findLastDocumentForUrl

GoogleChrome · Aug 9, 2023 · 86b9602 · 86b9602
1 parent 6cf368b
commit 86b9602
Show file tree

Hide file tree

Showing 3 changed files with 20 additions and 9 deletions.
diff --git a/cli/test/smokehouse/test-definitions/service-worker-reloaded.js b/cli/test/smokehouse/test-definitions/service-worker-reloaded.js
@@ -4,7 +4,7 @@
  * Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
  */
 
-/** @type {LH.Config.Json} */
+/** @type {LH.Config} */
 const config = {
   extends: 'lighthouse:default',
   settings: {

diff --git a/core/computed/main-resource.js b/core/computed/main-resource.js
@@ -6,7 +6,7 @@
 
 import {makeComputedArtifact} from './computed-artifact.js';
 import {NetworkRecords} from './network-records.js';
-import UrlUtils from '../lib/url-utils.js';
+import {NetworkAnalyzer} from '../lib/dependency-graph/simulator/network-analyzer.js';
 
 /**
  * @fileoverview This artifact identifies the main resource on the page. Current solution assumes
@@ -21,18 +21,14 @@ class MainResource {
   static async compute_(data, context) {
     const {mainDocumentUrl} = data.URL;
     if (!mainDocumentUrl) throw new Error('mainDocumentUrl must exist to get the main resource');
-    const requests = await NetworkRecords.request(data.devtoolsLog, context);
+    const records = await NetworkRecords.request(data.devtoolsLog, context);
 
-    const mainResourceRequests = requests.filter(request =>
-      request.resourceType === 'Document' &&
-      UrlUtils.equalWithExcludedFragments(request.url, mainDocumentUrl)
-    );
     // We could have more than one record matching the main doucment url,
     // if the page did `location.reload()`. Since `mainDocumentUrl` refers to the _last_
     // document request, we should return the last candidate here. Besides, the browser
     // would have evicted the first request by the time `MainDocumentRequest` (a consumer
     // of this computed artifact) attempts to fetch the contents, resulting in a protocol error.
-    const mainResource = mainResourceRequests[mainResourceRequests.length - 1];
+    const mainResource = NetworkAnalyzer.findLastDocumentForUrl(records, mainDocumentUrl);
     if (!mainResource) {
       throw new Error('Unable to identify the main resource');
     }

diff --git a/core/lib/dependency-graph/simulator/network-analyzer.js b/core/lib/dependency-graph/simulator/network-analyzer.js
@@ -506,9 +506,24 @@ class NetworkAnalyzer {
     );
   }
 
+  /**
+   * @param {Array<LH.Artifacts.NetworkRequest>} records
+   * @param {string} resourceUrl
+   * @return {LH.Artifacts.NetworkRequest|undefined}
+   */
+  static findLastDocumentForUrl(records, resourceUrl) {
+    // equalWithExcludedFragments is expensive, so check that the resourceUrl starts with the request url first
+    const matchingRequests = records.filter(request =>
+      request.resourceType === 'Document' &&
+      resourceUrl.startsWith(request.url) &&
+      UrlUtils.equalWithExcludedFragments(request.url, resourceUrl)
+    );
+    return matchingRequests[matchingRequests.length - 1];
+  }
+
   /**
    * Resolves redirect chain given a main document.
-   * See: {@link NetworkAnalyzer.findResourceForUrl}) for how to retrieve main document.
+   * See: {@link NetworkAnalyzer.findLastDocumentForUrl}) for how to retrieve main document.
    *
    * @param {LH.Artifacts.NetworkRequest} request
    * @return {LH.Artifacts.NetworkRequest}