core(entity-classification): classify chrome extensions into separate…

… entities (#15017)
GoogleChrome · May 16, 2023 · 8ecbb94 · 8ecbb94
1 parent fa2e00c
commit 8ecbb94
Show file tree

Hide file tree

Showing 5 changed files with 140 additions and 11 deletions.
diff --git a/core/computed/entity-classification.js b/core/computed/entity-classification.js
@@ -13,16 +13,51 @@ import thirdPartyWeb from '../lib/third-party-web.js';
 /** @typedef {Map<string, LH.Artifacts.Entity>} EntityCache */
 
 class EntityClassification {
+  /**
+   * @param {EntityCache} entityCache
+   * @param {string} url
+   * @param {string=} extensionName
+   * @return {LH.Artifacts.Entity}
+   */
+  static makeupChromeExtensionEntity_(entityCache, url, extensionName) {
+    const origin = Util.getChromeExtensionOrigin(url);
+    const host = new URL(origin).host;
+    const name = extensionName || host;
+
+    const cachedEntity = entityCache.get(origin);
+    if (cachedEntity) return cachedEntity;
+
+    const chromeExtensionEntity = {
+      name,
+      company: name,
+      category: 'Chrome Extension',
+      homepage: 'https://chromewebstore.google.com/detail/' + host,
+      categories: [],
+      domains: [],
+      averageExecutionTime: 0,
+      totalExecutionTime: 0,
+      totalOccurrences: 0,
+    };
+
+    entityCache.set(origin, chromeExtensionEntity);
+    return chromeExtensionEntity;
+  }
+
   /**
    * @param {EntityCache} entityCache
    * @param {string} url
    * @return {LH.Artifacts.Entity | undefined}
    */
-  static makeUpAnEntity(entityCache, url) {
+  static _makeUpAnEntity(entityCache, url) {
     if (!UrlUtils.isValid(url)) return;
-    // We can make up an entity only for those URLs with a valid domain attached.
-    // So we further restrict from allowed URLs to (http/https).
-    if (!Util.createOrReturnURL(url).protocol.startsWith('http')) return;
+
+    const parsedUrl = Util.createOrReturnURL(url);
+    if (parsedUrl.protocol === 'chrome-extension:') {
+      return EntityClassification.makeupChromeExtensionEntity_(entityCache, url);
+    }
+
+    // Make up an entity only for valid http/https URLs.
+    if (!parsedUrl.protocol.startsWith('http')) return;
 
     const rootDomain = Util.getRootDomain(url);
     if (!rootDomain) return;
@@ -43,6 +78,24 @@ class EntityClassification {
     return unrecognizedEntity;
   }
 
+  /**
+   * Preload Chrome extensions found in the devtoolsLog into cache.
+   * @param {EntityCache} entityCache
+   * @param {LH.DevtoolsLog} devtoolsLog
+   */
+  static _preloadChromeExtensionsToCache(entityCache, devtoolsLog) {
+    for (const entry of devtoolsLog.values()) {
+      if (entry.method !== 'Runtime.executionContextCreated') continue;
+
+      const origin = entry.params.context.origin;
+      if (!origin.startsWith('chrome-extension:')) continue;
+      if (entityCache.has(origin)) continue;
+
+      EntityClassification.makeupChromeExtensionEntity_(entityCache, origin,
+        entry.params.context.name);
+    }
+  }
+
   /**
    * @param {{URL: LH.Artifacts['URL'], devtoolsLog: LH.DevtoolsLog}} data
    * @param {LH.Artifacts.ComputedContext} context
@@ -57,12 +110,14 @@ class EntityClassification {
     /** @type {Map<LH.Artifacts.Entity, Set<string>>} */
     const urlsByEntity = new Map();
 
+    EntityClassification._preloadChromeExtensionsToCache(madeUpEntityCache, data.devtoolsLog);
+
     for (const record of networkRecords) {
       const {url} = record;
       if (entityByUrl.has(url)) continue;
 
       const entity = thirdPartyWeb.getEntity(url) ||
-        EntityClassification.makeUpAnEntity(madeUpEntityCache, url);
+        EntityClassification._makeUpAnEntity(madeUpEntityCache, url);
       if (!entity) continue;
 
       const entityURLs = urlsByEntity.get(entity) || new Set();
@@ -76,7 +131,7 @@ class EntityClassification {
     // See https://github.com/GoogleChrome/lighthouse/issues/13706
     const firstPartyUrl = data.URL.mainDocumentUrl || data.URL.finalDisplayedUrl;
     const firstParty = thirdPartyWeb.getEntity(firstPartyUrl) ||
-      EntityClassification.makeUpAnEntity(madeUpEntityCache, firstPartyUrl);
+      EntityClassification._makeUpAnEntity(madeUpEntityCache, firstPartyUrl);
 
     /**
      * Convenience function to check if a URL belongs to first party.

diff --git a/core/lib/url-utils.js b/core/lib/url-utils.js
@@ -87,6 +87,10 @@ class UrlUtils {
   static getOrigin(url) {
     try {
       const urlInfo = new URL(url);
+      if (urlInfo.protocol === 'chrome-extension:') {
+        // Chrome extensions return string "null" as origin, so we reconstruct the extension origin.
+        return Util.getChromeExtensionOrigin(url);
+      }
       // check for both host and origin since some URLs schemes like data and file set origin to the
       // string "null" instead of the object
       return (urlInfo.host && urlInfo.origin) || null;

diff --git a/core/test/computed/entity-classification-test.js b/core/test/computed/entity-classification-test.js
@@ -123,8 +123,58 @@ describe('Entity Classification computed artifact', () => {
     // Make sure only valid network urls with a domain is recognized.
     expect(entities).toEqual(['third-party.com']);
     expect(result.entityByUrl.size).toBe(1);
+    // First party check returns false for non-DT-log URLs.
+    expect(result.isFirstParty('chrome://version')).toEqual(false);
+  });
+
+  it('classifies chrome-extension URLs and resolves their names', async () => {
+    artifacts.URL = {
+      mainDocumentUrl: 'http://third-party.com',
+    };
+    artifacts.devtoolsLog = networkRecordsToDevtoolsLog([
+      {url: 'http://third-party.com'},
+      {url: 'data:foobar'},
+      {'url': 'chrome-extension://abcdefghijklmnopqrstuvwxyz/foo/bar.js'},
+      {'url': 'chrome-extension://nonresolvablechromextension/bar/baz.js'},
+    ]);
+
+    // Inject an executionContextCreated entry to resolve extension names
+    artifacts.devtoolsLog.push({
+      method: 'Runtime.executionContextCreated',
+      params: {
+        context: {
+          origin: 'chrome-extension://abcdefghijklmnopqrstuvwxyz',
+          name: 'Sample Chrome Extension',
+        },
+      },
+    });
+
+    const result = await EntityClassification.request(artifacts, context);
+    const entities = Array.from(result.urlsByEntity.keys()).map(e => e.name);
+    // Make sure first party is identified.
+    expect(result.firstParty.name).toBe('third-party.com');
+    // Make sure only valid network urls with a domain is recognized.
+    expect(entities).toEqual(['third-party.com', 'Sample Chrome Extension',
+      'nonresolvablechromextension']);
+
+    const extensionEntity = result.entityByUrl
+      .get('chrome-extension://abcdefghijklmnopqrstuvwxyz/foo/bar.js');
+    expect(extensionEntity).toHaveProperty('category', 'Chrome Extension');
+    expect(extensionEntity).toHaveProperty('name', 'Sample Chrome Extension');
+    expect(extensionEntity).toHaveProperty('homepage',
+      'https://chromewebstore.google.com/detail/abcdefghijklmnopqrstuvwxyz');
+
+    const extensionUnknownEntity = result.entityByUrl
+      .get('chrome-extension://nonresolvablechromextension/bar/baz.js');
+    expect(extensionUnknownEntity).toHaveProperty('category', 'Chrome Extension');
+    expect(extensionUnknownEntity).toHaveProperty('name', 'nonresolvablechromextension');
+    expect(extensionUnknownEntity).toHaveProperty('homepage',
+      'https://chromewebstore.google.com/detail/nonresolvablechromextension');
+
+    expect(result.entityByUrl.size).toBe(3);
     // First party check fails for non-DT-log URLs.
-    expect(result.isFirstParty('chrome-extension://abcdefghijklmnopqrstuvwxyz/foo/bar.js')).toEqual(false);
+    expect(result.isFirstParty('chrome-extension://abcdefghijklmnopqrstuvwxyz/foo/bar.js'))
+      .toEqual(false);
     expect(result.isFirstParty('chrome://new-tab-page')).toEqual(false);
   });
 });
diff --git a/report/test/renderer/details-renderer-test.js b/report/test/renderer/details-renderer-test.js
@@ -522,6 +522,8 @@ describe('DetailsRenderer', () => {
             entities: [
               {name: 'example.com', category: 'Cat', isFirstParty: true},
               {name: 'cdn.com', category: 'CDN'},
+              {name: 'Sample Chrome Extension', category: 'Chrome Extension',
+                origins: ['chrome-extension://abcdefghijklmnopqrstuvwxyz']},
             ],
           });
 
@@ -583,7 +585,7 @@ describe('DetailsRenderer', () => {
               {url: 'https://example.com/1', totalBytes: 100, wastedBytes: 500, entity: 'example.com'},
               {url: 'https://cdn.com/1', totalBytes: 300, wastedBytes: 700, entity: 'cdn.com'},
               {url: 'https://cdn.com/2', totalBytes: 400, wastedBytes: 800, entity: 'cdn.com'},
-              {url: 'chrome-extension://abcdefghijklmnopqrstuvwxyz/foo/bar.js', totalBytes: 300, wastedBytes: 700},
+              {url: 'chrome-extension://abcdefghijklmnopqrstuvwxyz/foo/bar.js', totalBytes: 300, wastedBytes: 700, entity: 'Sample Chrome Extension'},
               {url: 'chrome://new-tab-page', totalBytes: 300, wastedBytes: 700},
               {url: 'Unattributable', totalBytes: 500, wastedBytes: 500}, // entity not marked.
             ],
@@ -601,10 +603,15 @@ describe('DetailsRenderer', () => {
           );
           assert.deepStrictEqual(
             [...el.querySelectorAll('.lh-row--group')[2].children].map(td => td.textContent),
-            ['Unattributable', '1.1 KiB', '1.9 KiB'],
+            ['Sample Chrome Extension Chrome Extension', '0.3 KiB', '0.7 KiB'],
+            'did not render Chrome Extensions row'
+          );
+          assert.deepStrictEqual(
+            [...el.querySelectorAll('.lh-row--group')[3].children].map(td => td.textContent),
+            ['Unattributable', '0.8 KiB', '1.2 KiB'],
             'did not render all Unattributable row'
           );
-          assert.equal(el.querySelectorAll('tr').length, 10, `did not render ${tableType} rows`);
+          assert.equal(el.querySelectorAll('tr').length, 11, `did not render ${tableType} rows`);
         });
 
         it('does not group if entity classification is absent', () => {

diff --git a/shared/util.js b/shared/util.js
@@ -260,6 +260,16 @@ class Util {
     return name;
   }
 
+  /**
+   * Returns the origin portion of a Chrome extension URL.
+   * @param {string} url
+   * @return {string}
+   */
+  static getChromeExtensionOrigin(url) {
+    const parsedUrl = new URL(url);
+    return parsedUrl.protocol + '//' + parsedUrl.host;
+  }
+
   /**
    * Split a URL into a file, hostname and origin for easy display.
    * @param {string} url
@@ -270,7 +280,10 @@ class Util {
     return {
       file: Util.getURLDisplayName(parsedUrl),
       hostname: parsedUrl.hostname,
-      origin: parsedUrl.origin,
+      // Node's URL parsing behavior is different than Chrome and returns 'null'
+      // for chrome-extension:// URLs. See https://github.com/nodejs/node/issues/21955.
+      origin: parsedUrl.protocol === 'chrome-extension:' ?
+        Util.getChromeExtensionOrigin(url) : parsedUrl.origin,
     };
   }