Trim version and extraneus parts from AntiVirus product names.

chrome/browser/metrics/antivirus_metrics_provider_win.cc

 // Copyright 2016 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 "chrome/browser/metrics/antivirus_metrics_provider_win.h"
 
 #include <iwscapi.h>
 #include <stddef.h>
 #include <wbemidl.h>
 #include <windows.h>
 #include <wscapi.h>
 
+#include <algorithm>
 #include <string>
 
 #include "base/bind.h"
 #include "base/callback.h"
 #include "base/feature_list.h"
 #include "base/file_version_info_win.h"
 #include "base/files/file_path.h"
 #include "base/files/file_util.h"
 #include "base/metrics/field_trial.h"
 #include "base/metrics/histogram_macros.h"
 #include "base/path_service.h"
+#include "base/strings/string_split.h"
 #include "base/strings/string_util.h"
 #include "base/strings/sys_string_conversions.h"
 #include "base/task_runner_util.h"
 #include "base/threading/thread_restrictions.h"
 #include "base/version.h"
 #include "base/win/scoped_bstr.h"
 #include "base/win/scoped_com_initializer.h"
 #include "base/win/scoped_comptr.h"
 #include "base/win/scoped_variant.h"
 #include "base/win/windows_version.h"
@@ skipping 16 matching lines @@
   uint8_t definition_state : 4;  // 1 = Out of date, 0 = Up to date.
   uint8_t unknown_2 : 4;
   uint8_t security_state : 4;  //  0 = Inactive, 1 = Active, 2 = Snoozed.
   uint8_t security_provider;   // matches WSC_SECURITY_PROVIDER in wscapi.h.
   uint8_t unknown_3;
 };
 #pragma pack(pop)
 
 static_assert(sizeof(PRODUCT_STATE) == 4, "Wrong packing!");
 
+// Filter any part of a product string that looks like it might be a version
+// number.
+bool ShouldFilterPart(const std::string& str) {
+  if (str.empty())
+    return false;
+  // Special case for "360" (used by Norton) and "365" (used by Kaspersky).
+  if (str == "365" || str == "360")
+    return false;
+  for (const auto ch : str) {
+    if (ch == '.')
+      return true;

[rkaplow, 2017/01/12 23:18:56]

shouldn't this not return right away? i.e. this would mark "abc." as a version
string.
Could we verify that all elements of the string are digits/periods/etc?

[Will Harris, 2017/01/12 23:23:59]

this was specifically to handle stuff like

"nProtect Anti-Virus/Spyware V4.0"
"ESET NOD32 Antivirus 9.0.349.15P"

to be honest it doesn't actually matter too much if we're overzealous here as
long as some semblance of the name is retained.

I could add 'V' and 'P' to the blacklist but then it starts to get sort of
crazy... Happy to do what you think here...

[Will Harris, 2017/01/12 23:27:05]

hmm turns out this algorithm doesn't handle "nProtect Anti-Virus/Spyware V4.0"
anyway... working on it...

[Will Harris, 2017/01/12 23:39:06]

Done.

+    if (ch != '_' && !isdigit(ch))
+      return false;
+  }
+  return true;
+}
+
 bool ShouldReportFullNames() {
   // The expectation is that this will be disabled for the majority of users,
   // but this allows a small group to be enabled on other channels if there are
   // a large percentage of hashes collected on these channels that are not
   // resolved to names previously collected on Canary channel.
   bool enabled = base::FeatureList::IsEnabled(
       AntiVirusMetricsProvider::kReportNamesFeature);
 
   if (chrome::GetChannel() == version_info::Channel::CANARY)
     return true;
@@ skipping 108 matching lines @@
 
   MaybeAddUnregisteredAntiVirusProducts(&av_products);
 
   UMA_HISTOGRAM_ENUMERATION("UMA.AntiVirusMetricsProvider.Result",
                             result,
                             RESULT_COUNT);
 
   return av_products;
 }
 
+std::string AntiVirusMetricsProvider::TrimVersionOfAvProductName(
+    const std::string& av_product) {
+  auto av_product_parts = base::SplitString(
+      av_product, " ", base::TRIM_WHITESPACE, base::SPLIT_WANT_NONEMPTY);
+
+  if (av_product_parts.size() >= 2) {
+    // Skipping first element, remove any that look like version numbers.
+    av_product_parts.erase(
+        std::remove_if(av_product_parts.begin() + 1, av_product_parts.end(),
+                       ShouldFilterPart),
+        av_product_parts.end());
+  }
+
+  return base::JoinString(av_product_parts, " ");
+}
+
 void AntiVirusMetricsProvider::GotAntiVirusProducts(
     const base::Closure& done_callback,
     const std::vector<AvProduct>& av_products) {
   DCHECK(thread_checker_.CalledOnValidThread());
   av_products_ = av_products;
   done_callback.Run();
 }
 
 // static
 AntiVirusMetricsProvider::ResultCode
@@ skipping 268 matching lines @@
   std::string product_name("Trusteer Endpoint Protection");
   if (ShouldReportFullNames()) {
     av_product.set_product_name(product_name);
     av_product.set_product_version(product_version);
   }
   av_product.set_product_name_hash(metrics::HashName(product_name));
   av_product.set_product_version_hash(metrics::HashName(product_version));
 
   products->push_back(av_product);
 }