Avoid loading files that we know are binary ahead of time (#5612)

Author: mjesun

packages/jest-haste-map/src/__tests__/index.test.js

@@ -141,10 +141,12 @@ describe('HasteMap', () => {
       '/vegetables/melon.js': ['/**', ' * @providesModule Melon', ' */'].join(
         '\n',
       ),
+      '/video/video.mp4': Buffer.from([0xfa, 0xce, 0xb0, 0x0c]).toString(),
     });
     mockClocks = object({
       '/fruits': 'c:fake-clock:1',
       '/vegetables': 'c:fake-clock:2',
+      '/video': 'c:fake-clock:3',
     });
 
     mockChangedFiles = null;
@@ -305,6 +307,28 @@ describe('HasteMap', () => {
     });
   });
 
+  it('does not crawl native files even if requested to do so', async () => {
+    mockFs['/video/i-require-a-video.js'] = [
+      '/**',
+      ' * @providesModule IRequireAVideo',
+      ' */',
+      'module.exports = require("./video.mp4");',
+    ].join('\n');
+
+    const hasteMap = new HasteMap(
+      Object.assign({}, defaultConfig, {
+        extensions: [...defaultConfig.extensions],
+        roots: [...defaultConfig.roots, '/video'],
+      }),
+    );
+
+    const {__hasteMapForTest: data} = await hasteMap.build();
+
+    expect(data.map.IRequireAVideo).toBeDefined();
+    expect(data.files['/video/video.mp4']).toBeDefined();
+    expect(fs.readFileSync).not.toBeCalledWith('/video/video.mp4', 'utf8');
+  });
+
   it('retains all files if `retainAllFiles` is specified', () => {
     mockFs['/fruits/node_modules/fbjs/index.js'] = [
       '/**',

packages/jest-haste-map/src/blacklist.js

@@ -0,0 +1,65 @@
+/**
+ * Copyright (c) 2014-present, Facebook, Inc. All rights reserved.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ *
+ * @flow
+ */
+
+// This list is compiled after the MDN list of the most common MIME types (see
+// https://developer.mozilla.org/en-US/docs/Web/HTTP/Basics_of_HTTP/MIME_types/
+// Complete_list_of_MIME_types).
+//
+// Only MIME types starting with "image/", "video/", "audio/" and "font/" are
+// reflected in the list. Adding "application/" is too risky since some text
+// file formats (like ".js" and ".json") have an "application/" MIME type.
+//
+// Feel free to add any extensions that cannot contain any "@providesModule"
+// annotation.
+
+const extensions: Set<string> = new Set([
+  // JSONs are never haste modules, except for "package.json", which is handled.
+  '.json',
+
+  // Image extensions.
+  '.bmp',
+  '.gif',
+  '.ico',
+  '.jpeg',
+  '.jpg',
+  '.png',
+  '.svg',
+  '.tiff',
+  '.tif',
+  '.webp',
+
+  // Video extensions.
+  '.avi',
+  '.mp4',
+  '.mpeg',
+  '.mpg',
+  '.ogv',
+  '.webm',
+  '.3gp',
+  '.3g2',
+
+  // Audio extensions.
+  '.aac',
+  '.midi',
+  '.mid',
+  '.mp3',
+  '.oga',
+  '.wav',
+  '.3gp',
+  '.3g2',
+
+  // Font extensions.
+  '.eot',
+  '.otf',
+  '.ttf',
+  '.woff',
+  '.woff2',
+]);
+
+export default extensions;

packages/jest-haste-map/src/worker.js

@@ -12,11 +12,11 @@ import type {HasteImpl, WorkerMessage, WorkerMetadata} from './types';
 import path from 'path';
 import * as docblock from 'jest-docblock';
 import fs from 'graceful-fs';
+import blacklist from './blacklist';
 import H from './constants';
 import extractRequires from './lib/extract_requires';
 
-const JSON_EXTENSION = '.json';
-const PACKAGE_JSON = path.sep + 'package' + JSON_EXTENSION;
+const PACKAGE_JSON = path.sep + 'package.json';
 
 let hasteImpl: ?HasteImpl = null;
 let hasteImplModulePath: ?string = null;
@@ -35,26 +35,33 @@ export async function worker(data: WorkerMessage): Promise<WorkerMetadata> {
   }
 
   const filePath = data.filePath;
-  const content = fs.readFileSync(filePath, 'utf8');
   let module;
   let id: ?string;
   let dependencies;
 
   if (filePath.endsWith(PACKAGE_JSON)) {
-    const fileData = JSON.parse(content);
+    const fileData = JSON.parse(fs.readFileSync(filePath, 'utf8'));
+
     if (fileData.name) {
       id = fileData.name;
       module = [filePath, H.PACKAGE];
     }
-  } else if (!filePath.endsWith(JSON_EXTENSION)) {
+
+    return {dependencies, id, module};
+  }
+
+  if (!blacklist.has(filePath.substr(filePath.lastIndexOf('.')))) {
+    const content = fs.readFileSync(filePath, 'utf8');
+
     if (hasteImpl) {
       id = hasteImpl.getHasteName(filePath);
     } else {
       const doc = docblock.parse(docblock.extract(content));
-      const idPragmas = [].concat(doc.providesModule || doc.provides);
-      id = idPragmas[0];
+      id = [].concat(doc.providesModule || doc.provides)[0];
     }
+
     dependencies = extractRequires(content);
+
     if (id) {
       module = [filePath, H.MODULE];
     }