Add FontIndexer tool to enumerate locally installed fonts

This will aid in constructing a catalog of known font digests for the
identifiability study.

Change-Id: Ic0d4b603513aed6c5f1fec7d3606a371119e2707
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2399446Reviewed-by: John Abd-El-Malek <jam@chromium.org>
Reviewed-by: Dirk Pranke <dpranke@google.com>
Reviewed-by: Asanka Herath <asanka@chromium.org>
Commit-Queue: Alex Turner <alexmt@chromium.org>
Cr-Commit-Position: refs/heads/master@{#824015}

BUILD.gn

@@ -154,6 +154,7 @@ group("gn_all") {
       "//third_party/pdfium/samples:pdfium_test",
       "//tools/perf/clear_system_cache",
       "//tools/polymer:polymer_tools_python_unittests",
+      "//tools/privacy_budget:privacy_budget_tools",
       "//ui/accessibility:accessibility_perftests",
       "//ui/accessibility:accessibility_unittests",
       "//ui/accessibility/extensions:extension_tests",

third_party/blink/renderer/platform/BUILD.gn

@@ -253,6 +253,7 @@ component("platform") {
     "//services/device/public/mojom:mojom_blink",
     "//services/media_session/public/mojom:mojom_blink",
     "//third_party/blink/*",
+    "//tools/privacy_budget/font_indexer:*",
     "//url/mojom:url_mojom_gurl_blink",
     "//url/mojom:url_mojom_origin_blink",
   ]
@@ -1773,7 +1774,10 @@ component("platform") {
 }
 
 static_library("test_support") {
-  visibility += [ "//third_party/blink/*" ]
+  visibility += [
+    "//third_party/blink/*",
+    "//tools/privacy_budget/font_indexer:*",
+  ]
   testonly = true
 
   sources = [

tools/privacy_budget/BUILD.gn

+# 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.
+
+group("privacy_budget_tools") {
+  testonly = true
+  deps = [ "font_indexer:font_indexer" ]
+}

tools/privacy_budget/DEPS

+include_rules = [
+  "+base",
+  "+content/public",
+  "+mojo/core/embedder",
+  "+third_party/blink",
+]

tools/privacy_budget/OWNERS

+file://third_party/blink/public/common/privacy_budget/OWNERS
+
+# TEAM: privacy-sandbox-dev@chromium.org
+# COMPONENT: Privacy>Fingerprinting

tools/privacy_budget/font_indexer/BUILD.gn

+# 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.
+
+static_library("tools_lib") {
+  sources = [
+    "font_indexer.cc",
+    "font_indexer.h",
+  ]
+  deps = [
+    "//base:base",
+    "//content/public/browser:browser",
+    "//third_party/blink/renderer/platform:platform",
+  ]
+}
+
+executable("font_indexer") {
+  testonly = true
+  sources = [ "font_indexer_main.cc" ]
+  deps = [
+    ":tools_lib",
+    "//base:base",
+    "//base/test:test_support",
+    "//content/public/browser:browser",
+    "//mojo/core/embedder",
+    "//third_party/blink/renderer/platform:test_support",
+  ]
+}

tools/privacy_budget/font_indexer/font_indexer.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 TOOLS_PRIVACY_BUDGET_FONT_INDEXER_FONT_INDEXER_H_
+#define TOOLS_PRIVACY_BUDGET_FONT_INDEXER_FONT_INDEXER_H_
+
+#include <string>
+#include <utility>
+
+#include "base/callback_forward.h"
+#include "base/values.h"
+#include "third_party/blink/renderer/platform/fonts/font_cache.h"
+#include "third_party/blink/renderer/platform/fonts/font_description.h"
+#include "third_party/blink/renderer/platform/fonts/font_selection_types.h"
+#include "third_party/blink/renderer/platform/wtf/text/wtf_string.h"
+
+namespace privacy_budget {
+
+// Note that in the following constants, we prioritize the more common values as
+// we later associate each unique digest with the first set of settings for
+// which that digest is found.
+extern const std::pair<blink::FontSelectionValue, std::string> kFontWeights[];
+extern const std::pair<blink::FontSelectionValue, std::string> kFontWidths[];
+
+// Not as thorough as above, given its rarity and to reduce the speed impact.
+extern const std::pair<blink::FontSelectionValue, std::string> kFontSlopes[];
+
+// Following only used if |more_slope_checks_|
+extern const std::pair<blink::FontSelectionValue, std::string>
+    kAdditionalFontSlopes[];
+
+extern const char kOutputHeader[];
+extern const char kOutputSeparator[];
+
+// FontIndexer allows for enumerating all locally installed fonts and computing
+// identifiability digests of each typeface.
+class FontIndexer {
+ public:
+  FontIndexer();
+  ~FontIndexer();
+
+  // The main function that enumerates all fonts and prints a tab-separated file
+  // containing the fonts' details to stdout.
+  void PrintAllFonts();
+
+  // By default, this tool attempts to determine whether the fonts vary along
+  // each axis (i.e. width, weight and slope), skipping checks along the axes
+  // with no variation. This call disables this optimization, slowing the tool
+  // substantially, but possibly being more thorough (if the determinations are
+  // incorrect).
+  void SetNoSmartSkipping() { smart_skipping_ = false; }
+
+  // By default, the tool only checks a limited number of slope values as
+  // substantial slope variation is rare and slow to check for. This call adds
+  // more granularity when slopes are varied. This will slow down the tool, but
+  // will give more results if a font with many slope variations is available.
+  void SetMoreSlopeChecks() { more_slope_checks_ = true; }
+
+ private:
+  void FontListHasLoaded(std::unique_ptr<base::ListValue> list);
+  void WaitForFontListToLoad();
+
+  // Determines whether the fonts with |name| appear to vary along the specified
+  // axis, by comparing font digests at extreme values to |default_font_digest|.
+  bool DoFontsWithNameHaveVaryingWeights(WTF::AtomicString name,
+                                         int64_t default_font_digest);
+  bool DoFontsWithNameHaveVaryingWidths(WTF::AtomicString name,
+                                        int64_t default_font_digest);
+  bool DoFontsWithNameHaveVaryingSlopes(WTF::AtomicString name,
+                                        int64_t default_font_digest);
+
+  // Determines whether a font lookup for |name| with |font_description| results
+  // in a typeface with |digest|.
+  bool DoesFontHaveDigest(WTF::AtomicString name,
+                          blink::FontDescription font_description,
+                          int64_t digest);
+
+  // Enumerates fonts with |name| and prints tab-separated lines with each
+  // font's details.
+  void PrintAllFontsWithName(WTF::AtomicString name);
+
+  blink::FontCache* font_cache_;
+  bool has_font_list_loaded_ = false;
+  bool smart_skipping_ = true;
+  bool more_slope_checks_ = false;
+  base::OnceClosure quit_closure_;
+};
+
+}  // namespace privacy_budget
+
+#endif  // TOOLS_PRIVACY_BUDGET_FONT_INDEXER_FONT_INDEXER_H_

tools/privacy_budget/font_indexer/font_indexer_main.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 <iostream>
+#include <string>
+#include <utility>
+
+#include "base/command_line.h"
+#include "base/logging.h"
+#include "base/task/thread_pool/thread_pool_instance.h"
+#include "base/task_runner_util.h"
+#include "base/test/task_environment.h"
+#include "base/test/test_io_thread.h"
+#include "base/test/test_suite.h"
+#include "base/test/test_timeouts.h"
+#include "mojo/core/embedder/embedder.h"
+#include "mojo/core/embedder/scoped_ipc_support.h"
+#include "third_party/blink/public/platform/platform.h"
+#include "third_party/blink/renderer/platform/testing/testing_platform_support.h"
+#include "tools/privacy_budget/font_indexer/font_indexer.h"
+
+namespace {
+
+const char kHelpMsg[] = R"(
+  font_indexer [--no-smart-skipping] [--more-slope-checks]
+
+  --no-smart-skipping stops the tool from skipping checks along axes of
+  variation when it appears the font does not varying along those axes. This
+  will slow down the tool substantially, but may be more thorough if the checks
+  are incorrect.
+
+  --more-slope-checks gives more granular checking of different slopes. This
+  will slow down the tool, but will give more results if a font with many
+  slope variations is available.
+)";
+
+const char kNoSmartSkippingSwitch[] = "no-smart-skipping";
+const char kMoreSlopeChecksSwitch[] = "more-slope-checks";
+
+void PrintHelp() {
+  printf("%s\n\n", kHelpMsg);
+}
+
+bool ShouldPrintHelpAndQuit(const base::CommandLine::StringVector& args,
+                            const base::CommandLine::SwitchMap& switches) {
+  if (args.size() != 0U || switches.size() > 2) {
+    return true;
+  }
+  for (const auto& switch_entry : switches) {
+    std::string switch_name = switch_entry.first;
+    if (switch_name != kNoSmartSkippingSwitch &&
+        switch_name != kMoreSlopeChecksSwitch) {
+      return true;
+    }
+  }
+  return false;
+}
+
+}  // namespace
+
+int main(int argc, char* argv[]) {
+  base::CommandLine::Init(argc, argv);
+  base::CommandLine& command_line = *base::CommandLine::ForCurrentProcess();
+  base::CommandLine::StringVector args = command_line.GetArgs();
+
+  if (ShouldPrintHelpAndQuit(command_line.GetArgs(),
+                             command_line.GetSwitches())) {
+    PrintHelp();
+    return 1;
+  }
+
+  // Initialize a test environment to satisfy the expectations of
+  // content::GetFontListAsync().
+  blink::ScopedUnittestsEnvironmentSetup testEnvironmentSetup(argc, argv);
+  base::TestSuite testSuite(argc, argv);
+  mojo::core::Init();
+  base::TestIOThread testIoThread(base::TestIOThread::kAutoStart);
+  mojo::core::ScopedIPCSupport ipcSupport(
+      testIoThread.task_runner(),
+      mojo::core::ScopedIPCSupport::ShutdownPolicy::CLEAN);
+  TestTimeouts::Initialize();
+  base::test::TaskEnvironment env(
+      base::test::TaskEnvironment::TimeSource::MOCK_TIME);
+
+  // Set up and run tool.
+  privacy_budget::FontIndexer indexer;
+  if (command_line.HasSwitch(kNoSmartSkippingSwitch)) {
+    indexer.SetNoSmartSkipping();
+  }
+  if (command_line.HasSwitch(kMoreSlopeChecksSwitch)) {
+    indexer.SetMoreSlopeChecks();
+  }
+  indexer.PrintAllFonts();
+
+  return 0;
+}