Add missing licenses, and add a check (#3720)

Adds a new CI check that all code files have a copyright+license block (and that it's one we are expecting to see).

Fixes the ~350 files (!) that did not have them. This includes all of the files in the .../example/ directories, following the example of flutter/flutter. (This does mean some manual intervention will be needed when generating new example directories in the future, but it's one-time per example.)

Also standardized some variants that used different line breaks than most of the rest of the repo (likely added since I standardized them all a while ago, but didn't add a check for at the time to enforce going forward), to simplify the checks.

Fixes flutter/flutter#77114

script/tool/lib/src/license_check_command.dart

@@ -0,0 +1,209 @@
+// Copyright 2017 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.
+
+import 'dart:async';
+
+import 'package:file/file.dart';
+import 'package:path/path.dart' as p;
+
+import 'common.dart';
+
+const Set<String> _codeFileExtensions = <String>{
+  '.c',
+  '.cc',
+  '.cpp',
+  '.dart',
+  '.h',
+  '.html',
+  '.java',
+  '.m',
+  '.mm',
+  '.swift',
+  '.sh',
+};
+
+// Basenames without extensions of files to ignore.
+const Set<String> _ignoreBasenameList = <String>{
+  'flutter_export_environment',
+  'GeneratedPluginRegistrant',
+  'generated_plugin_registrant',
+};
+
+// File suffixes that otherwise match _codeFileExtensions to ignore.
+const Set<String> _ignoreSuffixList = <String>{
+  '.g.dart', // Generated API code.
+  '.mocks.dart', // Generated by Mockito.
+};
+
+// Full basenames of files to ignore.
+const Set<String> _ignoredFullBasenameList = <String>{
+  'resource.h', // Generated by VS.
+};
+
+// Copyright and license regexes.
+//
+// These are intentionally very simple, since almost all source in this
+// repository should be using the same license text, comment style, etc., so
+// they shouldn't need to be very flexible. Complexity can be added as-needed
+// on a case-by-case basis.
+final RegExp _copyrightRegex =
+    RegExp(r'^(?://|#|<!--) Copyright \d+,? ([^.]+)', multiLine: true);
+// Non-Flutter code. When adding license regexes here, include the copyright
+// info to ensure that any new additions are flagged for added scrutiny in
+// review.
+// -----
+// Third-party code used in url_launcher_web.
+final RegExp _workivaLicenseRegex = RegExp(
+    r'^// Copyright 2017 Workiva Inc..*'
+    '^// Licensed under the Apache License, Version 2.0',
+    multiLine: true,
+    dotAll: true);
+
+// TODO(stuartmorgan): Replace this with a single string once all the copyrights
+// are standardized.
+final List<String> _firstPartyAuthors = <String>[
+  'The Chromium Authors',
+  'the Chromium project authors',
+  'The Flutter Authors',
+  'the Flutter project authors',
+];
+
+/// Validates that code files have copyright and license blocks.
+class LicenseCheckCommand extends PluginCommand {
+  /// Creates a new license check command for [packagesDir].
+  LicenseCheckCommand(
+    Directory packagesDir,
+    FileSystem fileSystem, {
+    Print print = print,
+  })  : _print = print,
+        super(packagesDir, fileSystem);
+
+  final Print _print;
+
+  @override
+  final String name = 'license-check';
+
+  @override
+  final String description =
+      'Ensures that all code files have copyright/license blocks.';
+
+  @override
+  Future<Null> run() async {
+    Iterable<File> codeFiles = (await _getAllFiles()).where((File file) =>
+        _codeFileExtensions.contains(p.extension(file.path)) &&
+        !_shouldIgnoreFile(file));
+
+    bool succeeded = await _checkLicenses(codeFiles);
+
+    if (!succeeded) {
+      throw ToolExit(1);
+    }
+  }
+
+  // Creates the expected license block (without copyright) for first-party
+  // code.
+  String _generateLicense(String comment, {String suffix = ''}) {
+    return '${comment}Use of this source code is governed by a BSD-style license that can be\n'
+        '${comment}found in the LICENSE file.$suffix\n';
+  }
+
+  // Checks all license blocks for [codeFiles], returning false if any of them
+  // fail validation.
+  Future<bool> _checkLicenses(Iterable<File> codeFiles) async {
+    final List<File> filesWithoutDetectedCopyright = <File>[];
+    final List<File> filesWithoutDetectedLicense = <File>[];
+    final List<File> misplacedThirdPartyFiles = <File>[];
+
+    // Most code file types in the repository use '//' comments.
+    final String defaultBsdLicenseBlock = _generateLicense('// ');
+    // A few file types have a different comment structure.
+    final Map<String, String> bsdLicenseBlockByExtension = <String, String>{
+      '.sh': _generateLicense('# '),
+      '.html': _generateLicense('', suffix: ' -->'),
+    };
+
+    for (final File file in codeFiles) {
+      _print('Checking ${file.path}');
+      final String content = await file.readAsString();
+
+      final RegExpMatch copyright = _copyrightRegex.firstMatch(content);
+      if (copyright == null) {
+        filesWithoutDetectedCopyright.add(file);
+        continue;
+      }
+      final String author = copyright.group(1);
+      if (!_firstPartyAuthors.contains(author) &&
+          !p.split(file.path).contains('third_party')) {
+        misplacedThirdPartyFiles.add(file);
+      }
+
+      final String bsdLicense =
+          bsdLicenseBlockByExtension[p.extension(file.path)] ??
+              defaultBsdLicenseBlock;
+      if (!content.contains(bsdLicense) &&
+          !_workivaLicenseRegex.hasMatch(content)) {
+        filesWithoutDetectedLicense.add(file);
+      }
+    }
+    _print('\n\n');
+
+    // Sort by path for more usable output.
+    final pathCompare = (File a, File b) => a.path.compareTo(b.path);
+    filesWithoutDetectedCopyright.sort(pathCompare);
+    filesWithoutDetectedLicense.sort(pathCompare);
+    misplacedThirdPartyFiles.sort(pathCompare);
+
+    if (filesWithoutDetectedCopyright.isNotEmpty) {
+      _print('No copyright line was found for the following files:');
+      for (final File file in filesWithoutDetectedCopyright) {
+        _print('  ${file.path}');
+      }
+      _print('Please check that they have a copyright and license block. '
+          'If they do, the license check may need to be updated to recognize its '
+          'format.\n');
+    }
+
+    if (filesWithoutDetectedLicense.isNotEmpty) {
+      _print('No recognized license was found for the following files:');
+      for (final File file in filesWithoutDetectedLicense) {
+        _print('  ${file.path}');
+      }
+      _print('Please check that they have a license at the top of the file. '
+          'If they do, the license check may need to be updated to recognize '
+          'either the license or the specific format of the license '
+          'block.\n');
+    }
+
+    if (misplacedThirdPartyFiles.isNotEmpty) {
+      _print('The following files do not have a recognized first-party author '
+          'but are not in a "third_party/" directory:');
+      for (final File file in misplacedThirdPartyFiles) {
+        _print('  ${file.path}');
+      }
+      _print('Please move these files to "third_party/".\n');
+    }
+
+    bool succeeded = filesWithoutDetectedCopyright.isEmpty &&
+        filesWithoutDetectedLicense.isEmpty &&
+        misplacedThirdPartyFiles.isEmpty;
+    if (succeeded) {
+      _print('All files passed validation!');
+    }
+    return succeeded;
+  }
+
+  bool _shouldIgnoreFile(File file) {
+    final String path = file.path;
+    return _ignoreBasenameList.contains(p.basenameWithoutExtension(path)) ||
+        _ignoreSuffixList.any((String suffix) =>
+            path.endsWith(suffix) ||
+            _ignoredFullBasenameList.contains(p.basename(path)));
+  }
+
+  Future<List<File>> _getAllFiles() => packagesDir.parent
+      .list(recursive: true, followLinks: false)
+      .where((FileSystemEntity entity) => entity is File)
+      .map((FileSystemEntity file) => file as File)
+      .toList();
+}