The optional Malware Details also collects HTTP cache information. See design...

chrome/browser/safe_browsing/malware_details.cc

-// Copyright (c) 2010 The Chromium Authors. All rights reserved.
+// Copyright (c) 2011 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 //
 // Implementation of the MalwareDetails class.
 
 #include "chrome/browser/safe_browsing/malware_details.h"
 
+#include "base/callback.h"
 #include "base/lazy_instance.h"
+#include "base/md5.h"
+#include "chrome/browser/net/chrome_url_request_context.h"
+#include "chrome/browser/profiles/profile.h"
+#include "chrome/browser/safe_browsing/malware_details_cache.h"
+#include "chrome/browser/safe_browsing/report.pb.h"
 #include "chrome/browser/safe_browsing/safe_browsing_service.h"
-#include "chrome/browser/safe_browsing/report.pb.h"
+#include "chrome/common/net/url_request_context_getter.h"
 #include "chrome/common/safebrowsing_messages.h"
 #include "content/browser/browser_thread.h"
 #include "content/browser/renderer_host/render_view_host.h"
 #include "content/browser/tab_contents/navigation_entry.h"
 #include "content/browser/tab_contents/tab_contents.h"
+#include "net/base/io_buffer.h"
+#include "net/disk_cache/disk_cache.h"
+#include "net/http/http_cache.h"
+#include "net/http/http_response_headers.h"
+#include "net/http/http_response_info.h"
 
 using safe_browsing::ClientMalwareReportRequest;
 
 // Keep in sync with KMaxNodes in renderer/safe_browsing/malware_dom_details
 static const uint32 kMaxDomNodes = 500;
 
+// Only send tiny files for now, a better strategy would use the size
+// of the whole report and the user's bandwidth.
+static const int kMaxBodySize = 1024;
+
 // static
 MalwareDetailsFactory* MalwareDetails::factory_ = NULL;
 
 // The default MalwareDetailsFactory.  Global, made a singleton so we
 // don't leak it.
 class MalwareDetailsFactoryImpl
     : public MalwareDetailsFactory {
  public:
   MalwareDetails* CreateMalwareDetails(
+      SafeBrowsingService* sb_service,
       TabContents* tab_contents,
       const SafeBrowsingService::UnsafeResource& unsafe_resource) {
-    return new MalwareDetails(tab_contents, unsafe_resource);
+    return new MalwareDetails(sb_service, tab_contents, unsafe_resource);
   }
 
  private:
   friend struct base::DefaultLazyInstanceTraits<
       MalwareDetailsFactoryImpl>;
 
   MalwareDetailsFactoryImpl() { }
 
   DISALLOW_COPY_AND_ASSIGN(MalwareDetailsFactoryImpl);
 };
 
 static base::LazyInstance<MalwareDetailsFactoryImpl>
     g_malware_details_factory_impl(base::LINKER_INITIALIZED);
 
 // Create a MalwareDetails for the given tab.
 /* static */
 MalwareDetails* MalwareDetails::NewMalwareDetails(
+    SafeBrowsingService* sb_service,
     TabContents* tab_contents,
     const SafeBrowsingService::UnsafeResource& resource) {
   // Set up the factory if this has not been done already (tests do that
   // before this method is called).
   if (!factory_)
     factory_ = g_malware_details_factory_impl.Pointer();
-  return factory_->CreateMalwareDetails(tab_contents, resource);
+  return factory_->CreateMalwareDetails(sb_service, tab_contents, resource);
 }
 
 // Create a MalwareDetails for the given tab. Runs in the UI thread.
 MalwareDetails::MalwareDetails(
+    SafeBrowsingService* sb_service,
     TabContents* tab_contents,
     const SafeBrowsingService::UnsafeResource& resource)
     : TabContentsObserver(tab_contents),
-      resource_(resource) {
+      request_context_getter_(tab_contents->profile()->GetRequestContext()),
+      sb_service_(sb_service),
+      resource_(resource),
+      cache_collector_(new MalwareDetailsCacheCollector) {
   StartCollection();
 }
 
-MalwareDetails::~MalwareDetails() {}
+MalwareDetails::~MalwareDetails() {
+}
 
 bool MalwareDetails::OnMessageReceived(const IPC::Message& message) {
   bool handled = true;
   IPC_BEGIN_MESSAGE_MAP(MalwareDetails, message)
     IPC_MESSAGE_HANDLER(SafeBrowsingHostMsg_MalwareDOMDetails,
                         OnReceivedMalwareDOMDetails)
     IPC_MESSAGE_UNHANDLED(handled = false)
   IPC_END_MESSAGE_MAP()
   return handled;
 }
 
 bool MalwareDetails::IsPublicUrl(const GURL& url) const {
   return url.SchemeIs("http");  // TODO(panayiotis): also skip internal urls.
 }
 
 // Looks for a Resource for the given url in resources_.  If found, it
 // updates |resource|. Otherwise, it creates a new message, adds it to
 // resources_ and updates |resource| to point to it.
 ClientMalwareReportRequest::Resource* MalwareDetails::FindOrCreateResource(
     const GURL& url) {
-  ResourceMap::iterator it = resources_.find(url.spec());
+  safe_browsing::ResourceMap::iterator it = resources_.find(url.spec());
   if (it != resources_.end()) {
     return it->second.get();
   }
 
   // Create the resource for |url|.
   int id = resources_.size();
   linked_ptr<ClientMalwareReportRequest::Resource> new_resource(
       new ClientMalwareReportRequest::Resource());
   new_resource->set_url(url.spec());
   new_resource->set_id(id);
@@ skipping 98 matching lines @@
   // of our data structures (eg GetSerializedReport).
   BrowserThread::PostTask(
       BrowserThread::IO, FROM_HERE,
       NewRunnableMethod(
           this, &MalwareDetails::AddDOMDetails, params));
 }
 
 void MalwareDetails::AddDOMDetails(
     const std::vector<SafeBrowsingHostMsg_MalwareDOMDetails_Node>& params) {
   DCHECK(BrowserThread::CurrentlyOn(BrowserThread::IO));
+  DVLOG(1) << "Nodes from the DOM: " << params.size();
+
+  // If we have already started collecting data from the HTTP cache, don't
+  // modify our state.
+  if (cache_collector_->InProgress())
+    return;
+
   // Add the urls from the DOM to |resources_|.  The renderer could be
   // sending bogus messages, so limit the number of nodes we accept.
-  DVLOG(1) << "Nodes from the DOM: " << params.size();
   for (uint32 i = 0; i < params.size() && i < kMaxDomNodes; ++i) {
     SafeBrowsingHostMsg_MalwareDOMDetails_Node node = params[i];
     DVLOG(1) << node.url << ", " << node.tag_name << ", " << node.parent;
     AddUrl(node.url, node.parent, node.tag_name, &(node.children));
   }
 }
 
 // Called from the SB Service on the IO thread, after the user has
 // closed the tab, or clicked proceed or goback.  Since the user needs
 // to take an action, we expect this to be called after
 // OnReceivedMalwareDOMDetails in most cases. If not, we don't include
 // the DOM data in our report.
-const std::string* MalwareDetails::GetSerializedReport() {
+void MalwareDetails::FinishCollection() {
   DCHECK(BrowserThread::CurrentlyOn(BrowserThread::IO));
-  // The |report_| protocol buffer is now generated: We add all the
-  // urls in our |resources_| maps.
-  for (ResourceMap::const_iterator it = resources_.begin();
+
+  AddRef();  // Ensures we remain alive during cache collection.

[panayiotis, 2011/03/29 21:18:17]

I was able to get rid of this AddRef. Now the only one that remains is inside
the cache collector, due to the calls to http_cache->GetBackend() 

+  cache_collector_->StartCacheCollection(
+      request_context_getter_,
+      &resources_,
+      NewCallback(this, &MalwareDetails::OnCacheCollectionReady));
+}
+
+void MalwareDetails::OnCacheCollectionReady(int result) {
+  DVLOG(1) << "OnCacheCollectionReady.";
+  // Add all the urls in our |resources_| maps to the |report_| protocol buffer.
+  for (safe_browsing::ResourceMap::const_iterator it = resources_.begin();
        it != resources_.end(); it++) {
     ClientMalwareReportRequest::Resource* pb_resource =
         report_->add_resources();
     pb_resource->CopyFrom(*(it->second));
   }
+  report_->set_complete(result == net::OK);
 
-  scoped_ptr<std::string> request_data(new std::string());
-  if (!report_->SerializeToString(request_data.get())) {
+  // Send the report, using the SafeBrowsingService.
+  std::string serialized;
+  if (!report_->SerializeToString(&serialized)) {
     DLOG(ERROR) << "Unable to serialize the malware report.";
   }
 
-  return request_data.release();
+  sb_service_->SendSerializedMalwareDetails(serialized);
+  Release();
 }