[go: nahoru, domu]
Google Git
Sign in
chromium / chromium / src / 402f6fa36147aee3b3a16e89d6cf0aa2eda30a5b / . / PRESUBMIT.py
blob: 5daf10e5c8544600cc5a3ba39840770231013909 [file] [log] [blame]
# Copyright 2012 The Chromium Authors
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""Top-level presubmit script for Chromium.
See https://www.chromium.org/developers/how-tos/depottools/presubmit-scripts/
for more details about the presubmit API built into depot_tools.
"""
from typing import Callable
from typing import Optional
from typing import Sequence
from dataclasses import dataclass
PRESUBMIT_VERSION = '2.0.0'
# This line is 'magic' in that git-cl looks for it to decide whether to
# use Python3 instead of Python2 when running the code in this file.
USE_PYTHON3 = True
_EXCLUDED_PATHS = (
# Generated file
(r"chrome/android/webapk/shell_apk/src/org/chromium"
r"/webapk/lib/runtime_library/IWebApkApi.java"),
# File needs to write to stdout to emulate a tool it's replacing.
r"chrome/updater/mac/keystone/ksadmin.mm",
# Generated file.
(r"^components/variations/proto/devtools/"
r"client_variations.js"),
# These are video files, not typescript.
r"^media/test/data/.*.ts",
r"^native_client_sdksrc/build_tools/make_rules.py",
r"^native_client_sdk/src/build_tools/make_simple.py",
r"^native_client_sdk/src/tools/.*.mk",
r"^net/tools/spdyshark/.*",
r"^skia/.*",
r"^third_party/blink/.*",
r"^third_party/breakpad/.*",
# sqlite is an imported third party dependency.
r"^third_party/sqlite/.*",
r"^v8/.*",
r".*MakeFile$",
r".+_autogen\.h$",
r".+_pb2(_grpc)?\.py$",
r".+/pnacl_shim\.c$",
r"^gpu/config/.*_list_json\.cc$",
r"tools/md_browser/.*\.css$",
# Test pages for Maps telemetry tests.
r"tools/perf/page_sets/maps_perf_test.*",
# Test pages for WebRTC telemetry tests.
r"tools/perf/page_sets/webrtc_cases.*",
)
_EXCLUDED_SET_NO_PARENT_PATHS = (
# It's for historical reasons that blink isn't a top level directory, where
# it would be allowed to have "set noparent" to avoid top level owners
# accidentally +1ing changes.
'third_party/blink/OWNERS',
)
# Fragment of a regular expression that matches C++ and Objective-C++
# implementation files.
_IMPLEMENTATION_EXTENSIONS = r'\.(cc|cpp|cxx|mm)$'
# Fragment of a regular expression that matches C++ and Objective-C++
# header files.
_HEADER_EXTENSIONS = r'\.(h|hpp|hxx)$'
# Paths with sources that don't use //base.
_NON_BASE_DEPENDENT_PATHS = (
r"^chrome/browser/browser_switcher/bho/",
r"^tools/win/",
)
# Regular expression that matches code only used for test binaries
# (best effort).
_TEST_CODE_EXCLUDED_PATHS = (
r'.*/(fake_|test_|mock_).+%s' % _IMPLEMENTATION_EXTENSIONS,
r'.+_test_(base|support|util)%s' % _IMPLEMENTATION_EXTENSIONS,
# Test suite files, like:
# foo_browsertest.cc
# bar_unittest_mac.cc (suffix)
# baz_unittests.cc (plural)
r'.+_(api|browser|eg|int|perf|pixel|unit|ui)?test(s)?(_[a-z]+)?%s' %
_IMPLEMENTATION_EXTENSIONS,
r'.+_(fuzz|fuzzer)(_[a-z]+)?%s' % _IMPLEMENTATION_EXTENSIONS,
r'.+sync_service_impl_harness%s' % _IMPLEMENTATION_EXTENSIONS,
r'.*/(test|tool(s)?)/.*',
# content_shell is used for running content_browsertests.
r'content/shell/.*',
# Web test harness.
r'content/web_test/.*',
# Non-production example code.
r'mojo/examples/.*',
# Launcher for running iOS tests on the simulator.
r'testing/iossim/iossim\.mm$',
# EarlGrey app side code for tests.
r'ios/.*_app_interface\.mm$',
# Views Examples code
r'ui/views/examples/.*',
# Chromium Codelab
r'codelabs/*'
)
_THIRD_PARTY_EXCEPT_BLINK = 'third_party/(?!blink/)'
_TEST_ONLY_WARNING = (
'You might be calling functions intended only for testing from\n'
'production code. If you are doing this from inside another method\n'
'named as *ForTesting(), then consider exposing things to have tests\n'
'make that same call directly.\n'
'If that is not possible, you may put a comment on the same line with\n'
' // IN-TEST \n'
'to tell the PRESUBMIT script that the code is inside a *ForTesting()\n'
'method and can be ignored. Do not do this inside production code.\n'
'The android-binary-size trybot will block if the method exists in the\n'
'release apk.')
@dataclass
class BanRule:
# String pattern. If the pattern begins with a slash, the pattern will be
# treated as a regular expression instead.
pattern: str
# Explanation as a sequence of strings. Each string in the sequence will be
# printed on its own line.
explanation: Sequence[str]
# Whether or not to treat this ban as a fatal error. If unspecified,
# defaults to true.
treat_as_error: Optional[bool] = None
# Paths that should be excluded from the ban check. Each string is a regular
# expression that will be matched against the path of the file being checked
# relative to the root of the source tree.
excluded_paths: Optional[Sequence[str]] = None
_BANNED_JAVA_IMPORTS : Sequence[BanRule] = (
BanRule(
'import java.net.URI;',
(
'Use org.chromium.url.GURL instead of java.net.URI, where possible.',
),
excluded_paths=(
(r'net/android/javatests/src/org/chromium/net/'
'AndroidProxySelectorTest\.java'),
r'components/cronet/',
r'third_party/robolectric/local/',
),
),
BanRule(
'import android.annotation.TargetApi;',
(
'Do not use TargetApi, use @androidx.annotation.RequiresApi instead. '
'RequiresApi ensures that any calls are guarded by the appropriate '
'SDK_INT check. See https://crbug.com/1116486.',
),
),
BanRule(
'import android.support.test.rule.UiThreadTestRule;',
(
'Do not use UiThreadTestRule, just use '
'@org.chromium.base.test.UiThreadTest on test methods that should run '
'on the UI thread. See https://crbug.com/1111893.',
),
),
BanRule(
'import android.support.test.annotation.UiThreadTest;',
('Do not use android.support.test.annotation.UiThreadTest, use '
'org.chromium.base.test.UiThreadTest instead. See '
'https://crbug.com/1111893.',
),
),
BanRule(
'import android.support.test.rule.ActivityTestRule;',
(
'Do not use ActivityTestRule, use '
'org.chromium.base.test.BaseActivityTestRule instead.',
),
excluded_paths=(
'components/cronet/',
),
),
)
_BANNED_JAVA_FUNCTIONS : Sequence[BanRule] = (
BanRule(
'StrictMode.allowThreadDiskReads()',
(
'Prefer using StrictModeContext.allowDiskReads() to using StrictMode '
'directly.',
),
False,
),
BanRule(
'StrictMode.allowThreadDiskWrites()',
(
'Prefer using StrictModeContext.allowDiskWrites() to using StrictMode '
'directly.',
),
False,
),
BanRule(
'.waitForIdleSync()',
(
'Do not use waitForIdleSync as it masks underlying issues. There is '
'almost always something else you should wait on instead.',
),
False,
),
BanRule(
r'/(?<!\bsuper\.)(?<!\bIntent )\bregisterReceiver\(',
(
'Do not call android.content.Context.registerReceiver (or an override) '
'directly. Use one of the wrapper methods defined in '
'org.chromium.base.ContextUtils, such as '
'registerProtectedBroadcastReceiver, '
'registerExportedBroadcastReceiver, or '
'registerNonExportedBroadcastReceiver. See their documentation for '
'which one to use.',
),
True,
excluded_paths=(
r'.*Test[^a-z]',
r'third_party/',
'base/android/java/src/org/chromium/base/ContextUtils.java',
'chromecast/browser/android/apk/src/org/chromium/chromecast/shell/BroadcastReceiverScope.java',
),
),
BanRule(
r'/(?:extends|new)\s*(?:android.util.)?Property<[A-Za-z.]+,\s*(?:Integer|Float)>',
(
'Do not use Property<..., Integer|Float>, but use FloatProperty or '
'IntProperty because it will avoid unnecessary autoboxing of '
'primitives.',
),
),
BanRule(
'requestLayout()',
(
'Layouts can be expensive. Prefer using ViewUtils.requestLayout(), '
'which emits a trace event with additional information to help with '
'scroll jank investigations. See http://crbug.com/1354176.',
),
False,
excluded_paths=(
'ui/android/java/src/org/chromium/ui/base/ViewUtils.java',
),
),
)
_BANNED_JAVASCRIPT_FUNCTIONS : Sequence [BanRule] = (
BanRule(
r'/\bchrome\.send\b',
(
'The use of chrome.send is disallowed in Chrome (context: https://chromium.googlesource.com/chromium/src/+/refs/heads/main/docs/security/handling-messages-from-web-content.md).',
'Please use mojo instead for new webuis. https://docs.google.com/document/d/1RF-GSUoveYa37eoyZ9EhwMtaIwoW7Z88pIgNZ9YzQi4/edit#heading=h.gkk22wgk6wff',
),
True,
(
r'^(?!ash\/webui).+',
# TODO(crbug.com/1385601): pre-existing violations still need to be
# cleaned up.
'ash/webui/common/resources/cr.m.js',
'ash/webui/common/resources/multidevice_setup/multidevice_setup_browser_proxy.js',
'ash/webui/common/resources/quick_unlock/lock_screen_constants.js',
'ash/webui/common/resources/smb_shares/smb_browser_proxy.js',
'ash/webui/connectivity_diagnostics/resources/connectivity_diagnostics.js',
'ash/webui/diagnostics_ui/resources/diagnostics_browser_proxy.ts',
'ash/webui/multidevice_debug/resources/logs.js',
'ash/webui/multidevice_debug/resources/webui.js',
'ash/webui/projector_app/resources/annotator/trusted/annotator_browser_proxy.js',
'ash/webui/projector_app/resources/app/trusted/projector_browser_proxy.js',
'ash/webui/scanning/resources/scanning_browser_proxy.js',
),
),
)
_BANNED_OBJC_FUNCTIONS : Sequence[BanRule] = (
BanRule(
'addTrackingRect:',
(
'The use of -[NSView addTrackingRect:owner:userData:assumeInside:] is'
'prohibited. Please use CrTrackingArea instead.',
'http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts',
),
False,
),
BanRule(
r'/NSTrackingArea\W',
(
'The use of NSTrackingAreas is prohibited. Please use CrTrackingArea',
'instead.',
'http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts',
),
False,
),
BanRule(
'convertPointFromBase:',
(
'The use of -[NSView convertPointFromBase:] is almost certainly wrong.',
'Please use |convertPoint:(point) fromView:nil| instead.',
'http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts',
),
True,
),
BanRule(
'convertPointToBase:',
(
'The use of -[NSView convertPointToBase:] is almost certainly wrong.',
'Please use |convertPoint:(point) toView:nil| instead.',
'http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts',
),
True,
),
BanRule(
'convertRectFromBase:',
(
'The use of -[NSView convertRectFromBase:] is almost certainly wrong.',
'Please use |convertRect:(point) fromView:nil| instead.',
'http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts',
),
True,
),
BanRule(
'convertRectToBase:',
(
'The use of -[NSView convertRectToBase:] is almost certainly wrong.',
'Please use |convertRect:(point) toView:nil| instead.',
'http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts',
),
True,
),
BanRule(
'convertSizeFromBase:',
(
'The use of -[NSView convertSizeFromBase:] is almost certainly wrong.',
'Please use |convertSize:(point) fromView:nil| instead.',
'http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts',
),
True,
),
BanRule(
'convertSizeToBase:',
(
'The use of -[NSView convertSizeToBase:] is almost certainly wrong.',
'Please use |convertSize:(point) toView:nil| instead.',
'http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts',
),
True,
),
BanRule(
r"/\s+UTF8String\s*]",
(
'The use of -[NSString UTF8String] is dangerous as it can return null',
'even if |canBeConvertedToEncoding:NSUTF8StringEncoding| returns YES.',
'Please use |SysNSStringToUTF8| instead.',
),
True,
),
BanRule(
r'__unsafe_unretained',
(
'The use of __unsafe_unretained is almost certainly wrong, unless',
'when interacting with NSFastEnumeration or NSInvocation.',
'Please use __weak in files build with ARC, nothing otherwise.',
),
False,
),
BanRule(
'freeWhenDone:NO',
(
'The use of "freeWhenDone:NO" with the NoCopy creation of ',
'Foundation types is prohibited.',
),
True,
),
)
_BANNED_IOS_OBJC_FUNCTIONS = (
BanRule(
r'/\bTEST[(]',
(
'TEST() macro should not be used in Objective-C++ code as it does not ',
'drain the autorelease pool at the end of the test. Use TEST_F() ',
'macro instead with a fixture inheriting from PlatformTest (or a ',
'typedef).'
),
True,
),
BanRule(
r'/\btesting::Test\b',
(
'testing::Test should not be used in Objective-C++ code as it does ',
'not drain the autorelease pool at the end of the test. Use ',
'PlatformTest instead.'
),
True,
),
BanRule(
' systemImageNamed:',
(
'+[UIImage systemImageNamed:] should not be used to create symbols.',
'Instead use a wrapper defined in:',
'ios/chrome/browser/ui/icons/chrome_symbol.h'
),
True,
excluded_paths=(
'ios/chrome/browser/ui/icons/chrome_symbol.mm',
),
),
)
_BANNED_IOS_EGTEST_FUNCTIONS : Sequence[BanRule] = (
BanRule(
r'/\bEXPECT_OCMOCK_VERIFY\b',
(
'EXPECT_OCMOCK_VERIFY should not be used in EarlGrey tests because ',
'it is meant for GTests. Use [mock verify] instead.'
),
True,
),
)
_BANNED_CPP_FUNCTIONS : Sequence[BanRule] = (
BanRule(
r'/\busing namespace ',
(
'Using directives ("using namespace x") are banned by the Google Style',
'Guide ( http://google.github.io/styleguide/cppguide.html#Namespaces ).',
'Explicitly qualify symbols or use using declarations ("using x::foo").',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
# Make sure that gtest's FRIEND_TEST() macro is not used; the
# FRIEND_TEST_ALL_PREFIXES() macro from base/gtest_prod_util.h should be
# used instead since that allows for FLAKY_ and DISABLED_ prefixes.
BanRule(
'FRIEND_TEST(',
(
'Chromium code should not use gtest\'s FRIEND_TEST() macro. Include',
'base/gtest_prod_util.h and use FRIEND_TEST_ALL_PREFIXES() instead.',
),
False,
(),
),
BanRule(
'setMatrixClip',
(
'Overriding setMatrixClip() is prohibited; ',
'the base function is deprecated. ',
),
True,
(),
),
BanRule(
'SkRefPtr',
(
'The use of SkRefPtr is prohibited. ',
'Please use sk_sp<> instead.'
),
True,
(),
),
BanRule(
'SkAutoRef',
(
'The indirect use of SkRefPtr via SkAutoRef is prohibited. ',
'Please use sk_sp<> instead.'
),
True,
(),
),
BanRule(
'SkAutoTUnref',
(
'The use of SkAutoTUnref is dangerous because it implicitly ',
'converts to a raw pointer. Please use sk_sp<> instead.'
),
True,
(),
),
BanRule(
'SkAutoUnref',
(
'The indirect use of SkAutoTUnref through SkAutoUnref is dangerous ',
'because it implicitly converts to a raw pointer. ',
'Please use sk_sp<> instead.'
),
True,
(),
),
BanRule(
r'/HANDLE_EINTR\(.*close',
(
'HANDLE_EINTR(close) is invalid. If close fails with EINTR, the file',
'descriptor will be closed, and it is incorrect to retry the close.',
'Either call close directly and ignore its return value, or wrap close',
'in IGNORE_EINTR to use its return value. See http://crbug.com/269623'
),
True,
(),
),
BanRule(
r'/IGNORE_EINTR\((?!.*close)',
(
'IGNORE_EINTR is only valid when wrapping close. To wrap other system',
'calls, use HANDLE_EINTR. See http://crbug.com/269623',
),
True,
(
# Files that #define IGNORE_EINTR.
r'^base/posix/eintr_wrapper\.h$',
r'^ppapi/tests/test_broker\.cc$',
),
),
BanRule(
r'/v8::Extension\(',
(
'Do not introduce new v8::Extensions into the code base, use',
'gin::Wrappable instead. See http://crbug.com/334679',
),
True,
(
r'extensions/renderer/safe_builtins\.*',
),
),
BanRule(
'#pragma comment(lib,',
(
'Specify libraries to link with in build files and not in the source.',
),
True,
(
r'^base/third_party/symbolize/.*',
r'^third_party/abseil-cpp/.*',
),
),
BanRule(
r'/base::SequenceChecker\b',
(
'Consider using SEQUENCE_CHECKER macros instead of the class directly.',
),
False,
(),
),
BanRule(
r'/base::ThreadChecker\b',
(
'Consider using THREAD_CHECKER macros instead of the class directly.',
),
False,
(),
),
BanRule(
r'/\b(?!(Sequenced|SingleThread))\w*TaskRunner::(GetCurrentDefault|CurrentDefaultHandle)',
(
'It is not allowed to call these methods from the subclasses ',
'of Sequenced or SingleThread task runners.',
),
True,
(),
),
BanRule(
r'/(Time(|Delta|Ticks)|ThreadTicks)::FromInternalValue|ToInternalValue',
(
'base::TimeXXX::FromInternalValue() and ToInternalValue() are',
'deprecated (http://crbug.com/634507). Please avoid converting away',
'from the Time types in Chromium code, especially if any math is',
'being done on time values. For interfacing with platform/library',
'APIs, use FromMicroseconds() or InMicroseconds(), or one of the other',
'type converter methods instead. For faking TimeXXX values (for unit',
'testing only), use TimeXXX() + Microseconds(N). For',
'other use cases, please contact base/time/OWNERS.',
),
False,
(),
),
BanRule(
'CallJavascriptFunctionUnsafe',
(
"Don't use CallJavascriptFunctionUnsafe() in new code. Instead, use",
'AllowJavascript(), OnJavascriptAllowed()/OnJavascriptDisallowed(),',
'and CallJavascriptFunction(). See https://goo.gl/qivavq.',
),
False,
(
r'^content/browser/webui/web_ui_impl\.(cc|h)$',
r'^content/public/browser/web_ui\.h$',
r'^content/public/test/test_web_ui\.(cc|h)$',
),
),
BanRule(
'leveldb::DB::Open',
(
'Instead of leveldb::DB::Open() use leveldb_env::OpenDB() from',
'third_party/leveldatabase/env_chromium.h. It exposes databases to',
"Chrome's tracing, making their memory usage visible.",
),
True,
(
r'^third_party/leveldatabase/.*\.(cc|h)$',
),
),
BanRule(
'leveldb::NewMemEnv',
(
'Instead of leveldb::NewMemEnv() use leveldb_chrome::NewMemEnv() from',
'third_party/leveldatabase/leveldb_chrome.h. It exposes environments',
"to Chrome's tracing, making their memory usage visible.",
),
True,
(
r'^third_party/leveldatabase/.*\.(cc|h)$',
),
),
BanRule(
'RunLoop::QuitCurrent',
(
'Please migrate away from RunLoop::QuitCurrent*() methods. Use member',
'methods of a specific RunLoop instance instead.',
),
False,
(),
),
BanRule(
'base::ScopedMockTimeMessageLoopTaskRunner',
(
'ScopedMockTimeMessageLoopTaskRunner is deprecated. Prefer',
'TaskEnvironment::TimeSource::MOCK_TIME. There are still a',
'few cases that may require a ScopedMockTimeMessageLoopTaskRunner',
'(i.e. mocking the main MessageLoopForUI in browser_tests), but check',
'with gab@ first if you think you need it)',
),
False,
(),
),
BanRule(
'std::regex',
(
'Using std::regex adds unnecessary binary size to Chrome. Please use',
're2::RE2 instead (crbug.com/755321)',
),
True,
# Abseil's benchmarks never linked into chrome.
['third_party/abseil-cpp/.*_benchmark.cc'],
),
BanRule(
r'/\bstd::stoi\b',
(
'std::stoi uses exceptions to communicate results. ',
'Use base::StringToInt() instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::stol\b',
(
'std::stol uses exceptions to communicate results. ',
'Use base::StringToInt() instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::stoul\b',
(
'std::stoul uses exceptions to communicate results. ',
'Use base::StringToUint() instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::stoll\b',
(
'std::stoll uses exceptions to communicate results. ',
'Use base::StringToInt64() instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::stoull\b',
(
'std::stoull uses exceptions to communicate results. ',
'Use base::StringToUint64() instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::stof\b',
(
'std::stof uses exceptions to communicate results. ',
'For locale-independent values, e.g. reading numbers from disk',
'profiles, use base::StringToDouble().',
'For user-visible values, parse using ICU.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::stod\b',
(
'std::stod uses exceptions to communicate results. ',
'For locale-independent values, e.g. reading numbers from disk',
'profiles, use base::StringToDouble().',
'For user-visible values, parse using ICU.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::stold\b',
(
'std::stold uses exceptions to communicate results. ',
'For locale-independent values, e.g. reading numbers from disk',
'profiles, use base::StringToDouble().',
'For user-visible values, parse using ICU.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::to_string\b',
(
'std::to_string is locale dependent and slower than alternatives.',
'For locale-independent strings, e.g. writing numbers to disk',
'profiles, use base::NumberToString().',
'For user-visible strings, use base::FormatNumber() and',
'the related functions in base/i18n/number_formatting.h.',
),
False, # Only a warning since it is already used.
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\bstd::shared_ptr\b',
(
'std::shared_ptr should not be used. Use scoped_refptr instead.',
),
True,
[
# Needed for interop with third-party library.
'^third_party/blink/renderer/core/typed_arrays/array_buffer/' +
'array_buffer_contents\.(cc|h)',
'^third_party/blink/renderer/bindings/core/v8/' +
'v8_wasm_response_extensions.cc',
'^gin/array_buffer\.(cc|h)',
'^chrome/services/sharing/nearby/',
# Needed for interop with third-party library libunwindstack.
'^base/profiler/libunwindstack_unwinder_android\.(cc|h)',
# Needed for interop with third-party boringssl cert verifier
'^third_party/boringssl/',
'^net/cert/',
'^net/tools/cert_verify_tool/',
'^services/cert_verifier/',
'^components/certificate_transparency/',
'^components/media_router/common/providers/cast/certificate/',
# gRPC provides some C++ libraries that use std::shared_ptr<>.
'^chromeos/ash/services/libassistant/grpc/',
'^chromecast/cast_core/grpc',
'^chromecast/cast_core/runtime/browser',
'^ios/chrome/test/earl_grey/chrome_egtest_plugin_client\.(mm|h)',
# Fuchsia provides C++ libraries that use std::shared_ptr<>.
'^base/fuchsia/.*\.(cc|h)',
'.*fuchsia.*test\.(cc|h)',
# Needed for clang plugin tests
'^tools/clang/plugins/tests/',
_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\bstd::weak_ptr\b',
(
'std::weak_ptr should not be used. Use base::WeakPtr instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\blong long\b',
(
'long long is banned. Use stdint.h if you need a 64 bit number.',
),
False, # Only a warning since it is already used.
[_THIRD_PARTY_EXCEPT_BLINK], # Don't warn in third_party folders.
),
BanRule(
r'/\b(absl|std)::any\b',
(
'absl::any / std::any are not safe to use in a component build.',
),
True,
# Not an error in third party folders, though it probably should be :)
[_THIRD_PARTY_EXCEPT_BLINK],
),
BanRule(
r'/\bstd::bind\b',
(
'std::bind is banned because of lifetime risks.',
'Use base::BindOnce or base::BindRepeating instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
(
r'/\b(?:'
r'std::linear_congruential_engine|std::mersenne_twister_engine|'
r'std::subtract_with_carry_engine|std::discard_block_engine|'
r'std::independent_bits_engine|std::shuffle_order_engine|'
r'std::minstd_rand0|std::minstd_rand|'
r'std::mt19937|std::mt19937_64|'
r'std::ranlux24_base|std::ranlux48_base|std::ranlux24|std::ranlux48|'
r'std::knuth_b|'
r'std::default_random_engine|'
r'std::random_device'
r')\b'
),
(
'STL random number engines and generators are banned. Use the ',
'helpers in base/rand_util.h instead, e.g. base::RandBytes() or ',
'base::RandomBitGenerator.'
),
True,
[
# Not an error in third_party folders.
_THIRD_PARTY_EXCEPT_BLINK,
# Various tools which build outside of Chrome.
r'testing/libfuzzer',
r'tools/android/io_benchmark/',
# Fuzzers are allowed to use standard library random number generators
# since fuzzing speed + reproducibility is important.
r'tools/ipc_fuzzer/',
r'.+_fuzzer\.cc$',
r'.+_fuzzertest\.cc$',
# TODO(https://crbug.com/1380528): These are all unsanctioned uses of
# the standard library's random number generators, and should be
# migrated to the //base equivalent.
r'ash/ambient/model/ambient_topic_queue\.cc',
r'base/allocator/partition_allocator/partition_alloc_unittest\.cc',
r'base/ranges/algorithm_unittest\.cc',
r'base/test/launcher/test_launcher\.cc',
r'cc/metrics/video_playback_roughness_reporter_unittest\.cc',
r'chrome/browser/apps/app_service/metrics/website_metrics\.cc',
r'chrome/browser/ash/power/auto_screen_brightness/monotone_cubic_spline_unittest\.cc',
r'chrome/browser/ash/printing/zeroconf_printer_detector_unittest\.cc',
r'chrome/browser/nearby_sharing/contacts/nearby_share_contact_manager_impl_unittest\.cc',
r'chrome/browser/nearby_sharing/contacts/nearby_share_contacts_sorter_unittest\.cc',
r'chrome/browser/privacy_budget/mesa_distribution_unittest\.cc',
r'chrome/browser/web_applications/test/web_app_test_utils\.cc',
r'chrome/browser/web_applications/test/web_app_test_utils\.cc',
r'chrome/browser/win/conflicts/module_blocklist_cache_util_unittest\.cc',
r'chrome/chrome_cleaner/logging/detailed_info_sampler\.cc',
r'chromeos/ash/components/memory/userspace_swap/swap_storage_unittest\.cc',
r'chromeos/ash/components/memory/userspace_swap/userspace_swap\.cc',
r'components/metrics/metrics_state_manager\.cc',
r'components/omnibox/browser/history_quick_provider_performance_unittest\.cc',
r'components/zucchini/disassembler_elf_unittest\.cc',
r'content/browser/webid/federated_auth_request_impl\.cc',
r'content/browser/webid/federated_auth_request_impl\.cc',
r'media/cast/test/utility/udp_proxy\.h',
r'sql/recover_module/module_unittest\.cc',
],
),
BanRule(
r'/\babsl::bind_front\b',
(
'absl::bind_front is banned. Use base::BindOnce() or '
'base::BindRepeating() instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\bABSL_FLAG\b',
(
'ABSL_FLAG is banned. Use base::CommandLine instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\babsl::c_',
(
'Abseil container utilities are banned. Use base/ranges/algorithm.h',
'instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\babsl::FunctionRef\b',
(
'absl::FunctionRef is banned. Use base::FunctionRef instead.',
),
True,
[
# base::Bind{Once,Repeating} references absl::FunctionRef to disallow
# interoperability.
r'^base/functional/bind_internal\.h',
# base::FunctionRef is implemented on top of absl::FunctionRef.
r'^base/functional/function_ref.*\..+',
# Not an error in third_party folders.
_THIRD_PARTY_EXCEPT_BLINK,
],
),
BanRule(
r'/\babsl::(Insecure)?BitGen\b',
(
'absl random number generators are banned. Use the helpers in '
'base/rand_util.h instead, e.g. base::RandBytes() or ',
'base::RandomBitGenerator.'
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\babsl::Span\b',
(
'absl::Span is banned. Use base::span instead.',
),
True,
[
# Needed to use QUICHE API.
r'services/network/web_transport\.cc',
# Not an error in third_party folders.
_THIRD_PARTY_EXCEPT_BLINK
],
),
BanRule(
r'/\babsl::StatusOr\b',
(
'absl::StatusOr is banned. Use base::expected instead.',
),
True,
[
# Needed to use liburlpattern API.
r'third_party/blink/renderer/core/url_pattern/.*',
# Not an error in third_party folders.
_THIRD_PARTY_EXCEPT_BLINK
],
),
BanRule(
r'/\babsl::StrFormat\b',
(
'absl::StrFormat is banned for now. Use base::StringPrintf instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\babsl::string_view\b',
(
'absl::string_view is banned. Use base::StringPiece instead.',
),
True,
[
# Needed to use liburlpattern API.
r'third_party/blink/renderer/core/url_pattern/.*',
# Needed to use QUICHE API.
r'net/quic/.*',
r'net/spdy/.*',
r'net/test/embedded_test_server/.*',
r'net/third_party/quiche/.*',
r'services/network/web_transport\.cc',
# Not an error in third_party folders.
_THIRD_PARTY_EXCEPT_BLINK
],
),
BanRule(
r'/\babsl::(StrSplit|StrJoin|StrCat|StrAppend|Substitute|StrContains)\b',
(
'Abseil string utilities are banned. Use base/strings instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\babsl::(Mutex|CondVar|Notification|Barrier|BlockingCounter)\b',
(
'Abseil synchronization primitives are banned. Use',
'base/synchronization instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\babsl::(Duration|Time|TimeZone|CivilDay)\b',
(
'Abseil\'s time library is banned. Use base/time instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\bstd::optional\b',
(
'std::optional is banned. Use absl::optional instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\b#include <chrono>\b',
(
'<chrono> overlaps with Time APIs in base. Keep using',
'base classes.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\b#include <exception>\b',
(
'Exceptions are banned and disabled in Chromium.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\bstd::function\b',
(
'std::function is banned. Instead use base::OnceCallback or ',
'base::RepeatingCallback, which directly support Chromium\'s weak ',
'pointers, ref counting and more.',
),
True,
[
# Has tests that template trait helpers don't unintentionally match
# std::function.
r'base/functional/callback_helpers_unittest\.cc',
# Required to implement interfaces from the third-party perfetto
# library.
r'base/tracing/perfetto_task_runner\.cc',
r'base/tracing/perfetto_task_runner\.h',
# Needed for interop with the third-party nearby library type
# location::nearby::connections::ResultCallback.
'chrome/services/sharing/nearby/nearby_connections_conversions\.cc'
# Needed for interop with the internal libassistant library.
'chromeos/ash/services/libassistant/callback_utils\.h',
# Needed for interop with Fuchsia fidl APIs.
'fuchsia_web/webengine/browser/context_impl_browsertest\.cc',
'fuchsia_web/webengine/browser/cookie_manager_impl_unittest\.cc',
'fuchsia_web/webengine/browser/media_player_impl_unittest\.cc',
# Required to interop with interfaces from the third-party perfetto
# library.
'services/tracing/public/cpp/perfetto/custom_event_recorder\.cc',
'services/tracing/public/cpp/perfetto/perfetto_traced_process\.cc',
'services/tracing/public/cpp/perfetto/perfetto_traced_process\.h',
'services/tracing/public/cpp/perfetto/perfetto_tracing_backend\.cc',
'services/tracing/public/cpp/perfetto/producer_client\.cc',
'services/tracing/public/cpp/perfetto/producer_client\.h',
'services/tracing/public/cpp/perfetto/producer_test_utils\.cc',
'services/tracing/public/cpp/perfetto/producer_test_utils\.h',
# Required for interop with the third-party webrtc library.
'third_party/blink/renderer/modules/peerconnection/mock_peer_connection_impl\.cc',
'third_party/blink/renderer/modules/peerconnection/mock_peer_connection_impl\.h',
# This code is in the process of being extracted into a third-party library.
# See https://crbug.com/1322914
'^net/cert/pki/path_builder_unittest\.cc',
# TODO(https://crbug.com/1364577): Various uses that should be
# migrated to something else.
# Should use base::OnceCallback or base::RepeatingCallback.
'base/allocator/dispatcher/initializer_unittest\.cc',
'chrome/browser/ash/accessibility/speech_monitor\.cc',
'chrome/browser/ash/accessibility/speech_monitor\.h',
'chrome/browser/ash/login/ash_hud_login_browsertest\.cc',
'chromecast/base/observer_unittest\.cc',
'chromecast/browser/cast_web_view\.h',
'chromecast/public/cast_media_shlib\.h',
'device/bluetooth/floss/exported_callback_manager\.h',
'device/bluetooth/floss/floss_dbus_client\.h',
'device/fido/cable/v2_handshake_unittest\.cc',
'device/fido/pin\.cc',
'services/tracing/perfetto/test_utils\.h',
# Should use base::FunctionRef.
'chrome/browser/media/webrtc/test_stats_dictionary\.cc',
'chrome/browser/media/webrtc/test_stats_dictionary\.h',
'chromeos/ash/services/libassistant/device_settings_controller\.cc',
'components/browser_ui/client_certificate/android/ssl_client_certificate_request\.cc',
'components/gwp_asan/client/sampling_malloc_shims_unittest\.cc',
'content/browser/font_unique_name_lookup/font_unique_name_lookup_unittest\.cc',
# Does not need std::function at all.
'components/omnibox/browser/autocomplete_result\.cc',
'device/fido/win/webauthn_api\.cc',
'media/audio/alsa/alsa_util\.cc',
'media/remoting/stream_provider\.h',
'sql/vfs_wrapper\.cc',
# TODO(https://crbug.com/1364585): Remove usage and exception list
# entries.
'extensions/renderer/api/automation/automation_internal_custom_bindings\.cc',
'extensions/renderer/api/automation/automation_internal_custom_bindings\.h',
# TODO(https://crbug.com/1364579): Remove usage and exception list
# entry.
'ui/views/controls/focus_ring\.h',
# Various pre-existing uses in //tools that is low-priority to fix.
'tools/binary_size/libsupersize/viewer/caspian/diff\.cc',
'tools/binary_size/libsupersize/viewer/caspian/model\.cc',
'tools/binary_size/libsupersize/viewer/caspian/model\.h',
'tools/binary_size/libsupersize/viewer/caspian/tree_builder\.h',
'tools/clang/base_bind_rewriters/BaseBindRewriters\.cpp',
# Not an error in third_party folders.
_THIRD_PARTY_EXCEPT_BLINK
],
),
BanRule(
r'/\b#include <random>\b',
(
'Do not use any random number engines from <random>. Instead',
'use base::RandomBitGenerator.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\b#include <X11/',
(
'Do not use Xlib. Use xproto (from //ui/gfx/x:xproto) instead.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\bstd::ratio\b',
(
'std::ratio is banned by the Google Style Guide.',
),
True,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
r'/\bRunMessageLoop\b',
(
'RunMessageLoop is deprecated, use RunLoop instead.',
),
False,
(),
),
BanRule(
'RunAllPendingInMessageLoop()',
(
"Prefer RunLoop over RunAllPendingInMessageLoop, please contact gab@",
"if you're convinced you need this.",
),
False,
(),
),
BanRule(
'RunAllPendingInMessageLoop(BrowserThread',
(
'RunAllPendingInMessageLoop is deprecated. Use RunLoop for',
'BrowserThread::UI, BrowserTaskEnvironment::RunIOThreadUntilIdle',
'for BrowserThread::IO, and prefer RunLoop::QuitClosure to observe',
'async events instead of flushing threads.',
),
False,
(),
),
BanRule(
r'MessageLoopRunner',
(
'MessageLoopRunner is deprecated, use RunLoop instead.',
),
False,
(),
),
BanRule(
'GetDeferredQuitTaskForRunLoop',
(
"GetDeferredQuitTaskForRunLoop shouldn't be needed, please contact",
"gab@ if you found a use case where this is the only solution.",
),
False,
(),
),
BanRule(
'sqlite3_initialize(',
(
'Instead of calling sqlite3_initialize(), depend on //sql, ',
'#include "sql/initialize.h" and use sql::EnsureSqliteInitialized().',
),
True,
(
r'^sql/initialization\.(cc|h)$',
r'^third_party/sqlite/.*\.(c|cc|h)$',
),
),
BanRule(
'CREATE VIEW',
(
'SQL views are disabled in Chromium feature code',
'https://chromium.googlesource.com/chromium/src/+/HEAD/sql#no-views',
),
True,
(
_THIRD_PARTY_EXCEPT_BLINK,
# sql/ itself uses views when using memory-mapped IO.
r'^sql/.*',
# Various performance tools that do not build as part of Chrome.
r'^infra/.*',
r'^tools/perf.*',
r'.*perfetto.*',
),
),
BanRule(
'CREATE VIRTUAL TABLE',
(
'SQL virtual tables are disabled in Chromium feature code',
'https://chromium.googlesource.com/chromium/src/+/HEAD/sql#no-virtual-tables',
),
True,
(
_THIRD_PARTY_EXCEPT_BLINK,
# sql/ itself uses virtual tables in the recovery module and tests.
r'^sql/.*',
# TODO(https://crbug.com/695592): Remove once WebSQL is deprecated.
r'third_party/blink/web_tests/storage/websql/.*'
# Various performance tools that do not build as part of Chrome.
r'^tools/perf.*',
r'.*perfetto.*',
),
),
BanRule(
'std::random_shuffle',
(
'std::random_shuffle is deprecated in C++14, and removed in C++17. Use',
'base::RandomShuffle instead.'
),
True,
(),
),
BanRule(
'ios/web/public/test/http_server',
(
'web::HTTPserver is deprecated use net::EmbeddedTestServer instead.',
),
False,
(),
),
BanRule(
'GetAddressOf',
(
'Improper use of Microsoft::WRL::ComPtr<T>::GetAddressOf() has been ',
'implicated in a few leaks. ReleaseAndGetAddressOf() is safe but ',
'operator& is generally recommended. So always use operator& instead. ',
'See http://crbug.com/914910 for more conversion guidance.'
),
True,
(),
),
BanRule(
'SHFileOperation',
(
'SHFileOperation was deprecated in Windows Vista, and there are less ',
'complex functions to achieve the same goals. Use IFileOperation for ',
'any esoteric actions instead.'
),
True,
(),
),
BanRule(
'StringFromGUID2',
(
'StringFromGUID2 introduces an unnecessary dependency on ole32.dll.',
'Use base::win::WStringFromGUID instead.'
),
True,
(
r'/base/win/win_util_unittest.cc',
),
),
BanRule(
'StringFromCLSID',
(
'StringFromCLSID introduces an unnecessary dependency on ole32.dll.',
'Use base::win::WStringFromGUID instead.'
),
True,
(
r'/base/win/win_util_unittest.cc',
),
),
BanRule(
'kCFAllocatorNull',
(
'The use of kCFAllocatorNull with the NoCopy creation of ',
'CoreFoundation types is prohibited.',
),
True,
(),
),
BanRule(
'mojo::ConvertTo',
(
'mojo::ConvertTo and TypeConverter are deprecated. Please consider',
'StructTraits / UnionTraits / EnumTraits / ArrayTraits / MapTraits /',
'StringTraits if you would like to convert between custom types and',
'the wire format of mojom types.'
),
False,
(
r'^fuchsia_web/webengine/browser/url_request_rewrite_rules_manager\.cc$',
r'^fuchsia_web/webengine/url_request_rewrite_type_converters\.cc$',
r'^third_party/blink/.*\.(cc|h)$',
r'^content/renderer/.*\.(cc|h)$',
),
),
BanRule(
'GetInterfaceProvider',
(
'InterfaceProvider is deprecated.',
'Please use ExecutionContext::GetBrowserInterfaceBroker and overrides',
'or Platform::GetBrowserInterfaceBroker.'
),
False,
(),
),
BanRule(
'CComPtr',
(
'New code should use Microsoft::WRL::ComPtr from wrl/client.h as a ',
'replacement for CComPtr from ATL. See http://crbug.com/5027 for more ',
'details.'
),
False,
(),
),
BanRule(
r'/\b(IFACE|STD)METHOD_?\(',
(
'IFACEMETHOD() and STDMETHOD() make code harder to format and read.',
'Instead, always use IFACEMETHODIMP in the declaration.'
),
False,
[_THIRD_PARTY_EXCEPT_BLINK], # Not an error in third_party folders.
),
BanRule(
'set_owned_by_client',
(
'set_owned_by_client is deprecated.',
'views::View already owns the child views by default. This introduces ',
'a competing ownership model which makes the code difficult to reason ',
'about. See http://crbug.com/1044687 for more details.'
),
False,
(),
),
BanRule(
'RemoveAllChildViewsWithoutDeleting',
(
'RemoveAllChildViewsWithoutDeleting is deprecated.',
'This method is deemed dangerous as, unless raw pointers are re-added,',
'calls to this method introduce memory leaks.'
),
False,
(),
),
BanRule(
r'/\bTRACE_EVENT_ASYNC_',
(
'Please use TRACE_EVENT_NESTABLE_ASYNC_.. macros instead',
'of TRACE_EVENT_ASYNC_.. (crbug.com/1038710).',
),
False,
(
r'^base/trace_event/.*',
r'^base/tracing/.*',
),
),
BanRule(
r'/\bbase::debug::DumpWithoutCrashingUnthrottled[(][)]',
(
'base::debug::DumpWithoutCrashingUnthrottled() does not throttle',
'dumps and may spam crash reports. Consider if the throttled',
'variants suffice instead.',
),
False,
(),
),
BanRule(
'RoInitialize',
(
'Improper use of [base::win]::RoInitialize() has been implicated in a ',
'few COM initialization leaks. Use base::win::ScopedWinrtInitializer ',
'instead. See http://crbug.com/1197722 for more information.'
),
True,
(
r'^base/win/scoped_winrt_initializer\.cc$',
),
),
BanRule(
r'base::Watchdog',
(
'base::Watchdog is deprecated because it creates its own thread.',
'Instead, manually start a timer on a SequencedTaskRunner.',
),
False,
(),
),
BanRule(
'base::Passed',
(
'Do not use base::Passed. It is a legacy helper for capturing ',
'move-only types with base::BindRepeating, but invoking the ',
'resulting RepeatingCallback moves the captured value out of ',
'the callback storage, and subsequent invocations may pass the ',
'value in a valid but undefined state. Prefer base::BindOnce().',
'See http://crbug.com/1326449 for context.'
),
False,
(
# False positive, but it is also fine to let bind internals reference
# base::Passed.
r'^base[\\/]functional[\\/]bind\.h',
r'^base[\\/]functional[\\/]bind_internal\.h',
),
),
BanRule(
r'base::Feature k',
(
'Please use BASE_DECLARE_FEATURE() or BASE_FEATURE() instead of ',
'directly declaring/defining features.'
),
True,
[
_THIRD_PARTY_EXCEPT_BLINK,
],
),
BanRule(
r'\bchartorune\b',
(
'chartorune is not memory-safe, unless you can guarantee the input ',
'string is always null-terminated. Otherwise, please use charntorune ',
'from libphonenumber instead.'
),
True,
[
_THIRD_PARTY_EXCEPT_BLINK,
# Exceptions to this rule should have a fuzzer.
],
),
)
_BANNED_MOJOM_PATTERNS : Sequence[BanRule] = (
BanRule(
'handle<shared_buffer>',
(
'Please use one of the more specific shared memory types instead:',
' mojo_base.mojom.ReadOnlySharedMemoryRegion',
' mojo_base.mojom.WritableSharedMemoryRegion',
' mojo_base.mojom.UnsafeSharedMemoryRegion',
),
True,
),
)
_IPC_ENUM_TRAITS_DEPRECATED = (
'You are using IPC_ENUM_TRAITS() in your code. It has been deprecated.\n'
'See http://www.chromium.org/Home/chromium-security/education/'
'security-tips-for-ipc')
_LONG_PATH_ERROR = (
'Some files included in this CL have file names that are too long (> 200'
' characters). If committed, these files will cause issues on Windows. See'
' https://crbug.com/612667 for more details.'
)
_JAVA_MULTIPLE_DEFINITION_EXCLUDED_PATHS = [
r".*/AppHooksImpl\.java",
r".*/BuildHooksAndroidImpl\.java",
r".*/LicenseContentProvider\.java",
r".*/PlatformServiceBridgeImpl.java",
r".*chrome/android/feed/dummy/.*\.java",
]
# List of image extensions that are used as resources in chromium.
_IMAGE_EXTENSIONS = ['.svg', '.png', '.webp']
# These paths contain test data and other known invalid JSON files.
_KNOWN_TEST_DATA_AND_INVALID_JSON_FILE_PATTERNS = [
r'test/data/',
r'testing/buildbot/',
r'^components/policy/resources/policy_templates\.json$',
r'^third_party/protobuf/',
r'^third_party/blink/perf_tests/speedometer/resources/todomvc/learn.json',
r'^third_party/blink/renderer/devtools/protocol\.json$',
r'^third_party/blink/web_tests/external/wpt/',
r'^tools/perf/',
r'^tools/traceline/svgui/startup-release.json',
# vscode configuration files allow comments
r'^tools/vscode/',
]
# These are not checked on the public chromium-presubmit trybot.
# Add files here that rely on .py files that exists only for target_os="android"
# checkouts.
_ANDROID_SPECIFIC_PYDEPS_FILES = [
'chrome/android/features/create_stripped_java_factory.pydeps',
]
_GENERIC_PYDEPS_FILES = [
'android_webview/test/components/run_webview_component_smoketest.pydeps',
'android_webview/tools/run_cts.pydeps',
'base/android/jni_generator/jni_generator.pydeps',
'base/android/jni_generator/jni_registration_generator.pydeps',
'build/android/apk_operations.pydeps',
'build/android/devil_chromium.pydeps',
'build/android/gyp/aar.pydeps',
'build/android/gyp/aidl.pydeps',
'build/android/gyp/allot_native_libraries.pydeps',
'build/android/gyp/apkbuilder.pydeps',
'build/android/gyp/assert_static_initializers.pydeps',
'build/android/gyp/bytecode_processor.pydeps',
'build/android/gyp/bytecode_rewriter.pydeps',
'build/android/gyp/check_flag_expectations.pydeps',
'build/android/gyp/compile_java.pydeps',
'build/android/gyp/compile_kt.pydeps',
'build/android/gyp/compile_resources.pydeps',
'build/android/gyp/copy_ex.pydeps',
'build/android/gyp/create_apk_operations_script.pydeps',
'build/android/gyp/create_app_bundle.pydeps',
'build/android/gyp/create_app_bundle_apks.pydeps',
'build/android/gyp/create_bundle_wrapper_script.pydeps',
'build/android/gyp/create_java_binary_script.pydeps',
'build/android/gyp/create_r_java.pydeps',
'build/android/gyp/create_r_txt.pydeps',
'build/android/gyp/create_size_info_files.pydeps',
'build/android/gyp/create_test_apk_wrapper_script.pydeps',
'build/android/gyp/create_ui_locale_resources.pydeps',
'build/android/gyp/dex.pydeps',
'build/android/gyp/dist_aar.pydeps',
'build/android/gyp/filter_zip.pydeps',
'build/android/gyp/flatc_java.pydeps',
'build/android/gyp/gcc_preprocess.pydeps',
'build/android/gyp/generate_linker_version_script.pydeps',
'build/android/gyp/ijar.pydeps',
'build/android/gyp/jacoco_instr.pydeps',
'build/android/gyp/java_cpp_enum.pydeps',
'build/android/gyp/java_cpp_features.pydeps',
'build/android/gyp/java_cpp_strings.pydeps',
'build/android/gyp/java_google_api_keys.pydeps',
'build/android/gyp/jinja_template.pydeps',
'build/android/gyp/lint.pydeps',
'build/android/gyp/merge_manifest.pydeps',
'build/android/gyp/optimize_resources.pydeps',
'build/android/gyp/prepare_resources.pydeps',
'build/android/gyp/process_native_prebuilt.pydeps',
'build/android/gyp/proguard.pydeps',
'build/android/gyp/system_image_apks.pydeps',
'build/android/gyp/trace_event_bytecode_rewriter.pydeps',
'build/android/gyp/turbine.pydeps',
'build/android/gyp/unused_resources.pydeps',
'build/android/gyp/validate_static_library_dex_references.pydeps',
'build/android/gyp/write_build_config.pydeps',
'build/android/gyp/write_native_libraries_java.pydeps',
'build/android/gyp/zip.pydeps',
'build/android/incremental_install/generate_android_manifest.pydeps',
'build/android/incremental_install/write_installer_json.pydeps',
'build/android/pylib/results/presentation/test_results_presentation.pydeps',
'build/android/resource_sizes.pydeps',
'build/android/test_runner.pydeps',
'build/android/test_wrapper/logdog_wrapper.pydeps',
'build/lacros/lacros_resource_sizes.pydeps',
'build/protoc_java.pydeps',
'chrome/android/monochrome/scripts/monochrome_python_tests.pydeps',
'chrome/test/chromedriver/log_replay/client_replay_unittest.pydeps',
'chrome/test/chromedriver/test/run_py_tests.pydeps',
'chromecast/resource_sizes/chromecast_resource_sizes.pydeps',
'components/cronet/tools/generate_javadoc.pydeps',
'components/cronet/tools/jar_src.pydeps',
'components/module_installer/android/module_desc_java.pydeps',
'content/public/android/generate_child_service.pydeps',
'net/tools/testserver/testserver.pydeps',
'testing/scripts/run_isolated_script_test.pydeps',
'testing/merge_scripts/standard_isolated_script_merge.pydeps',
'testing/merge_scripts/standard_gtest_merge.pydeps',
'testing/merge_scripts/code_coverage/merge_results.pydeps',
'testing/merge_scripts/code_coverage/merge_steps.pydeps',
'third_party/android_platform/development/scripts/stack.pydeps',
'third_party/blink/renderer/bindings/scripts/build_web_idl_database.pydeps',
'third_party/blink/renderer/bindings/scripts/check_generated_file_list.pydeps',
'third_party/blink/renderer/bindings/scripts/collect_idl_files.pydeps',
'third_party/blink/renderer/bindings/scripts/generate_bindings.pydeps',
'third_party/blink/renderer/bindings/scripts/validate_web_idl.pydeps',
'third_party/blink/tools/blinkpy/web_tests/merge_results.pydeps',
'third_party/blink/tools/merge_web_test_results.pydeps',
'tools/binary_size/sizes.pydeps',
'tools/binary_size/supersize.pydeps',
'tools/perf/process_perf_results.pydeps',
]
_ALL_PYDEPS_FILES = _ANDROID_SPECIFIC_PYDEPS_FILES + _GENERIC_PYDEPS_FILES
# Bypass the AUTHORS check for these accounts.
_KNOWN_ROBOTS = set(
) | set('%s@appspot.gserviceaccount.com' % s for s in ('findit-for-me',)
) | set('%s@developer.gserviceaccount.com' % s for s in ('3su6n15k.default',)
) | set('%s@chops-service-accounts.iam.gserviceaccount.com' % s
for s in ('bling-autoroll-builder', 'v8-ci-autoroll-builder',
'wpt-autoroller', 'chrome-weblayer-builder',
'lacros-version-skew-roller', 'skylab-test-cros-roller',
'infra-try-recipes-tester', 'lacros-tracking-roller',
'lacros-sdk-version-roller', 'chrome-automated-expectation',
'chromium-automated-expectation')
) | set('%s@skia-public.iam.gserviceaccount.com' % s
for s in ('chromium-autoroll', 'chromium-release-autoroll')
) | set('%s@skia-corp.google.com.iam.gserviceaccount.com' % s
for s in ('chromium-internal-autoroll',)
) | set('%s@owners-cleanup-prod.google.com.iam.gserviceaccount.com' % s
for s in ('swarming-tasks',)
) | set('%s@fuchsia-infra.iam.gserviceaccount.com' % s
for s in ('global-integration-try-builder',
'global-integration-ci-builder'))
_INVALID_GRD_FILE_LINE = [
(r'<file lang=.* path=.*', 'Path should come before lang in GRD files.')
]
def _IsCPlusPlusFile(input_api, file_path):
"""Returns True if this file contains C++-like code (and not Python,
Go, Java, MarkDown, ...)"""
ext = input_api.os_path.splitext(file_path)[1]
# This list is compatible with CppChecker.IsCppFile but we should
# consider adding ".c" to it. If we do that we can use this function
# at more places in the code.
return ext in (
'.h',
'.cc',
'.cpp',
'.m',
'.mm',
)
def _IsCPlusPlusHeaderFile(input_api, file_path):
return input_api.os_path.splitext(file_path)[1] == ".h"
def _IsJavaFile(input_api, file_path):
return input_api.os_path.splitext(file_path)[1] == ".java"
def _IsProtoFile(input_api, file_path):
return input_api.os_path.splitext(file_path)[1] == ".proto"
def _IsXmlOrGrdFile(input_api, file_path):
ext = input_api.os_path.splitext(file_path)[1]
return ext in ('.grd', '.xml')
def _IsMojomFile(input_api, file_path):
return input_api.os_path.splitext(file_path)[1] == ".mojom"
def CheckNoUpstreamDepsOnClank(input_api, output_api):
"""Prevent additions of dependencies from the upstream repo on //clank."""
# clank can depend on clank
if input_api.change.RepositoryRoot().endswith('clank'):
return []
build_file_patterns = [
r'(.+/)?BUILD\.gn',
r'.+\.gni',
]
excluded_files = [r'build[/\\]config[/\\]android[/\\]config\.gni']
bad_pattern = input_api.re.compile(r'^[^#]*//clank')
error_message = 'Disallowed import on //clank in an upstream build file:'
def FilterFile(affected_file):
return input_api.FilterSourceFile(affected_file,
files_to_check=build_file_patterns,
files_to_skip=excluded_files)
problems = []
for f in input_api.AffectedSourceFiles(FilterFile):
local_path = f.LocalPath()
for line_number, line in f.ChangedContents():
if (bad_pattern.search(line)):
problems.append('%s:%d\n %s' %
(local_path, line_number, line.strip()))
if problems:
return [output_api.PresubmitPromptOrNotify(error_message, problems)]
else:
return []
def CheckNoProductionCodeUsingTestOnlyFunctions(input_api, output_api):
"""Attempts to prevent use of functions intended only for testing in
non-testing code. For now this is just a best-effort implementation
that ignores header files and may have some false positives. A
better implementation would probably need a proper C++ parser.
"""
# We only scan .cc files and the like, as the declaration of
# for-testing functions in header files are hard to distinguish from
# calls to such functions without a proper C++ parser.
file_inclusion_pattern = [r'.+%s' % _IMPLEMENTATION_EXTENSIONS]
base_function_pattern = r'[ :]test::[^\s]+|ForTest(s|ing)?|for_test(s|ing)?'
inclusion_pattern = input_api.re.compile(r'(%s)\s*\(' %
base_function_pattern)
comment_pattern = input_api.re.compile(r'//.*(%s)' % base_function_pattern)
allowlist_pattern = input_api.re.compile(r'// IN-TEST$')
exclusion_pattern = input_api.re.compile(
r'::[A-Za-z0-9_]+(%s)|(%s)[^;]+\{' %
(base_function_pattern, base_function_pattern))
# Avoid a false positive in this case, where the method name, the ::, and
# the closing { are all on different lines due to line wrapping.
# HelperClassForTesting::
# HelperClassForTesting(
# args)
# : member(0) {}
method_defn_pattern = input_api.re.compile(r'[A-Za-z0-9_]+::$')
def FilterFile(affected_file):
files_to_skip = (_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS +
input_api.DEFAULT_FILES_TO_SKIP)
return input_api.FilterSourceFile(
affected_file,
files_to_check=file_inclusion_pattern,
files_to_skip=files_to_skip)
problems = []
for f in input_api.AffectedSourceFiles(FilterFile):
local_path = f.LocalPath()
in_method_defn = False
for line_number, line in f.ChangedContents():
if (inclusion_pattern.search(line)
and not comment_pattern.search(line)
and not exclusion_pattern.search(line)
and not allowlist_pattern.search(line)
and not in_method_defn):
problems.append('%s:%d\n %s' %
(local_path, line_number, line.strip()))
in_method_defn = method_defn_pattern.search(line)
if problems:
return [
output_api.PresubmitPromptOrNotify(_TEST_ONLY_WARNING, problems)
]
else:
return []
def CheckNoProductionCodeUsingTestOnlyFunctionsJava(input_api, output_api):
"""This is a simplified version of
CheckNoProductionCodeUsingTestOnlyFunctions for Java files.
"""
javadoc_start_re = input_api.re.compile(r'^\s*/\*\*')
javadoc_end_re = input_api.re.compile(r'^\s*\*/')
name_pattern = r'ForTest(s|ing)?'
# Describes an occurrence of "ForTest*" inside a // comment.
comment_re = input_api.re.compile(r'//.*%s' % name_pattern)
# Describes @VisibleForTesting(otherwise = VisibleForTesting.PROTECTED)
annotation_re = input_api.re.compile(r'@VisibleForTesting\(')
# Catch calls.
inclusion_re = input_api.re.compile(r'(%s)\s*\(' % name_pattern)
# Ignore definitions. (Comments are ignored separately.)
exclusion_re = input_api.re.compile(r'(%s)[^;]+\{' % name_pattern)
problems = []
sources = lambda x: input_api.FilterSourceFile(
x,
files_to_skip=(('(?i).*test', r'.*\/junit\/') + input_api.
DEFAULT_FILES_TO_SKIP),
files_to_check=[r'.*\.java$'])
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=sources):
local_path = f.LocalPath()
is_inside_javadoc = False
for line_number, line in f.ChangedContents():
if is_inside_javadoc and javadoc_end_re.search(line):
is_inside_javadoc = False
if not is_inside_javadoc and javadoc_start_re.search(line):
is_inside_javadoc = True
if is_inside_javadoc:
continue
if (inclusion_re.search(line) and not comment_re.search(line)
and not annotation_re.search(line)
and not exclusion_re.search(line)):
problems.append('%s:%d\n %s' %
(local_path, line_number, line.strip()))
if problems:
return [
output_api.PresubmitPromptOrNotify(_TEST_ONLY_WARNING, problems)
]
else:
return []
def CheckNoIOStreamInHeaders(input_api, output_api):
"""Checks to make sure no .h files include <iostream>."""
files = []
pattern = input_api.re.compile(r'^#include\s*<iostream>',
input_api.re.MULTILINE)
for f in input_api.AffectedSourceFiles(input_api.FilterSourceFile):
if not f.LocalPath().endswith('.h'):
continue
contents = input_api.ReadFile(f)
if pattern.search(contents):
files.append(f)
if len(files):
return [
output_api.PresubmitError(
'Do not #include <iostream> in header files, since it inserts static '
'initialization into every file including the header. Instead, '
'#include <ostream>. See http://crbug.com/94794', files)
]
return []
def CheckNoStrCatRedefines(input_api, output_api):
"""Checks no windows headers with StrCat redefined are included directly."""
files = []
files_to_check = (r'.+%s' % _HEADER_EXTENSIONS,
r'.+%s' % _IMPLEMENTATION_EXTENSIONS)
files_to_skip = (input_api.DEFAULT_FILES_TO_SKIP +
_NON_BASE_DEPENDENT_PATHS)
sources_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=files_to_check, files_to_skip=files_to_skip)
pattern_deny = input_api.re.compile(
r'^#include\s*[<"](shlwapi|atlbase|propvarutil|sphelper).h[">]',
input_api.re.MULTILINE)
pattern_allow = input_api.re.compile(
r'^#include\s"base/win/windows_defines.inc"', input_api.re.MULTILINE)
for f in input_api.AffectedSourceFiles(sources_filter):
contents = input_api.ReadFile(f)
if pattern_deny.search(
contents) and not pattern_allow.search(contents):
files.append(f.LocalPath())
if len(files):
return [
output_api.PresubmitError(
'Do not #include shlwapi.h, atlbase.h, propvarutil.h or sphelper.h '
'directly since they pollute code with StrCat macro. Instead, '
'include matching header from base/win. See http://crbug.com/856536',
files)
]
return []
def CheckNoUNIT_TESTInSourceFiles(input_api, output_api):
"""Checks to make sure no source files use UNIT_TEST."""
problems = []
for f in input_api.AffectedFiles():
if (not f.LocalPath().endswith(('.cc', '.mm'))):
continue
for line_num, line in f.ChangedContents():
if 'UNIT_TEST ' in line or line.endswith('UNIT_TEST'):
problems.append(' %s:%d' % (f.LocalPath(), line_num))
if not problems:
return []
return [
output_api.PresubmitPromptWarning('UNIT_TEST is only for headers.\n' +
'\n'.join(problems))
]
def CheckNoDISABLETypoInTests(input_api, output_api):
"""Checks to prevent attempts to disable tests with DISABLE_ prefix.
This test warns if somebody tries to disable a test with the DISABLE_ prefix
instead of DISABLED_. To filter false positives, reports are only generated
if a corresponding MAYBE_ line exists.
"""
problems = []
# The following two patterns are looked for in tandem - is a test labeled
# as MAYBE_ followed by a DISABLE_ (instead of the correct DISABLED)
maybe_pattern = input_api.re.compile(r'MAYBE_([a-zA-Z0-9_]+)')
disable_pattern = input_api.re.compile(r'DISABLE_([a-zA-Z0-9_]+)')
# This is for the case that a test is disabled on all platforms.
full_disable_pattern = input_api.re.compile(
r'^\s*TEST[^(]*\([a-zA-Z0-9_]+,\s*DISABLE_[a-zA-Z0-9_]+\)',
input_api.re.MULTILINE)
for f in input_api.AffectedFiles(False):
if not 'test' in f.LocalPath() or not f.LocalPath().endswith('.cc'):
continue
# Search for MABYE_, DISABLE_ pairs.
disable_lines = {} # Maps of test name to line number.
maybe_lines = {}
for line_num, line in f.ChangedContents():
disable_match = disable_pattern.search(line)
if disable_match:
disable_lines[disable_match.group(1)] = line_num
maybe_match = maybe_pattern.search(line)
if maybe_match:
maybe_lines[maybe_match.group(1)] = line_num
# Search for DISABLE_ occurrences within a TEST() macro.
disable_tests = set(disable_lines.keys())
maybe_tests = set(maybe_lines.keys())
for test in disable_tests.intersection(maybe_tests):
problems.append(' %s:%d' % (f.LocalPath(), disable_lines[test]))
contents = input_api.ReadFile(f)
full_disable_match = full_disable_pattern.search(contents)
if full_disable_match:
problems.append(' %s' % f.LocalPath())
if not problems:
return []
return [
output_api.PresubmitPromptWarning(
'Attempt to disable a test with DISABLE_ instead of DISABLED_?\n' +
'\n'.join(problems))
]
def CheckForgettingMAYBEInTests(input_api, output_api):
"""Checks to make sure tests disabled conditionally are not missing a
corresponding MAYBE_ prefix.
"""
# Expect at least a lowercase character in the test name. This helps rule out
# false positives with macros wrapping the actual tests name.
define_maybe_pattern = input_api.re.compile(
r'^\#define MAYBE_(?P<test_name>\w*[a-z]\w*)')
# The test_maybe_pattern needs to handle all of these forms. The standard:
# IN_PROC_TEST_F(SyncTest, MAYBE_Start) {
# With a wrapper macro around the test name:
# IN_PROC_TEST_F(SyncTest, E2E_ENABLED(MAYBE_Start)) {
# And the odd-ball NACL_BROWSER_TEST_f format:
# NACL_BROWSER_TEST_F(NaClBrowserTest, SimpleLoad, {
# The optional E2E_ENABLED-style is handled with (\w*\()?
# The NACL_BROWSER_TEST_F pattern is handled by allowing a trailing comma or
# trailing ')'.
test_maybe_pattern = (
r'^\s*\w*TEST[^(]*\(\s*\w+,\s*(\w*\()?MAYBE_{test_name}[\),]')
suite_maybe_pattern = r'^\s*\w*TEST[^(]*\(\s*MAYBE_{test_name}[\),]'
warnings = []
# Read the entire files. We can't just read the affected lines, forgetting to
# add MAYBE_ on a change would not show up otherwise.
for f in input_api.AffectedFiles(False):
if not 'test' in f.LocalPath() or not f.LocalPath().endswith('.cc'):
continue
contents = input_api.ReadFile(f)
lines = contents.splitlines(True)
current_position = 0
warning_test_names = set()
for line_num, line in enumerate(lines, start=1):
current_position += len(line)
maybe_match = define_maybe_pattern.search(line)
if maybe_match:
test_name = maybe_match.group('test_name')
# Do not warn twice for the same test.
if (test_name in warning_test_names):
continue
warning_test_names.add(test_name)
# Attempt to find the corresponding MAYBE_ test or suite, starting from
# the current position.
test_match = input_api.re.compile(
test_maybe_pattern.format(test_name=test_name),
input_api.re.MULTILINE).search(contents, current_position)
suite_match = input_api.re.compile(
suite_maybe_pattern.format(test_name=test_name),
input_api.re.MULTILINE).search(contents, current_position)
if not test_match and not suite_match:
warnings.append(
output_api.PresubmitPromptWarning(
'%s:%d found MAYBE_ defined without corresponding test %s'
% (f.LocalPath(), line_num, test_name)))
return warnings
def CheckDCHECK_IS_ONHasBraces(input_api, output_api):
"""Checks to make sure DCHECK_IS_ON() does not skip the parentheses."""
errors = []
pattern = input_api.re.compile(r'\bDCHECK_IS_ON\b(?!\(\))',
input_api.re.MULTILINE)
for f in input_api.AffectedSourceFiles(input_api.FilterSourceFile):
if (not f.LocalPath().endswith(('.cc', '.mm', '.h'))):
continue
for lnum, line in f.ChangedContents():
if input_api.re.search(pattern, line):
errors.append(
output_api.PresubmitError((
'%s:%d: Use of DCHECK_IS_ON() must be written as "#if '
+ 'DCHECK_IS_ON()", not forgetting the parentheses.') %
(f.LocalPath(), lnum)))
return errors
# TODO(crbug/1138055): Reimplement CheckUmaHistogramChangesOnUpload check in a
# more reliable way. See
# https://chromium-review.googlesource.com/c/chromium/src/+/2500269
def CheckFlakyTestUsage(input_api, output_api):
"""Check that FlakyTest annotation is our own instead of the android one"""
pattern = input_api.re.compile(r'import android.test.FlakyTest;')
files = []
for f in input_api.AffectedSourceFiles(input_api.FilterSourceFile):
if f.LocalPath().endswith('Test.java'):
if pattern.search(input_api.ReadFile(f)):
files.append(f)
if len(files):
return [
output_api.PresubmitError(
'Use org.chromium.base.test.util.FlakyTest instead of '
'android.test.FlakyTest', files)
]
return []
def CheckNoDEPSGIT(input_api, output_api):
"""Make sure .DEPS.git is never modified manually."""
if any(f.LocalPath().endswith('.DEPS.git')
for f in input_api.AffectedFiles()):
return [
output_api.PresubmitError(
'Never commit changes to .DEPS.git. This file is maintained by an\n'
'automated system based on what\'s in DEPS and your changes will be\n'
'overwritten.\n'
'See https://sites.google.com/a/chromium.org/dev/developers/how-tos/'
'get-the-code#Rolling_DEPS\n'
'for more information')
]
return []
def CheckCrosApiNeedBrowserTest(input_api, output_api):
"""Check new crosapi should add browser test."""
has_new_crosapi = False
has_browser_test = False
for f in input_api.AffectedFiles():
path = f.LocalPath()
if (path.startswith('chromeos/crosapi/mojom') and
_IsMojomFile(input_api, path) and f.Action() == 'A'):
has_new_crosapi = True
if path.endswith('browsertest.cc') or path.endswith('browser_test.cc'):
has_browser_test = True
if has_new_crosapi and not has_browser_test:
return [
output_api.PresubmitPromptWarning(
'You are adding a new crosapi, but there is no file ends with '
'browsertest.cc file being added or modified. It is important '
'to add crosapi browser test coverage to avoid version '
' skew issues.\n'
'Check //docs/lacros/test_instructions.md for more information.'
)
]
return []
def CheckValidHostsInDEPSOnUpload(input_api, output_api):
"""Checks that DEPS file deps are from allowed_hosts."""
# Run only if DEPS file has been modified to annoy fewer bystanders.
if all(f.LocalPath() != 'DEPS' for f in input_api.AffectedFiles()):
return []
# Outsource work to gclient verify
try:
gclient_path = input_api.os_path.join(input_api.PresubmitLocalPath(),
'third_party', 'depot_tools',
'gclient.py')
input_api.subprocess.check_output(
[input_api.python3_executable, gclient_path, 'verify'],
stderr=input_api.subprocess.STDOUT)
return []
except input_api.subprocess.CalledProcessError as error:
return [
output_api.PresubmitError(
'DEPS file must have only git dependencies.',
long_text=error.output)
]
def _GetMessageForMatchingType(input_api, affected_file, line_number, line,
ban_rule):
"""Helper method for checking for banned constructs.
Returns an string composed of the name of the file, the line number where the
match has been found and the additional text passed as |message| in case the
target type name matches the text inside the line passed as parameter.
"""
result = []
# Ignore comments about banned types.
if input_api.re.search(r"^ *//", line):
return result
# A // nocheck comment will bypass this error.
if line.endswith(" nocheck"):
return result
matched = False
if ban_rule.pattern[0:1] == '/':
regex = ban_rule.pattern[1:]
if input_api.re.search(regex, line):
matched = True
elif ban_rule.pattern in line:
matched = True
if matched:
result.append(' %s:%d:' % (affected_file.LocalPath(), line_number))
for line in ban_rule.explanation:
result.append(' %s' % line)
return result
def CheckNoBannedFunctions(input_api, output_api):
"""Make sure that banned functions are not used."""
warnings = []
errors = []
def IsExcludedFile(affected_file, excluded_paths):
if not excluded_paths:
return False
local_path = affected_file.LocalPath()
# Consistently use / as path separator to simplify the writing of regex
# expressions.
local_path = local_path.replace(input_api.os_path.sep, '/')
for item in excluded_paths:
if input_api.re.match(item, local_path):
return True
return False
def IsIosObjcFile(affected_file):
local_path = affected_file.LocalPath()
if input_api.os_path.splitext(local_path)[-1] not in ('.mm', '.m',
'.h'):
return False
basename = input_api.os_path.basename(local_path)
if 'ios' in basename.split('_'):
return True
for sep in (input_api.os_path.sep, input_api.os_path.altsep):
if sep and 'ios' in local_path.split(sep):
return True
return False
def CheckForMatch(affected_file, line_num: int, line: str,
ban_rule: BanRule):
if IsExcludedFile(affected_file, ban_rule.excluded_paths):
return
problems = _GetMessageForMatchingType(input_api, f, line_num, line,
ban_rule)
if problems:
if ban_rule.treat_as_error is not None and ban_rule.treat_as_error:
errors.extend(problems)
else:
warnings.extend(problems)
file_filter = lambda f: f.LocalPath().endswith(('.java'))
for f in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in f.ChangedContents():
for ban_rule in _BANNED_JAVA_FUNCTIONS:
CheckForMatch(f, line_num, line, ban_rule)
file_filter = lambda f: f.LocalPath().endswith(('.js', '.ts'))
for f in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in f.ChangedContents():
for ban_rule in _BANNED_JAVASCRIPT_FUNCTIONS:
CheckForMatch(f, line_num, line, ban_rule)
file_filter = lambda f: f.LocalPath().endswith(('.mm', '.m', '.h'))
for f in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in f.ChangedContents():
for ban_rule in _BANNED_OBJC_FUNCTIONS:
CheckForMatch(f, line_num, line, ban_rule)
for f in input_api.AffectedFiles(file_filter=IsIosObjcFile):
for line_num, line in f.ChangedContents():
for ban_rule in _BANNED_IOS_OBJC_FUNCTIONS:
CheckForMatch(f, line_num, line, ban_rule)
egtest_filter = lambda f: f.LocalPath().endswith(('_egtest.mm'))
for f in input_api.AffectedFiles(file_filter=egtest_filter):
for line_num, line in f.ChangedContents():
for ban_rule in _BANNED_IOS_EGTEST_FUNCTIONS:
CheckForMatch(f, line_num, line, ban_rule)
file_filter = lambda f: f.LocalPath().endswith(('.cc', '.mm', '.h'))
for f in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in f.ChangedContents():
for ban_rule in _BANNED_CPP_FUNCTIONS:
CheckForMatch(f, line_num, line, ban_rule)
file_filter = lambda f: f.LocalPath().endswith(('.mojom'))
for f in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in f.ChangedContents():
for ban_rule in _BANNED_MOJOM_PATTERNS:
CheckForMatch(f, line_num, line, ban_rule)
result = []
if (warnings):
result.append(
output_api.PresubmitPromptWarning('Banned functions were used.\n' +
'\n'.join(warnings)))
if (errors):
result.append(
output_api.PresubmitError('Banned functions were used.\n' +
'\n'.join(errors)))
return result
def CheckNoLayoutCallsInTests(input_api, output_api):
"""Make sure there are no explicit calls to View::Layout() in tests"""
warnings = []
ban_rule = BanRule(
r'/(\.|->)Layout\(\);',
(
'Direct calls to View::Layout() are not allowed in tests. '
'If the view must be laid out here, use RunScheduledLayout(view). It '
'is found in //ui/views/test/views_test_utils.h. '
'See http://crbug.com/1350521 for more details.',
),
False,
)
file_filter = lambda f: input_api.re.search(
r'_(unittest|browsertest|ui_test).*\.(cc|mm)$', f.LocalPath())
for f in input_api.AffectedFiles(file_filter = file_filter):
for line_num, line in f.ChangedContents():
problems = _GetMessageForMatchingType(input_api, f,
line_num, line,
ban_rule)
if problems:
warnings.extend(problems)
result = []
if (warnings):
result.append(
output_api.PresubmitPromptWarning(
'Banned call to View::Layout() in tests.\n\n'.join(warnings)))
return result
def _CheckAndroidNoBannedImports(input_api, output_api):
"""Make sure that banned java imports are not used."""
errors = []
file_filter = lambda f: f.LocalPath().endswith(('.java'))
for f in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in f.ChangedContents():
for ban_rule in _BANNED_JAVA_IMPORTS:
# Consider merging this into the above function. There is no
# real difference anymore other than helping with a little
# bit of boilerplate text. Doing so means things like
# `treat_as_error` will also be uniformly handled.
problems = _GetMessageForMatchingType(input_api, f, line_num,
line, ban_rule)
if problems:
errors.extend(problems)
result = []
if (errors):
result.append(
output_api.PresubmitError('Banned imports were used.\n' +
'\n'.join(errors)))
return result
def CheckNoPragmaOnce(input_api, output_api):
"""Make sure that banned functions are not used."""
files = []
pattern = input_api.re.compile(r'^#pragma\s+once', input_api.re.MULTILINE)
for f in input_api.AffectedSourceFiles(input_api.FilterSourceFile):
if not f.LocalPath().endswith('.h'):
continue
if f.LocalPath().endswith('com_imported_mstscax.h'):
continue
contents = input_api.ReadFile(f)
if pattern.search(contents):
files.append(f)
if files:
return [
output_api.PresubmitError(
'Do not use #pragma once in header files.\n'
'See http://www.chromium.org/developers/coding-style#TOC-File-headers',
files)
]
return []
def CheckNoTrinaryTrueFalse(input_api, output_api):
"""Checks to make sure we don't introduce use of foo ? true : false."""
problems = []
pattern = input_api.re.compile(r'\?\s*(true|false)\s*:\s*(true|false)')
for f in input_api.AffectedFiles():
if not f.LocalPath().endswith(('.cc', '.h', '.inl', '.m', '.mm')):
continue
for line_num, line in f.ChangedContents():
if pattern.match(line):
problems.append(' %s:%d' % (f.LocalPath(), line_num))
if not problems:
return []
return [
output_api.PresubmitPromptWarning(
'Please consider avoiding the "? true : false" pattern if possible.\n'
+ '\n'.join(problems))
]
def CheckUnwantedDependencies(input_api, output_api):
"""Runs checkdeps on #include and import statements added in this
change. Breaking - rules is an error, breaking ! rules is a
warning.
"""
# Return early if no relevant file types were modified.
for f in input_api.AffectedFiles():
path = f.LocalPath()
if (_IsCPlusPlusFile(input_api, path) or _IsProtoFile(input_api, path)
or _IsJavaFile(input_api, path)):
break
else:
return []
import sys
# We need to wait until we have an input_api object and use this
# roundabout construct to import checkdeps because this file is
# eval-ed and thus doesn't have __file__.
original_sys_path = sys.path
try:
sys.path = sys.path + [
input_api.os_path.join(input_api.PresubmitLocalPath(),
'buildtools', 'checkdeps')
]
import checkdeps
from rules import Rule
finally:
# Restore sys.path to what it was before.
sys.path = original_sys_path
added_includes = []
added_imports = []
added_java_imports = []
for f in input_api.AffectedFiles():
if _IsCPlusPlusFile(input_api, f.LocalPath()):
changed_lines = [line for _, line in f.ChangedContents()]
added_includes.append([f.AbsoluteLocalPath(), changed_lines])
elif _IsProtoFile(input_api, f.LocalPath()):
changed_lines = [line for _, line in f.ChangedContents()]
added_imports.append([f.AbsoluteLocalPath(), changed_lines])
elif _IsJavaFile(input_api, f.LocalPath()):
changed_lines = [line for _, line in f.ChangedContents()]
added_java_imports.append([f.AbsoluteLocalPath(), changed_lines])
deps_checker = checkdeps.DepsChecker(input_api.PresubmitLocalPath())
error_descriptions = []
warning_descriptions = []
error_subjects = set()
warning_subjects = set()
for path, rule_type, rule_description in deps_checker.CheckAddedCppIncludes(
added_includes):
path = input_api.os_path.relpath(path, input_api.PresubmitLocalPath())
description_with_path = '%s\n %s' % (path, rule_description)
if rule_type == Rule.DISALLOW:
error_descriptions.append(description_with_path)
error_subjects.add("#includes")
else:
warning_descriptions.append(description_with_path)
warning_subjects.add("#includes")
for path, rule_type, rule_description in deps_checker.CheckAddedProtoImports(
added_imports):
path = input_api.os_path.relpath(path, input_api.PresubmitLocalPath())
description_with_path = '%s\n %s' % (path, rule_description)
if rule_type == Rule.DISALLOW:
error_descriptions.append(description_with_path)
error_subjects.add("imports")
else:
warning_descriptions.append(description_with_path)
warning_subjects.add("imports")
for path, rule_type, rule_description in deps_checker.CheckAddedJavaImports(
added_java_imports, _JAVA_MULTIPLE_DEFINITION_EXCLUDED_PATHS):
path = input_api.os_path.relpath(path, input_api.PresubmitLocalPath())
description_with_path = '%s\n %s' % (path, rule_description)
if rule_type == Rule.DISALLOW:
error_descriptions.append(description_with_path)
error_subjects.add("imports")
else:
warning_descriptions.append(description_with_path)
warning_subjects.add("imports")
results = []
if error_descriptions:
results.append(
output_api.PresubmitError(
'You added one or more %s that violate checkdeps rules.' %
" and ".join(error_subjects), error_descriptions))
if warning_descriptions:
results.append(
output_api.PresubmitPromptOrNotify(
'You added one or more %s of files that are temporarily\n'
'allowed but being removed. Can you avoid introducing the\n'
'%s? See relevant DEPS file(s) for details and contacts.' %
(" and ".join(warning_subjects), "/".join(warning_subjects)),
warning_descriptions))
return results
def CheckFilePermissions(input_api, output_api):
"""Check that all files have their permissions properly set."""
if input_api.platform == 'win32':
return []
checkperms_tool = input_api.os_path.join(input_api.PresubmitLocalPath(),
'tools', 'checkperms',
'checkperms.py')
args = [
input_api.python3_executable, checkperms_tool, '--root',
input_api.change.RepositoryRoot()
]
with input_api.CreateTemporaryFile() as file_list:
for f in input_api.AffectedFiles():
# checkperms.py file/directory arguments must be relative to the
# repository.
file_list.write((f.LocalPath() + '\n').encode('utf8'))
file_list.close()
args += ['--file-list', file_list.name]
try:
input_api.subprocess.check_output(args)
return []
except input_api.subprocess.CalledProcessError as error:
return [
output_api.PresubmitError('checkperms.py failed:',
long_text=error.output.decode(
'utf-8', 'ignore'))
]
def CheckNoAuraWindowPropertyHInHeaders(input_api, output_api):
"""Makes sure we don't include ui/aura/window_property.h
in header files.
"""
pattern = input_api.re.compile(r'^#include\s*"ui/aura/window_property.h"')
errors = []
for f in input_api.AffectedFiles():
if not f.LocalPath().endswith('.h'):
continue
for line_num, line in f.ChangedContents():
if pattern.match(line):
errors.append(' %s:%d' % (f.LocalPath(), line_num))
results = []
if errors:
results.append(
output_api.PresubmitError(
'Header files should not include ui/aura/window_property.h',
errors))
return results
def CheckNoInternalHeapIncludes(input_api, output_api):
"""Makes sure we don't include any headers from
third_party/blink/renderer/platform/heap/impl or
third_party/blink/renderer/platform/heap/v8_wrapper from files outside of
third_party/blink/renderer/platform/heap
"""
impl_pattern = input_api.re.compile(
r'^\s*#include\s*"third_party/blink/renderer/platform/heap/impl/.*"')
v8_wrapper_pattern = input_api.re.compile(
r'^\s*#include\s*"third_party/blink/renderer/platform/heap/v8_wrapper/.*"'
)
# Consistently use / as path separator to simplify the writing of regex
# expressions.
file_filter = lambda f: not input_api.re.match(
r"^third_party/blink/renderer/platform/heap/.*",
f.LocalPath().replace(input_api.os_path.sep, '/'))
errors = []
for f in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in f.ChangedContents():
if impl_pattern.match(line) or v8_wrapper_pattern.match(line):
errors.append(' %s:%d' % (f.LocalPath(), line_num))
results = []
if errors:
results.append(
output_api.PresubmitError(
'Do not include files from third_party/blink/renderer/platform/heap/impl'
' or third_party/blink/renderer/platform/heap/v8_wrapper. Use the '
'relevant counterparts from third_party/blink/renderer/platform/heap',
errors))
return results
def _CheckForVersionControlConflictsInFile(input_api, f):
pattern = input_api.re.compile('^(?:<<<<<<<|>>>>>>>) |^=======$')
errors = []
for line_num, line in f.ChangedContents():
if f.LocalPath().endswith(('.md', '.rst', '.txt')):
# First-level headers in markdown look a lot like version control
# conflict markers. http://daringfireball.net/projects/markdown/basics
continue
if pattern.match(line):
errors.append(' %s:%d %s' % (f.LocalPath(), line_num, line))
return errors
def CheckForVersionControlConflicts(input_api, output_api):
"""Usually this is not intentional and will cause a compile failure."""
errors = []
for f in input_api.AffectedFiles():
errors.extend(_CheckForVersionControlConflictsInFile(input_api, f))
results = []
if errors:
results.append(
output_api.PresubmitError(
'Version control conflict markers found, please resolve.',
errors))
return results
def CheckGoogleSupportAnswerUrlOnUpload(input_api, output_api):
pattern = input_api.re.compile('support\.google\.com\/chrome.*/answer')
errors = []
for f in input_api.AffectedFiles():
for line_num, line in f.ChangedContents():
if pattern.search(line):
errors.append(' %s:%d %s' % (f.LocalPath(), line_num, line))
results = []
if errors:
results.append(
output_api.PresubmitPromptWarning(
'Found Google support URL addressed by answer number. Please replace '
'with a p= identifier instead. See crbug.com/679462\n',
errors))
return results
def CheckHardcodedGoogleHostsInLowerLayers(input_api, output_api):
def FilterFile(affected_file):
"""Filter function for use with input_api.AffectedSourceFiles,
below. This filters out everything except non-test files from
top-level directories that generally speaking should not hard-code
service URLs (e.g. src/android_webview/, src/content/ and others).
"""
return input_api.FilterSourceFile(
affected_file,
files_to_check=[r'^(android_webview|base|content|net)/.*'],
files_to_skip=(_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS +
input_api.DEFAULT_FILES_TO_SKIP))
base_pattern = ('"[^"]*(google|googleapis|googlezip|googledrive|appspot)'
'\.(com|net)[^"]*"')
comment_pattern = input_api.re.compile('//.*%s' % base_pattern)
pattern = input_api.re.compile(base_pattern)
problems = [] # items are (filename, line_number, line)
for f in input_api.AffectedSourceFiles(FilterFile):
for line_num, line in f.ChangedContents():
if not comment_pattern.search(line) and pattern.search(line):
problems.append((f.LocalPath(), line_num, line))
if problems:
return [
output_api.PresubmitPromptOrNotify(
'Most layers below src/chrome/ should not hardcode service URLs.\n'
'Are you sure this is correct?', [
' %s:%d: %s' % (problem[0], problem[1], problem[2])
for problem in problems
])
]
else:
return []
def CheckChromeOsSyncedPrefRegistration(input_api, output_api):
"""Warns if Chrome OS C++ files register syncable prefs as browser prefs."""
def FileFilter(affected_file):
"""Includes directories known to be Chrome OS only."""
return input_api.FilterSourceFile(
affected_file,
files_to_check=(
'^ash/',
'^chromeos/', # Top-level src/chromeos.
'.*/chromeos/', # Any path component.
'^components/arc',
'^components/exo'),
files_to_skip=(input_api.DEFAULT_FILES_TO_SKIP))
prefs = []
priority_prefs = []
for f in input_api.AffectedFiles(file_filter=FileFilter):
for line_num, line in f.ChangedContents():
if input_api.re.search('PrefRegistrySyncable::SYNCABLE_PREF',
line):
prefs.append(' %s:%d:' % (f.LocalPath(), line_num))
prefs.append(' %s' % line)
if input_api.re.search(
'PrefRegistrySyncable::SYNCABLE_PRIORITY_PREF', line):
priority_prefs.append(' %s:%d' % (f.LocalPath(), line_num))
priority_prefs.append(' %s' % line)
results = []
if (prefs):
results.append(
output_api.PresubmitPromptWarning(
'Preferences were registered as SYNCABLE_PREF and will be controlled '
'by browser sync settings. If these prefs should be controlled by OS '
'sync settings use SYNCABLE_OS_PREF instead.\n' +
'\n'.join(prefs)))
if (priority_prefs):
results.append(
output_api.PresubmitPromptWarning(
'Preferences were registered as SYNCABLE_PRIORITY_PREF and will be '
'controlled by browser sync settings. If these prefs should be '
'controlled by OS sync settings use SYNCABLE_OS_PRIORITY_PREF '
'instead.\n' + '\n'.join(prefs)))
return results
def CheckNoAbbreviationInPngFileName(input_api, output_api):
"""Makes sure there are no abbreviations in the name of PNG files.
The native_client_sdk directory is excluded because it has auto-generated PNG
files for documentation.
"""
errors = []
files_to_check = [r'.*\.png$']
files_to_skip = [r'^native_client_sdk/',
r'^services/test/',
r'^third_party/blink/web_tests/',
]
file_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=files_to_check, files_to_skip=files_to_skip)
abbreviation = input_api.re.compile('.+_[a-z]\.png|.+_[a-z]_.*\.png')
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=file_filter):
file_name = input_api.os_path.split(f.LocalPath())[1]
if abbreviation.search(file_name):
errors.append(' %s' % f.LocalPath())
results = []
if errors:
results.append(
output_api.PresubmitError(
'The name of PNG files should not have abbreviations. \n'
'Use _hover.png, _center.png, instead of _h.png, _c.png.\n'
'Contact oshima@chromium.org if you have questions.', errors))
return results
def CheckNoProductIconsAddedToPublicRepo(input_api, output_api):
"""Heuristically identifies product icons based on their file name and reminds
contributors not to add them to the Chromium repository.
"""
errors = []
files_to_check = [r'.*google.*\.png$|.*google.*\.svg$|.*google.*\.icon$']
file_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=files_to_check)
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=file_filter):
errors.append(' %s' % f.LocalPath())
results = []
if errors:
# Give warnings instead of errors on presubmit --all and presubmit
# --files.
message_type = (output_api.PresubmitNotifyResult if input_api.no_diffs
else output_api.PresubmitError)
results.append(
message_type(
'Trademarked images should not be added to the public repo. '
'See crbug.com/944754', errors))
return results
def _ExtractAddRulesFromParsedDeps(parsed_deps):
"""Extract the rules that add dependencies from a parsed DEPS file.
Args:
parsed_deps: the locals dictionary from evaluating the DEPS file."""
add_rules = set()
add_rules.update([
rule[1:] for rule in parsed_deps.get('include_rules', [])
if rule.startswith('+') or rule.startswith('!')
])
for _, rules in parsed_deps.get('specific_include_rules', {}).items():
add_rules.update([
rule[1:] for rule in rules
if rule.startswith('+') or rule.startswith('!')
])
return add_rules
def _ParseDeps(contents):
"""Simple helper for parsing DEPS files."""
# Stubs for handling special syntax in the root DEPS file.
class _VarImpl:
def __init__(self, local_scope):
self._local_scope = local_scope
def Lookup(self, var_name):
"""Implements the Var syntax."""
try:
return self._local_scope['vars'][var_name]
except KeyError:
raise Exception('Var is not defined: %s' % var_name)
local_scope = {}
global_scope = {
'Var': _VarImpl(local_scope).Lookup,
'Str': str,
}
exec(contents, global_scope, local_scope)
return local_scope
def _CalculateAddedDeps(os_path, old_contents, new_contents):
"""Helper method for CheckAddedDepsHaveTargetApprovals. Returns
a set of DEPS entries that we should look up.
For a directory (rather than a specific filename) we fake a path to
a specific filename by adding /DEPS. This is chosen as a file that
will seldom or never be subject to per-file include_rules.
"""
# We ignore deps entries on auto-generated directories.
AUTO_GENERATED_DIRS = ['grit', 'jni']
old_deps = _ExtractAddRulesFromParsedDeps(_ParseDeps(old_contents))
new_deps = _ExtractAddRulesFromParsedDeps(_ParseDeps(new_contents))
added_deps = new_deps.difference(old_deps)
results = set()
for added_dep in added_deps:
if added_dep.split('/')[0] in AUTO_GENERATED_DIRS:
continue
# Assume that a rule that ends in .h is a rule for a specific file.
if added_dep.endswith('.h'):
results.add(added_dep)
else:
results.add(os_path.join(added_dep, 'DEPS'))
return results
def CheckAddedDepsHaveTargetApprovals(input_api, output_api):
"""When a dependency prefixed with + is added to a DEPS file, we
want to make sure that the change is reviewed by an OWNER of the
target file or directory, to avoid layering violations from being
introduced. This check verifies that this happens.
"""
# We rely on Gerrit's code-owners to check approvals.
# input_api.gerrit is always set for Chromium, but other projects
# might not use Gerrit.
if not input_api.gerrit or input_api.no_diffs:
return []
if 'PRESUBMIT_SKIP_NETWORK' in input_api.environ:
return []
try:
if (input_api.change.issue and
input_api.gerrit.IsOwnersOverrideApproved(
input_api.change.issue)):
# Skip OWNERS check when Owners-Override label is approved. This is
# intended for global owners, trusted bots, and on-call sheriffs.
# Review is still required for these changes.
return []
except Exception as e:
return [output_api.PresubmitPromptWarning(
'Failed to retrieve owner override status - %s' % str(e))]
virtual_depended_on_files = set()
# Consistently use / as path separator to simplify the writing of regex
# expressions.
file_filter = lambda f: not input_api.re.match(
r"^third_party/blink/.*",
f.LocalPath().replace(input_api.os_path.sep, '/'))
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=file_filter):
filename = input_api.os_path.basename(f.LocalPath())
if filename == 'DEPS':
virtual_depended_on_files.update(
_CalculateAddedDeps(input_api.os_path,
'\n'.join(f.OldContents()),
'\n'.join(f.NewContents())))
if not virtual_depended_on_files:
return []
if input_api.is_committing:
if input_api.tbr:
return [
output_api.PresubmitNotifyResult(
'--tbr was specified, skipping OWNERS check for DEPS additions'
)
]
# TODO(dcheng): Make this generate an error on dry runs if the reviewer
# is not added, to prevent review serialization.
if input_api.dry_run:
return [
output_api.PresubmitNotifyResult(
'This is a dry run, skipping OWNERS check for DEPS additions'
)
]
if not input_api.change.issue:
return [
output_api.PresubmitError(
"DEPS approval by OWNERS check failed: this change has "
"no change number, so we can't check it for approvals.")
]
output = output_api.PresubmitError
else:
output = output_api.PresubmitNotifyResult
owner_email, reviewers = (
input_api.canned_checks.GetCodereviewOwnerAndReviewers(
input_api, None, approval_needed=input_api.is_committing))
owner_email = owner_email or input_api.change.author_email
approval_status = input_api.owners_client.GetFilesApprovalStatus(
virtual_depended_on_files, reviewers.union([owner_email]), [])
missing_files = [
f for f in virtual_depended_on_files
if approval_status[f] != input_api.owners_client.APPROVED
]
# We strip the /DEPS part that was added by
# _FilesToCheckForIncomingDeps to fake a path to a file in a
# directory.
def StripDeps(path):
start_deps = path.rfind('/DEPS')
if start_deps != -1:
return path[:start_deps]
else:
return path
unapproved_dependencies = [
"'+%s'," % StripDeps(path) for path in missing_files
]
if unapproved_dependencies:
output_list = [
output(
'You need LGTM from owners of depends-on paths in DEPS that were '
'modified in this CL:\n %s' %
'\n '.join(sorted(unapproved_dependencies)))
]
suggested_owners = input_api.owners_client.SuggestOwners(
missing_files, exclude=[owner_email])
output_list.append(
output('Suggested missing target path OWNERS:\n %s' %
'\n '.join(suggested_owners or [])))
return output_list
return []
# TODO: add unit tests.
def CheckSpamLogging(input_api, output_api):
file_inclusion_pattern = [r'.+%s' % _IMPLEMENTATION_EXTENSIONS]
files_to_skip = (
_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS +
input_api.DEFAULT_FILES_TO_SKIP + (
r"^base/fuchsia/scoped_fx_logger\.cc$",
r"^base/logging\.h$",
r"^base/logging\.cc$",
r"^base/task/thread_pool/task_tracker\.cc$",
r"^chrome/app/chrome_main_delegate\.cc$",
r"^chrome/browser/chrome_browser_main\.cc$",
r"^chrome/browser/ui/startup/startup_browser_creator\.cc$",
r"^chrome/browser/browser_switcher/bho/.*",
r"^chrome/browser/diagnostics/diagnostics_writer\.cc$",
r"^chrome/chrome_cleaner/.*",
r"^chrome/chrome_elf/dll_hash/dll_hash_main\.cc$",
r"^chrome/installer/setup/.*",
r"^chromecast/",
r"^components/browser_watcher/dump_stability_report_main_win\.cc$",
r"^components/media_control/renderer/media_playback_options\.cc$",
r"^components/viz/service/display/"
r"overlay_strategy_underlay_cast\.cc$",
r"^components/zucchini/.*",
# TODO(peter): Remove exception. https://crbug.com/534537
r"^content/browser/notifications/"
r"notification_event_dispatcher_impl\.cc$",
r"^content/common/gpu/client/gl_helper_benchmark\.cc$",
r"^courgette/courgette_minimal_tool\.cc$",
r"^courgette/courgette_tool\.cc$",
r"^extensions/renderer/logging_native_handler\.cc$",
r"^fuchsia_web/common/init_logging\.cc$",
r"^fuchsia_web/runners/common/web_component\.cc$",
r"^fuchsia_web/shell/.*\.cc$",
r"^headless/app/headless_shell\.cc$",
r"^ipc/ipc_logging\.cc$",
r"^native_client_sdk/",
r"^remoting/base/logging\.h$",
r"^remoting/host/.*",
r"^sandbox/linux/.*",
r"^storage/browser/file_system/dump_file_system\.cc$",
r"^tools/",
r"^ui/base/resource/data_pack\.cc$",
r"^ui/aura/bench/bench_main\.cc$",
r"^ui/ozone/platform/cast/",
r"^ui/base/x/xwmstartupcheck/"
r"xwmstartupcheck\.cc$"))
source_file_filter = lambda x: input_api.FilterSourceFile(
x, files_to_check=file_inclusion_pattern, files_to_skip=files_to_skip)
log_info = set([])
printf = set([])
for f in input_api.AffectedSourceFiles(source_file_filter):
for _, line in f.ChangedContents():
if input_api.re.search(r"\bD?LOG\s*\(\s*INFO\s*\)", line):
log_info.add(f.LocalPath())
elif input_api.re.search(r"\bD?LOG_IF\s*\(\s*INFO\s*,", line):
log_info.add(f.LocalPath())
if input_api.re.search(r"\bprintf\(", line):
printf.add(f.LocalPath())
elif input_api.re.search(r"\bfprintf\((stdout|stderr)", line):
printf.add(f.LocalPath())
if log_info:
return [
output_api.PresubmitError(
'These files spam the console log with LOG(INFO):',
items=log_info)
]
if printf:
return [
output_api.PresubmitError(
'These files spam the console log with printf/fprintf:',
items=printf)
]
return []
def CheckForAnonymousVariables(input_api, output_api):
"""These types are all expected to hold locks while in scope and
so should never be anonymous (which causes them to be immediately
destroyed)."""
they_who_must_be_named = [
'base::AutoLock',
'base::AutoReset',
'base::AutoUnlock',
'SkAutoAlphaRestore',
'SkAutoBitmapShaderInstall',
'SkAutoBlitterChoose',
'SkAutoBounderCommit',
'SkAutoCallProc',
'SkAutoCanvasRestore',
'SkAutoCommentBlock',
'SkAutoDescriptor',
'SkAutoDisableDirectionCheck',
'SkAutoDisableOvalCheck',
'SkAutoFree',
'SkAutoGlyphCache',
'SkAutoHDC',
'SkAutoLockColors',
'SkAutoLockPixels',
'SkAutoMalloc',
'SkAutoMaskFreeImage',
'SkAutoMutexAcquire',
'SkAutoPathBoundsUpdate',
'SkAutoPDFRelease',
'SkAutoRasterClipValidate',
'SkAutoRef',
'SkAutoTime',
'SkAutoTrace',
'SkAutoUnref',
]
anonymous = r'(%s)\s*[({]' % '|'.join(they_who_must_be_named)
# bad: base::AutoLock(lock.get());
# not bad: base::AutoLock lock(lock.get());
bad_pattern = input_api.re.compile(anonymous)
# good: new base::AutoLock(lock.get())
good_pattern = input_api.re.compile(r'\bnew\s*' + anonymous)
errors = []
for f in input_api.AffectedFiles():
if not f.LocalPath().endswith(('.cc', '.h', '.inl', '.m', '.mm')):
continue
for linenum, line in f.ChangedContents():
if bad_pattern.search(line) and not good_pattern.search(line):
errors.append('%s:%d' % (f.LocalPath(), linenum))
if errors:
return [
output_api.PresubmitError(
'These lines create anonymous variables that need to be named:',
items=errors)
]
return []
def CheckUniquePtrOnUpload(input_api, output_api):
# Returns whether |template_str| is of the form <T, U...> for some types T
# and U. Assumes that |template_str| is already in the form <...>.
def HasMoreThanOneArg(template_str):
# Level of <...> nesting.
nesting = 0
for c in template_str:
if c == '<':
nesting += 1
elif c == '>':
nesting -= 1
elif c == ',' and nesting == 1:
return True
return False
file_inclusion_pattern = [r'.+%s' % _IMPLEMENTATION_EXTENSIONS]
sources = lambda affected_file: input_api.FilterSourceFile(
affected_file,
files_to_skip=(_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS + input_api.
DEFAULT_FILES_TO_SKIP),
files_to_check=file_inclusion_pattern)
# Pattern to capture a single "<...>" block of template arguments. It can
# handle linearly nested blocks, such as "<std::vector<std::set<T>>>", but
# cannot handle branching structures, such as "<pair<set<T>,set<U>>". The
# latter would likely require counting that < and > match, which is not
# expressible in regular languages. Should the need arise, one can introduce
# limited counting (matching up to a total number of nesting depth), which
# should cover all practical cases for already a low nesting limit.
template_arg_pattern = (
r'<[^>]*' # Opening block of <.
r'>([^<]*>)?') # Closing block of >.
# Prefix expressing that whatever follows is not already inside a <...>
# block.
not_inside_template_arg_pattern = r'(^|[^<,\s]\s*)'
null_construct_pattern = input_api.re.compile(
not_inside_template_arg_pattern + r'\bstd::unique_ptr' +
template_arg_pattern + r'\(\)')
# Same as template_arg_pattern, but excluding type arrays, e.g., <T[]>.
template_arg_no_array_pattern = (
r'<[^>]*[^]]' # Opening block of <.
r'>([^(<]*[^]]>)?') # Closing block of >.
# Prefix saying that what follows is the start of an expression.
start_of_expr_pattern = r'(=|\breturn|^)\s*'
# Suffix saying that what follows are call parentheses with a non-empty list
# of arguments.
nonempty_arg_list_pattern = r'\(([^)]|$)'
# Put the template argument into a capture group for deeper examination later.
return_construct_pattern = input_api.re.compile(
start_of_expr_pattern + r'std::unique_ptr' + '(?P<template_arg>' +
template_arg_no_array_pattern + ')' + nonempty_arg_list_pattern)
problems_constructor = []
problems_nullptr = []
for f in input_api.AffectedSourceFiles(sources):
for line_number, line in f.ChangedContents():
# Disallow:
# return std::unique_ptr<T>(foo);
# bar = std::unique_ptr<T>(foo);
# But allow:
# return std::unique_ptr<T[]>(foo);
# bar = std::unique_ptr<T[]>(foo);
# And also allow cases when the second template argument is present. Those
# cases cannot be handled by std::make_unique:
# return std::unique_ptr<T, U>(foo);
# bar = std::unique_ptr<T, U>(foo);
local_path = f.LocalPath()
return_construct_result = return_construct_pattern.search(line)
if return_construct_result and not HasMoreThanOneArg(
return_construct_result.group('template_arg')):
problems_constructor.append(
'%s:%d\n %s' % (local_path, line_number, line.strip()))
# Disallow:
# std::unique_ptr<T>()
if null_construct_pattern.search(line):
problems_nullptr.append(
'%s:%d\n %s' % (local_path, line_number, line.strip()))
errors = []
if problems_nullptr:
errors.append(
output_api.PresubmitPromptWarning(
'The following files use std::unique_ptr<T>(). Use nullptr instead.',
problems_nullptr))
if problems_constructor:
errors.append(
output_api.PresubmitError(
'The following files use explicit std::unique_ptr constructor. '
'Use std::make_unique<T>() instead, or use base::WrapUnique if '
'std::make_unique is not an option.', problems_constructor))
return errors
def CheckUserActionUpdate(input_api, output_api):
"""Checks if any new user action has been added."""
if any('actions.xml' == input_api.os_path.basename(f)
for f in input_api.LocalPaths()):
# If actions.xml is already included in the changelist, the PRESUBMIT
# for actions.xml will do a more complete presubmit check.
return []
file_inclusion_pattern = [r'.*\.(cc|mm)$']
files_to_skip = (_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS +
input_api.DEFAULT_FILES_TO_SKIP)
file_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=file_inclusion_pattern, files_to_skip=files_to_skip)
action_re = r'[^a-zA-Z]UserMetricsAction\("([^"]*)'
current_actions = None
for f in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in f.ChangedContents():
match = input_api.re.search(action_re, line)
if match:
# Loads contents in tools/metrics/actions/actions.xml to memory. It's
# loaded only once.
if not current_actions:
with open(
'tools/metrics/actions/actions.xml') as actions_f:
current_actions = actions_f.read()
# Search for the matched user action name in |current_actions|.
for action_name in match.groups():
action = 'name="{0}"'.format(action_name)
if action not in current_actions:
return [
output_api.PresubmitPromptWarning(
'File %s line %d: %s is missing in '
'tools/metrics/actions/actions.xml. Please run '
'tools/metrics/actions/extract_actions.py to update.'
% (f.LocalPath(), line_num, action_name))
]
return []
def _ImportJSONCommentEater(input_api):
import sys
sys.path = sys.path + [
input_api.os_path.join(input_api.PresubmitLocalPath(), 'tools',
'json_comment_eater')
]
import json_comment_eater
return json_comment_eater
def _GetJSONParseError(input_api, filename, eat_comments=True):
try:
contents = input_api.ReadFile(filename)
if eat_comments:
json_comment_eater = _ImportJSONCommentEater(input_api)
contents = json_comment_eater.Nom(contents)
input_api.json.loads(contents)
except ValueError as e:
return e
return None
def _GetIDLParseError(input_api, filename):
try:
contents = input_api.ReadFile(filename)
for i, char in enumerate(contents):
if not char.isascii():
return (
'Non-ascii character "%s" (ord %d) found at offset %d.' %
(char, ord(char), i))
idl_schema = input_api.os_path.join(input_api.PresubmitLocalPath(),
'tools', 'json_schema_compiler',
'idl_schema.py')
process = input_api.subprocess.Popen(
[input_api.python3_executable, idl_schema],
stdin=input_api.subprocess.PIPE,
stdout=input_api.subprocess.PIPE,
stderr=input_api.subprocess.PIPE,
universal_newlines=True)
(_, error) = process.communicate(input=contents)
return error or None
except ValueError as e:
return e
def CheckParseErrors(input_api, output_api):
"""Check that IDL and JSON files do not contain syntax errors."""
actions = {
'.idl': _GetIDLParseError,
'.json': _GetJSONParseError,
}
# Most JSON files are preprocessed and support comments, but these do not.
json_no_comments_patterns = [
r'^testing/',
]
# Only run IDL checker on files in these directories.
idl_included_patterns = [
r'^chrome/common/extensions/api/',
r'^extensions/common/api/',
]
def get_action(affected_file):
filename = affected_file.LocalPath()
return actions.get(input_api.os_path.splitext(filename)[1])
def FilterFile(affected_file):
action = get_action(affected_file)
if not action:
return False
path = affected_file.LocalPath()
if _MatchesFile(input_api,
_KNOWN_TEST_DATA_AND_INVALID_JSON_FILE_PATTERNS, path):
return False
if (action == _GetIDLParseError
and not _MatchesFile(input_api, idl_included_patterns, path)):
return False
return True
results = []
for affected_file in input_api.AffectedFiles(file_filter=FilterFile,
include_deletes=False):
action = get_action(affected_file)
kwargs = {}
if (action == _GetJSONParseError
and _MatchesFile(input_api, json_no_comments_patterns,
affected_file.LocalPath())):
kwargs['eat_comments'] = False
parse_error = action(input_api, affected_file.AbsoluteLocalPath(),
**kwargs)
if parse_error:
results.append(
output_api.PresubmitError(
'%s could not be parsed: %s' %
(affected_file.LocalPath(), parse_error)))
return results
def CheckJavaStyle(input_api, output_api):
"""Runs checkstyle on changed java files and returns errors if any exist."""
# Return early if no java files were modified.
if not any(
_IsJavaFile(input_api, f.LocalPath())
for f in input_api.AffectedFiles()):
return []
import sys
original_sys_path = sys.path
try:
sys.path = sys.path + [
input_api.os_path.join(input_api.PresubmitLocalPath(), 'tools',
'android', 'checkstyle')
]
import checkstyle
finally:
# Restore sys.path to what it was before.
sys.path = original_sys_path
return checkstyle.run_presubmit(
input_api,
output_api,
files_to_skip=_EXCLUDED_PATHS + input_api.DEFAULT_FILES_TO_SKIP)
def CheckPythonDevilInit(input_api, output_api):
"""Checks to make sure devil is initialized correctly in python scripts."""
script_common_initialize_pattern = input_api.re.compile(
r'script_common\.InitializeEnvironment\(')
devil_env_config_initialize = input_api.re.compile(
r'devil_env\.config\.Initialize\(')
errors = []
sources = lambda affected_file: input_api.FilterSourceFile(
affected_file,
files_to_skip=(_EXCLUDED_PATHS + input_api.DEFAULT_FILES_TO_SKIP + (
r'^build/android/devil_chromium\.py',
r'^third_party/.*',
)),
files_to_check=[r'.*\.py$'])
for f in input_api.AffectedSourceFiles(sources):
for line_num, line in f.ChangedContents():
if (script_common_initialize_pattern.search(line)
or devil_env_config_initialize.search(line)):
errors.append("%s:%d" % (f.LocalPath(), line_num))
results = []
if errors:
results.append(
output_api.PresubmitError(
'Devil initialization should always be done using '
'devil_chromium.Initialize() in the chromium project, to use better '
'defaults for dependencies (ex. up-to-date version of adb).',
errors))
return results
def _MatchesFile(input_api, patterns, path):
# Consistently use / as path separator to simplify the writing of regex
# expressions.
path = path.replace(input_api.os_path.sep, '/')
for pattern in patterns:
if input_api.re.search(pattern, path):
return True
return False
def _ChangeHasSecurityReviewer(input_api, owners_file):
"""Returns True iff the CL has a reviewer from SECURITY_OWNERS.
Args:
input_api: The presubmit input API.
owners_file: OWNERS file with required reviewers. Typically, this is
something like ipc/SECURITY_OWNERS.
Note: if the presubmit is running for commit rather than for upload, this
only returns True if a security reviewer has also approved the CL.
"""
# Owners-Override should bypass all additional OWNERS enforcement checks.
# A CR+1 vote will still be required to land this change.
if (input_api.change.issue and input_api.gerrit.IsOwnersOverrideApproved(
input_api.change.issue)):
return True
owner_email, reviewers = (
input_api.canned_checks.GetCodereviewOwnerAndReviewers(
input_api,
None,
approval_needed=input_api.is_committing and not input_api.dry_run))
security_owners = input_api.owners_client.ListOwners(owners_file)
return any(owner in reviewers for owner in security_owners)
@dataclass
class _SecurityProblemWithItems:
problem: str
items: Sequence[str]
@dataclass
class _MissingSecurityOwnersResult:
owners_file_problems: Sequence[_SecurityProblemWithItems]
has_security_sensitive_files: bool
missing_reviewer_problem: Optional[_SecurityProblemWithItems]
def _FindMissingSecurityOwners(input_api,
output_api,
file_patterns: Sequence[str],
excluded_patterns: Sequence[str],
required_owners_file: str,
custom_rule_function: Optional[Callable] = None
) -> _MissingSecurityOwnersResult:
"""Find OWNERS files missing per-file rules for security-sensitive files.
Args:
input_api: the PRESUBMIT input API object.
output_api: the PRESUBMIT output API object.
file_patterns: basename patterns that require a corresponding per-file
security restriction.
excluded_patterns: path patterns that should be exempted from
requiring a security restriction.
required_owners_file: path to the required OWNERS file, e.g.
ipc/SECURITY_OWNERS
cc_alias: If not None, email that will be CCed automatically if the
change contains security-sensitive files, as determined by
`file_patterns` and `excluded_patterns`.
custom_rule_function: If not None, will be called with `input_api` and
the current file under consideration. Returning True will add an
exact match per-file rule check for the current file.
"""
# `to_check` is a mapping of an OWNERS file path to Patterns.
#
# Patterns is a dictionary mapping glob patterns (suitable for use in
# per-file rules) to a PatternEntry.
#
# PatternEntry is a dictionary with two keys:
# - 'files': the files that are matched by this pattern
# - 'rules': the per-file rules needed for this pattern
#
# For example, if we expect OWNERS file to contain rules for *.mojom and
# *_struct_traits*.*, Patterns might look like this:
# {
# '*.mojom': {
# 'files': ...,
# 'rules': [
# 'per-file *.mojom=set noparent',
# 'per-file *.mojom=file://ipc/SECURITY_OWNERS',
# ],
# },
# '*_struct_traits*.*': {
# 'files': ...,
# 'rules': [
# 'per-file *_struct_traits*.*=set noparent',
# 'per-file *_struct_traits*.*=file://ipc/SECURITY_OWNERS',
# ],
# },
# }
to_check = {}
files_to_review = []
def AddPatternToCheck(file, pattern):
owners_file = input_api.os_path.join(
input_api.os_path.dirname(file.LocalPath()), 'OWNERS')
if owners_file not in to_check:
to_check[owners_file] = {}
if pattern not in to_check[owners_file]:
to_check[owners_file][pattern] = {
'files': [],
'rules': [
f'per-file {pattern}=set noparent',
f'per-file {pattern}=file://{required_owners_file}',
]
}
to_check[owners_file][pattern]['files'].append(file.LocalPath())
files_to_review.append(file.LocalPath())
# Only enforce security OWNERS rules for a directory if that directory has a
# file that matches `file_patterns`. For example, if a directory only
# contains *.mojom files and no *_messages*.h files, the check should only
# ensure that rules for *.mojom files are present.
for file in input_api.AffectedFiles(include_deletes=False):
file_basename = input_api.os_path.basename(file.LocalPath())
if custom_rule_function is not None and custom_rule_function(
input_api, file):
AddPatternToCheck(file, file_basename)
continue
if any(
input_api.fnmatch.fnmatch(file.LocalPath(), pattern)
for pattern in excluded_patterns):
continue
for pattern in file_patterns:
# Unlike `excluded_patterns`, `file_patterns` is checked only against the
# file's basename.
if input_api.fnmatch.fnmatch(file_basename, pattern):
AddPatternToCheck(file, pattern)
break
has_security_sensitive_files = bool(to_check)
# Check if any newly added lines in OWNERS files intersect with required
# per-file OWNERS lines. If so, ensure that a security reviewer is included.
# This is a hack, but is needed because the OWNERS check (by design) ignores
# new OWNERS entries; otherwise, a non-owner could add someone as a new
# OWNER and have that newly-added OWNER self-approve their own addition.
newly_covered_files = []
for file in input_api.AffectedFiles(include_deletes=False):
if not file.LocalPath() in to_check:
continue
for _, line in file.ChangedContents():
for _, entry in to_check[file.LocalPath()].items():
if line in entry['rules']:
newly_covered_files.extend(entry['files'])
missing_reviewer_problems = None
if newly_covered_files and not _ChangeHasSecurityReviewer(
input_api, required_owners_file):
missing_reviewer_problems = _SecurityProblemWithItems(
f'Review from an owner in {required_owners_file} is required for '
'the following newly-added files:',
[f'{file}' for file in sorted(set(newly_covered_files))])
# Go through the OWNERS files to check, filtering out rules that are already
# present in that OWNERS file.
for owners_file, patterns in to_check.items():
try:
lines = set(
input_api.ReadFile(
input_api.os_path.join(input_api.change.RepositoryRoot(),
owners_file)).splitlines())
for entry in patterns.values():
entry['rules'] = [
rule for rule in entry['rules'] if rule not in lines
]
except IOError:
# No OWNERS file, so all the rules are definitely missing.
continue
# All the remaining lines weren't found in OWNERS files, so emit an error.
owners_file_problems = []
for owners_file, patterns in to_check.items():
missing_lines = []
files = []
for _, entry in patterns.items():
files.extend(entry['files'])
missing_lines.extend(entry['rules'])
if missing_lines:
joined_missing_lines = '\n'.join(line for line in missing_lines)
owners_file_problems.append(
_SecurityProblemWithItems(
'Found missing OWNERS lines for security-sensitive files. '
f'Please add the following lines to {owners_file}:\n'
f'{joined_missing_lines}\n\nTo ensure security review for:',
files))
return _MissingSecurityOwnersResult(owners_file_problems,
has_security_sensitive_files,
missing_reviewer_problems)
def _CheckChangeForIpcSecurityOwners(input_api, output_api):
# Whether or not a file affects IPC is (mostly) determined by a simple list
# of filename patterns.
file_patterns = [
# Legacy IPC:
'*_messages.cc',
'*_messages*.h',
'*_param_traits*.*',
# Mojo IPC:
'*.mojom',
'*_mojom_traits*.*',
'*_type_converter*.*',
# Android native IPC:
'*.aidl',
]
excluded_patterns = [
# These third_party directories do not contain IPCs, but contain files
# matching the above patterns, which trigger false positives.
'third_party/crashpad/*',
'third_party/blink/renderer/platform/bindings/*',
'third_party/protobuf/benchmarks/python/*',
'third_party/win_build_output/*',
# Enum-only mojoms used for web metrics, so no security review needed.
'third_party/blink/public/mojom/use_counter/metrics/*',
# These files are just used to communicate between class loaders running
# in the same process.
'weblayer/browser/java/org/chromium/weblayer_private/interfaces/*',
'weblayer/browser/java/org/chromium/weblayer_private/test_interfaces/*',
]
def IsMojoServiceManifestFile(input_api, file):
manifest_pattern = input_api.re.compile('manifests?\.(cc|h)$')
test_manifest_pattern = input_api.re.compile('test_manifests?\.(cc|h)')
if not manifest_pattern.search(file.LocalPath()):
return False
if test_manifest_pattern.search(file.LocalPath()):
return False
# All actual service manifest files should contain at least one
# qualified reference to service_manager::Manifest.
return any('service_manager::Manifest' in line
for line in file.NewContents())
return _FindMissingSecurityOwners(
input_api,
output_api,
file_patterns,
excluded_patterns,
'ipc/SECURITY_OWNERS',
custom_rule_function=IsMojoServiceManifestFile)
def _CheckChangeForFuchsiaSecurityOwners(input_api, output_api):
file_patterns = [
# Component specifications.
'*.cml', # Component Framework v2.
'*.cmx', # Component Framework v1.
# Fuchsia IDL protocol specifications.
'*.fidl',
]
# Don't check for owners files for changes in these directories.
excluded_patterns = [
'third_party/crashpad/*',
]
return _FindMissingSecurityOwners(input_api, output_api, file_patterns,
excluded_patterns,
'build/fuchsia/SECURITY_OWNERS')
def CheckSecurityOwners(input_api, output_api):
"""Checks that various security-sensitive files have an IPC OWNERS rule."""
ipc_results = _CheckChangeForIpcSecurityOwners(input_api, output_api)
fuchsia_results = _CheckChangeForFuchsiaSecurityOwners(
input_api, output_api)
if ipc_results.has_security_sensitive_files:
output_api.AppendCC('ipc-security-reviews@chromium.org')
results = []
missing_reviewer_problems = []
if ipc_results.missing_reviewer_problem:
missing_reviewer_problems.append(ipc_results.missing_reviewer_problem)
if fuchsia_results.missing_reviewer_problem:
missing_reviewer_problems.append(
fuchsia_results.missing_reviewer_problem)
# Missing reviewers are an error unless there's no issue number
# associated with this branch; in that case, the presubmit is being run
# with --all or --files.
#
# Note that upload should never be an error; otherwise, it would be
# impossible to upload changes at all.
if input_api.is_committing and input_api.change.issue:
make_presubmit_message = output_api.PresubmitError
else:
make_presubmit_message = output_api.PresubmitNotifyResult
for problem in missing_reviewer_problems:
results.append(
make_presubmit_message(problem.problem, items=problem.items))
owners_file_problems = []
owners_file_problems.extend(ipc_results.owners_file_problems)
owners_file_problems.extend(fuchsia_results.owners_file_problems)
for problem in owners_file_problems:
# Missing per-file rules are always an error. While swarming and caching
# means that uploading a patchset with updated OWNERS files and sending
# it to the CQ again should not have a large incremental cost, it is
# still frustrating to discover the error only after the change has
# already been uploaded.
results.append(
output_api.PresubmitError(problem.problem, items=problem.items))
return results
def _GetFilesUsingSecurityCriticalFunctions(input_api):
"""Checks affected files for changes to security-critical calls. This
function checks the full change diff, to catch both additions/changes
and removals.
Returns a dict keyed by file name, and the value is a set of detected
functions.
"""
# Map of function pretty name (displayed in an error) to the pattern to
# match it with.
_PATTERNS_TO_CHECK = {
'content::GetServiceSandboxType<>()': 'GetServiceSandboxType\\<'
}
_PATTERNS_TO_CHECK = {
k: input_api.re.compile(v)
for k, v in _PATTERNS_TO_CHECK.items()
}
# We don't want to trigger on strings within this file.
def presubmit_file_filter(f):
return 'PRESUBMIT.py' != input_api.os_path.split(f.LocalPath())[1]
# Scan all affected files for changes touching _FUNCTIONS_TO_CHECK.
files_to_functions = {}
for f in input_api.AffectedFiles(file_filter=presubmit_file_filter):
diff = f.GenerateScmDiff()
for line in diff.split('\n'):
# Not using just RightHandSideLines() because removing a
# call to a security-critical function can be just as important
# as adding or changing the arguments.
if line.startswith('-') or (line.startswith('+')
and not line.startswith('++')):
for name, pattern in _PATTERNS_TO_CHECK.items():
if pattern.search(line):
path = f.LocalPath()
if not path in files_to_functions:
files_to_functions[path] = set()
files_to_functions[path].add(name)
return files_to_functions
def CheckSecurityChanges(input_api, output_api):
"""Checks that changes involving security-critical functions are reviewed
by the security team.
"""
files_to_functions = _GetFilesUsingSecurityCriticalFunctions(input_api)
if not len(files_to_functions):
return []
owners_file = 'ipc/SECURITY_OWNERS'
if _ChangeHasSecurityReviewer(input_api, owners_file):
return []
msg = 'The following files change calls to security-sensitive functions\n' \
'that need to be reviewed by {}.\n'.format(owners_file)
for path, names in files_to_functions.items():
msg += ' {}\n'.format(path)
for name in names:
msg += ' {}\n'.format(name)
msg += '\n'
if input_api.is_committing:
output = output_api.PresubmitError
else:
output = output_api.PresubmitNotifyResult
return [output(msg)]
def CheckSetNoParent(input_api, output_api):
"""Checks that set noparent is only used together with an OWNERS file in
//build/OWNERS.setnoparent (see also
//docs/code_reviews.md#owners-files-details)
"""
# Return early if no OWNERS files were modified.
if not any(f.LocalPath().endswith('OWNERS')
for f in input_api.AffectedFiles(include_deletes=False)):
return []
errors = []
allowed_owners_files_file = 'build/OWNERS.setnoparent'
allowed_owners_files = set()
with open(allowed_owners_files_file, 'r') as f:
for line in f:
line = line.strip()
if not line or line.startswith('#'):
continue
allowed_owners_files.add(line)
per_file_pattern = input_api.re.compile('per-file (.+)=(.+)')
for f in input_api.AffectedFiles(include_deletes=False):
if not f.LocalPath().endswith('OWNERS'):
continue
found_owners_files = set()
found_set_noparent_lines = dict()
# Parse the OWNERS file.
for lineno, line in enumerate(f.NewContents(), 1):
line = line.strip()
if line.startswith('set noparent'):
found_set_noparent_lines[''] = lineno
if line.startswith('file://'):
if line in allowed_owners_files:
found_owners_files.add('')
if line.startswith('per-file'):
match = per_file_pattern.match(line)
if match:
glob = match.group(1).strip()
directive = match.group(2).strip()
if directive == 'set noparent':
found_set_noparent_lines[glob] = lineno
if directive.startswith('file://'):
if directive in allowed_owners_files:
found_owners_files.add(glob)
# Check that every set noparent line has a corresponding file:// line
# listed in build/OWNERS.setnoparent. An exception is made for top level
# directories since src/OWNERS shouldn't review them.
linux_path = f.LocalPath().replace(input_api.os_path.sep, '/')
if (linux_path.count('/') != 1
and (not linux_path in _EXCLUDED_SET_NO_PARENT_PATHS)):
for set_noparent_line in found_set_noparent_lines:
if set_noparent_line in found_owners_files:
continue
errors.append(' %s:%d' %
(linux_path,
found_set_noparent_lines[set_noparent_line]))
results = []
if errors:
if input_api.is_committing:
output = output_api.PresubmitError
else:
output = output_api.PresubmitPromptWarning
results.append(
output(
'Found the following "set noparent" restrictions in OWNERS files that '
'do not include owners from build/OWNERS.setnoparent:',
long_text='\n\n'.join(errors)))
return results
def CheckUselessForwardDeclarations(input_api, output_api):
"""Checks that added or removed lines in non third party affected
header files do not lead to new useless class or struct forward
declaration.
"""
results = []
class_pattern = input_api.re.compile(r'^class\s+(\w+);$',
input_api.re.MULTILINE)
struct_pattern = input_api.re.compile(r'^struct\s+(\w+);$',
input_api.re.MULTILINE)
for f in input_api.AffectedFiles(include_deletes=False):
if (f.LocalPath().startswith('third_party')
and not f.LocalPath().startswith('third_party/blink')
and not f.LocalPath().startswith('third_party\\blink')):
continue
if not f.LocalPath().endswith('.h'):
continue
contents = input_api.ReadFile(f)
fwd_decls = input_api.re.findall(class_pattern, contents)
fwd_decls.extend(input_api.re.findall(struct_pattern, contents))
useless_fwd_decls = []
for decl in fwd_decls:
count = sum(1 for _ in input_api.re.finditer(
r'\b%s\b' % input_api.re.escape(decl), contents))
if count == 1:
useless_fwd_decls.append(decl)
if not useless_fwd_decls:
continue
for line in f.GenerateScmDiff().splitlines():
if (line.startswith('-') and not line.startswith('--')
or line.startswith('+') and not line.startswith('++')):
for decl in useless_fwd_decls:
if input_api.re.search(r'\b%s\b' % decl, line[1:]):
results.append(
output_api.PresubmitPromptWarning(
'%s: %s forward declaration is no longer needed'
% (f.LocalPath(), decl)))
useless_fwd_decls.remove(decl)
return results
def _CheckAndroidDebuggableBuild(input_api, output_api):
"""Checks that code uses BuildInfo.isDebugAndroid() instead of
Build.TYPE.equals('') or ''.equals(Build.TYPE) to check if
this is a debuggable build of Android.
"""
build_type_check_pattern = input_api.re.compile(
r'\bBuild\.TYPE\.equals\(|\.equals\(\s*\bBuild\.TYPE\)')
errors = []
sources = lambda affected_file: input_api.FilterSourceFile(
affected_file,
files_to_skip=(
_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS + input_api.
DEFAULT_FILES_TO_SKIP + (
r"^android_webview/support_library/boundary_interfaces/",
r"^chrome/android/webapk/.*",
r'^third_party/.*',
r"tools/android/customtabs_benchmark/.*",
r"webview/chromium/License.*",
)),
files_to_check=[r'.*\.java$'])
for f in input_api.AffectedSourceFiles(sources):
for line_num, line in f.ChangedContents():
if build_type_check_pattern.search(line):
errors.append("%s:%d" % (f.LocalPath(), line_num))
results = []
if errors:
results.append(
output_api.PresubmitPromptWarning(
'Build.TYPE.equals or .equals(Build.TYPE) usage is detected.'
' Please use BuildInfo.isDebugAndroid() instead.', errors))
return results
# TODO: add unit tests
def _CheckAndroidToastUsage(input_api, output_api):
"""Checks that code uses org.chromium.ui.widget.Toast instead of
android.widget.Toast (Chromium Toast doesn't force hardware
acceleration on low-end devices, saving memory).
"""
toast_import_pattern = input_api.re.compile(
r'^import android\.widget\.Toast;$')
errors = []
sources = lambda affected_file: input_api.FilterSourceFile(
affected_file,
files_to_skip=(_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS + input_api.
DEFAULT_FILES_TO_SKIP + (r'^chromecast/.*',
r'^remoting/.*')),
files_to_check=[r'.*\.java$'])
for f in input_api.AffectedSourceFiles(sources):
for line_num, line in f.ChangedContents():
if toast_import_pattern.search(line):
errors.append("%s:%d" % (f.LocalPath(), line_num))
results = []
if errors:
results.append(
output_api.PresubmitError(
'android.widget.Toast usage is detected. Android toasts use hardware'
' acceleration, and can be\ncostly on low-end devices. Please use'
' org.chromium.ui.widget.Toast instead.\n'
'Contact dskiba@chromium.org if you have any questions.',
errors))
return results
def _CheckAndroidCrLogUsage(input_api, output_api):
"""Checks that new logs using org.chromium.base.Log:
- Are using 'TAG' as variable name for the tags (warn)
- Are using a tag that is shorter than 20 characters (error)
"""
# Do not check format of logs in the given files
cr_log_check_excluded_paths = [
# //chrome/android/webapk cannot depend on //base
r"^chrome/android/webapk/.*",
# WebView license viewer code cannot depend on //base; used in stub APK.
r"^android_webview/glue/java/src/com/android/"
r"webview/chromium/License.*",
# The customtabs_benchmark is a small app that does not depend on Chromium
# java pieces.
r"tools/android/customtabs_benchmark/.*",
]
cr_log_import_pattern = input_api.re.compile(
r'^import org\.chromium\.base\.Log;$', input_api.re.MULTILINE)
class_in_base_pattern = input_api.re.compile(
r'^package org\.chromium\.base;$', input_api.re.MULTILINE)
has_some_log_import_pattern = input_api.re.compile(r'^import .*\.Log;$',
input_api.re.MULTILINE)
# Extract the tag from lines like `Log.d(TAG, "*");` or `Log.d("TAG", "*");`
log_call_pattern = input_api.re.compile(r'\bLog\.\w\((?P<tag>\"?\w+)')
log_decl_pattern = input_api.re.compile(
r'static final String TAG = "(?P<name>(.*))"')
rough_log_decl_pattern = input_api.re.compile(r'\bString TAG\s*=')
REF_MSG = ('See docs/android_logging.md for more info.')
sources = lambda x: input_api.FilterSourceFile(
x,
files_to_check=[r'.*\.java$'],
files_to_skip=cr_log_check_excluded_paths)
tag_decl_errors = []
tag_length_errors = []
tag_errors = []
tag_with_dot_errors = []
util_log_errors = []
for f in input_api.AffectedSourceFiles(sources):
file_content = input_api.ReadFile(f)
has_modified_logs = False
# Per line checks
if (cr_log_import_pattern.search(file_content)
or (class_in_base_pattern.search(file_content)
and not has_some_log_import_pattern.search(file_content))):
# Checks to run for files using cr log
for line_num, line in f.ChangedContents():
if rough_log_decl_pattern.search(line):
has_modified_logs = True
# Check if the new line is doing some logging
match = log_call_pattern.search(line)
if match:
has_modified_logs = True
# Make sure it uses "TAG"
if not match.group('tag') == 'TAG':
tag_errors.append("%s:%d" % (f.LocalPath(), line_num))
else:
# Report non cr Log function calls in changed lines
for line_num, line in f.ChangedContents():
if log_call_pattern.search(line):
util_log_errors.append("%s:%d" % (f.LocalPath(), line_num))
# Per file checks
if has_modified_logs:
# Make sure the tag is using the "cr" prefix and is not too long
match = log_decl_pattern.search(file_content)
tag_name = match.group('name') if match else None
if not tag_name:
tag_decl_errors.append(f.LocalPath())
elif len(tag_name) > 20:
tag_length_errors.append(f.LocalPath())
elif '.' in tag_name:
tag_with_dot_errors.append(f.LocalPath())
results = []
if tag_decl_errors:
results.append(
output_api.PresubmitPromptWarning(
'Please define your tags using the suggested format: .\n'
'"private static final String TAG = "<package tag>".\n'
'They will be prepended with "cr_" automatically.\n' + REF_MSG,
tag_decl_errors))
if tag_length_errors:
results.append(
output_api.PresubmitError(
'The tag length is restricted by the system to be at most '
'20 characters.\n' + REF_MSG, tag_length_errors))
if tag_errors:
results.append(
output_api.PresubmitPromptWarning(
'Please use a variable named "TAG" for your log tags.\n' +
REF_MSG, tag_errors))
if util_log_errors:
results.append(
output_api.PresubmitPromptWarning(
'Please use org.chromium.base.Log for new logs.\n' + REF_MSG,
util_log_errors))
if tag_with_dot_errors:
results.append(
output_api.PresubmitPromptWarning(
'Dot in log tags cause them to be elided in crash reports.\n' +
REF_MSG, tag_with_dot_errors))
return results
def _CheckAndroidTestJUnitFrameworkImport(input_api, output_api):
"""Checks that junit.framework.* is no longer used."""
deprecated_junit_framework_pattern = input_api.re.compile(
r'^import junit\.framework\..*;', input_api.re.MULTILINE)
sources = lambda x: input_api.FilterSourceFile(
x, files_to_check=[r'.*\.java$'], files_to_skip=None)
errors = []
for f in input_api.AffectedFiles(file_filter=sources):
for line_num, line in f.ChangedContents():
if deprecated_junit_framework_pattern.search(line):
errors.append("%s:%d" % (f.LocalPath(), line_num))
results = []
if errors:
results.append(
output_api.PresubmitError(
'APIs from junit.framework.* are deprecated, please use JUnit4 framework'
'(org.junit.*) from //third_party/junit. Contact yolandyan@chromium.org'
' if you have any question.', errors))
return results
def _CheckAndroidTestJUnitInheritance(input_api, output_api):
"""Checks that if new Java test classes have inheritance.
Either the new test class is JUnit3 test or it is a JUnit4 test class
with a base class, either case is undesirable.
"""
class_declaration_pattern = input_api.re.compile(r'^public class \w*Test ')
sources = lambda x: input_api.FilterSourceFile(
x, files_to_check=[r'.*Test\.java$'], files_to_skip=None)
errors = []
for f in input_api.AffectedFiles(file_filter=sources):
if not f.OldContents():
class_declaration_start_flag = False
for line_num, line in f.ChangedContents():
if class_declaration_pattern.search(line):
class_declaration_start_flag = True
if class_declaration_start_flag and ' extends ' in line:
errors.append('%s:%d' % (f.LocalPath(), line_num))
if '{' in line:
class_declaration_start_flag = False
results = []
if errors:
results.append(
output_api.PresubmitPromptWarning(
'The newly created files include Test classes that inherits from base'
' class. Please do not use inheritance in JUnit4 tests or add new'
' JUnit3 tests. Contact yolandyan@chromium.org if you have any'
' questions.', errors))
return results
def _CheckAndroidTestAnnotationUsage(input_api, output_api):
"""Checks that android.test.suitebuilder.annotation.* is no longer used."""
deprecated_annotation_import_pattern = input_api.re.compile(
r'^import android\.test\.suitebuilder\.annotation\..*;',
input_api.re.MULTILINE)
sources = lambda x: input_api.FilterSourceFile(
x, files_to_check=[r'.*\.java$'], files_to_skip=None)
errors = []
for f in input_api.AffectedFiles(file_filter=sources):
for line_num, line in f.ChangedContents():
if deprecated_annotation_import_pattern.search(line):
errors.append("%s:%d" % (f.LocalPath(), line_num))
results = []
if errors:
results.append(
output_api.PresubmitError(
'Annotations in android.test.suitebuilder.annotation have been'
' deprecated since API level 24. Please use android.support.test.filters'
' from //third_party/android_support_test_runner:runner_java instead.'
' Contact yolandyan@chromium.org if you have any questions.',
errors))
return results
def _CheckAndroidNewMdpiAssetLocation(input_api, output_api):
"""Checks if MDPI assets are placed in a correct directory."""
file_filter = lambda f: (f.LocalPath().endswith(
'.png') and ('/res/drawable/'.replace('/', input_api.os_path.sep) in f.
LocalPath() or '/res/drawable-ldrtl/'.replace(
'/', input_api.os_path.sep) in f.LocalPath()))
errors = []
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=file_filter):
errors.append(' %s' % f.LocalPath())
results = []
if errors:
results.append(
output_api.PresubmitError(
'MDPI assets should be placed in /res/drawable-mdpi/ or '
'/res/drawable-ldrtl-mdpi/\ninstead of /res/drawable/ and'
'/res/drawable-ldrtl/.\n'
'Contact newt@chromium.org if you have questions.', errors))
return results
def _CheckAndroidWebkitImports(input_api, output_api):
"""Checks that code uses org.chromium.base.Callback instead of
android.webview.ValueCallback except in the WebView glue layer
and WebLayer.
"""
valuecallback_import_pattern = input_api.re.compile(
r'^import android\.webkit\.ValueCallback;$')
errors = []
sources = lambda affected_file: input_api.FilterSourceFile(
affected_file,
files_to_skip=(_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS + input_api.
DEFAULT_FILES_TO_SKIP + (
r'^android_webview/glue/.*',
r'^weblayer/.*',
)),
files_to_check=[r'.*\.java$'])
for f in input_api.AffectedSourceFiles(sources):
for line_num, line in f.ChangedContents():
if valuecallback_import_pattern.search(line):
errors.append("%s:%d" % (f.LocalPath(), line_num))
results = []
if errors:
results.append(
output_api.PresubmitError(
'android.webkit.ValueCallback usage is detected outside of the glue'
' layer. To stay compatible with the support library, android.webkit.*'
' classes should only be used inside the glue layer and'
' org.chromium.base.Callback should be used instead.', errors))
return results
def _CheckAndroidXmlStyle(input_api, output_api, is_check_on_upload):
"""Checks Android XML styles """
# Return early if no relevant files were modified.
if not any(
_IsXmlOrGrdFile(input_api, f.LocalPath())
for f in input_api.AffectedFiles(include_deletes=False)):
return []
import sys
original_sys_path = sys.path
try:
sys.path = sys.path + [
input_api.os_path.join(input_api.PresubmitLocalPath(), 'tools',
'android', 'checkxmlstyle')
]
import checkxmlstyle
finally:
# Restore sys.path to what it was before.
sys.path = original_sys_path
if is_check_on_upload:
return checkxmlstyle.CheckStyleOnUpload(input_api, output_api)
else:
return checkxmlstyle.CheckStyleOnCommit(input_api, output_api)
def _CheckAndroidInfoBarDeprecation(input_api, output_api):
"""Checks Android Infobar Deprecation """
import sys
original_sys_path = sys.path
try:
sys.path = sys.path + [
input_api.os_path.join(input_api.PresubmitLocalPath(), 'tools',
'android', 'infobar_deprecation')
]
import infobar_deprecation
finally:
# Restore sys.path to what it was before.
sys.path = original_sys_path
return infobar_deprecation.CheckDeprecationOnUpload(input_api, output_api)
class _PydepsCheckerResult:
def __init__(self, cmd, pydeps_path, process, old_contents):
self._cmd = cmd
self._pydeps_path = pydeps_path
self._process = process
self._old_contents = old_contents
def GetError(self):
"""Returns an error message, or None."""
import difflib
if self._process.wait() != 0:
# STDERR should already be printed.
return 'Command failed: ' + self._cmd
new_contents = self._process.stdout.read().splitlines()[2:]
if self._old_contents != new_contents:
diff = '\n'.join(
difflib.context_diff(self._old_contents, new_contents))
return ('File is stale: {}\n'
'Diff (apply to fix):\n'
'{}\n'
'To regenerate, run:\n\n'
' {}').format(self._pydeps_path, diff, self._cmd)
return None
class PydepsChecker:
def __init__(self, input_api, pydeps_files):
self._file_cache = {}
self._input_api = input_api
self._pydeps_files = pydeps_files
def _LoadFile(self, path):
"""Returns the list of paths within a .pydeps file relative to //."""
if path not in self._file_cache:
with open(path, encoding='utf-8') as f:
self._file_cache[path] = f.read()
return self._file_cache[path]
def _ComputeNormalizedPydepsEntries(self, pydeps_path):
"""Returns an iterable of paths within the .pydep, relativized to //."""
pydeps_data = self._LoadFile(pydeps_path)
uses_gn_paths = '--gn-paths' in pydeps_data
entries = (l for l in pydeps_data.splitlines()
if not l.startswith('#'))
if uses_gn_paths:
# Paths look like: //foo/bar/baz
return (e[2:] for e in entries)
else:
# Paths look like: path/relative/to/file.pydeps
os_path = self._input_api.os_path
pydeps_dir = os_path.dirname(pydeps_path)
return (os_path.normpath(os_path.join(pydeps_dir, e))
for e in entries)
def _CreateFilesToPydepsMap(self):
"""Returns a map of local_path -> list_of_pydeps."""
ret = {}
for pydep_local_path in self._pydeps_files:
for path in self._ComputeNormalizedPydepsEntries(pydep_local_path):
ret.setdefault(path, []).append(pydep_local_path)
return ret
def ComputeAffectedPydeps(self):
"""Returns an iterable of .pydeps files that might need regenerating."""
affected_pydeps = set()
file_to_pydeps_map = None
for f in self._input_api.AffectedFiles(include_deletes=True):
local_path = f.LocalPath()
# Changes to DEPS can lead to .pydeps changes if any .py files are in
# subrepositories. We can't figure out which files change, so re-check
# all files.
# Changes to print_python_deps.py affect all .pydeps.
if local_path in ('DEPS', 'PRESUBMIT.py'
) or local_path.endswith('print_python_deps.py'):
return self._pydeps_files
elif local_path.endswith('.pydeps'):
if local_path in self._pydeps_files:
affected_pydeps.add(local_path)
elif local_path.endswith('.py'):
if file_to_pydeps_map is None:
file_to_pydeps_map = self._CreateFilesToPydepsMap()
affected_pydeps.update(file_to_pydeps_map.get(local_path, ()))
return affected_pydeps
def DetermineIfStaleAsync(self, pydeps_path):
"""Runs print_python_deps.py to see if the files is stale."""
import os
old_pydeps_data = self._LoadFile(pydeps_path).splitlines()
if old_pydeps_data:
cmd = old_pydeps_data[1][1:].strip()
if '--output' not in cmd:
cmd += ' --output ' + pydeps_path
old_contents = old_pydeps_data[2:]
else:
# A default cmd that should work in most cases (as long as pydeps filename
# matches the script name) so that PRESUBMIT.py does not crash if pydeps
# file is empty/new.
cmd = 'build/print_python_deps.py {} --root={} --output={}'.format(
pydeps_path[:-4], os.path.dirname(pydeps_path), pydeps_path)
old_contents = []
env = dict(os.environ)
env['PYTHONDONTWRITEBYTECODE'] = '1'
process = self._input_api.subprocess.Popen(
cmd + ' --output ""',
shell=True,
env=env,
stdout=self._input_api.subprocess.PIPE,
encoding='utf-8')
return _PydepsCheckerResult(cmd, pydeps_path, process, old_contents)
def _ParseGclientArgs():
args = {}
with open('build/config/gclient_args.gni', 'r') as f:
for line in f:
line = line.strip()
if not line or line.startswith('#'):
continue
attribute, value = line.split('=')
args[attribute.strip()] = value.strip()
return args
def CheckPydepsNeedsUpdating(input_api, output_api, checker_for_tests=None):
"""Checks if a .pydeps file needs to be regenerated."""
# This check is for Python dependency lists (.pydeps files), and involves
# paths not only in the PRESUBMIT.py, but also in the .pydeps files. It
# doesn't work on Windows and Mac, so skip it on other platforms.
if not input_api.platform.startswith('linux'):
return []
results = []
# First, check for new / deleted .pydeps.
for f in input_api.AffectedFiles(include_deletes=True):
# Check whether we are running the presubmit check for a file in src.
# f.LocalPath is relative to repo (src, or internal repo).
# os_path.exists is relative to src repo.
# Therefore if os_path.exists is true, it means f.LocalPath is relative
# to src and we can conclude that the pydeps is in src.
if f.LocalPath().endswith('.pydeps'):
if input_api.os_path.exists(f.LocalPath()):
if f.Action() == 'D' and f.LocalPath() in _ALL_PYDEPS_FILES:
results.append(
output_api.PresubmitError(
'Please update _ALL_PYDEPS_FILES within //PRESUBMIT.py to '
'remove %s' % f.LocalPath()))
elif f.Action() != 'D' and f.LocalPath(
) not in _ALL_PYDEPS_FILES:
results.append(
output_api.PresubmitError(
'Please update _ALL_PYDEPS_FILES within //PRESUBMIT.py to '
'include %s' % f.LocalPath()))
if results:
return results
is_android = _ParseGclientArgs().get('checkout_android', 'false') == 'true'
checker = checker_for_tests or PydepsChecker(input_api, _ALL_PYDEPS_FILES)
affected_pydeps = set(checker.ComputeAffectedPydeps())
affected_android_pydeps = affected_pydeps.intersection(
set(_ANDROID_SPECIFIC_PYDEPS_FILES))
if affected_android_pydeps and not is_android:
results.append(
output_api.PresubmitPromptOrNotify(
'You have changed python files that may affect pydeps for android\n'
'specific scripts. However, the relevant presubmit check cannot be\n'
'run because you are not using an Android checkout. To validate that\n'
'the .pydeps are correct, re-run presubmit in an Android checkout, or\n'
'use the android-internal-presubmit optional trybot.\n'
'Possibly stale pydeps files:\n{}'.format(
'\n'.join(affected_android_pydeps))))
all_pydeps = _ALL_PYDEPS_FILES if is_android else _GENERIC_PYDEPS_FILES
pydeps_to_check = affected_pydeps.intersection(all_pydeps)
# Process these concurrently, as each one takes 1-2 seconds.
pydep_results = [checker.DetermineIfStaleAsync(p) for p in pydeps_to_check]
for result in pydep_results:
error_msg = result.GetError()
if error_msg:
results.append(output_api.PresubmitError(error_msg))
return results
def CheckSingletonInHeaders(input_api, output_api):
"""Checks to make sure no header files have |Singleton<|."""
def FileFilter(affected_file):
# It's ok for base/memory/singleton.h to have |Singleton<|.
files_to_skip = (_EXCLUDED_PATHS + input_api.DEFAULT_FILES_TO_SKIP +
(r"^base/memory/singleton\.h$",
r"^net/quic/platform/impl/quic_singleton_impl\.h$"))
return input_api.FilterSourceFile(affected_file,
files_to_skip=files_to_skip)
pattern = input_api.re.compile(r'(?<!class\sbase::)Singleton\s*<')
files = []
for f in input_api.AffectedSourceFiles(FileFilter):
if (f.LocalPath().endswith('.h') or f.LocalPath().endswith('.hxx')
or f.LocalPath().endswith('.hpp')
or f.LocalPath().endswith('.inl')):
contents = input_api.ReadFile(f)
for line in contents.splitlines(False):
if (not line.lstrip().startswith('//')
and # Strip C++ comment.
pattern.search(line)):
files.append(f)
break
if files:
return [
output_api.PresubmitError(
'Found base::Singleton<T> in the following header files.\n' +
'Please move them to an appropriate source file so that the ' +
'template gets instantiated in a single compilation unit.',
files)
]
return []
_DEPRECATED_CSS = [
# Values
( "-webkit-box", "flex" ),
( "-webkit-inline-box", "inline-flex" ),
( "-webkit-flex", "flex" ),
( "-webkit-inline-flex", "inline-flex" ),
( "-webkit-min-content", "min-content" ),
( "-webkit-max-content", "max-content" ),
# Properties
( "-webkit-background-clip", "background-clip" ),
( "-webkit-background-origin", "background-origin" ),
( "-webkit-background-size", "background-size" ),
( "-webkit-box-shadow", "box-shadow" ),
( "-webkit-user-select", "user-select" ),
# Functions
( "-webkit-gradient", "gradient" ),
( "-webkit-repeating-gradient", "repeating-gradient" ),
( "-webkit-linear-gradient", "linear-gradient" ),
( "-webkit-repeating-linear-gradient", "repeating-linear-gradient" ),
( "-webkit-radial-gradient", "radial-gradient" ),
( "-webkit-repeating-radial-gradient", "repeating-radial-gradient" ),
]
# TODO: add unit tests
def CheckNoDeprecatedCss(input_api, output_api):
""" Make sure that we don't use deprecated CSS
properties, functions or values. Our external
documentation and iOS CSS for dom distiller
(reader mode) are ignored by the hooks as it
needs to be consumed by WebKit. """
results = []
file_inclusion_pattern = [r".+\.css$"]
files_to_skip = (_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS +
input_api.DEFAULT_FILES_TO_SKIP +
(r"^chrome/common/extensions/docs", r"^chrome/docs",
r"^native_client_sdk"))
file_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=file_inclusion_pattern, files_to_skip=files_to_skip)
for fpath in input_api.AffectedFiles(file_filter=file_filter):
for line_num, line in fpath.ChangedContents():
for (deprecated_value, value) in _DEPRECATED_CSS:
if deprecated_value in line:
results.append(
output_api.PresubmitError(
"%s:%d: Use of deprecated CSS %s, use %s instead" %
(fpath.LocalPath(), line_num, deprecated_value,
value)))
return results
def CheckForRelativeIncludes(input_api, output_api):
bad_files = {}
for f in input_api.AffectedFiles(include_deletes=False):
if (f.LocalPath().startswith('third_party')
and not f.LocalPath().startswith('third_party/blink')
and not f.LocalPath().startswith('third_party\\blink')):
continue
if not _IsCPlusPlusFile(input_api, f.LocalPath()):
continue
relative_includes = [
line for _, line in f.ChangedContents()
if "#include" in line and "../" in line
]
if not relative_includes:
continue
bad_files[f.LocalPath()] = relative_includes
if not bad_files:
return []
error_descriptions = []
for file_path, bad_lines in bad_files.items():
error_description = file_path
for line in bad_lines:
error_description += '\n ' + line
error_descriptions.append(error_description)
results = []
results.append(
output_api.PresubmitError(
'You added one or more relative #include paths (including "../").\n'
'These shouldn\'t be used because they can be used to include headers\n'
'from code that\'s not correctly specified as a dependency in the\n'
'relevant BUILD.gn file(s).', error_descriptions))
return results
def CheckForCcIncludes(input_api, output_api):
"""Check that nobody tries to include a cc file. It's a relatively
common error which results in duplicate symbols in object
files. This may not always break the build until someone later gets
very confusing linking errors."""
results = []
for f in input_api.AffectedFiles(include_deletes=False):
# We let third_party code do whatever it wants
if (f.LocalPath().startswith('third_party')
and not f.LocalPath().startswith('third_party/blink')
and not f.LocalPath().startswith('third_party\\blink')):
continue
if not _IsCPlusPlusFile(input_api, f.LocalPath()):
continue
for _, line in f.ChangedContents():
if line.startswith('#include "'):
included_file = line.split('"')[1]
if _IsCPlusPlusFile(input_api, included_file):
# The most common naming for external files with C++ code,
# apart from standard headers, is to call them foo.inc, but
# Chromium sometimes uses foo-inc.cc so allow that as well.
if not included_file.endswith(('.h', '-inc.cc')):
results.append(
output_api.PresubmitError(
'Only header files or .inc files should be included in other\n'
'C++ files. Compiling the contents of a cc file more than once\n'
'will cause duplicate information in the build which may later\n'
'result in strange link_errors.\n' +
f.LocalPath() + ':\n ' + line))
return results
def _CheckWatchlistDefinitionsEntrySyntax(key, value, ast):
if not isinstance(key, ast.Str):
return 'Key at line %d must be a string literal' % key.lineno
if not isinstance(value, ast.Dict):
return 'Value at line %d must be a dict' % value.lineno
if len(value.keys) != 1:
return 'Dict at line %d must have single entry' % value.lineno
if not isinstance(value.keys[0], ast.Str) or value.keys[0].s != 'filepath':
return (
'Entry at line %d must have a string literal \'filepath\' as key' %
value.lineno)
return None
def _CheckWatchlistsEntrySyntax(key, value, ast, email_regex):
if not isinstance(key, ast.Str):
return 'Key at line %d must be a string literal' % key.lineno
if not isinstance(value, ast.List):
return 'Value at line %d must be a list' % value.lineno
for element in value.elts:
if not isinstance(element, ast.Str):
return 'Watchlist elements on line %d is not a string' % key.lineno
if not email_regex.match(element.s):
return ('Watchlist element on line %d doesn\'t look like a valid '
+ 'email: %s') % (key.lineno, element.s)
return None
def _CheckWATCHLISTSEntries(wd_dict, w_dict, input_api):
mismatch_template = (
'Mismatch between WATCHLIST_DEFINITIONS entry (%s) and WATCHLISTS '
'entry (%s)')
email_regex = input_api.re.compile(
r"^[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]+$")
ast = input_api.ast
i = 0
last_key = ''
while True:
if i >= len(wd_dict.keys):
if i >= len(w_dict.keys):
return None
return mismatch_template % ('missing',
'line %d' % w_dict.keys[i].lineno)
elif i >= len(w_dict.keys):
return (mismatch_template %
('line %d' % wd_dict.keys[i].lineno, 'missing'))
wd_key = wd_dict.keys[i]
w_key = w_dict.keys[i]
result = _CheckWatchlistDefinitionsEntrySyntax(wd_key,
wd_dict.values[i], ast)
if result is not None:
return 'Bad entry in WATCHLIST_DEFINITIONS dict: %s' % result
result = _CheckWatchlistsEntrySyntax(w_key, w_dict.values[i], ast,
email_regex)
if result is not None:
return 'Bad entry in WATCHLISTS dict: %s' % result
if wd_key.s != w_key.s:
return mismatch_template % ('%s at line %d' %
(wd_key.s, wd_key.lineno),
'%s at line %d' %
(w_key.s, w_key.lineno))
if wd_key.s < last_key:
return (
'WATCHLISTS dict is not sorted lexicographically at line %d and %d'
% (wd_key.lineno, w_key.lineno))
last_key = wd_key.s
i = i + 1
def _CheckWATCHLISTSSyntax(expression, input_api):
ast = input_api.ast
if not isinstance(expression, ast.Expression):
return 'WATCHLISTS file must contain a valid expression'
dictionary = expression.body
if not isinstance(dictionary, ast.Dict) or len(dictionary.keys) != 2:
return 'WATCHLISTS file must have single dict with exactly two entries'
first_key = dictionary.keys[0]
first_value = dictionary.values[0]
second_key = dictionary.keys[1]
second_value = dictionary.values[1]
if (not isinstance(first_key, ast.Str)
or first_key.s != 'WATCHLIST_DEFINITIONS'
or not isinstance(first_value, ast.Dict)):
return ('The first entry of the dict in WATCHLISTS file must be '
'WATCHLIST_DEFINITIONS dict')
if (not isinstance(second_key, ast.Str) or second_key.s != 'WATCHLISTS'
or not isinstance(second_value, ast.Dict)):
return ('The second entry of the dict in WATCHLISTS file must be '
'WATCHLISTS dict')
return _CheckWATCHLISTSEntries(first_value, second_value, input_api)
def CheckWATCHLISTS(input_api, output_api):
for f in input_api.AffectedFiles(include_deletes=False):
if f.LocalPath() == 'WATCHLISTS':
contents = input_api.ReadFile(f, 'r')
try:
# First, make sure that it can be evaluated.
input_api.ast.literal_eval(contents)
# Get an AST tree for it and scan the tree for detailed style checking.
expression = input_api.ast.parse(contents,
filename='WATCHLISTS',
mode='eval')
except ValueError as e:
return [
output_api.PresubmitError('Cannot parse WATCHLISTS file',
long_text=repr(e))
]
except SyntaxError as e:
return [
output_api.PresubmitError('Cannot parse WATCHLISTS file',
long_text=repr(e))
]
except TypeError as e:
return [
output_api.PresubmitError('Cannot parse WATCHLISTS file',
long_text=repr(e))
]
result = _CheckWATCHLISTSSyntax(expression, input_api)
if result is not None:
return [output_api.PresubmitError(result)]
break
return []
def CheckGnRebasePath(input_api, output_api):
"""Checks that target_gen_dir is not used wtih "//" in rebase_path().
Developers should use root_build_dir instead of "//" when using target_gen_dir because
Chromium is sometimes built outside of the source tree.
"""
def gn_files(f):
return input_api.FilterSourceFile(f, files_to_check=(r'.+\.gn', ))
rebase_path_regex = input_api.re.compile(r'rebase_path\(("\$target_gen_dir"|target_gen_dir), ("/"|"//")\)')
problems = []
for f in input_api.AffectedSourceFiles(gn_files):
for line_num, line in f.ChangedContents():
if rebase_path_regex.search(line):
problems.append(
'Absolute path in rebase_path() in %s:%d' %
(f.LocalPath(), line_num))
if problems:
return [
output_api.PresubmitPromptWarning(
'Using an absolute path in rebase_path()',
items=sorted(problems),
long_text=(
'rebase_path() should use root_build_dir instead of "/" ',
'since builds can be initiated from outside of the source ',
'root.'))
]
return []
def CheckGnGlobForward(input_api, output_api):
"""Checks that forward_variables_from(invoker, "*") follows best practices.
As documented at //build/docs/writing_gn_templates.md
"""
def gn_files(f):
return input_api.FilterSourceFile(f, files_to_check=(r'.+\.gni', ))
problems = []
for f in input_api.AffectedSourceFiles(gn_files):
for line_num, line in f.ChangedContents():
if 'forward_variables_from(invoker, "*")' in line:
problems.append(
'Bare forward_variables_from(invoker, "*") in %s:%d' %
(f.LocalPath(), line_num))
if problems:
return [
output_api.PresubmitPromptWarning(
'forward_variables_from("*") without exclusions',
items=sorted(problems),
long_text=(
'The variables "visibility" and "test_only" should be '
'explicitly listed in forward_variables_from(). For more '
'details, see:\n'
'https://chromium.googlesource.com/chromium/src/+/HEAD/'
'build/docs/writing_gn_templates.md'
'#Using-forward_variables_from'))
]
return []
def CheckNewHeaderWithoutGnChangeOnUpload(input_api, output_api):
"""Checks that newly added header files have corresponding GN changes.
Note that this is only a heuristic. To be precise, run script:
build/check_gn_headers.py.
"""
def headers(f):
return input_api.FilterSourceFile(
f, files_to_check=(r'.+%s' % _HEADER_EXTENSIONS, ))
new_headers = []
for f in input_api.AffectedSourceFiles(headers):
if f.Action() != 'A':
continue
new_headers.append(f.LocalPath())
def gn_files(f):
return input_api.FilterSourceFile(f, files_to_check=(r'.+\.gn', ))
all_gn_changed_contents = ''
for f in input_api.AffectedSourceFiles(gn_files):
for _, line in f.ChangedContents():
all_gn_changed_contents += line
problems = []
for header in new_headers:
basename = input_api.os_path.basename(header)
if basename not in all_gn_changed_contents:
problems.append(header)
if problems:
return [
output_api.PresubmitPromptWarning(
'Missing GN changes for new header files',
items=sorted(problems),
long_text=
'Please double check whether newly added header files need '
'corresponding changes in gn or gni files.\nThis checking is only a '
'heuristic. Run build/check_gn_headers.py to be precise.\n'
'Read https://crbug.com/661774 for more info.')
]
return []
def CheckCorrectProductNameInMessages(input_api, output_api):
"""Check that Chromium-branded strings don't include "Chrome" or vice versa.
This assumes we won't intentionally reference one product from the other
product.
"""
all_problems = []
test_cases = [{
"filename_postfix": "google_chrome_strings.grd",
"correct_name": "Chrome",
"incorrect_name": "Chromium",
}, {
"filename_postfix": "chromium_strings.grd",
"correct_name": "Chromium",
"incorrect_name": "Chrome",
}]
for test_case in test_cases:
problems = []
filename_filter = lambda x: x.LocalPath().endswith(test_case[
"filename_postfix"])
# Check each new line. Can yield false positives in multiline comments, but
# easier than trying to parse the XML because messages can have nested
# children, and associating message elements with affected lines is hard.
for f in input_api.AffectedSourceFiles(filename_filter):
for line_num, line in f.ChangedContents():
if "<message" in line or "<!--" in line or "-->" in line:
continue
if test_case["incorrect_name"] in line:
problems.append("Incorrect product name in %s:%d" %
(f.LocalPath(), line_num))
if problems:
message = (
"Strings in %s-branded string files should reference \"%s\", not \"%s\""
% (test_case["correct_name"], test_case["correct_name"],
test_case["incorrect_name"]))
all_problems.append(
output_api.PresubmitPromptWarning(message, items=problems))
return all_problems
def CheckForTooLargeFiles(input_api, output_api):
"""Avoid large files, especially binary files, in the repository since
git doesn't scale well for those. They will be in everyone's repo
clones forever, forever making Chromium slower to clone and work
with."""
# Uploading files to cloud storage is not trivial so we don't want
# to set the limit too low, but the upper limit for "normal" large
# files seems to be 1-2 MB, with a handful around 5-8 MB, so
# anything over 20 MB is exceptional.
TOO_LARGE_FILE_SIZE_LIMIT = 20 * 1024 * 1024
# Special exemption for a file that is slightly over the limit.
SPECIAL_FILE_SIZE_LIMIT = 25 * 1024 * 1024
SPECIAL_FILE_NAME = 'transport_security_state_static.json'
too_large_files = []
for f in input_api.AffectedFiles():
# Check both added and modified files (but not deleted files).
if f.Action() in ('A', 'M'):
size = input_api.os_path.getsize(f.AbsoluteLocalPath())
limit = (SPECIAL_FILE_SIZE_LIMIT if
f.AbsoluteLocalPath().endswith(SPECIAL_FILE_NAME) else
TOO_LARGE_FILE_SIZE_LIMIT)
if size > limit:
too_large_files.append("%s: %d bytes" % (f.LocalPath(), size))
if too_large_files:
message = (
'Do not commit large files to git since git scales badly for those.\n'
+
'Instead put the large files in cloud storage and use DEPS to\n' +
'fetch them.\n' + '\n'.join(too_large_files))
return [
output_api.PresubmitError('Too large files found in commit',
long_text=message + '\n')
]
else:
return []
def CheckFuzzTargetsOnUpload(input_api, output_api):
"""Checks specific for fuzz target sources."""
EXPORTED_SYMBOLS = [
'LLVMFuzzerInitialize',
'LLVMFuzzerCustomMutator',
'LLVMFuzzerCustomCrossOver',
'LLVMFuzzerMutate',
]
REQUIRED_HEADER = '#include "testing/libfuzzer/libfuzzer_exports.h"'
def FilterFile(affected_file):
"""Ignore libFuzzer source code."""
files_to_check = r'.*fuzz.*\.(h|hpp|hcc|cc|cpp|cxx)$'
files_to_skip = r"^third_party/libFuzzer"
return input_api.FilterSourceFile(affected_file,
files_to_check=[files_to_check],
files_to_skip=[files_to_skip])
files_with_missing_header = []
for f in input_api.AffectedSourceFiles(FilterFile):
contents = input_api.ReadFile(f, 'r')
if REQUIRED_HEADER in contents:
continue
if any(symbol in contents for symbol in EXPORTED_SYMBOLS):
files_with_missing_header.append(f.LocalPath())
if not files_with_missing_header:
return []
long_text = (
'If you define any of the libFuzzer optional functions (%s), it is '
'recommended to add \'%s\' directive. Otherwise, the fuzz target may '
'work incorrectly on Mac (crbug.com/687076).\nNote that '
'LLVMFuzzerInitialize should not be used, unless your fuzz target needs '
'to access command line arguments passed to the fuzzer. Instead, prefer '
'static initialization and shared resources as documented in '
'https://chromium.googlesource.com/chromium/src/+/main/testing/'
'libfuzzer/efficient_fuzzing.md#simplifying-initialization_cleanup.\n'
% (', '.join(EXPORTED_SYMBOLS), REQUIRED_HEADER))
return [
output_api.PresubmitPromptWarning(message="Missing '%s' in:" %
REQUIRED_HEADER,
items=files_with_missing_header,
long_text=long_text)
]
def _CheckNewImagesWarning(input_api, output_api):
"""
Warns authors who add images into the repo to make sure their images are
optimized before committing.
"""
images_added = False
image_paths = []
errors = []
filter_lambda = lambda x: input_api.FilterSourceFile(
x,
files_to_skip=(('(?i).*test', r'.*\/junit\/') + input_api.
DEFAULT_FILES_TO_SKIP),
files_to_check=[r'.*\/(drawable|mipmap)'])
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=filter_lambda):
local_path = f.LocalPath().lower()
if any(
local_path.endswith(extension)
for extension in _IMAGE_EXTENSIONS):
images_added = True
image_paths.append(f)
if images_added:
errors.append(
output_api.PresubmitPromptWarning(
'It looks like you are trying to commit some images. If these are '
'non-test-only images, please make sure to read and apply the tips in '
'https://chromium.googlesource.com/chromium/src/+/HEAD/docs/speed/'
'binary_size/optimization_advice.md#optimizing-images\nThis check is '
'FYI only and will not block your CL on the CQ.', image_paths))
return errors
def ChecksAndroidSpecificOnUpload(input_api, output_api):
"""Groups upload checks that target android code."""
results = []
results.extend(_CheckAndroidCrLogUsage(input_api, output_api))
results.extend(_CheckAndroidDebuggableBuild(input_api, output_api))
results.extend(_CheckAndroidNewMdpiAssetLocation(input_api, output_api))
results.extend(_CheckAndroidToastUsage(input_api, output_api))
results.extend(_CheckAndroidTestJUnitInheritance(input_api, output_api))
results.extend(_CheckAndroidTestJUnitFrameworkImport(
input_api, output_api))
results.extend(_CheckAndroidTestAnnotationUsage(input_api, output_api))
results.extend(_CheckAndroidWebkitImports(input_api, output_api))
results.extend(_CheckAndroidXmlStyle(input_api, output_api, True))
results.extend(_CheckNewImagesWarning(input_api, output_api))
results.extend(_CheckAndroidNoBannedImports(input_api, output_api))
results.extend(_CheckAndroidInfoBarDeprecation(input_api, output_api))
return results
def ChecksAndroidSpecificOnCommit(input_api, output_api):
"""Groups commit checks that target android code."""
results = []
results.extend(_CheckAndroidXmlStyle(input_api, output_api, False))
return results
# TODO(chrishall): could we additionally match on any path owned by
# ui/accessibility/OWNERS ?
_ACCESSIBILITY_PATHS = (
r"^chrome/browser.*/accessibility/",
r"^chrome/browser/extensions/api/automation.*/",
r"^chrome/renderer/extensions/accessibility_.*",
r"^chrome/tests/data/accessibility/",
r"^components/services/screen_ai/",
r"^content/browser/accessibility/",
r"^content/renderer/accessibility/",
r"^content/tests/data/accessibility/",
r"^extensions/renderer/api/automation/",
r"^services/accessibility/",
r"^ui/accessibility/",
r"^ui/views/accessibility/",
)
def CheckAccessibilityRelnotesField(input_api, output_api):
"""Checks that commits to accessibility code contain an AX-Relnotes field in
their commit message."""
def FileFilter(affected_file):
paths = _ACCESSIBILITY_PATHS
return input_api.FilterSourceFile(affected_file, files_to_check=paths)
# Only consider changes affecting accessibility paths.
if not any(input_api.AffectedFiles(file_filter=FileFilter)):
return []
# AX-Relnotes can appear in either the description or the footer.
# When searching the description, require 'AX-Relnotes:' to appear at the
# beginning of a line.
ax_regex = input_api.re.compile('ax-relnotes[:=]')
description_has_relnotes = any(
ax_regex.match(line)
for line in input_api.change.DescriptionText().lower().splitlines())
footer_relnotes = input_api.change.GitFootersFromDescription().get(
'AX-Relnotes', [])
if description_has_relnotes or footer_relnotes:
return []
# TODO(chrishall): link to Relnotes documentation in message.
message = (
"Missing 'AX-Relnotes:' field required for accessibility changes"
"\n please add 'AX-Relnotes: [release notes].' to describe any "
"user-facing changes"
"\n otherwise add 'AX-Relnotes: n/a.' if this change has no "
"user-facing effects"
"\n if this is confusing or annoying then please contact members "
"of ui/accessibility/OWNERS.")
return [output_api.PresubmitNotifyResult(message)]
_ACCESSIBILITY_EVENTS_TEST_PATH = (
r"^content/test/data/accessibility/event/.*\.html",
)
_ACCESSIBILITY_TREE_TEST_PATH = (
r"^content/test/data/accessibility/accname/.*\.html",
r"^content/test/data/accessibility/aria/.*\.html",
r"^content/test/data/accessibility/css/.*\.html",
r"^content/test/data/accessibility/html/.*\.html",
)
_ACCESSIBILITY_ANDROID_EVENTS_TEST_PATH = (
r"^.*/WebContentsAccessibilityEventsTest\.java",
)
_ACCESSIBILITY_ANDROID_TREE_TEST_PATH = (
r"^.*/WebContentsAccessibilityTreeTest\.java",
)
def CheckAccessibilityEventsTestsAreIncludedForAndroid(input_api, output_api):
"""Checks that commits that include a newly added, renamed/moved, or deleted
test in the DumpAccessibilityEventsTest suite also includes a corresponding
change to the Android test."""
def FilePathFilter(affected_file):
paths = _ACCESSIBILITY_EVENTS_TEST_PATH
return input_api.FilterSourceFile(affected_file, files_to_check=paths)
def AndroidFilePathFilter(affected_file):
paths = _ACCESSIBILITY_ANDROID_EVENTS_TEST_PATH
return input_api.FilterSourceFile(affected_file, files_to_check=paths)
# Only consider changes in the events test data path with html type.
if not any(
input_api.AffectedFiles(include_deletes=True,
file_filter=FilePathFilter)):
return []
# If the commit contains any change to the Android test file, ignore.
if any(
input_api.AffectedFiles(include_deletes=True,
file_filter=AndroidFilePathFilter)):
return []
# Only consider changes that are adding/renaming or deleting a file
message = []
for f in input_api.AffectedFiles(include_deletes=True,
file_filter=FilePathFilter):
if f.Action() == 'A' or f.Action() == 'D':
message = (
"It appears that you are adding, renaming or deleting"
"\na dump_accessibility_events* test, but have not included"
"\na corresponding change for Android."
"\nPlease include (or remove) the test from:"
"\n content/public/android/javatests/src/org/chromium/"
"content/browser/accessibility/"
"WebContentsAccessibilityEventsTest.java"
"\nIf this message is confusing or annoying, please contact"
"\nmembers of ui/accessibility/OWNERS.")
# If no message was set, return empty.
if not len(message):
return []
return [output_api.PresubmitPromptWarning(message)]
def CheckAccessibilityTreeTestsAreIncludedForAndroid(input_api, output_api):
"""Checks that commits that include a newly added, renamed/moved, or deleted
test in the DumpAccessibilityTreeTest suite also includes a corresponding
change to the Android test."""
def FilePathFilter(affected_file):
paths = _ACCESSIBILITY_TREE_TEST_PATH
return input_api.FilterSourceFile(affected_file, files_to_check=paths)
def AndroidFilePathFilter(affected_file):
paths = _ACCESSIBILITY_ANDROID_TREE_TEST_PATH
return input_api.FilterSourceFile(affected_file, files_to_check=paths)
# Only consider changes in the various tree test data paths with html type.
if not any(
input_api.AffectedFiles(include_deletes=True,
file_filter=FilePathFilter)):
return []
# If the commit contains any change to the Android test file, ignore.
if any(
input_api.AffectedFiles(include_deletes=True,
file_filter=AndroidFilePathFilter)):
return []
# Only consider changes that are adding/renaming or deleting a file
message = []
for f in input_api.AffectedFiles(include_deletes=True,
file_filter=FilePathFilter):
if f.Action() == 'A' or f.Action() == 'D':
message = (
"It appears that you are adding, renaming or deleting"
"\na dump_accessibility_tree* test, but have not included"
"\na corresponding change for Android."
"\nPlease include (or remove) the test from:"
"\n content/public/android/javatests/src/org/chromium/"
"content/browser/accessibility/"
"WebContentsAccessibilityTreeTest.java"
"\nIf this message is confusing or annoying, please contact"
"\nmembers of ui/accessibility/OWNERS.")
# If no message was set, return empty.
if not len(message):
return []
return [output_api.PresubmitPromptWarning(message)]
def CheckEsLintConfigChanges(input_api, output_api):
"""Suggest using "git cl presubmit --files" when .eslintrc.js files are
modified. This is important because enabling an error in .eslintrc.js can
trigger errors in any .js or .ts files in its directory, leading to hidden
presubmit errors."""
results = []
eslint_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=[r'.*\.eslintrc\.js$'])
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=eslint_filter):
local_dir = input_api.os_path.dirname(f.LocalPath())
# Use / characters so that the commands printed work on any OS.
local_dir = local_dir.replace(input_api.os_path.sep, '/')
if local_dir:
local_dir += '/'
results.append(
output_api.PresubmitNotifyResult(
'%(file)s modified. Consider running \'git cl presubmit --files '
'"%(dir)s*.js;%(dir)s*.ts"\' in order to check and fix the affected '
'files before landing this change.' %
{ 'file' : f.LocalPath(), 'dir' : local_dir}))
return results
# string pattern, sequence of strings to show when pattern matches,
# error flag. True if match is a presubmit error, otherwise it's a warning.
_NON_INCLUSIVE_TERMS = (
(
# Note that \b pattern in python re is pretty particular. In this
# regexp, 'class WhiteList ...' will match, but 'class FooWhiteList
# ...' will not. This may require some tweaking to catch these cases
# without triggering a lot of false positives. Leaving it naive and
# less matchy for now.
r'/\b(?i)((black|white)list|master|slave)\b', # nocheck
(
'Please don\'t use blacklist, whitelist, ' # nocheck
'or slave in your', # nocheck
'code and make every effort to use other terms. Using "// nocheck"',
'"# nocheck" or "<!-- nocheck -->"',
'at the end of the offending line will bypass this PRESUBMIT error',
'but avoid using this whenever possible. Reach out to',
'community@chromium.org if you have questions'),
True),)
def ChecksCommon(input_api, output_api):
"""Checks common to both upload and commit."""
results = []
results.extend(
input_api.canned_checks.PanProjectChecks(
input_api, output_api, excluded_paths=_EXCLUDED_PATHS))
author = input_api.change.author_email
if author and author not in _KNOWN_ROBOTS:
results.extend(
input_api.canned_checks.CheckAuthorizedAuthor(
input_api, output_api))
results.extend(
input_api.canned_checks.CheckChangeHasNoTabs(
input_api,
output_api,
source_file_filter=lambda x: x.LocalPath().endswith('.grd')))
results.extend(
input_api.RunTests(
input_api.canned_checks.CheckVPythonSpec(input_api, output_api)))
dirmd = 'dirmd.bat' if input_api.is_windows else 'dirmd'
dirmd_bin = input_api.os_path.join(input_api.PresubmitLocalPath(),
'third_party', 'depot_tools', dirmd)
results.extend(
input_api.RunTests(
input_api.canned_checks.CheckDirMetadataFormat(
input_api, output_api, dirmd_bin)))
results.extend(
input_api.canned_checks.CheckOwnersDirMetadataExclusive(
input_api, output_api))
results.extend(
input_api.canned_checks.CheckNoNewMetadataInOwners(
input_api, output_api))
results.extend(
input_api.canned_checks.CheckInclusiveLanguage(
input_api,
output_api,
excluded_directories_relative_path=[
'infra', 'inclusive_language_presubmit_exempt_dirs.txt'
],
non_inclusive_terms=_NON_INCLUSIVE_TERMS))
presubmit_py_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=[r'.*PRESUBMIT\.py$'])
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=presubmit_py_filter):
full_path = input_api.os_path.dirname(f.AbsoluteLocalPath())
test_file = input_api.os_path.join(full_path, 'PRESUBMIT_test.py')
# The PRESUBMIT.py file (and the directory containing it) might have
# been affected by being moved or removed, so only try to run the tests
# if they still exist.
if not input_api.os_path.exists(test_file):
continue
use_python3 = False
with open(f.LocalPath()) as fp:
use_python3 = any(
line.startswith('USE_PYTHON3 = True')
for line in fp.readlines())
results.extend(
input_api.canned_checks.RunUnitTestsInDirectory(
input_api,
output_api,
full_path,
files_to_check=[r'^PRESUBMIT_test\.py$'],
run_on_python2=not use_python3,
run_on_python3=use_python3,
skip_shebang_check=True))
return results
def CheckPatchFiles(input_api, output_api):
problems = [
f.LocalPath() for f in input_api.AffectedFiles()
if f.LocalPath().endswith(('.orig', '.rej'))
]
# Cargo.toml.orig files are part of third-party crates downloaded from
# crates.io and should be included.
problems = [f for f in problems if not f.endswith('Cargo.toml.orig')]
if problems:
return [
output_api.PresubmitError("Don't commit .rej and .orig files.",
problems)
]
else:
return []
def CheckBuildConfigMacrosWithoutInclude(input_api, output_api):
# Excludes OS_CHROMEOS, which is not defined in build_config.h.
macro_re = input_api.re.compile(
r'^\s*#(el)?if.*\bdefined\(((COMPILER_|ARCH_CPU_|WCHAR_T_IS_)[^)]*)')
include_re = input_api.re.compile(r'^#include\s+"build/build_config.h"',
input_api.re.MULTILINE)
extension_re = input_api.re.compile(r'\.[a-z]+$')
errors = []
config_h_file = input_api.os_path.join('build', 'build_config.h')
for f in input_api.AffectedFiles(include_deletes=False):
# The build-config macros are allowed to be used in build_config.h
# without including itself.
if f.LocalPath() == config_h_file:
continue
if not f.LocalPath().endswith(
('.h', '.c', '.cc', '.cpp', '.m', '.mm')):
continue
found_line_number = None
found_macro = None
all_lines = input_api.ReadFile(f, 'r').splitlines()
for line_num, line in enumerate(all_lines):
match = macro_re.search(line)
if match:
found_line_number = line_num
found_macro = match.group(2)
break
if not found_line_number:
continue
found_include_line = -1
for line_num, line in enumerate(all_lines):
if include_re.search(line):
found_include_line = line_num
break
if found_include_line >= 0 and found_include_line < found_line_number:
continue
if not f.LocalPath().endswith('.h'):
primary_header_path = extension_re.sub('.h', f.AbsoluteLocalPath())
try:
content = input_api.ReadFile(primary_header_path, 'r')
if include_re.search(content):
continue
except IOError:
pass
errors.append('%s:%d %s macro is used without first including build/'
'build_config.h.' %
(f.LocalPath(), found_line_number, found_macro))
if errors:
return [output_api.PresubmitPromptWarning('\n'.join(errors))]
return []
def CheckForSuperfluousStlIncludesInHeaders(input_api, output_api):
stl_include_re = input_api.re.compile(r'^#include\s+<('
r'algorithm|'
r'array|'
r'limits|'
r'list|'
r'map|'
r'memory|'
r'queue|'
r'set|'
r'string|'
r'unordered_map|'
r'unordered_set|'
r'utility|'
r'vector)>')
std_namespace_re = input_api.re.compile(r'std::')
errors = []
for f in input_api.AffectedFiles():
if not _IsCPlusPlusHeaderFile(input_api, f.LocalPath()):
continue
uses_std_namespace = False
has_stl_include = False
for line in f.NewContents():
if has_stl_include and uses_std_namespace:
break
if not has_stl_include and stl_include_re.search(line):
has_stl_include = True
continue
if not uses_std_namespace and (std_namespace_re.search(line)
or 'no-std-usage-because-pch-file' in line):
uses_std_namespace = True
continue
if has_stl_include and not uses_std_namespace:
errors.append(
'%s: Includes STL header(s) but does not reference std::' %
f.LocalPath())
if errors:
return [output_api.PresubmitPromptWarning('\n'.join(errors))]
return []
def _CheckForDeprecatedOSMacrosInFile(input_api, f):
"""Check for sensible looking, totally invalid OS macros."""
preprocessor_statement = input_api.re.compile(r'^\s*#')
os_macro = input_api.re.compile(r'defined\(OS_([^)]+)\)')
results = []
for lnum, line in f.ChangedContents():
if preprocessor_statement.search(line):
for match in os_macro.finditer(line):
results.append(
' %s:%d: %s' %
(f.LocalPath(), lnum, 'defined(OS_' + match.group(1) +
') -> BUILDFLAG(IS_' + match.group(1) + ')'))
return results
def CheckForDeprecatedOSMacros(input_api, output_api):
"""Check all affected files for invalid OS macros."""
bad_macros = []
# The OS_ macros are allowed to be used in build/build_config.h.
config_h_file = input_api.os_path.join('build', 'build_config.h')
for f in input_api.AffectedSourceFiles(None):
if not f.LocalPath().endswith(('.py', '.js', '.html', '.css', '.md')) \
and f.LocalPath() != config_h_file:
bad_macros.extend(_CheckForDeprecatedOSMacrosInFile(input_api, f))
if not bad_macros:
return []
return [
output_api.PresubmitError(
'OS macros have been deprecated. Please use BUILDFLAGs instead (still '
'defined in build_config.h):', bad_macros)
]
def _CheckForInvalidIfDefinedMacrosInFile(input_api, f):
"""Check all affected files for invalid "if defined" macros."""
ALWAYS_DEFINED_MACROS = (
"TARGET_CPU_PPC",
"TARGET_CPU_PPC64",
"TARGET_CPU_68K",
"TARGET_CPU_X86",
"TARGET_CPU_ARM",
"TARGET_CPU_MIPS",
"TARGET_CPU_SPARC",
"TARGET_CPU_ALPHA",
"TARGET_IPHONE_SIMULATOR",
"TARGET_OS_EMBEDDED",
"TARGET_OS_IPHONE",
"TARGET_OS_MAC",
"TARGET_OS_UNIX",
"TARGET_OS_WIN32",
)
ifdef_macro = input_api.re.compile(
r'^\s*#.*(?:ifdef\s|defined\()([^\s\)]+)')
results = []
for lnum, line in f.ChangedContents():
for match in ifdef_macro.finditer(line):
if match.group(1) in ALWAYS_DEFINED_MACROS:
always_defined = ' %s is always defined. ' % match.group(1)
did_you_mean = 'Did you mean \'#if %s\'?' % match.group(1)
results.append(
' %s:%d %s\n\t%s' %
(f.LocalPath(), lnum, always_defined, did_you_mean))
return results
def CheckForInvalidIfDefinedMacros(input_api, output_api):
"""Check all affected files for invalid "if defined" macros."""
bad_macros = []
skipped_paths = ['third_party/sqlite/', 'third_party/abseil-cpp/']
for f in input_api.AffectedFiles():
if any([f.LocalPath().startswith(path) for path in skipped_paths]):
continue
if f.LocalPath().endswith(('.h', '.c', '.cc', '.m', '.mm')):
bad_macros.extend(
_CheckForInvalidIfDefinedMacrosInFile(input_api, f))
if not bad_macros:
return []
return [
output_api.PresubmitError(
'Found ifdef check on always-defined macro[s]. Please fix your code\n'
'or check the list of ALWAYS_DEFINED_MACROS in src/PRESUBMIT.py.',
bad_macros)
]
def CheckForIPCRules(input_api, output_api):
"""Check for same IPC rules described in
http://www.chromium.org/Home/chromium-security/education/security-tips-for-ipc
"""
base_pattern = r'IPC_ENUM_TRAITS\('
inclusion_pattern = input_api.re.compile(r'(%s)' % base_pattern)
comment_pattern = input_api.re.compile(r'//.*(%s)' % base_pattern)
problems = []
for f in input_api.AffectedSourceFiles(None):
local_path = f.LocalPath()
if not local_path.endswith('.h'):
continue
for line_number, line in f.ChangedContents():
if inclusion_pattern.search(
line) and not comment_pattern.search(line):
problems.append('%s:%d\n %s' %
(local_path, line_number, line.strip()))
if problems:
return [
output_api.PresubmitPromptWarning(_IPC_ENUM_TRAITS_DEPRECATED,
problems)
]
else:
return []
def CheckForLongPathnames(input_api, output_api):
"""Check to make sure no files being submitted have long paths.
This causes issues on Windows.
"""
problems = []
for f in input_api.AffectedTestableFiles():
local_path = f.LocalPath()
# Windows has a path limit of 260 characters. Limit path length to 200 so
# that we have some extra for the prefix on dev machines and the bots.
if len(local_path) > 200:
problems.append(local_path)
if problems:
return [output_api.PresubmitError(_LONG_PATH_ERROR, problems)]
else:
return []
def CheckForIncludeGuards(input_api, output_api):
"""Check that header files have proper guards against multiple inclusion.
If a file should not have such guards (and it probably should) then it
should include the string "no-include-guard-because-multiply-included" or
"no-include-guard-because-pch-file".
"""
def is_chromium_header_file(f):
# We only check header files under the control of the Chromium
# project. That is, those outside third_party apart from
# third_party/blink.
# We also exclude *_message_generator.h headers as they use
# include guards in a special, non-typical way.
file_with_path = input_api.os_path.normpath(f.LocalPath())
return (file_with_path.endswith('.h')
and not file_with_path.endswith('_message_generator.h')
and not file_with_path.endswith('com_imported_mstscax.h')
and (not file_with_path.startswith('third_party')
or file_with_path.startswith(
input_api.os_path.join('third_party', 'blink'))))
def replace_special_with_underscore(string):
return input_api.re.sub(r'[+\\/.-]', '_', string)
errors = []
for f in input_api.AffectedSourceFiles(is_chromium_header_file):
guard_name = None
guard_line_number = None
seen_guard_end = False
file_with_path = input_api.os_path.normpath(f.LocalPath())
base_file_name = input_api.os_path.splitext(
input_api.os_path.basename(file_with_path))[0]
upper_base_file_name = base_file_name.upper()
expected_guard = replace_special_with_underscore(
file_with_path.upper() + '_')
# For "path/elem/file_name.h" we should really only accept
# PATH_ELEM_FILE_NAME_H_ per coding style. Unfortunately there
# are too many (1000+) files with slight deviations from the
# coding style. The most important part is that the include guard
# is there, and that it's unique, not the name so this check is
# forgiving for existing files.
#
# As code becomes more uniform, this could be made stricter.
guard_name_pattern_list = [
# Anything with the right suffix (maybe with an extra _).
r'\w+_H__?',
# To cover include guards with old Blink style.
r'\w+_h',
# Anything including the uppercase name of the file.
r'\w*' + input_api.re.escape(
replace_special_with_underscore(upper_base_file_name)) +
r'\w*',
]
guard_name_pattern = '|'.join(guard_name_pattern_list)
guard_pattern = input_api.re.compile(r'#ifndef\s+(' +
guard_name_pattern + ')')
for line_number, line in enumerate(f.NewContents()):
if ('no-include-guard-because-multiply-included' in line
or 'no-include-guard-because-pch-file' in line):
guard_name = 'DUMMY' # To not trigger check outside the loop.
break
if guard_name is None:
match = guard_pattern.match(line)
if match:
guard_name = match.group(1)
guard_line_number = line_number
# We allow existing files to use include guards whose names
# don't match the chromium style guide, but new files should
# get it right.
if guard_name != expected_guard:
if f.Action() == 'A': # If file was just 'A'dded
errors.append(
output_api.PresubmitPromptWarning(
'Header using the wrong include guard name %s'
% guard_name, [
'%s:%d' %
(f.LocalPath(), line_number + 1)
], 'Expected: %r\nFound: %r' %
(expected_guard, guard_name)))
else:
# The line after #ifndef should have a #define of the same name.
if line_number == guard_line_number + 1:
expected_line = '#define %s' % guard_name
if line != expected_line:
errors.append(
output_api.PresubmitPromptWarning(
'Missing "%s" for include guard' %
expected_line,
['%s:%d' % (f.LocalPath(), line_number + 1)],
'Expected: %r\nGot: %r' %
(expected_line, line)))
if not seen_guard_end and line == '#endif // %s' % guard_name:
seen_guard_end = True
elif seen_guard_end:
if line.strip() != '':
errors.append(
output_api.PresubmitPromptWarning(
'Include guard %s not covering the whole file'
% (guard_name), [f.LocalPath()]))
break # Nothing else to check and enough to warn once.
if guard_name is None:
errors.append(
output_api.PresubmitPromptWarning(
'Missing include guard in %s\n'
'Recommended name: %s\n'
'This check can be disabled by having the string\n'
'"no-include-guard-because-multiply-included" or\n'
'"no-include-guard-because-pch-file" in the header.'
% (f.LocalPath(), expected_guard)))
return errors
def CheckForWindowsLineEndings(input_api, output_api):
"""Check source code and known ascii text files for Windows style line
endings.
"""
known_text_files = r'.*\.(txt|html|htm|py|gyp|gypi|gn|isolate|icon)$'
file_inclusion_pattern = (known_text_files,
r'.+%s' % _IMPLEMENTATION_EXTENSIONS,
r'.+%s' % _HEADER_EXTENSIONS)
problems = []
source_file_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=file_inclusion_pattern, files_to_skip=None)
for f in input_api.AffectedSourceFiles(source_file_filter):
# Ignore test files that contain crlf intentionally.
if f.LocalPath().endswith('crlf.txt'):
continue
include_file = False
for line in input_api.ReadFile(f, 'r').splitlines(True):
if line.endswith('\r\n'):
include_file = True
if include_file:
problems.append(f.LocalPath())
if problems:
return [
output_api.PresubmitPromptWarning(
'Are you sure that you want '
'these files to contain Windows style line endings?\n' +
'\n'.join(problems))
]
return []
def CheckIconFilesForLicenseHeaders(input_api, output_api):
"""Check that .icon files (which are fragments of C++) have license headers.
"""
icon_files = (r'.*\.icon$', )
icons = lambda x: input_api.FilterSourceFile(x, files_to_check=icon_files)
return input_api.canned_checks.CheckLicense(input_api,
output_api,
source_file_filter=icons)
def CheckForUseOfChromeAppsDeprecations(input_api, output_api):
"""Check source code for use of Chrome App technologies being
deprecated.
"""
def _CheckForDeprecatedTech(input_api,
output_api,
detection_list,
files_to_check=None,
files_to_skip=None):
if (files_to_check or files_to_skip):
source_file_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=files_to_check, files_to_skip=files_to_skip)
else:
source_file_filter = None
problems = []
for f in input_api.AffectedSourceFiles(source_file_filter):
if f.Action() == 'D':
continue
for _, line in f.ChangedContents():
if any(detect in line for detect in detection_list):
problems.append(f.LocalPath())
return problems
# to avoid this presubmit script triggering warnings
files_to_skip = ['PRESUBMIT.py', 'PRESUBMIT_test.py']
problems = []
# NMF: any files with extensions .nmf or NMF
_NMF_FILES = r'\.(nmf|NMF)$'
problems += _CheckForDeprecatedTech(
input_api,
output_api,
detection_list=[''], # any change to the file will trigger warning
files_to_check=[r'.+%s' % _NMF_FILES])
# MANIFEST: any manifest.json that in its diff includes "app":
_MANIFEST_FILES = r'(manifest\.json)$'
problems += _CheckForDeprecatedTech(
input_api,
output_api,
detection_list=['"app":'],
files_to_check=[r'.*%s' % _MANIFEST_FILES])
# NaCl / PNaCl: any file that in its diff contains the strings in the list
problems += _CheckForDeprecatedTech(
input_api,
output_api,
detection_list=['config=nacl', 'enable-nacl', 'cpu=pnacl', 'nacl_io'],
files_to_skip=files_to_skip + [r"^native_client_sdk/"])
# PPAPI: any C/C++ file that in its diff includes a ppapi library
problems += _CheckForDeprecatedTech(
input_api,
output_api,
detection_list=['#include "ppapi', '#include <ppapi'],
files_to_check=(r'.+%s' % _HEADER_EXTENSIONS,
r'.+%s' % _IMPLEMENTATION_EXTENSIONS),
files_to_skip=[r"^ppapi/"])
if problems:
return [
output_api.PresubmitPromptWarning(
'You are adding/modifying code'
'related to technologies which will soon be deprecated (Chrome Apps, NaCl,'
' PNaCl, PPAPI). See this blog post for more details:\n'
'https://blog.chromium.org/2020/08/changes-to-chrome-app-support-timeline.html\n'
'and this documentation for options to replace these technologies:\n'
'https://developer.chrome.com/docs/apps/migration/\n' +
'\n'.join(problems))
]
return []
def CheckSyslogUseWarningOnUpload(input_api, output_api, src_file_filter=None):
"""Checks that all source files use SYSLOG properly."""
syslog_files = []
for f in input_api.AffectedSourceFiles(src_file_filter):
for line_number, line in f.ChangedContents():
if 'SYSLOG' in line:
syslog_files.append(f.LocalPath() + ':' + str(line_number))
if syslog_files:
return [
output_api.PresubmitPromptWarning(
'Please make sure there are no privacy sensitive bits of data in SYSLOG'
' calls.\nFiles to check:\n',
items=syslog_files)
]
return []
def CheckChangeOnUpload(input_api, output_api):
if input_api.version < [2, 0, 0]:
return [
output_api.PresubmitError(
"Your depot_tools is out of date. "
"This PRESUBMIT.py requires at least presubmit_support version 2.0.0, "
"but your version is %d.%d.%d" % tuple(input_api.version))
]
results = []
results.extend(
input_api.canned_checks.CheckPatchFormatted(input_api, output_api))
return results
def CheckChangeOnCommit(input_api, output_api):
if input_api.version < [2, 0, 0]:
return [
output_api.PresubmitError(
"Your depot_tools is out of date. "
"This PRESUBMIT.py requires at least presubmit_support version 2.0.0, "
"but your version is %d.%d.%d" % tuple(input_api.version))
]
results = []
# Make sure the tree is 'open'.
results.extend(
input_api.canned_checks.CheckTreeIsOpen(
input_api,
output_api,
json_url='http://chromium-status.appspot.com/current?format=json'))
results.extend(
input_api.canned_checks.CheckPatchFormatted(input_api, output_api))
results.extend(
input_api.canned_checks.CheckChangeHasBugField(input_api, output_api))
results.extend(
input_api.canned_checks.CheckChangeHasNoUnwantedTags(
input_api, output_api))
return results
def CheckStrings(input_api, output_api):
"""Check string ICU syntax validity and if translation screenshots exist."""
# Skip translation screenshots check if a SkipTranslationScreenshotsCheck
# footer is set to true.
git_footers = input_api.change.GitFootersFromDescription()
skip_screenshot_check_footer = [
footer.lower() for footer in git_footers.get(
u'Skip-Translation-Screenshots-Check', [])
]
run_screenshot_check = u'true' not in skip_screenshot_check_footer
import os
import re
import sys
from io import StringIO
new_or_added_paths = set(f.LocalPath() for f in input_api.AffectedFiles()
if (f.Action() == 'A' or f.Action() == 'M'))
removed_paths = set(f.LocalPath()
for f in input_api.AffectedFiles(include_deletes=True)
if f.Action() == 'D')
affected_grds = [
f for f in input_api.AffectedFiles()
if f.LocalPath().endswith(('.grd', '.grdp'))
]
affected_grds = [
f for f in affected_grds if not 'testdata' in f.LocalPath()
]
if not affected_grds:
return []
affected_png_paths = [
f.AbsoluteLocalPath() for f in input_api.AffectedFiles()
if (f.LocalPath().endswith('.png'))
]
# Check for screenshots. Developers can upload screenshots using
# tools/translation/upload_screenshots.py which finds and uploads
# images associated with .grd files (e.g. test_grd/IDS_STRING.png for the
# message named IDS_STRING in test.grd) and produces a .sha1 file (e.g.
# test_grd/IDS_STRING.png.sha1) for each png when the upload is successful.
#
# The logic here is as follows:
#
# - If the CL has a .png file under the screenshots directory for a grd
# file, warn the developer. Actual images should never be checked into the
# Chrome repo.
#
# - If the CL contains modified or new messages in grd files and doesn't
# contain the corresponding .sha1 files, warn the developer to add images
# and upload them via tools/translation/upload_screenshots.py.
#
# - If the CL contains modified or new messages in grd files and the
# corresponding .sha1 files, everything looks good.
#
# - If the CL contains removed messages in grd files but the corresponding
# .sha1 files aren't removed, warn the developer to remove them.
unnecessary_screenshots = []
missing_sha1 = []
missing_sha1_modified = []
unnecessary_sha1_files = []
# This checks verifies that the ICU syntax of messages this CL touched is
# valid, and reports any found syntax errors.
# Without this presubmit check, ICU syntax errors in Chromium strings can land
# without developers being aware of them. Later on, such ICU syntax errors
# break message extraction for translation, hence would block Chromium
# translations until they are fixed.
icu_syntax_errors = []
def _CheckScreenshotAdded(screenshots_dir, message_id):
sha1_path = input_api.os_path.join(screenshots_dir,
message_id + '.png.sha1')
if sha1_path not in new_or_added_paths:
missing_sha1.append(sha1_path)
def _CheckScreenshotModified(screenshots_dir, message_id):
sha1_path = input_api.os_path.join(screenshots_dir,
message_id + '.png.sha1')
if sha1_path not in new_or_added_paths:
missing_sha1_modified.append(sha1_path)
def _CheckScreenshotRemoved(screenshots_dir, message_id):
sha1_path = input_api.os_path.join(screenshots_dir,
message_id + '.png.sha1')
if input_api.os_path.exists(
sha1_path) and sha1_path not in removed_paths:
unnecessary_sha1_files.append(sha1_path)
def _ValidateIcuSyntax(text, level, signatures):
"""Validates ICU syntax of a text string.
Check if text looks similar to ICU and checks for ICU syntax correctness
in this case. Reports various issues with ICU syntax and values of
variants. Supports checking of nested messages. Accumulate information of
each ICU messages found in the text for further checking.
Args:
text: a string to check.
level: a number of current nesting level.
signatures: an accumulator, a list of tuple of (level, variable,
kind, variants).
Returns:
None if a string is not ICU or no issue detected.
A tuple of (message, start index, end index) if an issue detected.
"""
valid_types = {
'plural': (frozenset(
['=0', '=1', 'zero', 'one', 'two', 'few', 'many',
'other']), frozenset(['=1', 'other'])),
'selectordinal': (frozenset(
['=0', '=1', 'zero', 'one', 'two', 'few', 'many',
'other']), frozenset(['one', 'other'])),
'select': (frozenset(), frozenset(['other'])),
}
# Check if the message looks like an attempt to use ICU
# plural. If yes - check if its syntax strictly matches ICU format.
like = re.match(r'^[^{]*\{[^{]*\b(plural|selectordinal|select)\b',
text)
if not like:
signatures.append((level, None, None, None))
return
# Check for valid prefix and suffix
m = re.match(
r'^([^{]*\{)([a-zA-Z0-9_]+),\s*'
r'(plural|selectordinal|select),\s*'
r'(?:offset:\d+)?\s*(.*)', text, re.DOTALL)
if not m:
return (('This message looks like an ICU plural, '
'but does not follow ICU syntax.'), like.start(),
like.end())
starting, variable, kind, variant_pairs = m.groups()
variants, depth, last_pos = _ParseIcuVariants(variant_pairs,
m.start(4))
if depth:
return ('Invalid ICU format. Unbalanced opening bracket', last_pos,
len(text))
first = text[0]
ending = text[last_pos:]
if not starting:
return ('Invalid ICU format. No initial opening bracket',
last_pos - 1, last_pos)
if not ending or '}' not in ending:
return ('Invalid ICU format. No final closing bracket',
last_pos - 1, last_pos)
elif first != '{':
return ((
'Invalid ICU format. Extra characters at the start of a complex '
'message (go/icu-message-migration): "%s"') % starting, 0,
len(starting))
elif ending != '}':
return ((
'Invalid ICU format. Extra characters at the end of a complex '
'message (go/icu-message-migration): "%s"') % ending,
last_pos - 1, len(text) - 1)
if kind not in valid_types:
return (('Unknown ICU message type %s. '
'Valid types are: plural, select, selectordinal') % kind,
0, 0)
known, required = valid_types[kind]
defined_variants = set()
for variant, variant_range, value, value_range in variants:
start, end = variant_range
if variant in defined_variants:
return ('Variant "%s" is defined more than once' % variant,
start, end)
elif known and variant not in known:
return ('Variant "%s" is not valid for %s message' %
(variant, kind), start, end)
defined_variants.add(variant)
# Check for nested structure
res = _ValidateIcuSyntax(value[1:-1], level + 1, signatures)
if res:
return (res[0], res[1] + value_range[0] + 1,
res[2] + value_range[0] + 1)
missing = required - defined_variants
if missing:
return ('Required variants missing: %s' % ', '.join(missing), 0,
len(text))
signatures.append((level, variable, kind, defined_variants))
def _ParseIcuVariants(text, offset=0):
"""Parse variants part of ICU complex message.
Builds a tuple of variant names and values, as well as
their offsets in the input string.
Args:
text: a string to parse
offset: additional offset to add to positions in the text to get correct
position in the complete ICU string.
Returns:
List of tuples, each tuple consist of four fields: variant name,
variant name span (tuple of two integers), variant value, value
span (tuple of two integers).
"""
depth, start, end = 0, -1, -1
variants = []
key = None
for idx, char in enumerate(text):
if char == '{':
if not depth:
start = idx
chunk = text[end + 1:start]
key = chunk.strip()
pos = offset + end + 1 + chunk.find(key)
span = (pos, pos + len(key))
depth += 1
elif char == '}':
if not depth:
return variants, depth, offset + idx
depth -= 1
if not depth:
end = idx
variants.append((key, span, text[start:end + 1],
(offset + start, offset + end + 1)))
return variants, depth, offset + end + 1
try:
old_sys_path = sys.path
sys.path = sys.path + [
input_api.os_path.join(input_api.PresubmitLocalPath(), 'tools',
'translation')
]
from helper import grd_helper
finally:
sys.path = old_sys_path
for f in affected_grds:
file_path = f.LocalPath()
old_id_to_msg_map = {}
new_id_to_msg_map = {}
# Note that this code doesn't check if the file has been deleted. This is
# OK because it only uses the old and new file contents and doesn't load
# the file via its path.
# It's also possible that a file's content refers to a renamed or deleted
# file via a <part> tag, such as <part file="now-deleted-file.grdp">. This
# is OK as well, because grd_helper ignores <part> tags when loading .grd or
# .grdp files.
if file_path.endswith('.grdp'):
if f.OldContents():
old_id_to_msg_map = grd_helper.GetGrdpMessagesFromString(
'\n'.join(f.OldContents()))
if f.NewContents():
new_id_to_msg_map = grd_helper.GetGrdpMessagesFromString(
'\n'.join(f.NewContents()))
else:
file_dir = input_api.os_path.dirname(file_path) or '.'
if f.OldContents():
old_id_to_msg_map = grd_helper.GetGrdMessages(
StringIO('\n'.join(f.OldContents())), file_dir)
if f.NewContents():
new_id_to_msg_map = grd_helper.GetGrdMessages(
StringIO('\n'.join(f.NewContents())), file_dir)
grd_name, ext = input_api.os_path.splitext(
input_api.os_path.basename(file_path))
screenshots_dir = input_api.os_path.join(
input_api.os_path.dirname(file_path),
grd_name + ext.replace('.', '_'))
# Compute added, removed and modified message IDs.
old_ids = set(old_id_to_msg_map)
new_ids = set(new_id_to_msg_map)
added_ids = new_ids - old_ids
removed_ids = old_ids - new_ids
modified_ids = set([])
for key in old_ids.intersection(new_ids):
if (old_id_to_msg_map[key].ContentsAsXml('', True) !=
new_id_to_msg_map[key].ContentsAsXml('', True)):
# The message content itself changed. Require an updated screenshot.
modified_ids.add(key)
elif old_id_to_msg_map[key].attrs['meaning'] != \
new_id_to_msg_map[key].attrs['meaning']:
# The message meaning changed. Ensure there is a screenshot for it.
sha1_path = input_api.os_path.join(screenshots_dir,
key + '.png.sha1')
if sha1_path not in new_or_added_paths and not \
input_api.os_path.exists(sha1_path):
# There is neither a previous screenshot nor is a new one added now.
# Require a screenshot.
modified_ids.add(key)
if run_screenshot_check:
# Check the screenshot directory for .png files. Warn if there is any.
for png_path in affected_png_paths:
if png_path.startswith(screenshots_dir):
unnecessary_screenshots.append(png_path)
for added_id in added_ids:
_CheckScreenshotAdded(screenshots_dir, added_id)
for modified_id in modified_ids:
_CheckScreenshotModified(screenshots_dir, modified_id)
for removed_id in removed_ids:
_CheckScreenshotRemoved(screenshots_dir, removed_id)
# Check new and changed strings for ICU syntax errors.
for key in added_ids.union(modified_ids):
msg = new_id_to_msg_map[key].ContentsAsXml('', True)
err = _ValidateIcuSyntax(msg, 0, [])
if err is not None:
icu_syntax_errors.append(str(key) + ': ' + str(err[0]))
results = []
if run_screenshot_check:
if unnecessary_screenshots:
results.append(
output_api.PresubmitError(
'Do not include actual screenshots in the changelist. Run '
'tools/translate/upload_screenshots.py to upload them instead:',
sorted(unnecessary_screenshots)))
if missing_sha1:
results.append(
output_api.PresubmitError(
'You are adding UI strings.\n'
'To ensure the best translations, take screenshots of the relevant UI '
'(https://g.co/chrome/translation) and add these files to your '
'changelist:', sorted(missing_sha1)))
if missing_sha1_modified:
results.append(
output_api.PresubmitError(
'You are modifying UI strings or their meanings.\n'
'To ensure the best translations, take screenshots of the relevant UI '
'(https://g.co/chrome/translation) and add these files to your '
'changelist:', sorted(missing_sha1_modified)))
if unnecessary_sha1_files:
results.append(
output_api.PresubmitError(
'You removed strings associated with these files. Remove:',
sorted(unnecessary_sha1_files)))
else:
results.append(
output_api.PresubmitPromptOrNotify('Skipping translation '
'screenshots check.'))
if icu_syntax_errors:
results.append(
output_api.PresubmitPromptWarning(
'ICU syntax errors were found in the following strings (problems or '
'feedback? Contact rainhard@chromium.org):',
items=icu_syntax_errors))
return results
def CheckTranslationExpectations(input_api, output_api,
repo_root=None,
translation_expectations_path=None,
grd_files=None):
import sys
affected_grds = [
f for f in input_api.AffectedFiles()
if (f.LocalPath().endswith('.grd') or f.LocalPath().endswith('.grdp'))
]
if not affected_grds:
return []
try:
old_sys_path = sys.path
sys.path = sys.path + [
input_api.os_path.join(input_api.PresubmitLocalPath(), 'tools',
'translation')
]
from helper import git_helper
from helper import translation_helper
finally:
sys.path = old_sys_path
# Check that translation expectations can be parsed and we can get a list of
# translatable grd files. |repo_root| and |translation_expectations_path| are
# only passed by tests.
if not repo_root:
repo_root = input_api.PresubmitLocalPath()
if not translation_expectations_path:
translation_expectations_path = input_api.os_path.join(
repo_root, 'tools', 'gritsettings', 'translation_expectations.pyl')
if not grd_files:
grd_files = git_helper.list_grds_in_repository(repo_root)
# Ignore bogus grd files used only for testing
# ui/webui/resources/tools/generate_grd.py.
ignore_path = input_api.os_path.join('ui', 'webui', 'resources', 'tools',
'tests')
grd_files = [p for p in grd_files if ignore_path not in p]
try:
translation_helper.get_translatable_grds(
repo_root, grd_files, translation_expectations_path)
except Exception as e:
return [
output_api.PresubmitNotifyResult(
'Failed to get a list of translatable grd files. This happens when:\n'
' - One of the modified grd or grdp files cannot be parsed or\n'
' - %s is not updated.\n'
'Stack:\n%s' % (translation_expectations_path, str(e)))
]
return []
def CheckStableMojomChanges(input_api, output_api):
"""Changes to [Stable] mojom types must preserve backward-compatibility."""
changed_mojoms = input_api.AffectedFiles(
include_deletes=True,
file_filter=lambda f: f.LocalPath().endswith(('.mojom')))
if not changed_mojoms or input_api.no_diffs:
return []
delta = []
for mojom in changed_mojoms:
delta.append({
'filename': mojom.LocalPath(),
'old': '\n'.join(mojom.OldContents()) or None,
'new': '\n'.join(mojom.NewContents()) or None,
})
process = input_api.subprocess.Popen([
input_api.python3_executable,
input_api.os_path.join(
input_api.PresubmitLocalPath(), 'mojo', 'public', 'tools', 'mojom',
'check_stable_mojom_compatibility.py'), '--src-root',
input_api.PresubmitLocalPath()
],
stdin=input_api.subprocess.PIPE,
stdout=input_api.subprocess.PIPE,
stderr=input_api.subprocess.PIPE,
universal_newlines=True)
(x, error) = process.communicate(input=input_api.json.dumps(delta))
if process.returncode:
return [
output_api.PresubmitError(
'One or more [Stable] mojom definitions appears to have been changed '
'in a way that is not backward-compatible.',
long_text=error)
]
return []
def CheckDeprecationOfPreferences(input_api, output_api):
"""Removing a preference should come with a deprecation."""
def FilterFile(affected_file):
"""Accept only .cc files and the like."""
file_inclusion_pattern = [r'.+%s' % _IMPLEMENTATION_EXTENSIONS]
files_to_skip = (_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS +
input_api.DEFAULT_FILES_TO_SKIP)
return input_api.FilterSourceFile(
affected_file,
files_to_check=file_inclusion_pattern,
files_to_skip=files_to_skip)
def ModifiedLines(affected_file):
"""Returns a list of tuples (line number, line text) of added and removed
lines.
Deleted lines share the same line number as the previous line.
This relies on the scm diff output describing each changed code section
with a line of the form
^@@ <old line num>,<old size> <new line num>,<new size> @@$
"""
line_num = 0
modified_lines = []
for line in affected_file.GenerateScmDiff().splitlines():
# Extract <new line num> of the patch fragment (see format above).
m = input_api.re.match(r'^@@ [0-9\,\+\-]+ \+([0-9]+)\,[0-9]+ @@',
line)
if m:
line_num = int(m.groups(1)[0])
continue
if ((line.startswith('+') and not line.startswith('++'))
or (line.startswith('-') and not line.startswith('--'))):
modified_lines.append((line_num, line))
if not line.startswith('-'):
line_num += 1
return modified_lines
def FindLineWith(lines, needle):
"""Returns the line number (i.e. index + 1) in `lines` containing `needle`.
If 0 or >1 lines contain `needle`, -1 is returned.
"""
matching_line_numbers = [
# + 1 for 1-based counting of line numbers.
i + 1 for i, line in enumerate(lines) if needle in line
]
return matching_line_numbers[0] if len(
matching_line_numbers) == 1 else -1
def ModifiedPrefMigration(affected_file):
"""Returns whether the MigrateObsolete.*Pref functions were modified."""
# Determine first and last lines of MigrateObsolete.*Pref functions.
new_contents = affected_file.NewContents()
range_1 = (FindLineWith(new_contents,
'BEGIN_MIGRATE_OBSOLETE_LOCAL_STATE_PREFS'),
FindLineWith(new_contents,
'END_MIGRATE_OBSOLETE_LOCAL_STATE_PREFS'))
range_2 = (FindLineWith(new_contents,
'BEGIN_MIGRATE_OBSOLETE_PROFILE_PREFS'),
FindLineWith(new_contents,
'END_MIGRATE_OBSOLETE_PROFILE_PREFS'))
if (-1 in range_1 + range_2):
raise Exception(
'Broken .*MIGRATE_OBSOLETE_.*_PREFS markers in browser_prefs.cc.'
)
# Check whether any of the modified lines are part of the
# MigrateObsolete.*Pref functions.
for line_nr, line in ModifiedLines(affected_file):
if (range_1[0] <= line_nr <= range_1[1]
or range_2[0] <= line_nr <= range_2[1]):
return True
return False
register_pref_pattern = input_api.re.compile(r'Register.+Pref')
browser_prefs_file_pattern = input_api.re.compile(
r'chrome/browser/prefs/browser_prefs.cc')
changes = input_api.AffectedFiles(include_deletes=True,
file_filter=FilterFile)
potential_problems = []
for f in changes:
for line in f.GenerateScmDiff().splitlines():
# Check deleted lines for pref registrations.
if (line.startswith('-') and not line.startswith('--')
and register_pref_pattern.search(line)):
potential_problems.append('%s: %s' % (f.LocalPath(), line))
if browser_prefs_file_pattern.search(f.LocalPath()):
# If the developer modified the MigrateObsolete.*Prefs() functions, we
# assume that they knew that they have to deprecate preferences and don't
# warn.
try:
if ModifiedPrefMigration(f):
return []
except Exception as e:
return [output_api.PresubmitError(str(e))]
if potential_problems:
return [
output_api.PresubmitPromptWarning(
'Discovered possible removal of preference registrations.\n\n'
'Please make sure to properly deprecate preferences by clearing their\n'
'value for a couple of milestones before finally removing the code.\n'
'Otherwise data may stay in the preferences files forever. See\n'
'Migrate*Prefs() in chrome/browser/prefs/browser_prefs.cc and\n'
'chrome/browser/prefs/README.md for examples.\n'
'This may be a false positive warning (e.g. if you move preference\n'
'registrations to a different place).\n', potential_problems)
]
return []
def CheckConsistentGrdChanges(input_api, output_api):
"""Changes to GRD files must be consistent for tools to read them."""
changed_grds = input_api.AffectedFiles(
include_deletes=False,
file_filter=lambda f: f.LocalPath().endswith(('.grd')))
errors = []
invalid_file_regexes = [(input_api.re.compile(matcher), msg)
for matcher, msg in _INVALID_GRD_FILE_LINE]
for grd in changed_grds:
for i, line in enumerate(grd.NewContents()):
for matcher, msg in invalid_file_regexes:
if matcher.search(line):
errors.append(
output_api.PresubmitError(
'Problem on {grd}:{i} - {msg}'.format(
grd=grd.LocalPath(), i=i + 1, msg=msg)))
return errors
def CheckMPArchApiUsage(input_api, output_api):
"""CC the MPArch watchlist if the CL uses an API that is ambiguous in the
presence of MPArch features such as bfcache, prerendering, and fenced frames.
"""
# Only consider top-level directories that (1) can use content APIs or
# problematic blink APIs, (2) apply to desktop or android chrome, and (3)
# are known to have a significant number of uses of the APIs of concern.
files_to_check = (
r'^(chrome|components|content|extensions|third_party/blink/renderer)/.+%s' %
_IMPLEMENTATION_EXTENSIONS,
r'^(chrome|components|content|extensions|third_party/blink/renderer)/.+%s' %
_HEADER_EXTENSIONS,
)
files_to_skip = (_EXCLUDED_PATHS + _TEST_CODE_EXCLUDED_PATHS +
input_api.DEFAULT_FILES_TO_SKIP)
source_file_filter = lambda f: input_api.FilterSourceFile(
f, files_to_check=files_to_check, files_to_skip=files_to_skip)
# Here we list the classes/methods we're monitoring. For the "fyi" cases,
# we add the CL to the watchlist, but we don't omit a warning or have it be
# included in the triage rotation.
# Note that since these are are just regular expressions and we don't have
# the compiler's AST, we could have spurious matches (e.g. an unrelated class
# could have a method named IsInMainFrame).
fyi_concerning_class_pattern = input_api.re.compile(
r'WebContentsObserver|WebContentsUserData')
# A subset of WebContentsObserver overrides where there's particular risk for
# confusing tab and page level operations and data (e.g. incorrectly
# resetting page state in DidFinishNavigation).
fyi_concerning_wco_methods = [
'DidStartNavigation',
'ReadyToCommitNavigation',
'DidFinishNavigation',
'RenderViewReady',
'RenderViewDeleted',
'RenderViewHostChanged',
'DOMContentLoaded',
'DidFinishLoad',
]
concerning_nav_handle_methods = [
'IsInMainFrame',
]
concerning_web_contents_methods = [
'FromRenderFrameHost',
'FromRenderViewHost',
]
fyi_concerning_web_contents_methods = [
'GetRenderViewHost',
]
concerning_rfh_methods = [
'GetParent',
'GetMainFrame',
]
fyi_concerning_rfh_methods = [
'GetFrameTreeNodeId',
]
concerning_rfhi_methods = [
'is_main_frame',
]
concerning_ftn_methods = [
'IsMainFrame',
]
concerning_blink_frame_methods = [
'IsCrossOriginToNearestMainFrame',
]
concerning_method_pattern = input_api.re.compile(r'(' + r'|'.join(
item for sublist in [
concerning_nav_handle_methods,
concerning_web_contents_methods, concerning_rfh_methods,
concerning_rfhi_methods, concerning_ftn_methods,
concerning_blink_frame_methods,
] for item in sublist) + r')\(')
fyi_concerning_method_pattern = input_api.re.compile(r'(' + r'|'.join(
item for sublist in [
fyi_concerning_wco_methods, fyi_concerning_web_contents_methods,
fyi_concerning_rfh_methods,
] for item in sublist) + r')\(')
used_apis = set()
used_fyi_methods = False
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=source_file_filter):
for line_num, line in f.ChangedContents():
fyi_class_match = fyi_concerning_class_pattern.search(line)
if fyi_class_match:
used_fyi_methods = True
fyi_method_match = fyi_concerning_method_pattern.search(line)
if fyi_method_match:
used_fyi_methods = True
method_match = concerning_method_pattern.search(line)
if method_match:
used_apis.add(method_match[1])
if not used_apis:
if used_fyi_methods:
output_api.AppendCC('mparch-reviews+watchfyi@chromium.org')
return []
output_api.AppendCC('mparch-reviews+watch@chromium.org')
message = ('This change uses API(s) that are ambiguous in the presence of '
'MPArch features such as bfcache, prerendering, and fenced '
'frames.')
explanation = (
'Please double check whether new code assumes that a WebContents only '
'contains a single page at a time. Notably, checking whether a frame '
'is the \"main frame\" is not specific enough to determine whether it '
'corresponds to the document reflected in the omnibox. A WebContents '
'may have additional main frames for prerendered pages, bfcached '
'pages, fenced frames, etc. '
'See this doc [1] and the comments on the individual APIs '
'for guidance and this doc [2] for context. The MPArch review '
'watchlist has been CC\'d on this change to help identify any issues.\n'
'[1] https://docs.google.com/document/d/13l16rWTal3o5wce4i0RwdpMP5ESELLKr439Faj2BBRo/edit?usp=sharing\n'
'[2] https://docs.google.com/document/d/1NginQ8k0w3znuwTiJ5qjYmBKgZDekvEPC22q0I4swxQ/edit?usp=sharing'
)
return [
output_api.PresubmitNotifyResult(message,
items=list(used_apis),
long_text=explanation)
]
def CheckAssertAshOnlyCode(input_api, output_api):
"""Errors if a BUILD.gn file in an ash/ directory doesn't include
assert(is_chromeos_ash).
"""
def FileFilter(affected_file):
"""Includes directories known to be Ash only."""
return input_api.FilterSourceFile(
affected_file,
files_to_check=(
r'^ash/.*BUILD\.gn', # Top-level src/ash/.
r'.*/ash/.*BUILD\.gn'), # Any path component.
files_to_skip=(input_api.DEFAULT_FILES_TO_SKIP))
errors = []
pattern = input_api.re.compile(r'assert\(is_chromeos_ash')
for f in input_api.AffectedFiles(include_deletes=False,
file_filter=FileFilter):
if (not pattern.search(input_api.ReadFile(f))):
errors.append(
output_api.PresubmitError(
'Please add assert(is_chromeos_ash) to %s. If that\'s not '
'possible, please create and issue and add a comment such '
'as:\n # TODO(https://crbug.com/XXX): add '
'assert(is_chromeos_ash) when ...' % f.LocalPath()))
return errors
def _IsRendererOnlyCppFile(input_api, affected_file):
path = affected_file.LocalPath()
if not _IsCPlusPlusFile(input_api, path):
return False
# Any code under a "renderer" subdirectory is assumed to be Renderer-only.
if "/renderer/" in path:
return True
# Blink's public/web API is only used/included by Renderer-only code. Note
# that public/platform API may be used in non-Renderer processes (e.g. there
# are some includes in code used by Utility, PDF, or Plugin processes).
if "/blink/public/web/" in path:
return True
# We assume that everything else may be used outside of Renderer processes.
return False
# TODO(https://crbug.com/1273182): Remove these checks, once they are replaced
# by the Chromium Clang Plugin (which will be preferable because it will
# 1) report errors earlier - at compile-time and 2) cover more rules).
def CheckRawPtrUsage(input_api, output_api):
"""Rough checks that raw_ptr<T> usage guidelines are followed."""
errors = []
# The regex below matches "raw_ptr<" following a word boundary, but not in a
# C++ comment.
raw_ptr_matcher = input_api.re.compile(r'^((?!//).)*\braw_ptr<')
file_filter = lambda f: _IsRendererOnlyCppFile(input_api, f)
for f, line_num, line in input_api.RightHandSideLines(file_filter):
if raw_ptr_matcher.search(line):
errors.append(
output_api.PresubmitError(
'Problem on {path}:{line} - '\
'raw_ptr<T> should not be used in Renderer-only code '\
'(as documented in the "Pointers to unprotected memory" '\
'section in //base/memory/raw_ptr.md)'.format(
path=f.LocalPath(), line=line_num)))
return errors
def CheckPythonShebang(input_api, output_api):
"""Checks that python scripts use #!/usr/bin/env instead of hardcoding a
system-wide python.
"""
errors = []
sources = lambda affected_file: input_api.FilterSourceFile(
affected_file,
files_to_skip=((_THIRD_PARTY_EXCEPT_BLINK,
r'third_party/blink/web_tests/external/') + input_api.
DEFAULT_FILES_TO_SKIP),
files_to_check=[r'.*\.py$'])
for f in input_api.AffectedSourceFiles(sources):
for line_num, line in f.ChangedContents():
if line_num == 1 and line.startswith('#!/usr/bin/python'):
errors.append(f.LocalPath())
break
result = []
for file in errors:
result.append(
output_api.PresubmitError(
"Please use '#!/usr/bin/env python/2/3' as the shebang of %s" %
file))
return result
def CheckBatchAnnotation(input_api, output_api):
"""Checks that tests have either @Batch or @DoNotBatch annotation. If this
is not an instrumentation test, disregard."""
batch_annotation = input_api.re.compile(r'^\s*@Batch')
do_not_batch_annotation = input_api.re.compile(r'^\s*@DoNotBatch')
robolectric_test = input_api.re.compile(r'[rR]obolectric')
test_class_declaration = input_api.re.compile(r'^\s*public\sclass.*Test')
uiautomator_test = input_api.re.compile(r'[uU]i[aA]utomator')
missing_annotation_errors = []
extra_annotation_errors = []
def _FilterFile(affected_file):
return input_api.FilterSourceFile(
affected_file,
files_to_skip=input_api.DEFAULT_FILES_TO_SKIP,
files_to_check=[r'.*Test\.java$'])
for f in input_api.AffectedSourceFiles(_FilterFile):
batch_matched = None
do_not_batch_matched = None
is_instrumentation_test = True
for line in f.NewContents():
if robolectric_test.search(line) or uiautomator_test.search(line):
# Skip Robolectric and UiAutomator tests.
is_instrumentation_test = False
break
if not batch_matched:
batch_matched = batch_annotation.search(line)
if not do_not_batch_matched:
do_not_batch_matched = do_not_batch_annotation.search(line)
test_class_declaration_matched = test_class_declaration.search(
line)
if test_class_declaration_matched:
break
if (is_instrumentation_test and
not batch_matched and
not do_not_batch_matched):
missing_annotation_errors.append(str(f.LocalPath()))
if (not is_instrumentation_test and
(batch_matched or
do_not_batch_matched)):
extra_annotation_errors.append(str(f.LocalPath()))
results = []
if missing_annotation_errors:
results.append(
output_api.PresubmitPromptWarning(
"""
Instrumentation tests should use either @Batch or @DoNotBatch. If tests are not
safe to run in batch, please use @DoNotBatch with reasons.
""", missing_annotation_errors))
if extra_annotation_errors:
results.append(
output_api.PresubmitPromptWarning(
"""
Robolectric tests do not need a @Batch or @DoNotBatch annotations.
""", extra_annotation_errors))
return results
def CheckMockAnnotation(input_api, output_api):
"""Checks that we have annotated all Mockito.mock()-ed or Mockito.spy()-ed
classes with @Mock or @Spy. If this is not an instrumentation test,
disregard."""
# This is just trying to be approximately correct. We are not writing a
# Java parser, so special cases like statically importing mock() then
# calling an unrelated non-mockito spy() function will cause a false
# positive.
package_name = input_api.re.compile(r'^package\s+(\w+(?:\.\w+)+);')
mock_static_import = input_api.re.compile(
r'^import\s+static\s+org.mockito.Mockito.(?:mock|spy);')
import_class = input_api.re.compile(r'import\s+((?:\w+\.)+)(\w+);')
mock_annotation = input_api.re.compile(r'^\s*@(?:Mock|Spy)')
field_type = input_api.re.compile(r'(\w+)(?:<\w+>)?\s+\w+\s*(?:;|=)')
mock_or_spy_function_call = r'(?:mock|spy)\(\s*(?:new\s*)?(\w+)(?:\.class|\()'
fully_qualified_mock_function = input_api.re.compile(
r'Mockito\.' + mock_or_spy_function_call)
statically_imported_mock_function = input_api.re.compile(
r'\W' + mock_or_spy_function_call)
robolectric_test = input_api.re.compile(r'[rR]obolectric')
uiautomator_test = input_api.re.compile(r'[uU]i[aA]utomator')
def _DoClassLookup(class_name, class_name_map, package):
found = class_name_map.get(class_name)
if found is not None:
return found
else:
return package + '.' + class_name
def _FilterFile(affected_file):
return input_api.FilterSourceFile(
affected_file,
files_to_skip=input_api.DEFAULT_FILES_TO_SKIP,
files_to_check=[r'.*Test\.java$'])
mocked_by_function_classes = set()
mocked_by_annotation_classes = set()
class_to_filename = {}
for f in input_api.AffectedSourceFiles(_FilterFile):
mock_function_regex = fully_qualified_mock_function
next_line_is_annotated = False
fully_qualified_class_map = {}
package = None
for line in f.NewContents():
if robolectric_test.search(line) or uiautomator_test.search(line):
# Skip Robolectric and UiAutomator tests.
break
m = package_name.search(line)
if m:
package = m.group(1)
continue
if mock_static_import.search(line):
mock_function_regex = statically_imported_mock_function
continue
m = import_class.search(line)
if m:
fully_qualified_class_map[m.group(2)] = m.group(1) + m.group(2)
continue
if next_line_is_annotated:
next_line_is_annotated = False
fully_qualified_class = _DoClassLookup(
field_type.search(line).group(1), fully_qualified_class_map,
package)
mocked_by_annotation_classes.add(fully_qualified_class)
continue
if mock_annotation.search(line):
next_line_is_annotated = True
continue
m = mock_function_regex.search(line)
if m:
fully_qualified_class = _DoClassLookup(m.group(1),
fully_qualified_class_map, package)
# Skipping builtin classes, since they don't get optimized.
if fully_qualified_class.startswith(
'android.') or fully_qualified_class.startswith(
'java.'):
continue
class_to_filename[fully_qualified_class] = str(f.LocalPath())
mocked_by_function_classes.add(fully_qualified_class)
results = []
missed_classes = mocked_by_function_classes - mocked_by_annotation_classes
if missed_classes:
error_locations = []
for c in missed_classes:
error_locations.append(c + ' in ' + class_to_filename[c])
results.append(
output_api.PresubmitPromptWarning(
"""
Mockito.mock()/spy() cause issues with our Java optimizer. You have 3 options:
1) If the mocked variable can be a class member, annotate the member with
@Mock/@Spy.
2) If the mocked variable cannot be a class member, create a dummy member
variable of that type, annotated with @Mock/@Spy. This dummy does not need
to be used or initialized in any way.
3) If the mocked type is definitely not going to be optimized, whether it's a
builtin type which we don't ship, or a class you know R8 will treat
specially, you can ignore this warning.
""", error_locations))
return results
def CheckNoJsInIos(input_api, output_api):
"""Checks to make sure that JavaScript files are not used on iOS."""
def _FilterFile(affected_file):
return input_api.FilterSourceFile(
affected_file,
files_to_skip=input_api.DEFAULT_FILES_TO_SKIP +
(r'^ios/third_party/*', r'^third_party/*'),
files_to_check=[r'^ios/.*\.js$', r'.*/ios/.*\.js$'])
error_paths = []
warning_paths = []
for f in input_api.AffectedSourceFiles(_FilterFile):
local_path = f.LocalPath()
if input_api.os_path.splitext(local_path)[1] == '.js':
if f.Action() == 'A':
error_paths.append(local_path)
elif f.Action() != 'D':
warning_paths.append(local_path)
results = []
if warning_paths:
results.append(output_api.PresubmitPromptWarning(
'TypeScript is now fully supported for iOS feature scripts. '
'Consider converting JavaScript files to TypeScript. See '
'//ios/web/public/js_messaging/README.md for more details.',
warning_paths))
if error_paths:
results.append(output_api.PresubmitError(
'Do not use JavaScript on iOS as TypeScript is fully supported. '
'See //ios/web/public/js_messaging/README.md for help using '
'scripts on iOS.', error_paths))
return results