Issue 4599 - Update the page structure more eagerly

chrome/ext/background.js

 /*
  * This file is part of Adblock Plus <https://adblockplus.org/>,
  * Copyright (C) 2006-2016 Eyeo GmbH
  *
  * Adblock Plus is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 3 as
  * published by the Free Software Foundation.
  *
  * Adblock Plus is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ skipping 107 matching lines @@
 
     var frame = frames[frameId];
     if (!frame)
       frame = frames[frameId] = {};
 
     return frame;
   }
 
   var eagerlyUpdatedPages = new ext.PageMap();
 
-  ext._updatePageFrameStructure = function(frameId, tabId, url, eager,
-                                           parentFrameId)
+  ext._updatePageFrameStructure = function(frameId, tabId, url, eager)
   {
     if (frameId == 0)
     {
       let page = new Page({id: tabId, url: url});
 
       if (eagerlyUpdatedPages.get(page) != url)
       {
         ext._removeFromAllPageMaps(tabId);
 
         // When a sitekey header is received we must immediately update the page
@@ skipping 13 matching lines @@
           // when Chrome automatically resets them on navigation.
           if (chrome.runtime.lastError)
             ext.pages.onLoading._dispatch(page);
         });
       }
     }
 
     // Update frame URL in frame structure
     var frame = createFrame(tabId, frameId);
     frame.url = new URL(url);
-
-    // We only get the parentFrameId for onBeforeNavigate, not for onCommitted
-    if (typeof parentFrameId != "undefined")
-      frame.parent = framesOfTabs[tabId][parentFrameId] || null;
   };
 
   chrome.webNavigation.onBeforeNavigate.addListener(function(details)
   {
-    // Since requests are sometimes made by a page before its onCommitted event
-    // fires we also need to update the page structure here. Otherwise those
-    // requests will be considered third party since a URL for the page wasn't
-    // yet determined.
-    ext._updatePageFrameStructure(details.frameId, details.tabId,
-                                  details.url, false, details.parentFrameId);

[Wladimir Palant, 2016/11/14 08:18:44]

You are essentially reverting changes from issue 4386, the onCommitted handler
no longer serves any purpose. This is bad: onBeforeNavigate is often called when
no navigation occurs in the end. In this scenario you will end up misattributing
all requests from a tab to a page that never loaded.

+    // Capture parent frame here because onCommitted doesn't get this info.
+    var frame = createFrame(details.tabId, details.frameId);
+    frame.parent = framesOfTabs[details.tabId][details.parentFrameId] || null;
+
+    // Chrome will sometimes perform requests for pre-rendered pages before
+    // navigation was confirmed and the onComitted event was fired. Those
+    // requests will often wrongly be considered to be third party, since we
+    // don't know the correct document domain for requests until after we've
+    // updated the page structure. As a work around we update the page
+    // structure eagerly for new tabs (which includes pre-rendered pages).
+    if (!(details.tabId in framesOfTabs))

[Wladimir Palant, 2016/11/14 15:56:29]

While this is probably not wrong, I think that you are only solving a small part
of the issue with this approach...

[kzar, 2016/11/15 12:19:28]

Well so far the issue only described a problem with pre-rendered tabs, but I
have just tested and confirmed you're indeed right that the problem is not
limited to them. When clicking a link to http://www.apptap.com I saw requests
being made just before its onComitted event fired. Back to the drawing board...

+      ext._updatePageFrameStructure(details.frameId, details.tabId, details.url);
   });
 
   chrome.webNavigation.onCommitted.addListener(function(details)
   {
-    ext._updatePageFrameStructure(details.frameId, details.tabId,
-                                  details.url, false);
+    ext._updatePageFrameStructure(details.frameId, details.tabId, details.url);
   });
 
   function forgetTab(tabId)
   {
     ext.pages.onRemoved._dispatch(tabId);
 
     ext._removeFromAllPageMaps(tabId);
     delete framesOfTabs[tabId];
   }
 
@@ skipping 455 matching lines @@
   ext.windows = {
     create: function(createData, callback)
     {
       chrome.windows.create(createData, function(createdWindow)
       {
         afterTabLoaded(callback)(createdWindow.tabs[0]);
       });
     }
   };
 })();