Only track http and https in duplicate resource handler metrics

content/browser/renderer_host/duplicate_content_resource_handler.cc

 // Copyright (c) 2012 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.
 
 #include "content/browser/renderer_host/duplicate_content_resource_handler.h"
 
 #include <set>
 
 #include "base/lazy_instance.h"
 #include "base/logging.h"
 #include "base/metrics/histogram.h"
+#include "base/string_util.h"
 #include "content/browser/renderer_host/resource_request_info_impl.h"
 #include "net/base/io_buffer.h"
 #include "net/url_request/url_request.h"
 
 namespace content {
 namespace {
 
 class GlobalDuplicateRecords {
  public:
   static GlobalDuplicateRecords* GetInstance() {
@@ skipping 60 matching lines @@
     const net::URLRequestStatus& status,
     const std::string& security_info) {
 
   if (status.is_success())
     RecordContentMetrics();
 
   return next_handler_->OnResponseCompleted(request_id, status, security_info);
 }
 
 void DuplicateContentResourceHandler::RecordContentMetrics() {
+  // Ignore data and blob URLs.
+  const std::string& url_spec = request_->url().spec();
+  if (StartsWithASCII(url_spec, "data:", false) ||

[gavinp, 2012/08/09 17:28:05]

So since data URLs will usually have the same URL for the same content , the
total cost from these should be about 8 bytes per unique data URL, right?

For the blob: URLs, I wonder.

Frank, I think none of these resources are being retrieved from the web. So if
your goal is to track web retrievals, this exclusion makes sense. What I don't
quite understand is how many of these there are in practice.

+      StartsWithASCII(url_spec, "blob:", false)) {
+    bytes_read_ = 0;
+    read_buffer_ = NULL;
+    return;
+  }
   MH_UINT32 contents_hash = PMurHash32_Result(pmurhash_ph1_,
                                               pmurhash_pcarry_, bytes_read_);
-
   // Combine the contents_hash with the url, so we can test if future content
   // identical resources have the same original url or not.
-  MH_UINT32 hashed_with_url;
-  const std::string& url_spec = request_->url().spec();
   PMurHash32_Process(&pmurhash_ph1_, &pmurhash_pcarry_,
                      url_spec.data(), url_spec.length());
-  hashed_with_url = PMurHash32_Result(pmurhash_ph1_, pmurhash_pcarry_,
+  MH_UINT32 hashed_with_url = PMurHash32_Result(pmurhash_ph1_, pmurhash_pcarry_,
                                       url_spec.length() + bytes_read_);

[frankwang, 2012/08/09 17:21:12]

nit: indentation

 
   DVLOG(4) << "url: " << url_spec;
   DVLOG(4) << "contents hash: " << contents_hash;
   DVLOG(4) << "hash with url: " << hashed_with_url;
 
   std::set<MH_UINT32>* content_matches =
       GlobalDuplicateRecords::GetInstance()->content_matches();
   std::set<MH_UINT32>* content_and_url_matches =
       GlobalDuplicateRecords::GetInstance()->content_and_url_matches();
 
@@ skipping 12 matching lines @@
                                resource_type_, ResourceType::LAST_TYPE);
   }
   content_matches->insert(contents_hash);
   content_and_url_matches->insert(hashed_with_url);
 
   bytes_read_ = 0;
   read_buffer_ = NULL;
 }
 
 }  // namespace content