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

chrome/browser/safe_browsing/malware_details.h

 // Copyright (c) 2010 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.
 
 #ifndef CHROME_BROWSER_SAFE_BROWSING_MALWARE_DETAILS_H_
 #define CHROME_BROWSER_SAFE_BROWSING_MALWARE_DETAILS_H_
 #pragma once
 
 // A class that encapsulates the detailed malware reports sent when
 // users opt-in to do so from the malware warning page.
 
 // An instance of this class is generated when a malware warning page
 // is shown (SafeBrowsingBlockingPage). It is passed on to the
 // SafeBrowsing service when the warning goes away.
 
 #include <string>
 #include <vector>
 
 #include "base/hash_tables.h"
 #include "base/linked_ptr.h"
 #include "base/scoped_ptr.h"
 #include "chrome/browser/safe_browsing/report.pb.h"
 #include "chrome/browser/safe_browsing/safe_browsing_service.h"
 #include "content/browser/tab_contents/tab_contents_observer.h"
+#include "net/base/completion_callback.h"
 
 class TabContents;
 struct ViewHostMsg_MalwareDOMDetails_Params;
 
+namespace disk_cache {
+class Backend;
+class Entry;
+}
+namespace net {
+class IOBuffer;
+class URLRequestContext;
+}
+
 class MalwareDetailsFactory;
 
 class MalwareDetails : public base::RefCountedThreadSafe<MalwareDetails>,
                        public TabContentsObserver {
  public:
   // Constructs a new MalwareDetails instance, using the factory.
   static MalwareDetails* NewMalwareDetails(
       TabContents* tab_contents,
       const SafeBrowsingService::UnsafeResource& resource);
 
   // Makes the passed |factory| the factory used to instanciate
   // SafeBrowsingBlockingPage objects. Useful for tests.
   static void RegisterFactory(MalwareDetailsFactory* factory) {
     factory_ = factory;
   }
 
+  // Callback with the malware details, serialized. See GetSerializedReport.
+  typedef Callback1<scoped_refptr<MalwareDetails> >::Type
+      MalwareDetailsCallback;
+
   // The SafeBrowsingService calls this from the IO thread, to get the
   // serialized report as a string and send it over.
+  // |callback| will be notified when the operation completes.
+  void StartCacheCollection(MalwareDetailsCallback* callback);
+
   const std::string* GetSerializedReport();
 
   // TabContentsObserver implementation.
   virtual bool OnMessageReceived(const IPC::Message& message);
 
  protected:
   friend class MalwareDetailsFactoryImpl;
 
   MalwareDetails(TabContents* tab_contents,
                  const SafeBrowsingService::UnsafeResource resource);
 
   // Called on the IO thread with the DOM details.
   virtual void AddDOMDetails(
       const ViewHostMsg_MalwareDOMDetails_Params& params);
 
+  virtual net::URLRequestContext* GetURLRequestContext();
+
   virtual ~MalwareDetails();

[lzheng, 2011/03/05 00:37:50]

I think ~MalwareDetails() should be right after MalwareDetails(xxx).

[panayiotis, 2011/03/29 18:14:55]

Done.

 
+  // The report protocol buffer.
+  scoped_ptr<safe_browsing::ClientMalwareReportRequest> report_;
+
  private:
   friend class base::RefCountedThreadSafe<MalwareDetails>;
 
   // Maps a URL to its Resource.
   typedef base::hash_map<
     std::string,
     linked_ptr<safe_browsing::ClientMalwareReportRequest::Resource> >
   ResourceMap;

[lzheng, 2011/03/05 00:37:50]

I would suggest to move the cache-mining code to its own class. It will make the
MalwareDetails class easier to follow. Besides the cache relate code is relative
complicated and also self-containing.

You might want to add some comments regarding the relations among the threads.

[panayiotis, 2011/03/29 18:14:55]

Done.

 
+  enum CacheState {
+    STATE_NONE,
+    STATE_OPEN_CACHE,
+    STATE_OPEN_CACHE_COMPLETE,
+    STATE_OPEN_ENTRY,
+    STATE_OPEN_ENTRY_COMPLETE,
+    STATE_READ_RESPONSE,
+    STATE_READ_RESPONSE_COMPLETE,
+    STATE_READ_DATA,
+    STATE_READ_DATA_COMPLETE,
+    STATE_CACHE_DONE,
+  };
+
   // Starts the collection of the report.
   void StartCollection();
 
   // Whether the url is "public" so we can add it to the report.
   bool IsPublicUrl(const GURL& url) const;
 
   // Finds an existing Resource for the given url, or creates a new
   // one if not found, and adds it to |resources_|. Returns the
   // found/created resource.
   safe_browsing::ClientMalwareReportRequest::Resource* FindOrCreateResource(
       const GURL& url);
 
   // Adds a Resource to resources_ with the given parent-child
   // relationship. |parent| and |tagname| can be empty, |children| can be NULL.
   void AddUrl(const GURL& url,
               const GURL& parent,
               const std::string& tagname,
               const std::vector<GURL>* children);
 
   // Message handler.
   void OnReceivedMalwareDOMDetails(
       const ViewHostMsg_MalwareDOMDetails_Params& params);
 
+  // Methods for getting the HTML body and HTTP headers from the HTTP cache.
+  // They run on the IO thread.
+  void CacheLoop(int result);
+  int DoOpenCache();
+  int DoOpenCacheComplete(int result);
+  int DoOpenEntry();
+  int DoOpenEntryComplete(int result);
+  int DoReadResponse();
+  int DoReadResponseComplete(int result);
+  int DoReadData();
+  int DoReadDataComplete(int result);
+  void DoCacheDone(int result);
+
   const SafeBrowsingService::UnsafeResource resource_;
 
   // For every Url we collect we create a Resource message. We keep
   // them in a map so we can avoid duplicates.
   ResourceMap resources_;
 
-  // The report protocol buffer.
-  scoped_ptr<safe_browsing::ClientMalwareReportRequest> report_;
-
   // The factory used to instanciate SafeBrowsingBlockingPage objects.
   // Usefull for tests, so they can provide their own implementation of
   // SafeBrowsingBlockingPage.
   static MalwareDetailsFactory* factory_;
 
+  // The next cache loop state.
+  CacheState cache_state_;
+
+  // Used to get a pointer to the HTTP cache.
+  URLRequestContextGetter* request_context_getter_;
+
+  // The disk cache.
+  disk_cache::Backend* cache_;
+
+  // Callback for when the cache entry is ready.
+  net::CompletionCallbackImpl<MalwareDetails> cache_callback_;
+
+  // Callback for when the cache entry is opened.
+  scoped_refptr<
+    net::CancelableCompletionCallback<MalwareDetails> > entry_callback_;
+
+  // SafeBrowsingService provides us a method to call when we are done
+  // with the cache. The service outlives the MalwareDetails (in fact, it
+  // takes ownership of MalwareDetails).
+  MalwareDetailsCallback* callback_;
+
+  // Points to the url for which we are fetching the HTTP cache entry or
+  // redirect chain.
+  ResourceMap::iterator resources_it_;
+
+  // The current disk entry.
+  disk_cache::Entry* entry_;
+
+  // The current buffer for reading the contents of the disk entry.
+  scoped_refptr<net::IOBuffer> buf_;
+  int buf_len_;
+
   FRIEND_TEST_ALL_PREFIXES(MalwareDetailsTest, MalwareDOMDetails);
+  FRIEND_TEST_ALL_PREFIXES(MalwareDetailsTest, HTTPCache);
+  FRIEND_TEST_ALL_PREFIXES(MalwareDetailsTest, HTTPCacheNoEntries);
 
   DISALLOW_COPY_AND_ASSIGN(MalwareDetails);
 };
 
 // Factory for creating MalwareDetails.  Useful for tests.
 class MalwareDetailsFactory {
  public:
   virtual ~MalwareDetailsFactory() { }
 
   virtual MalwareDetails* CreateMalwareDetails(
       TabContents* tab_contents,
       const SafeBrowsingService::UnsafeResource& unsafe_resource) = 0;
 };
 
 #endif  // CHROME_BROWSER_SAFE_BROWSING_MALWARE_DETAILS_H_