URL Formatter: Add StripSubdomain method that preserves eTLD + 1.

components/url_formatter/url_formatter.cc

 // Copyright 2015 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 "components/url_formatter/url_formatter.h"
 
 #include <algorithm>
 #include <utility>
 #include <vector>
 
 #include "base/lazy_instance.h"
 #include "base/macros.h"
 #include "base/numerics/safe_conversions.h"
 #include "base/strings/string_piece.h"
 #include "base/strings/string_util.h"
 #include "base/strings/utf_offset_string_conversions.h"
 #include "base/strings/utf_string_conversions.h"
 #include "base/threading/thread_local_storage.h"
 #include "components/url_formatter/idn_spoof_checker.h"
+#include "net/base/registry_controlled_domains/registry_controlled_domain.h"
 #include "third_party/icu/source/common/unicode/uidna.h"
 #include "third_party/icu/source/common/unicode/utypes.h"
 #include "url/gurl.h"
 #include "url/third_party/mozilla/url_parse.h"
 
 namespace url_formatter {
 
 namespace {
 
 base::string16 IDNToUnicodeWithAdjustments(
@@ skipping 541 matching lines @@
 base::string16 IDNToUnicode(base::StringPiece host) {
   return IDNToUnicodeWithAdjustments(host, nullptr);
 }
 
 base::string16 StripWWW(const base::string16& text) {
   const base::string16 www(base::ASCIIToUTF16("www."));
   return base::StartsWith(text, www, base::CompareCase::SENSITIVE)
       ? text.substr(www.length()) : text;
 }
 
-base::string16 StripWWWFromHost(const GURL& url) {
+base::StringPiece StripSubdomains(const GURL& url,
+                                  StripSubdomainTypes subdomains) {
   DCHECK(url.is_valid());
-  return StripWWW(base::ASCIIToUTF16(url.host_piece()));
+
+  base::StringPiece host_piece = url.host_piece();
+  std::string etld_plus_one =

[Peter Kasting, 2017/06/17 02:46:55]

Nit: Prefer to use the same terminology as the RCDS, "domain and registry".

[tommycli, 2017/06/19 23:03:07]

Done.

+      net::registry_controlled_domains::GetDomainAndRegistry(
+          host_piece,
+          net::registry_controlled_domains::INCLUDE_PRIVATE_REGISTRIES);
+
+  if (host_piece == etld_plus_one)
+    return host_piece;
+
+  if ((subdomains & kStripWWW) != 0 &&
+      base::StartsWith(host_piece, "www.", base::CompareCase::SENSITIVE)) {
+    return host_piece.substr(4);

[Peter Kasting, 2017/06/17 02:46:55]

This early-returns if the host is www.m.foo.com, which seems wrong.

[tommycli, 2017/06/19 23:03:08]

Done.

+  }
+
+  if ((subdomains & kStripM) != 0 &&
+      base::StartsWith(host_piece, "m.", base::CompareCase::SENSITIVE)) {

[Peter Kasting, 2017/06/17 02:46:55]

This fails to fix up the most common mobile host I encounter,
en.m.wikipedia.org.

It seems like the solution to these last two comments is to remove these
subdomains anywhere they occur prior to the eTLD+1.

[tommycli, 2017/06/19 23:03:07]

Done.

+    return host_piece.substr(2);
+  }
+
+  return host_piece;
 }
 
 }  // namespace url_formatter