[SH iOS] Add component code for disabling per-site

This adds a helper function that platforms can use to determine if the
current site's URL should prevent the Link to Text feature from being
offered. This will be called in a follow-up by platform-specific code
that enables/disables the option. For the time being, we expect the
number of sites affected to be very small and will reevaluate if the
list gets longer than ~a dozen.

Eventually we'd like to make this more dynamic (possibly delivering the
URLs and regexes from a server), but this version is meant to get a very
quick spot fix out the door.

Bug: 1157981
Change-Id: I1785459b298d93528415d8735aaa01c4d7a7769e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2587432Reviewed-by: Sebastien Lalancette <seblalancette@chromium.org>
Reviewed-by: Nico Weber <thakis@chromium.org>
Commit-Queue: Tommy Martino <tmartino@chromium.org>
Cr-Commit-Position: refs/heads/master@{#842655}

components/shared_highlighting/core/common/BUILD.gn

@@ -4,6 +4,8 @@
 
 static_library("common") {
   sources = [
+    "disabled_sites.cc",
+    "disabled_sites.h",
     "shared_highlighting_metrics.cc",
     "shared_highlighting_metrics.h",
     "text_fragment.cc",
@@ -19,12 +21,14 @@ static_library("common") {
     "//components/search_engines:search_engine_utils",
     "//services/metrics/public/cpp:metrics_cpp",
     "//services/metrics/public/cpp:ukm_builders",
+    "//third_party/re2",
   ]
 }
 
 source_set("unit_tests") {
   testonly = true
   sources = [
+    "disabled_sites_unittest.cc",
     "shared_highlighting_metrics_unittest.cc",
     "text_fragment_unittest.cc",
     "text_fragments_utils_unittest.cc",

components/shared_highlighting/core/common/DEPS

@@ -4,5 +4,6 @@ include_rules = [
   "+components/ukm/test_ukm_recorder.h",
   "+net/base",
   "+services/metrics/public",
+  "+third_party/re2",
   "+url",
 ]
\ No newline at end of file

components/shared_highlighting/core/common/disabled_sites.cc

+// Copyright 2020 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/shared_highlighting/core/common/disabled_sites.h"
+
+#include "third_party/re2/src/re2/re2.h"
+
+#include <map>
+#include <utility>
+
+namespace shared_highlighting {
+
+bool ShouldOfferLinkToText(const GURL& url) {
+  // If a URL's host matches a key in this map, then the path will be tested
+  // against the RE stored in the value. For example, {"foo.com", ".*"} means
+  // any page on the foo.com domain.
+  const static std::map<std::string, std::string> kBlocklist = {
+      {"youtube.com", ".*"},
+      // TODO(crbug.com/1157981): special case this to cover other Google TLDs
+      {"google.com", "^\\/amp\\/.*"}};
+
+  const std::string domain =
+      url.host().compare(0, 4, "www.") == 0 ? url.host().substr(4) : url.host();
+  auto it = kBlocklist.find(domain);
+  if (it != kBlocklist.end()) {
+    return !re2::RE2::FullMatch(url.path(), it->second);
+  }
+  return true;
+}
+
+}  // namespace shared_highlighting

components/shared_highlighting/core/common/disabled_sites.h

+// Copyright 2020 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.
+
+#ifndef COMPONENTS_SHARED_HIGHLIGHTING_CORE_COMMON_DISABLED_SITES_H_
+#define COMPONENTS_SHARED_HIGHLIGHTING_CORE_COMMON_DISABLED_SITES_H_
+
+#include "url/gurl.h"
+
+namespace shared_highlighting {
+
+// Returns true iff Link to Text menu options should be enabled on this page.
+// Uses a blocklist to identify certain sites where personalized or dynamic
+// content make it unlikely that a generated URL will actually work when
+// shared.
+bool ShouldOfferLinkToText(const GURL& url);
+
+}  // namespace shared_highlighting
+
+#endif  // COMPONENTS_SHARED_HIGHLIGHTING_CORE_COMMON_DISABLED_SITES_H_

components/shared_highlighting/core/common/disabled_sites_unittest.cc

+// Copyright 2020 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/shared_highlighting/core/common/disabled_sites.h"
+
+#include "testing/gtest/include/gtest/gtest.h"
+#include "url/gurl.h"
+
+namespace shared_highlighting {
+namespace {
+
+TEST(DisabledSitesTest, AllPaths) {
+  EXPECT_FALSE(ShouldOfferLinkToText(GURL("https://www.youtube.com")));
+  EXPECT_FALSE(ShouldOfferLinkToText(GURL("https://www.youtube.com/somepage")));
+  EXPECT_FALSE(ShouldOfferLinkToText(GURL("https://youtube.com")));
+  EXPECT_FALSE(ShouldOfferLinkToText(GURL("https://youtube.com/somepage")));
+}
+
+TEST(DisabledSitesTest, SpecificPages) {
+  // Paths starting with /amp/ are disabled.
+  EXPECT_FALSE(ShouldOfferLinkToText(GURL("https://www.google.com/amp/")));
+  EXPECT_FALSE(ShouldOfferLinkToText(GURL("https://www.google.com/amp/foo")));
+  EXPECT_FALSE(ShouldOfferLinkToText(GURL("https://google.com/amp/")));
+  EXPECT_FALSE(ShouldOfferLinkToText(GURL("https://google.com/amp/foo")));
+
+  // Other paths are not.
+  EXPECT_TRUE(ShouldOfferLinkToText(GURL("https://www.google.com")));
+  EXPECT_TRUE(ShouldOfferLinkToText(GURL("https://www.google.com/somepage")));
+  EXPECT_TRUE(ShouldOfferLinkToText(GURL("https://google.com")));
+  EXPECT_TRUE(ShouldOfferLinkToText(GURL("https://google.com/somepage")));
+
+  // Paths with /amp/ later on are also not affected.
+  EXPECT_TRUE(ShouldOfferLinkToText(GURL("https://google.com/foo/amp/")));
+  EXPECT_TRUE(ShouldOfferLinkToText(GURL("https://google.com/foo/amp/bar")));
+}
+
+TEST(DisabledSitesTest, NonMatchingHost) {
+  EXPECT_TRUE(ShouldOfferLinkToText(GURL("https://www.example.com")));
+}
+
+}  // namespace
+}  // namespace shared_highlighting