Crawler frontend

lib/crawler.js

 Cu.import("resource://gre/modules/Services.jsm");
 
 function abprequire(module)
 {
   let result = {};
   result.wrappedJSObject = result;
   Services.obs.notifyObservers(result, "adblockplus-require", module);
   return result.exports;
 }
 
 let {Storage} = require("storage");
 let {Client} = require("client");
 
 let {Policy} = abprequire("contentPolicy");
 let {Filter} = abprequire("filterClasses");
+let {Utils} = abprequire("utils");
 
 let origProcessNode = Policy.processNode;
 
 let siteTabs;
 let currentTabs;
 
 function processNode(wnd, node, contentType, location, collapse)
 {
   let result = origProcessNode.apply(this, arguments);
   let url = (contentType === Policy.type.ELEMHIDE) ? location.text :
       location.spec;
-  let site = siteTabs[wnd.top.location.href];
+
+  let topWindow = wnd.top;
+  if (!topWindow.document)
+  {
+    Cu.reportError("No document associated with the node's top window");
+    return result;
+  }
+
+  let tabbrowser = Utils.getChromeWindow(topWindow).gBrowser;
+  if (!tabbrowser)
+  {
+    Cu.reportError("Unable to get a tabbrowser reference");
+    return result;
+  }
+
+  let browser = tabbrowser.getBrowserForDocument(topWindow.document);
+  if (!browser)
+  {
+    Cu.reportError("Unable to get browser for the tab");
+    return result;
+  }
+
+  let site = siteTabs.get(browser);
   let filtered = !result;
   Storage.write([url, site, filtered]);
   return result;
 }
 
 function loadSite(site, window, callback)
 {
   if (!site)
     return;
 
   let tabbrowser = window.gBrowser;
   let tab = tabbrowser.addTab(site);
   let browser = tabbrowser.getBrowserForTab(tab);
 
+  siteTabs.set(browser, site);
+
   let progressListener = {
     onStateChange: function(aBrowser, aWebProgress, aRequest, aStateFlags, aStatus)
     {
       if (browser !== aBrowser)
         return;
 
       if (!(aStateFlags & Ci.nsIWebProgressListener.STATE_STOP))
         return;
 
       tabbrowser.removeTabsProgressListener(progressListener);
       tabbrowser.removeTab(tab);
       callback();
-    },
-    onLocationChange: function(aBrowser, aWebProgress, aRequest, aLocation, aFlags)
-    {
-      if (browser === aBrowser)
-        siteTabs[aLocation.spec] = site;

[Wladimir Palant, 2012/09/25 09:40:39]

Taking over comment from previous review: this needs to be done properly,
attaching data to the <browser> element should work.

     }
   };
   tabbrowser.addTabsProgressListener(progressListener);
 }
 
 function loadSites(backendUrl, parallelTabs, window, sites, callback)
 {
   while (currentTabs < parallelTabs && sites.length)
   {
     currentTabs++;
     let site = sites.shift();
     loadSite(site, window, function()
     {
       currentTabs--;
       if (!sites.length && !currentTabs)
       {
         Storage.finish();
         let dataFilePath = Storage.dataFile.path;
-        Client.sendCrawlerDataFile(backendUrl, window, dataFilePath, function()
+        Client.sendCrawlerDataFile(backendUrl, dataFilePath, function()
         {
           Storage.destroy();
           callback();
         });
       }
       else
         loadSites(backendUrl, parallelTabs, window, sites, callback);
     });
   }
 }
 
 let Crawler = exports.Crawler = {};
 
 Crawler.crawl = function(backendUrl, parallelTabs, window, callback)
 {
   if (Policy.processNode != origProcessNode)
     return;
 
   Policy.processNode = processNode;
 
-  siteTabs = {};
+  siteTabs = new WeakMap();
   currentTabs = 0;
 
   Storage.init();
 
   Client.fetchCrawlableSites(backendUrl, function(sites)
   {
     loadSites(backendUrl, parallelTabs, window, sites, function()
     {
       Policy.processNode = origProcessNode;
+      siteTabs = null;
       callback();
     });
   });
 };