Reduce buggy usage of the registry controlled domain service.

url/url_canon_unittest.cc

 // Copyright 2013 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 <errno.h>
 #include <stddef.h>
 
 #include "base/macros.h"
 #include "testing/gtest/include/gtest/gtest.h"
 #include "url/third_party/mozilla/url_parse.h"
@@ skipping 869 matching lines @@
 
   // This tests tests.
   const char spec[] = "192.168.0.1";
   CanonicalizeIPAddress(spec, Component(), &output1, &host_info);
   EXPECT_FALSE(host_info.IsIPAddress());
 
   CanonicalizeIPAddress(spec, Component(0, 0), &output1, &host_info);
   EXPECT_FALSE(host_info.IsIPAddress());
 }
 
+// Verifies that CanonicalizeHostSubstring produces the expected output and
+// does not "fix" IP addresses. Because this code is a subset of
+// CanonicalizeHost, the shared functionality is not tested.
+TEST(URLCanonTest, CanonicalizeHostSubstring) {
+  // Basic sanity check.
+  {
+    std::string out_str;
+    StdStringCanonOutput output(&out_str);
+    EXPECT_TRUE(CanonicalizeHostSubstring(
+        "M\xc3\x9cNCHEN.com", Component(0, 12), &output));
+    output.Complete();
+    EXPECT_EQ("xn--mnchen-3ya.com", out_str);
+  }
+
+  // Failure case.
+  {
+    std::string out_str;
+    StdStringCanonOutput output(&out_str);
+    EXPECT_FALSE(CanonicalizeHostSubstring(
+        WStringToUTF16(L"\xfdd0zyx.com").c_str(), Component(0, 8), &output));
+    output.Complete();
+    EXPECT_EQ("%EF%BF%BDzyx.com", out_str);
+  }
+
+  // Should return true for empty input strings.
+  {
+    std::string out_str;
+    StdStringCanonOutput output(&out_str);
+    EXPECT_TRUE(CanonicalizeHostSubstring("", Component(0, 0), &output));
+    output.Complete();
+    EXPECT_EQ("", out_str);

[Peter Kasting, 2016/10/22 05:04:20]

Nit: Since |out_str| is a std::string I tend to prefer EXPECT_EQ(std::string(),
out_str) (or EXPECT_TRUE(out_str.empty())), but whatever

+  }
+
+  // Numbers that look like IP addresses should not be changed.
+  {
+    std::string out_str;
+    StdStringCanonOutput output(&out_str);
+    EXPECT_TRUE(CanonicalizeHostSubstring(
+        "01.02.03.04", Component(0, 11), &output));
+    output.Complete();
+    EXPECT_EQ("01.02.03.04", out_str);
+  }
+}
+
 TEST(URLCanonTest, UserInfo) {
   // Note that the canonicalizer should escape and treat empty components as
   // not being there.
 
   // We actually parse a full input URL so we can get the initial components.
   struct UserComponentCase {
     const char* input;
     const char* expected;
     Component expected_username;
     Component expected_password;
@@ skipping 1294 matching lines @@
   };
 
   for (auto& test_case : cases) {
     SCOPED_TRACE(test_case.scheme);
     EXPECT_EQ(test_case.expected_port,
               DefaultPortForScheme(test_case.scheme, strlen(test_case.scheme)));
   }
 }
 
 }  // namespace url