Add ftp protocol to the safebrowsing checks. Fixed the crash with canonicalization of non-standard urls in CanonicalizeUrl function. Add CanCheckUrl to a more central location. Added the null check in GeneratePathsToCheck. Remove redundant checks from resource dispatcher host.

BUG=40605
TEST=SafeBrowsingUtilTest.CanonicalizeUrl

Review URL: http://codereview.chromium.org/2471002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@48720 0039d316-1c4b-4281-b951-d872f2087c98

3 files changed, 32 insertions, 2 deletions

diff --git a/chrome/browser/safe_browsing/safe_browsing_service.cc b/chrome/browser/safe_browsing/safe_browsing_service.cc
index ede1587..7999296 100644
--- a/chrome/browser/safe_browsing/safe_browsing_service.cc
+++ b/chrome/browser/safe_browsing/safe_browsing_service.cc
@@ -60,7 +60,8 @@ void SafeBrowsingService::ShutDown() {
 }
 
 bool SafeBrowsingService::CanCheckUrl(const GURL& url) const {
-  return url.SchemeIs(chrome::kHttpScheme) ||
+  return url.SchemeIs(chrome::kFtpScheme) ||
+         url.SchemeIs(chrome::kHttpScheme) ||
          url.SchemeIs(chrome::kHttpsScheme);
 }
 
@@ -69,6 +70,9 @@ bool SafeBrowsingService::CheckUrl(const GURL& url, Client* client) {
   if (!enabled_)
     return true;
 
+  if (!CanCheckUrl(url))
+    return true;
+
   if (!MakeDatabaseAvailable()) {
     QueuedCheck check;
     check.client = client;
diff --git a/chrome/browser/safe_browsing/safe_browsing_util.cc b/chrome/browser/safe_browsing/safe_browsing_util.cc
index 3fa3a27..80c3d3ad 100644
--- a/chrome/browser/safe_browsing/safe_browsing_util.cc
+++ b/chrome/browser/safe_browsing/safe_browsing_util.cc
@@ -214,6 +214,12 @@ void CanonicalizeUrl(const GURL& url,
                      std::string* canonicalized_hostname,
                      std::string* canonicalized_path,
                      std::string* canonicalized_query) {
+  DCHECK(url.is_valid());
+
+  // We only canonicalize "normal" URLs.
+  if (!url.IsStandard())
+    return;
+
   // Following canonicalization steps are excluded since url parsing takes care
   // of those :-
   // 1. Remove any tab (0x09), CR (0x0d), and LF (0x0a) chars from url.
@@ -354,7 +360,7 @@ void GeneratePathsToCheck(const GURL& url, std::vector<std::string>* paths) {
       paths->push_back(std::string(path.begin(), i + 1));
   }
 
-  if (paths->back() != path)
+  if (!paths->empty() && paths->back() != path)
     paths->push_back(path);
 
   if (!query.empty())
diff --git a/chrome/browser/safe_browsing/safe_browsing_util_unittest.cc b/chrome/browser/safe_browsing/safe_browsing_util_unittest.cc
index d74982c..5204bfe 100644
--- a/chrome/browser/safe_browsing/safe_browsing_util_unittest.cc
+++ b/chrome/browser/safe_browsing/safe_browsing_util_unittest.cc
@@ -239,6 +239,26 @@ TEST(SafeBrowsingUtilTest, CanonicalizeUrl) {
       "host.com",
       "/abc/def/xyz",
       ""
+    }, {
+      "ftp://host.com/foo?bar",
+      "host.com",
+      "/foo",
+      "bar"
+    }, {
+      "data:text/html;charset=utf-8,%0D%0A",
+      "",
+      "",
+      ""
+    }, {
+      "javascript:alert()",
+      "",
+      "",
+      ""
+    }, {
+      "mailto:abc@example.com",
+      "",
+      "",
+      ""
     },
   };
   for (size_t i = 0; i < ARRAYSIZE_UNSAFE(tests); ++i) {