixPath = posixPathImport;
+
 export * from './codeTranslationsUtils';
 export * from './markdownParser';
 export * from './markdownLinks';
+export * from './escapePath';
 
 const fileHash = new Map();
 export async function generate(
@@ -129,20 +134,6 @@ export function genComponentName(pagePath: string): string {
   return upperFirst(camelCase(pageHash));
 }
 
-/**
- * Convert Windows backslash paths to posix style paths.
- * E.g: endi\\lie -> endi/lie
- */
-export function posixPath(str: string): string {
-  const isExtendedLengthPath = /^\\\\\?\\/.test(str);
-  const hasNonAscii = /[^\u0000-\u0080]+/.test(str); // eslint-disable-line
-
-  if (isExtendedLengthPath || hasNonAscii) {
-    return str;
-  }
-  return str.replace(/\\/g, '/');
-}
-
 // When you want to display a path in a message/warning/error,
 // it's more convenient to:
 // - make it relative to cwd()

packages/docusaurus-utils/src/posixPath.ts

@@ -0,0 +1,27 @@
+/**
+ * Copyright (c) Facebook, Inc. and its affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+/**
+ * Convert Windows backslash paths to posix style paths.
+ * E.g: endi\\lie -> endi/lie
+ *
+ * Looks like this code was originally copied from https://github.com/sindresorhus/slash/blob/main/index.js
+ *
+ */
+export function posixPath(str: string): string {
+  const isExtendedLengthPath = /^\\\\\?\\/.test(str);
+
+  // TODO not sure why we need this
+  // See https://github.com/sindresorhus/slash/pull/16#issuecomment-833528479
+  // See https://github.com/facebook/docusaurus/issues/4730#issuecomment-833530370
+  const hasNonAscii = /[^\u0000-\u0080]+/.test(str); // eslint-disable-line
+
+  if (isExtendedLengthPath || hasNonAscii) {
+    return str;
+  }
+  return str.replace(/\\/g, '/');
+}