diff options
Diffstat (limited to 'Tools/Scripts/webkitpy/style')
29 files changed, 11960 insertions, 0 deletions
diff --git a/Tools/Scripts/webkitpy/style/__init__.py b/Tools/Scripts/webkitpy/style/__init__.py new file mode 100644 index 0000000..ef65bee --- /dev/null +++ b/Tools/Scripts/webkitpy/style/__init__.py @@ -0,0 +1 @@ +# Required for Python to search this directory for module files diff --git a/Tools/Scripts/webkitpy/style/checker.py b/Tools/Scripts/webkitpy/style/checker.py new file mode 100644 index 0000000..6f1beb0 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checker.py @@ -0,0 +1,749 @@ +# Copyright (C) 2009 Google Inc. All rights reserved. +# Copyright (C) 2010 Chris Jerdonek (chris.jerdonek@gmail.com) +# Copyright (C) 2010 ProFUSION embedded systems +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Front end of some style-checker modules.""" + +import logging +import os.path +import sys + +from checkers.common import categories as CommonCategories +from checkers.common import CarriageReturnChecker +from checkers.cpp import CppChecker +from checkers.python import PythonChecker +from checkers.test_expectations import TestExpectationsChecker +from checkers.text import TextChecker +from checkers.xml import XMLChecker +from error_handlers import DefaultStyleErrorHandler +from filter import FilterConfiguration +from optparser import ArgumentParser +from optparser import DefaultCommandOptionValues +from webkitpy.style_references import configure_logging as _configure_logging + +_log = logging.getLogger("webkitpy.style.checker") + +# These are default option values for the command-line option parser. +_DEFAULT_MIN_CONFIDENCE = 1 +_DEFAULT_OUTPUT_FORMAT = 'emacs' + + +# FIXME: For style categories we will never want to have, remove them. +# For categories for which we want to have similar functionality, +# modify the implementation and enable them. +# +# Throughout this module, we use "filter rule" rather than "filter" +# for an individual boolean filter flag like "+foo". This allows us to +# reserve "filter" for what one gets by collectively applying all of +# the filter rules. +# +# The base filter rules are the filter rules that begin the list of +# filter rules used to check style. For example, these rules precede +# any user-specified filter rules. Since by default all categories are +# checked, this list should normally include only rules that begin +# with a "-" sign. +_BASE_FILTER_RULES = [ + '-build/endif_comment', + '-build/include_what_you_use', # <string> for std::string + '-build/storage_class', # const static + '-legal/copyright', + '-readability/multiline_comment', + '-readability/braces', # int foo() {}; + '-readability/fn_size', + '-readability/casting', + '-readability/function', + '-runtime/arrays', # variable length array + '-runtime/casting', + '-runtime/sizeof', + '-runtime/explicit', # explicit + '-runtime/virtual', # virtual dtor + '-runtime/printf', + '-runtime/threadsafe_fn', + '-runtime/rtti', + '-whitespace/blank_line', + '-whitespace/end_of_line', + '-whitespace/labels', + # List Python pep8 categories last. + # + # Because much of WebKit's Python code base does not abide by the + # PEP8 79 character limit, we ignore the 79-character-limit category + # pep8/E501 for now. + # + # FIXME: Consider bringing WebKit's Python code base into conformance + # with the 79 character limit, or some higher limit that is + # agreeable to the WebKit project. + '-pep8/E501', + ] + + +# The path-specific filter rules. +# +# This list is order sensitive. Only the first path substring match +# is used. See the FilterConfiguration documentation in filter.py +# for more information on this list. +# +# Each string appearing in this nested list should have at least +# one associated unit test assertion. These assertions are located, +# for example, in the test_path_rules_specifier() unit test method of +# checker_unittest.py. +_PATH_RULES_SPECIFIER = [ + # Files in these directories are consumers of the WebKit + # API and therefore do not follow the same header including + # discipline as WebCore. + + ([# TestNetscapePlugIn has no config.h and uses funny names like + # NPP_SetWindow. + "Tools/DumpRenderTree/TestNetscapePlugIn/", + # The API test harnesses have no config.h and use funny macros like + # TEST_CLASS_NAME. + "Tools/WebKitAPITest/", + "Tools/TestWebKitAPI/"], + ["-build/include", + "-readability/naming"]), + ([# The EFL APIs use EFL naming style, which includes + # both lower-cased and camel-cased, underscore-sparated + # values. + "WebKit/efl/ewk/", + # There is no clean way to avoid "yy_*" names used by flex. + "WebCore/css/CSSParser.cpp", + # Qt code uses '_' in some places (such as private slots + # and on test xxx_data methos on tests) + "JavaScriptCore/qt/api/", + "WebKit/qt/Api/", + "WebKit/qt/tests/", + "WebKit/qt/declarative/", + "WebKit/qt/examples/"], + ["-readability/naming"]), + ([# The GTK+ APIs use GTK+ naming style, which includes + # lower-cased, underscore-separated values. + # Also, GTK+ allows the use of NULL. + "WebCore/bindings/scripts/test/GObject", + "WebKit/gtk/webkit/", + "Tools/DumpRenderTree/gtk/"], + ["-readability/naming", + "-readability/null"]), + ([# Header files in ForwardingHeaders have no header guards or + # exceptional header guards (e.g., WebCore_FWD_Debugger_h). + "/ForwardingHeaders/"], + ["-build/header_guard"]), + ([# assembler has lots of opcodes that use underscores, so + # we don't check for underscores in that directory. + "/JavaScriptCore/assembler/"], + ["-readability/naming"]), + + # WebKit2 rules: + # WebKit2 doesn't use config.h, and certain directories have other + # idiosyncracies. + ([# NPAPI has function names with underscores. + "WebKit2/WebProcess/Plugins/Netscape"], + ["-build/include_order", + "-readability/naming"]), + ([# The WebKit2 C API has names with underscores and whitespace-aligned + # struct members. + "WebKit2/UIProcess/API/C/", + "WebKit2/WebProcess/InjectedBundle/API/c/"], + ["-build/include_order", + "-readability/naming", + "-whitespace/declaration"]), + ([# Nothing in WebKit2 uses config.h. + "WebKit2/"], + ["-build/include_order"]), + + # For third-party Python code, keep only the following checks-- + # + # No tabs: to avoid having to set the SVN allow-tabs property. + # No trailing white space: since this is easy to correct. + # No carriage-return line endings: since this is easy to correct. + # + (["webkitpy/thirdparty/"], + ["-", + "+pep8/W191", # Tabs + "+pep8/W291", # Trailing white space + "+whitespace/carriage_return"]), +] + + +_CPP_FILE_EXTENSIONS = [ + 'c', + 'cpp', + 'h', + ] + +_PYTHON_FILE_EXTENSION = 'py' + +_TEXT_FILE_EXTENSIONS = [ + 'ac', + 'cc', + 'cgi', + 'css', + 'exp', + 'flex', + 'gyp', + 'gypi', + 'html', + 'idl', + 'in', + 'js', + 'mm', + 'php', + 'pl', + 'pm', + 'pri', + 'pro', + 'rb', + 'sh', + 'txt', + 'wm', + 'xhtml', + 'y', + ] + +_XML_FILE_EXTENSIONS = [ + 'vcproj', + 'vsprops', + ] + +# Files to skip that are less obvious. +# +# Some files should be skipped when checking style. For example, +# WebKit maintains some files in Mozilla style on purpose to ease +# future merges. +_SKIPPED_FILES_WITH_WARNING = [ + "gtk2drawing.c", # WebCore/platform/gtk/gtk2drawing.c + "gtkdrawing.h", # WebCore/platform/gtk/gtkdrawing.h + "WebKit/gtk/tests/", + # Soup API that is still being cooked, will be removed from WebKit + # in a few months when it is merged into soup proper. The style + # follows the libsoup style completely. + "WebCore/platform/network/soup/cache/", + ] + + +# Files to skip that are more common or obvious. +# +# This list should be in addition to files with FileType.NONE. Files +# with FileType.NONE are automatically skipped without warning. +_SKIPPED_FILES_WITHOUT_WARNING = [ + "LayoutTests/", + ] + +# Extensions of files which are allowed to contain carriage returns. +_CARRIAGE_RETURN_ALLOWED_FILE_EXTENSIONS = [ + 'vcproj', + 'vsprops', + ] + +# The maximum number of errors to report per file, per category. +# If a category is not a key, then it has no maximum. +_MAX_REPORTS_PER_CATEGORY = { + "whitespace/carriage_return": 1 +} + + +def _all_categories(): + """Return the set of all categories used by check-webkit-style.""" + # Take the union across all checkers. + categories = CommonCategories.union(CppChecker.categories) + categories = categories.union(TestExpectationsChecker.categories) + + # FIXME: Consider adding all of the pep8 categories. Since they + # are not too meaningful for documentation purposes, for + # now we add only the categories needed for the unit tests + # (which validate the consistency of the configuration + # settings against the known categories, etc). + categories = categories.union(["pep8/W191", "pep8/W291", "pep8/E501"]) + + return categories + + +def _check_webkit_style_defaults(): + """Return the default command-line options for check-webkit-style.""" + return DefaultCommandOptionValues(min_confidence=_DEFAULT_MIN_CONFIDENCE, + output_format=_DEFAULT_OUTPUT_FORMAT) + + +# This function assists in optparser not having to import from checker. +def check_webkit_style_parser(): + all_categories = _all_categories() + default_options = _check_webkit_style_defaults() + return ArgumentParser(all_categories=all_categories, + base_filter_rules=_BASE_FILTER_RULES, + default_options=default_options) + + +def check_webkit_style_configuration(options): + """Return a StyleProcessorConfiguration instance for check-webkit-style. + + Args: + options: A CommandOptionValues instance. + + """ + filter_configuration = FilterConfiguration( + base_rules=_BASE_FILTER_RULES, + path_specific=_PATH_RULES_SPECIFIER, + user_rules=options.filter_rules) + + return StyleProcessorConfiguration(filter_configuration=filter_configuration, + max_reports_per_category=_MAX_REPORTS_PER_CATEGORY, + min_confidence=options.min_confidence, + output_format=options.output_format, + stderr_write=sys.stderr.write) + + +def _create_log_handlers(stream): + """Create and return a default list of logging.Handler instances. + + Format WARNING messages and above to display the logging level, and + messages strictly below WARNING not to display it. + + Args: + stream: See the configure_logging() docstring. + + """ + # Handles logging.WARNING and above. + error_handler = logging.StreamHandler(stream) + error_handler.setLevel(logging.WARNING) + formatter = logging.Formatter("%(levelname)s: %(message)s") + error_handler.setFormatter(formatter) + + # Create a logging.Filter instance that only accepts messages + # below WARNING (i.e. filters out anything WARNING or above). + non_error_filter = logging.Filter() + # The filter method accepts a logging.LogRecord instance. + non_error_filter.filter = lambda record: record.levelno < logging.WARNING + + non_error_handler = logging.StreamHandler(stream) + non_error_handler.addFilter(non_error_filter) + formatter = logging.Formatter("%(message)s") + non_error_handler.setFormatter(formatter) + + return [error_handler, non_error_handler] + + +def _create_debug_log_handlers(stream): + """Create and return a list of logging.Handler instances for debugging. + + Args: + stream: See the configure_logging() docstring. + + """ + handler = logging.StreamHandler(stream) + formatter = logging.Formatter("%(name)s: %(levelname)-8s %(message)s") + handler.setFormatter(formatter) + + return [handler] + + +def configure_logging(stream, logger=None, is_verbose=False): + """Configure logging, and return the list of handlers added. + + Returns: + A list of references to the logging handlers added to the root + logger. This allows the caller to later remove the handlers + using logger.removeHandler. This is useful primarily during unit + testing where the caller may want to configure logging temporarily + and then undo the configuring. + + Args: + stream: A file-like object to which to log. The stream must + define an "encoding" data attribute, or else logging + raises an error. + logger: A logging.logger instance to configure. This parameter + should be used only in unit tests. Defaults to the + root logger. + is_verbose: A boolean value of whether logging should be verbose. + + """ + # If the stream does not define an "encoding" data attribute, the + # logging module can throw an error like the following: + # + # Traceback (most recent call last): + # File "/System/Library/Frameworks/Python.framework/Versions/2.6/... + # lib/python2.6/logging/__init__.py", line 761, in emit + # self.stream.write(fs % msg.encode(self.stream.encoding)) + # LookupError: unknown encoding: unknown + if logger is None: + logger = logging.getLogger() + + if is_verbose: + logging_level = logging.DEBUG + handlers = _create_debug_log_handlers(stream) + else: + logging_level = logging.INFO + handlers = _create_log_handlers(stream) + + handlers = _configure_logging(logging_level=logging_level, logger=logger, + handlers=handlers) + + return handlers + + +# Enum-like idiom +class FileType: + + NONE = 0 # FileType.NONE evaluates to False. + # Alphabetize remaining types + CPP = 1 + PYTHON = 2 + TEXT = 3 + XML = 4 + + +class CheckerDispatcher(object): + + """Supports determining whether and how to check style, based on path.""" + + def _file_extension(self, file_path): + """Return the file extension without the leading dot.""" + return os.path.splitext(file_path)[1].lstrip(".") + + def should_skip_with_warning(self, file_path): + """Return whether the given file should be skipped with a warning.""" + for skipped_file in _SKIPPED_FILES_WITH_WARNING: + if file_path.find(skipped_file) >= 0: + return True + return False + + def should_skip_without_warning(self, file_path): + """Return whether the given file should be skipped without a warning.""" + if not self._file_type(file_path): # FileType.NONE. + return True + # Since "LayoutTests" is in _SKIPPED_FILES_WITHOUT_WARNING, make + # an exception to prevent files like "LayoutTests/ChangeLog" and + # "LayoutTests/ChangeLog-2009-06-16" from being skipped. + # Files like 'test_expectations.txt' and 'drt_expectations.txt' + # are also should not be skipped. + # + # FIXME: Figure out a good way to avoid having to add special logic + # for this special case. + basename = os.path.basename(file_path) + if basename.startswith('ChangeLog'): + return False + elif basename == 'test_expectations.txt' or basename == 'drt_expectations.txt': + return False + for skipped_file in _SKIPPED_FILES_WITHOUT_WARNING: + if file_path.find(skipped_file) >= 0: + return True + return False + + def should_check_and_strip_carriage_returns(self, file_path): + return self._file_extension(file_path) not in _CARRIAGE_RETURN_ALLOWED_FILE_EXTENSIONS + + def _file_type(self, file_path): + """Return the file type corresponding to the given file.""" + file_extension = self._file_extension(file_path) + + if (file_extension in _CPP_FILE_EXTENSIONS) or (file_path == '-'): + # FIXME: Do something about the comment below and the issue it + # raises since cpp_style already relies on the extension. + # + # Treat stdin as C++. Since the extension is unknown when + # reading from stdin, cpp_style tests should not rely on + # the extension. + return FileType.CPP + elif file_extension == _PYTHON_FILE_EXTENSION: + return FileType.PYTHON + elif file_extension in _XML_FILE_EXTENSIONS: + return FileType.XML + elif (os.path.basename(file_path).startswith('ChangeLog') or + (not file_extension and "Tools/Scripts/" in file_path) or + file_extension in _TEXT_FILE_EXTENSIONS): + return FileType.TEXT + else: + return FileType.NONE + + def _create_checker(self, file_type, file_path, handle_style_error, + min_confidence): + """Instantiate and return a style checker based on file type.""" + if file_type == FileType.NONE: + checker = None + elif file_type == FileType.CPP: + file_extension = self._file_extension(file_path) + checker = CppChecker(file_path, file_extension, + handle_style_error, min_confidence) + elif file_type == FileType.PYTHON: + checker = PythonChecker(file_path, handle_style_error) + elif file_type == FileType.XML: + checker = XMLChecker(file_path, handle_style_error) + elif file_type == FileType.TEXT: + basename = os.path.basename(file_path) + if basename == 'test_expectations.txt' or basename == 'drt_expectations.txt': + checker = TestExpectationsChecker(file_path, handle_style_error) + else: + checker = TextChecker(file_path, handle_style_error) + else: + raise ValueError('Invalid file type "%(file_type)s": the only valid file types ' + "are %(NONE)s, %(CPP)s, and %(TEXT)s." + % {"file_type": file_type, + "NONE": FileType.NONE, + "CPP": FileType.CPP, + "TEXT": FileType.TEXT}) + + return checker + + def dispatch(self, file_path, handle_style_error, min_confidence): + """Instantiate and return a style checker based on file path.""" + file_type = self._file_type(file_path) + + checker = self._create_checker(file_type, + file_path, + handle_style_error, + min_confidence) + return checker + + +# FIXME: Remove the stderr_write attribute from this class and replace +# its use with calls to a logging module logger. +class StyleProcessorConfiguration(object): + + """Stores configuration values for the StyleProcessor class. + + Attributes: + min_confidence: An integer between 1 and 5 inclusive that is the + minimum confidence level of style errors to report. + + max_reports_per_category: The maximum number of errors to report + per category, per file. + + stderr_write: A function that takes a string as a parameter and + serves as stderr.write. + + """ + + def __init__(self, + filter_configuration, + max_reports_per_category, + min_confidence, + output_format, + stderr_write): + """Create a StyleProcessorConfiguration instance. + + Args: + filter_configuration: A FilterConfiguration instance. The default + is the "empty" filter configuration, which + means that all errors should be checked. + + max_reports_per_category: The maximum number of errors to report + per category, per file. + + min_confidence: An integer between 1 and 5 inclusive that is the + minimum confidence level of style errors to report. + The default is 1, which reports all style errors. + + output_format: A string that is the output format. The supported + output formats are "emacs" which emacs can parse + and "vs7" which Microsoft Visual Studio 7 can parse. + + stderr_write: A function that takes a string as a parameter and + serves as stderr.write. + + """ + self._filter_configuration = filter_configuration + self._output_format = output_format + + self.max_reports_per_category = max_reports_per_category + self.min_confidence = min_confidence + self.stderr_write = stderr_write + + def is_reportable(self, category, confidence_in_error, file_path): + """Return whether an error is reportable. + + An error is reportable if both the confidence in the error is + at least the minimum confidence level and the current filter + says the category should be checked for the given path. + + Args: + category: A string that is a style category. + confidence_in_error: An integer between 1 and 5 inclusive that is + the application's confidence in the error. + A higher number means greater confidence. + file_path: The path of the file being checked + + """ + if confidence_in_error < self.min_confidence: + return False + + return self._filter_configuration.should_check(category, file_path) + + def write_style_error(self, + category, + confidence_in_error, + file_path, + line_number, + message): + """Write a style error to the configured stderr.""" + if self._output_format == 'vs7': + format_string = "%s(%s): %s [%s] [%d]\n" + else: + format_string = "%s:%s: %s [%s] [%d]\n" + + self.stderr_write(format_string % (file_path, + line_number, + message, + category, + confidence_in_error)) + + +class ProcessorBase(object): + + """The base class for processors of lists of lines.""" + + def should_process(self, file_path): + """Return whether the file at file_path should be processed. + + The TextFileReader class calls this method prior to reading in + the lines of a file. Use this method, for example, to prevent + the style checker from reading binary files into memory. + + """ + raise NotImplementedError('Subclasses should implement.') + + def process(self, lines, file_path, **kwargs): + """Process lines of text read from a file. + + Args: + lines: A list of lines of text to process. + file_path: The path from which the lines were read. + **kwargs: This argument signifies that the process() method of + subclasses of ProcessorBase may support additional + keyword arguments. + For example, a style checker's check() method + may support a "reportable_lines" parameter that represents + the line numbers of the lines for which style errors + should be reported. + + """ + raise NotImplementedError('Subclasses should implement.') + + +class StyleProcessor(ProcessorBase): + + """A ProcessorBase for checking style. + + Attributes: + error_count: An integer that is the total number of reported + errors for the lifetime of this instance. + + """ + + def __init__(self, configuration, mock_dispatcher=None, + mock_increment_error_count=None, + mock_carriage_checker_class=None): + """Create an instance. + + Args: + configuration: A StyleProcessorConfiguration instance. + mock_dispatcher: A mock CheckerDispatcher instance. This + parameter is for unit testing. Defaults to a + CheckerDispatcher instance. + mock_increment_error_count: A mock error-count incrementer. + mock_carriage_checker_class: A mock class for checking and + transforming carriage returns. + This parameter is for unit testing. + Defaults to CarriageReturnChecker. + + """ + if mock_dispatcher is None: + dispatcher = CheckerDispatcher() + else: + dispatcher = mock_dispatcher + + if mock_increment_error_count is None: + # The following blank line is present to avoid flagging by pep8.py. + + def increment_error_count(): + """Increment the total count of reported errors.""" + self.error_count += 1 + else: + increment_error_count = mock_increment_error_count + + if mock_carriage_checker_class is None: + # This needs to be a class rather than an instance since the + # process() method instantiates one using parameters. + carriage_checker_class = CarriageReturnChecker + else: + carriage_checker_class = mock_carriage_checker_class + + self.error_count = 0 + + self._carriage_checker_class = carriage_checker_class + self._configuration = configuration + self._dispatcher = dispatcher + self._increment_error_count = increment_error_count + + def should_process(self, file_path): + """Return whether the file should be checked for style.""" + if self._dispatcher.should_skip_without_warning(file_path): + return False + if self._dispatcher.should_skip_with_warning(file_path): + _log.warn('File exempt from style guide. Skipping: "%s"' + % file_path) + return False + return True + + def process(self, lines, file_path, line_numbers=None): + """Check the given lines for style. + + Arguments: + lines: A list of all lines in the file to check. + file_path: The path of the file to process. If possible, the path + should be relative to the source root. Otherwise, + path-specific logic may not behave as expected. + line_numbers: A list of line numbers of the lines for which + style errors should be reported, or None if errors + for all lines should be reported. When not None, this + list normally contains the line numbers corresponding + to the modified lines of a patch. + + """ + _log.debug("Checking style: " + file_path) + + style_error_handler = DefaultStyleErrorHandler( + configuration=self._configuration, + file_path=file_path, + increment_error_count=self._increment_error_count, + line_numbers=line_numbers) + + carriage_checker = self._carriage_checker_class(style_error_handler) + + # Check for and remove trailing carriage returns ("\r"). + if self._dispatcher.should_check_and_strip_carriage_returns(file_path): + lines = carriage_checker.check(lines) + + min_confidence = self._configuration.min_confidence + checker = self._dispatcher.dispatch(file_path, + style_error_handler, + min_confidence) + + if checker is None: + raise AssertionError("File should not be checked: '%s'" % file_path) + + _log.debug("Using class: " + checker.__class__.__name__) + + checker.check(lines) diff --git a/Tools/Scripts/webkitpy/style/checker_unittest.py b/Tools/Scripts/webkitpy/style/checker_unittest.py new file mode 100755 index 0000000..d9057a8 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checker_unittest.py @@ -0,0 +1,832 @@ +#!/usr/bin/python +# -*- coding: utf-8; -*- +# +# Copyright (C) 2009 Google Inc. All rights reserved. +# Copyright (C) 2009 Torch Mobile Inc. +# Copyright (C) 2009 Apple Inc. All rights reserved. +# Copyright (C) 2010 Chris Jerdonek (chris.jerdonek@gmail.com) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit tests for style.py.""" + +import logging +import os +import unittest + +import checker as style +from webkitpy.style_references import LogTesting +from webkitpy.style_references import TestLogStream +from checker import _BASE_FILTER_RULES +from checker import _MAX_REPORTS_PER_CATEGORY +from checker import _PATH_RULES_SPECIFIER as PATH_RULES_SPECIFIER +from checker import _all_categories +from checker import check_webkit_style_configuration +from checker import check_webkit_style_parser +from checker import configure_logging +from checker import CheckerDispatcher +from checker import ProcessorBase +from checker import StyleProcessor +from checker import StyleProcessorConfiguration +from checkers.cpp import CppChecker +from checkers.python import PythonChecker +from checkers.text import TextChecker +from checkers.xml import XMLChecker +from error_handlers import DefaultStyleErrorHandler +from filter import validate_filter_rules +from filter import FilterConfiguration +from optparser import ArgumentParser +from optparser import CommandOptionValues +from webkitpy.common.system.logtesting import LoggingTestCase +from webkitpy.style.filereader import TextFileReader + + +class ConfigureLoggingTestBase(unittest.TestCase): + + """Base class for testing configure_logging(). + + Sub-classes should implement: + + is_verbose: The is_verbose value to pass to configure_logging(). + + """ + + def setUp(self): + is_verbose = self.is_verbose + + log_stream = TestLogStream(self) + + # Use a logger other than the root logger or one prefixed with + # webkit so as not to conflict with test-webkitpy logging. + logger = logging.getLogger("unittest") + + # Configure the test logger not to pass messages along to the + # root logger. This prevents test messages from being + # propagated to loggers used by test-webkitpy logging (e.g. + # the root logger). + logger.propagate = False + + self._handlers = configure_logging(stream=log_stream, logger=logger, + is_verbose=is_verbose) + self._log = logger + self._log_stream = log_stream + + def tearDown(self): + """Reset logging to its original state. + + This method ensures that the logging configuration set up + for a unit test does not affect logging in other unit tests. + + """ + logger = self._log + for handler in self._handlers: + logger.removeHandler(handler) + + def assert_log_messages(self, messages): + """Assert that the logged messages equal the given messages.""" + self._log_stream.assertMessages(messages) + + +class ConfigureLoggingTest(ConfigureLoggingTestBase): + + """Tests the configure_logging() function.""" + + is_verbose = False + + def test_warning_message(self): + self._log.warn("test message") + self.assert_log_messages(["WARNING: test message\n"]) + + def test_below_warning_message(self): + # We test the boundary case of a logging level equal to 29. + # In practice, we will probably only be calling log.info(), + # which corresponds to a logging level of 20. + level = logging.WARNING - 1 # Equals 29. + self._log.log(level, "test message") + self.assert_log_messages(["test message\n"]) + + def test_debug_message(self): + self._log.debug("test message") + self.assert_log_messages([]) + + def test_two_messages(self): + self._log.info("message1") + self._log.info("message2") + self.assert_log_messages(["message1\n", "message2\n"]) + + +class ConfigureLoggingVerboseTest(ConfigureLoggingTestBase): + + """Tests the configure_logging() function with is_verbose True.""" + + is_verbose = True + + def test_debug_message(self): + self._log.debug("test message") + self.assert_log_messages(["unittest: DEBUG test message\n"]) + + +class GlobalVariablesTest(unittest.TestCase): + + """Tests validity of the global variables.""" + + def _all_categories(self): + return _all_categories() + + def defaults(self): + return style._check_webkit_style_defaults() + + def test_webkit_base_filter_rules(self): + base_filter_rules = _BASE_FILTER_RULES + defaults = self.defaults() + already_seen = [] + validate_filter_rules(base_filter_rules, self._all_categories()) + # Also do some additional checks. + for rule in base_filter_rules: + # Check no leading or trailing white space. + self.assertEquals(rule, rule.strip()) + # All categories are on by default, so defaults should + # begin with -. + self.assertTrue(rule.startswith('-')) + # Check no rule occurs twice. + self.assertFalse(rule in already_seen) + already_seen.append(rule) + + def test_defaults(self): + """Check that default arguments are valid.""" + default_options = self.defaults() + + # FIXME: We should not need to call parse() to determine + # whether the default arguments are valid. + parser = ArgumentParser(all_categories=self._all_categories(), + base_filter_rules=[], + default_options=default_options) + # No need to test the return value here since we test parse() + # on valid arguments elsewhere. + # + # The default options are valid: no error or SystemExit. + parser.parse(args=[]) + + def test_path_rules_specifier(self): + all_categories = self._all_categories() + for (sub_paths, path_rules) in PATH_RULES_SPECIFIER: + validate_filter_rules(path_rules, self._all_categories()) + + config = FilterConfiguration(path_specific=PATH_RULES_SPECIFIER) + + def assertCheck(path, category): + """Assert that the given category should be checked.""" + message = ('Should check category "%s" for path "%s".' + % (category, path)) + self.assertTrue(config.should_check(category, path)) + + def assertNoCheck(path, category): + """Assert that the given category should not be checked.""" + message = ('Should not check category "%s" for path "%s".' + % (category, path)) + self.assertFalse(config.should_check(category, path), message) + + assertCheck("random_path.cpp", + "build/include") + assertNoCheck("Tools/WebKitAPITest/main.cpp", + "build/include") + assertCheck("random_path.cpp", + "readability/naming") + assertNoCheck("WebKit/gtk/webkit/webkit.h", + "readability/naming") + assertNoCheck("Tools/DumpRenderTree/gtk/DumpRenderTree.cpp", + "readability/null") + assertNoCheck("WebKit/efl/ewk/ewk_view.h", + "readability/naming") + assertNoCheck("WebCore/css/CSSParser.cpp", + "readability/naming") + + # Test if Qt exceptions are indeed working + assertCheck("JavaScriptCore/qt/api/qscriptengine.cpp", + "readability/braces") + assertCheck("WebKit/qt/Api/qwebpage.cpp", + "readability/braces") + assertCheck("WebKit/qt/tests/qwebelement/tst_qwebelement.cpp", + "readability/braces") + assertCheck("WebKit/qt/declarative/platformplugin/WebPlugin.cpp", + "readability/braces") + assertCheck("WebKit/qt/examples/platformplugin/WebPlugin.cpp", + "readability/braces") + assertNoCheck("JavaScriptCore/qt/api/qscriptengine.cpp", + "readability/naming") + assertNoCheck("WebKit/qt/Api/qwebpage.cpp", + "readability/naming") + assertNoCheck("WebKit/qt/tests/qwebelement/tst_qwebelement.cpp", + "readability/naming") + assertNoCheck("WebKit/qt/declarative/platformplugin/WebPlugin.cpp", + "readability/naming") + assertNoCheck("WebKit/qt/examples/platformplugin/WebPlugin.cpp", + "readability/naming") + + assertNoCheck("WebCore/ForwardingHeaders/debugger/Debugger.h", + "build/header_guard") + + # Third-party Python code: webkitpy/thirdparty + path = "Tools/Scripts/webkitpy/thirdparty/mock.py" + assertNoCheck(path, "build/include") + assertNoCheck(path, "pep8/E401") # A random pep8 category. + assertCheck(path, "pep8/W191") + assertCheck(path, "pep8/W291") + assertCheck(path, "whitespace/carriage_return") + + def test_max_reports_per_category(self): + """Check that _MAX_REPORTS_PER_CATEGORY is valid.""" + all_categories = self._all_categories() + for category in _MAX_REPORTS_PER_CATEGORY.iterkeys(): + self.assertTrue(category in all_categories, + 'Key "%s" is not a category' % category) + + +class CheckWebKitStyleFunctionTest(unittest.TestCase): + + """Tests the functions with names of the form check_webkit_style_*.""" + + def test_check_webkit_style_configuration(self): + # Exercise the code path to make sure the function does not error out. + option_values = CommandOptionValues() + configuration = check_webkit_style_configuration(option_values) + + def test_check_webkit_style_parser(self): + # Exercise the code path to make sure the function does not error out. + parser = check_webkit_style_parser() + + +class CheckerDispatcherSkipTest(unittest.TestCase): + + """Tests the "should skip" methods of the CheckerDispatcher class.""" + + def setUp(self): + self._dispatcher = CheckerDispatcher() + + def test_should_skip_with_warning(self): + """Test should_skip_with_warning().""" + # Check a non-skipped file. + self.assertFalse(self._dispatcher.should_skip_with_warning("foo.txt")) + + # Check skipped files. + paths_to_skip = [ + "gtk2drawing.c", + "gtkdrawing.h", + "WebCore/platform/gtk/gtk2drawing.c", + "WebCore/platform/gtk/gtkdrawing.h", + "WebKit/gtk/tests/testatk.c", + ] + + for path in paths_to_skip: + self.assertTrue(self._dispatcher.should_skip_with_warning(path), + "Checking: " + path) + + def _assert_should_skip_without_warning(self, path, is_checker_none, + expected): + # Check the file type before asserting the return value. + checker = self._dispatcher.dispatch(file_path=path, + handle_style_error=None, + min_confidence=3) + message = 'while checking: %s' % path + self.assertEquals(checker is None, is_checker_none, message) + self.assertEquals(self._dispatcher.should_skip_without_warning(path), + expected, message) + + def test_should_skip_without_warning__true(self): + """Test should_skip_without_warning() for True return values.""" + # Check a file with NONE file type. + path = 'foo.asdf' # Non-sensical file extension. + self._assert_should_skip_without_warning(path, + is_checker_none=True, + expected=True) + + # Check files with non-NONE file type. These examples must be + # drawn from the _SKIPPED_FILES_WITHOUT_WARNING configuration + # variable. + path = os.path.join('LayoutTests', 'foo.txt') + self._assert_should_skip_without_warning(path, + is_checker_none=False, + expected=True) + + def test_should_skip_without_warning__false(self): + """Test should_skip_without_warning() for False return values.""" + paths = ['foo.txt', + os.path.join('LayoutTests', 'ChangeLog'), + ] + + for path in paths: + self._assert_should_skip_without_warning(path, + is_checker_none=False, + expected=False) + + +class CheckerDispatcherCarriageReturnTest(unittest.TestCase): + def test_should_check_and_strip_carriage_returns(self): + files = { + 'foo.txt': True, + 'foo.cpp': True, + 'foo.vcproj': False, + 'foo.vsprops': False, + } + + dispatcher = CheckerDispatcher() + for file_path, expected_result in files.items(): + self.assertEquals(dispatcher.should_check_and_strip_carriage_returns(file_path), expected_result, 'Checking: %s' % file_path) + + +class CheckerDispatcherDispatchTest(unittest.TestCase): + + """Tests dispatch() method of CheckerDispatcher class.""" + + def mock_handle_style_error(self): + pass + + def dispatch(self, file_path): + """Call dispatch() with the given file path.""" + dispatcher = CheckerDispatcher() + checker = dispatcher.dispatch(file_path, + self.mock_handle_style_error, + min_confidence=3) + return checker + + def assert_checker_none(self, file_path): + """Assert that the dispatched checker is None.""" + checker = self.dispatch(file_path) + self.assertTrue(checker is None, 'Checking: "%s"' % file_path) + + def assert_checker(self, file_path, expected_class): + """Assert the type of the dispatched checker.""" + checker = self.dispatch(file_path) + got_class = checker.__class__ + self.assertEquals(got_class, expected_class, + 'For path "%(file_path)s" got %(got_class)s when ' + "expecting %(expected_class)s." + % {"file_path": file_path, + "got_class": got_class, + "expected_class": expected_class}) + + def assert_checker_cpp(self, file_path): + """Assert that the dispatched checker is a CppChecker.""" + self.assert_checker(file_path, CppChecker) + + def assert_checker_python(self, file_path): + """Assert that the dispatched checker is a PythonChecker.""" + self.assert_checker(file_path, PythonChecker) + + def assert_checker_text(self, file_path): + """Assert that the dispatched checker is a TextChecker.""" + self.assert_checker(file_path, TextChecker) + + def assert_checker_xml(self, file_path): + """Assert that the dispatched checker is a XMLChecker.""" + self.assert_checker(file_path, XMLChecker) + + def test_cpp_paths(self): + """Test paths that should be checked as C++.""" + paths = [ + "-", + "foo.c", + "foo.cpp", + "foo.h", + ] + + for path in paths: + self.assert_checker_cpp(path) + + # Check checker attributes on a typical input. + file_base = "foo" + file_extension = "c" + file_path = file_base + "." + file_extension + self.assert_checker_cpp(file_path) + checker = self.dispatch(file_path) + self.assertEquals(checker.file_extension, file_extension) + self.assertEquals(checker.file_path, file_path) + self.assertEquals(checker.handle_style_error, self.mock_handle_style_error) + self.assertEquals(checker.min_confidence, 3) + # Check "-" for good measure. + file_base = "-" + file_extension = "" + file_path = file_base + self.assert_checker_cpp(file_path) + checker = self.dispatch(file_path) + self.assertEquals(checker.file_extension, file_extension) + self.assertEquals(checker.file_path, file_path) + + def test_python_paths(self): + """Test paths that should be checked as Python.""" + paths = [ + "foo.py", + "Tools/Scripts/modules/text_style.py", + ] + + for path in paths: + self.assert_checker_python(path) + + # Check checker attributes on a typical input. + file_base = "foo" + file_extension = "css" + file_path = file_base + "." + file_extension + self.assert_checker_text(file_path) + checker = self.dispatch(file_path) + self.assertEquals(checker.file_path, file_path) + self.assertEquals(checker.handle_style_error, + self.mock_handle_style_error) + + def test_text_paths(self): + """Test paths that should be checked as text.""" + paths = [ + "ChangeLog", + "ChangeLog-2009-06-16", + "foo.ac", + "foo.cc", + "foo.cgi", + "foo.css", + "foo.exp", + "foo.flex", + "foo.gyp", + "foo.gypi", + "foo.html", + "foo.idl", + "foo.in", + "foo.js", + "foo.mm", + "foo.php", + "foo.pl", + "foo.pm", + "foo.pri", + "foo.pro", + "foo.rb", + "foo.sh", + "foo.txt", + "foo.wm", + "foo.xhtml", + "foo.y", + os.path.join("WebCore", "ChangeLog"), + os.path.join("WebCore", "inspector", "front-end", "inspector.js"), + os.path.join("Tools", "Scripts", "check-webkit-style"), + ] + + for path in paths: + self.assert_checker_text(path) + + # Check checker attributes on a typical input. + file_base = "foo" + file_extension = "css" + file_path = file_base + "." + file_extension + self.assert_checker_text(file_path) + checker = self.dispatch(file_path) + self.assertEquals(checker.file_path, file_path) + self.assertEquals(checker.handle_style_error, self.mock_handle_style_error) + + def test_xml_paths(self): + """Test paths that should be checked as XML.""" + paths = [ + "WebCore/WebCore.vcproj/WebCore.vcproj", + "WebKitLibraries/win/tools/vsprops/common.vsprops", + ] + + for path in paths: + self.assert_checker_xml(path) + + # Check checker attributes on a typical input. + file_base = "foo" + file_extension = "vcproj" + file_path = file_base + "." + file_extension + self.assert_checker_xml(file_path) + checker = self.dispatch(file_path) + self.assertEquals(checker.file_path, file_path) + self.assertEquals(checker.handle_style_error, + self.mock_handle_style_error) + + def test_none_paths(self): + """Test paths that have no file type..""" + paths = [ + "Makefile", + "foo.asdf", # Non-sensical file extension. + "foo.png", + "foo.exe", + ] + + for path in paths: + self.assert_checker_none(path) + + +class StyleProcessorConfigurationTest(unittest.TestCase): + + """Tests the StyleProcessorConfiguration class.""" + + def setUp(self): + self._error_messages = [] + """The messages written to _mock_stderr_write() of this class.""" + + def _mock_stderr_write(self, message): + self._error_messages.append(message) + + def _style_checker_configuration(self, output_format="vs7"): + """Return a StyleProcessorConfiguration instance for testing.""" + base_rules = ["-whitespace", "+whitespace/tab"] + filter_configuration = FilterConfiguration(base_rules=base_rules) + + return StyleProcessorConfiguration( + filter_configuration=filter_configuration, + max_reports_per_category={"whitespace/newline": 1}, + min_confidence=3, + output_format=output_format, + stderr_write=self._mock_stderr_write) + + def test_init(self): + """Test the __init__() method.""" + configuration = self._style_checker_configuration() + + # Check that __init__ sets the "public" data attributes correctly. + self.assertEquals(configuration.max_reports_per_category, + {"whitespace/newline": 1}) + self.assertEquals(configuration.stderr_write, self._mock_stderr_write) + self.assertEquals(configuration.min_confidence, 3) + + def test_is_reportable(self): + """Test the is_reportable() method.""" + config = self._style_checker_configuration() + + self.assertTrue(config.is_reportable("whitespace/tab", 3, "foo.txt")) + + # Test the confidence check code path by varying the confidence. + self.assertFalse(config.is_reportable("whitespace/tab", 2, "foo.txt")) + + # Test the category check code path by varying the category. + self.assertFalse(config.is_reportable("whitespace/line", 4, "foo.txt")) + + def _call_write_style_error(self, output_format): + config = self._style_checker_configuration(output_format=output_format) + config.write_style_error(category="whitespace/tab", + confidence_in_error=5, + file_path="foo.h", + line_number=100, + message="message") + + def test_write_style_error_emacs(self): + """Test the write_style_error() method.""" + self._call_write_style_error("emacs") + self.assertEquals(self._error_messages, + ["foo.h:100: message [whitespace/tab] [5]\n"]) + + def test_write_style_error_vs7(self): + """Test the write_style_error() method.""" + self._call_write_style_error("vs7") + self.assertEquals(self._error_messages, + ["foo.h(100): message [whitespace/tab] [5]\n"]) + + +class StyleProcessor_EndToEndTest(LoggingTestCase): + + """Test the StyleProcessor class with an emphasis on end-to-end tests.""" + + def setUp(self): + LoggingTestCase.setUp(self) + self._messages = [] + + def _mock_stderr_write(self, message): + """Save a message so it can later be asserted.""" + self._messages.append(message) + + def test_init(self): + """Test __init__ constructor.""" + configuration = StyleProcessorConfiguration( + filter_configuration=FilterConfiguration(), + max_reports_per_category={}, + min_confidence=3, + output_format="vs7", + stderr_write=self._mock_stderr_write) + processor = StyleProcessor(configuration) + + self.assertEquals(processor.error_count, 0) + self.assertEquals(self._messages, []) + + def test_process(self): + configuration = StyleProcessorConfiguration( + filter_configuration=FilterConfiguration(), + max_reports_per_category={}, + min_confidence=3, + output_format="vs7", + stderr_write=self._mock_stderr_write) + processor = StyleProcessor(configuration) + + processor.process(lines=['line1', 'Line with tab:\t'], + file_path='foo.txt') + self.assertEquals(processor.error_count, 1) + expected_messages = ['foo.txt(2): Line contains tab character. ' + '[whitespace/tab] [5]\n'] + self.assertEquals(self._messages, expected_messages) + + +class StyleProcessor_CodeCoverageTest(LoggingTestCase): + + """Test the StyleProcessor class with an emphasis on code coverage. + + This class makes heavy use of mock objects. + + """ + + class MockDispatchedChecker(object): + + """A mock checker dispatched by the MockDispatcher.""" + + def __init__(self, file_path, min_confidence, style_error_handler): + self.file_path = file_path + self.min_confidence = min_confidence + self.style_error_handler = style_error_handler + + def check(self, lines): + self.lines = lines + + class MockDispatcher(object): + + """A mock CheckerDispatcher class.""" + + def __init__(self): + self.dispatched_checker = None + + def should_skip_with_warning(self, file_path): + return file_path.endswith('skip_with_warning.txt') + + def should_skip_without_warning(self, file_path): + return file_path.endswith('skip_without_warning.txt') + + def should_check_and_strip_carriage_returns(self, file_path): + return not file_path.endswith('carriage_returns_allowed.txt') + + def dispatch(self, file_path, style_error_handler, min_confidence): + if file_path.endswith('do_not_process.txt'): + return None + + checker = StyleProcessor_CodeCoverageTest.MockDispatchedChecker( + file_path, + min_confidence, + style_error_handler) + + # Save the dispatched checker so the current test case has a + # way to access and check it. + self.dispatched_checker = checker + + return checker + + def setUp(self): + LoggingTestCase.setUp(self) + # We can pass an error-message swallower here because error message + # output is tested instead in the end-to-end test case above. + configuration = StyleProcessorConfiguration( + filter_configuration=FilterConfiguration(), + max_reports_per_category={"whitespace/newline": 1}, + min_confidence=3, + output_format="vs7", + stderr_write=self._swallow_stderr_message) + + mock_carriage_checker_class = self._create_carriage_checker_class() + mock_dispatcher = self.MockDispatcher() + # We do not need to use a real incrementer here because error-count + # incrementing is tested instead in the end-to-end test case above. + mock_increment_error_count = self._do_nothing + + processor = StyleProcessor(configuration=configuration, + mock_carriage_checker_class=mock_carriage_checker_class, + mock_dispatcher=mock_dispatcher, + mock_increment_error_count=mock_increment_error_count) + + self._configuration = configuration + self._mock_dispatcher = mock_dispatcher + self._processor = processor + + def _do_nothing(self): + # We provide this function so the caller can pass it to the + # StyleProcessor constructor. This lets us assert the equality of + # the DefaultStyleErrorHandler instance generated by the process() + # method with an expected instance. + pass + + def _swallow_stderr_message(self, message): + """Swallow a message passed to stderr.write().""" + # This is a mock stderr.write() for passing to the constructor + # of the StyleProcessorConfiguration class. + pass + + def _create_carriage_checker_class(self): + + # Create a reference to self with a new name so its name does not + # conflict with the self introduced below. + test_case = self + + class MockCarriageChecker(object): + + """A mock carriage-return checker.""" + + def __init__(self, style_error_handler): + self.style_error_handler = style_error_handler + + # This gives the current test case access to the + # instantiated carriage checker. + test_case.carriage_checker = self + + def check(self, lines): + # Save the lines so the current test case has a way to access + # and check them. + self.lines = lines + + return lines + + return MockCarriageChecker + + def test_should_process__skip_without_warning(self): + """Test should_process() for a skip-without-warning file.""" + file_path = "foo/skip_without_warning.txt" + + self.assertFalse(self._processor.should_process(file_path)) + + def test_should_process__skip_with_warning(self): + """Test should_process() for a skip-with-warning file.""" + file_path = "foo/skip_with_warning.txt" + + self.assertFalse(self._processor.should_process(file_path)) + + self.assertLog(['WARNING: File exempt from style guide. ' + 'Skipping: "foo/skip_with_warning.txt"\n']) + + def test_should_process__true_result(self): + """Test should_process() for a file that should be processed.""" + file_path = "foo/skip_process.txt" + + self.assertTrue(self._processor.should_process(file_path)) + + def test_process__checker_dispatched(self): + """Test the process() method for a path with a dispatched checker.""" + file_path = 'foo.txt' + lines = ['line1', 'line2'] + line_numbers = [100] + + expected_error_handler = DefaultStyleErrorHandler( + configuration=self._configuration, + file_path=file_path, + increment_error_count=self._do_nothing, + line_numbers=line_numbers) + + self._processor.process(lines=lines, + file_path=file_path, + line_numbers=line_numbers) + + # Check that the carriage-return checker was instantiated correctly + # and was passed lines correctly. + carriage_checker = self.carriage_checker + self.assertEquals(carriage_checker.style_error_handler, + expected_error_handler) + self.assertEquals(carriage_checker.lines, ['line1', 'line2']) + + # Check that the style checker was dispatched correctly and was + # passed lines correctly. + checker = self._mock_dispatcher.dispatched_checker + self.assertEquals(checker.file_path, 'foo.txt') + self.assertEquals(checker.min_confidence, 3) + self.assertEquals(checker.style_error_handler, expected_error_handler) + + self.assertEquals(checker.lines, ['line1', 'line2']) + + def test_process__no_checker_dispatched(self): + """Test the process() method for a path with no dispatched checker.""" + path = os.path.join('foo', 'do_not_process.txt') + self.assertRaises(AssertionError, self._processor.process, + lines=['line1', 'line2'], file_path=path, + line_numbers=[100]) + + def test_process__carriage_returns_not_stripped(self): + """Test that carriage returns aren't stripped from files that are allowed to contain them.""" + file_path = 'carriage_returns_allowed.txt' + lines = ['line1\r', 'line2\r'] + line_numbers = [100] + self._processor.process(lines=lines, + file_path=file_path, + line_numbers=line_numbers) + # The carriage return checker should never have been invoked, and so + # should not have saved off any lines. + self.assertFalse(hasattr(self.carriage_checker, 'lines')) diff --git a/Tools/Scripts/webkitpy/style/checkers/__init__.py b/Tools/Scripts/webkitpy/style/checkers/__init__.py new file mode 100644 index 0000000..ef65bee --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/__init__.py @@ -0,0 +1 @@ +# Required for Python to search this directory for module files diff --git a/Tools/Scripts/webkitpy/style/checkers/common.py b/Tools/Scripts/webkitpy/style/checkers/common.py new file mode 100644 index 0000000..76aa956 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/common.py @@ -0,0 +1,74 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY +# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Supports style checking not specific to any one file type.""" + + +# FIXME: Test this list in the same way that the list of CppChecker +# categories is tested, for example by checking that all of its +# elements appear in the unit tests. This should probably be done +# after moving the relevant cpp_unittest.ErrorCollector code +# into a shared location and refactoring appropriately. +categories = set([ + "whitespace/carriage_return", + "whitespace/tab"]) + + +class CarriageReturnChecker(object): + + """Supports checking for and handling carriage returns.""" + + def __init__(self, handle_style_error): + self._handle_style_error = handle_style_error + + def check(self, lines): + """Check for and strip trailing carriage returns from lines.""" + for line_number in range(len(lines)): + if not lines[line_number].endswith("\r"): + continue + + self._handle_style_error(line_number + 1, # Correct for offset. + "whitespace/carriage_return", + 1, + "One or more unexpected \\r (^M) found; " + "better to use only a \\n") + + lines[line_number] = lines[line_number].rstrip("\r") + + return lines + + +class TabChecker(object): + + """Supports checking for and handling tabs.""" + + def __init__(self, file_path, handle_style_error): + self.file_path = file_path + self.handle_style_error = handle_style_error + + def check(self, lines): + # FIXME: share with cpp_style. + for line_number, line in enumerate(lines): + if "\t" in line: + self.handle_style_error(line_number + 1, + "whitespace/tab", 5, + "Line contains tab character.") diff --git a/Tools/Scripts/webkitpy/style/checkers/common_unittest.py b/Tools/Scripts/webkitpy/style/checkers/common_unittest.py new file mode 100644 index 0000000..1fe1263 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/common_unittest.py @@ -0,0 +1,124 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY +# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit tests for common.py.""" + +import unittest + +from common import CarriageReturnChecker +from common import TabChecker + +# FIXME: The unit tests for the cpp, text, and common checkers should +# share supporting test code. This can include, for example, the +# mock style error handling code and the code to check that all +# of a checker's categories are covered by the unit tests. +# Such shared code can be located in a shared test file, perhaps +# even this file. +class CarriageReturnCheckerTest(unittest.TestCase): + + """Tests check_no_carriage_return().""" + + _category = "whitespace/carriage_return" + _confidence = 1 + _expected_message = ("One or more unexpected \\r (^M) found; " + "better to use only a \\n") + + def setUp(self): + self._style_errors = [] # The list of accumulated style errors. + + def _mock_style_error_handler(self, line_number, category, confidence, + message): + """Append the error information to the list of style errors.""" + error = (line_number, category, confidence, message) + self._style_errors.append(error) + + def assert_carriage_return(self, input_lines, expected_lines, error_lines): + """Process the given line and assert that the result is correct.""" + handle_style_error = self._mock_style_error_handler + + checker = CarriageReturnChecker(handle_style_error) + output_lines = checker.check(input_lines) + + # Check both the return value and error messages. + self.assertEquals(output_lines, expected_lines) + + expected_errors = [(line_number, self._category, self._confidence, + self._expected_message) + for line_number in error_lines] + self.assertEquals(self._style_errors, expected_errors) + + def test_ends_with_carriage(self): + self.assert_carriage_return(["carriage return\r"], + ["carriage return"], + [1]) + + def test_ends_with_nothing(self): + self.assert_carriage_return(["no carriage return"], + ["no carriage return"], + []) + + def test_ends_with_newline(self): + self.assert_carriage_return(["no carriage return\n"], + ["no carriage return\n"], + []) + + def test_carriage_in_middle(self): + # The CarriageReturnChecker checks only the final character + # of each line. + self.assert_carriage_return(["carriage\r in a string"], + ["carriage\r in a string"], + []) + + def test_multiple_errors(self): + self.assert_carriage_return(["line1", "line2\r", "line3\r"], + ["line1", "line2", "line3"], + [2, 3]) + + +class TabCheckerTest(unittest.TestCase): + + """Tests for TabChecker.""" + + def assert_tab(self, input_lines, error_lines): + """Assert when the given lines contain tabs.""" + self._error_lines = [] + + def style_error_handler(line_number, category, confidence, message): + self.assertEqual(category, 'whitespace/tab') + self.assertEqual(confidence, 5) + self.assertEqual(message, 'Line contains tab character.') + self._error_lines.append(line_number) + + checker = TabChecker('', style_error_handler) + checker.check(input_lines) + self.assertEquals(self._error_lines, error_lines) + + def test_notab(self): + self.assert_tab([''], []) + self.assert_tab(['foo', 'bar'], []) + + def test_tab(self): + self.assert_tab(['\tfoo'], [1]) + self.assert_tab(['line1', '\tline2', 'line3\t'], [2, 3]) + +if __name__ == '__main__': + unittest.main() diff --git a/Tools/Scripts/webkitpy/style/checkers/cpp.py b/Tools/Scripts/webkitpy/style/checkers/cpp.py new file mode 100644 index 0000000..94e5bdd --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/cpp.py @@ -0,0 +1,3171 @@ +#!/usr/bin/python +# -*- coding: utf-8 -*- +# +# Copyright (C) 2009 Google Inc. All rights reserved. +# Copyright (C) 2009 Torch Mobile Inc. +# Copyright (C) 2009 Apple Inc. All rights reserved. +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +# This is the modified version of Google's cpplint. The original code is +# http://google-styleguide.googlecode.com/svn/trunk/cpplint/cpplint.py + +"""Support for check-webkit-style.""" + +import codecs +import math # for log +import os +import os.path +import re +import sre_compile +import string +import sys +import unicodedata + +# The key to use to provide a class to fake loading a header file. +INCLUDE_IO_INJECTION_KEY = 'include_header_io' + +# Headers that we consider STL headers. +_STL_HEADERS = frozenset([ + 'algobase.h', 'algorithm', 'alloc.h', 'bitset', 'deque', 'exception', + 'function.h', 'functional', 'hash_map', 'hash_map.h', 'hash_set', + 'hash_set.h', 'iterator', 'list', 'list.h', 'map', 'memory', 'pair.h', + 'pthread_alloc', 'queue', 'set', 'set.h', 'sstream', 'stack', + 'stl_alloc.h', 'stl_relops.h', 'type_traits.h', + 'utility', 'vector', 'vector.h', + ]) + + +# Non-STL C++ system headers. +_CPP_HEADERS = frozenset([ + 'algo.h', 'builtinbuf.h', 'bvector.h', 'cassert', 'cctype', + 'cerrno', 'cfloat', 'ciso646', 'climits', 'clocale', 'cmath', + 'complex', 'complex.h', 'csetjmp', 'csignal', 'cstdarg', 'cstddef', + 'cstdio', 'cstdlib', 'cstring', 'ctime', 'cwchar', 'cwctype', + 'defalloc.h', 'deque.h', 'editbuf.h', 'exception', 'fstream', + 'fstream.h', 'hashtable.h', 'heap.h', 'indstream.h', 'iomanip', + 'iomanip.h', 'ios', 'iosfwd', 'iostream', 'iostream.h', 'istream.h', + 'iterator.h', 'limits', 'map.h', 'multimap.h', 'multiset.h', + 'numeric', 'ostream.h', 'parsestream.h', 'pfstream.h', 'PlotFile.h', + 'procbuf.h', 'pthread_alloc.h', 'rope', 'rope.h', 'ropeimpl.h', + 'SFile.h', 'slist', 'slist.h', 'stack.h', 'stdexcept', + 'stdiostream.h', 'streambuf.h', 'stream.h', 'strfile.h', 'string', + 'strstream', 'strstream.h', 'tempbuf.h', 'tree.h', 'typeinfo', 'valarray', + ]) + + +# Assertion macros. These are defined in base/logging.h and +# testing/base/gunit.h. Note that the _M versions need to come first +# for substring matching to work. +_CHECK_MACROS = [ + 'DCHECK', 'CHECK', + 'EXPECT_TRUE_M', 'EXPECT_TRUE', + 'ASSERT_TRUE_M', 'ASSERT_TRUE', + 'EXPECT_FALSE_M', 'EXPECT_FALSE', + 'ASSERT_FALSE_M', 'ASSERT_FALSE', + ] + +# Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE +_CHECK_REPLACEMENT = dict([(m, {}) for m in _CHECK_MACROS]) + +for op, replacement in [('==', 'EQ'), ('!=', 'NE'), + ('>=', 'GE'), ('>', 'GT'), + ('<=', 'LE'), ('<', 'LT')]: + _CHECK_REPLACEMENT['DCHECK'][op] = 'DCHECK_%s' % replacement + _CHECK_REPLACEMENT['CHECK'][op] = 'CHECK_%s' % replacement + _CHECK_REPLACEMENT['EXPECT_TRUE'][op] = 'EXPECT_%s' % replacement + _CHECK_REPLACEMENT['ASSERT_TRUE'][op] = 'ASSERT_%s' % replacement + _CHECK_REPLACEMENT['EXPECT_TRUE_M'][op] = 'EXPECT_%s_M' % replacement + _CHECK_REPLACEMENT['ASSERT_TRUE_M'][op] = 'ASSERT_%s_M' % replacement + +for op, inv_replacement in [('==', 'NE'), ('!=', 'EQ'), + ('>=', 'LT'), ('>', 'LE'), + ('<=', 'GT'), ('<', 'GE')]: + _CHECK_REPLACEMENT['EXPECT_FALSE'][op] = 'EXPECT_%s' % inv_replacement + _CHECK_REPLACEMENT['ASSERT_FALSE'][op] = 'ASSERT_%s' % inv_replacement + _CHECK_REPLACEMENT['EXPECT_FALSE_M'][op] = 'EXPECT_%s_M' % inv_replacement + _CHECK_REPLACEMENT['ASSERT_FALSE_M'][op] = 'ASSERT_%s_M' % inv_replacement + + +# These constants define types of headers for use with +# _IncludeState.check_next_include_order(). +_CONFIG_HEADER = 0 +_PRIMARY_HEADER = 1 +_OTHER_HEADER = 2 +_MOC_HEADER = 3 + + +# A dictionary of items customize behavior for unit test. For example, +# INCLUDE_IO_INJECTION_KEY allows providing a custom io class which allows +# for faking a header file. +_unit_test_config = {} + + +# The regexp compilation caching is inlined in all regexp functions for +# performance reasons; factoring it out into a separate function turns out +# to be noticeably expensive. +_regexp_compile_cache = {} + + +def match(pattern, s): + """Matches the string with the pattern, caching the compiled regexp.""" + if not pattern in _regexp_compile_cache: + _regexp_compile_cache[pattern] = sre_compile.compile(pattern) + return _regexp_compile_cache[pattern].match(s) + + +def search(pattern, s): + """Searches the string for the pattern, caching the compiled regexp.""" + if not pattern in _regexp_compile_cache: + _regexp_compile_cache[pattern] = sre_compile.compile(pattern) + return _regexp_compile_cache[pattern].search(s) + + +def sub(pattern, replacement, s): + """Substitutes occurrences of a pattern, caching the compiled regexp.""" + if not pattern in _regexp_compile_cache: + _regexp_compile_cache[pattern] = sre_compile.compile(pattern) + return _regexp_compile_cache[pattern].sub(replacement, s) + + +def subn(pattern, replacement, s): + """Substitutes occurrences of a pattern, caching the compiled regexp.""" + if not pattern in _regexp_compile_cache: + _regexp_compile_cache[pattern] = sre_compile.compile(pattern) + return _regexp_compile_cache[pattern].subn(replacement, s) + + +def iteratively_replace_matches_with_char(pattern, char_replacement, s): + """Returns the string with replacement done. + + Every character in the match is replaced with char. + Due to the iterative nature, pattern should not match char or + there will be an infinite loop. + + Example: + pattern = r'<[^>]>' # template parameters + char_replacement = '_' + s = 'A<B<C, D>>' + Returns 'A_________' + + Args: + pattern: The regex to match. + char_replacement: The character to put in place of every + character of the match. + s: The string on which to do the replacements. + + Returns: + True, if the given line is blank. + """ + while True: + matched = search(pattern, s) + if not matched: + return s + start_match_index = matched.start(0) + end_match_index = matched.end(0) + match_length = end_match_index - start_match_index + s = s[:start_match_index] + char_replacement * match_length + s[end_match_index:] + + +def up_to_unmatched_closing_paren(s): + """Splits a string into two parts up to first unmatched ')'. + + Args: + s: a string which is a substring of line after '(' + (e.g., "a == (b + c))"). + + Returns: + A pair of strings (prefix before first unmatched ')', + remainder of s after first unmatched ')'), e.g., + up_to_unmatched_closing_paren("a == (b + c)) { ") + returns "a == (b + c)", " {". + Returns None, None if there is no unmatched ')' + + """ + i = 1 + for pos, c in enumerate(s): + if c == '(': + i += 1 + elif c == ')': + i -= 1 + if i == 0: + return s[:pos], s[pos + 1:] + return None, None + +class _IncludeState(dict): + """Tracks line numbers for includes, and the order in which includes appear. + + As a dict, an _IncludeState object serves as a mapping between include + filename and line number on which that file was included. + + Call check_next_include_order() once for each header in the file, passing + in the type constants defined above. Calls in an illegal order will + raise an _IncludeError with an appropriate error message. + + """ + # self._section will move monotonically through this set. If it ever + # needs to move backwards, check_next_include_order will raise an error. + _INITIAL_SECTION = 0 + _CONFIG_SECTION = 1 + _PRIMARY_SECTION = 2 + _OTHER_SECTION = 3 + + _TYPE_NAMES = { + _CONFIG_HEADER: 'WebCore config.h', + _PRIMARY_HEADER: 'header this file implements', + _OTHER_HEADER: 'other header', + _MOC_HEADER: 'moc file', + } + _SECTION_NAMES = { + _INITIAL_SECTION: "... nothing.", + _CONFIG_SECTION: "WebCore config.h.", + _PRIMARY_SECTION: 'a header this file implements.', + _OTHER_SECTION: 'other header.', + } + + def __init__(self): + dict.__init__(self) + self._section = self._INITIAL_SECTION + self._visited_primary_section = False + self.header_types = dict(); + + def visited_primary_section(self): + return self._visited_primary_section + + def check_next_include_order(self, header_type, file_is_header): + """Returns a non-empty error message if the next header is out of order. + + This function also updates the internal state to be ready to check + the next include. + + Args: + header_type: One of the _XXX_HEADER constants defined above. + file_is_header: Whether the file that owns this _IncludeState is itself a header + + Returns: + The empty string if the header is in the right order, or an + error message describing what's wrong. + + """ + if header_type == _CONFIG_HEADER and file_is_header: + return 'Header file should not contain WebCore config.h.' + if header_type == _PRIMARY_HEADER and file_is_header: + return 'Header file should not contain itself.' + if header_type == _MOC_HEADER: + return '' + + error_message = '' + if self._section != self._OTHER_SECTION: + before_error_message = ('Found %s before %s' % + (self._TYPE_NAMES[header_type], + self._SECTION_NAMES[self._section + 1])) + after_error_message = ('Found %s after %s' % + (self._TYPE_NAMES[header_type], + self._SECTION_NAMES[self._section])) + + if header_type == _CONFIG_HEADER: + if self._section >= self._CONFIG_SECTION: + error_message = after_error_message + self._section = self._CONFIG_SECTION + elif header_type == _PRIMARY_HEADER: + if self._section >= self._PRIMARY_SECTION: + error_message = after_error_message + elif self._section < self._CONFIG_SECTION: + error_message = before_error_message + self._section = self._PRIMARY_SECTION + self._visited_primary_section = True + else: + assert header_type == _OTHER_HEADER + if not file_is_header and self._section < self._PRIMARY_SECTION: + error_message = before_error_message + self._section = self._OTHER_SECTION + + return error_message + + +class _FunctionState(object): + """Tracks current function name and the number of lines in its body. + + Attributes: + min_confidence: The minimum confidence level to use while checking style. + + """ + + _NORMAL_TRIGGER = 250 # for --v=0, 500 for --v=1, etc. + _TEST_TRIGGER = 400 # about 50% more than _NORMAL_TRIGGER. + + def __init__(self, min_confidence): + self.min_confidence = min_confidence + self.current_function = '' + self.in_a_function = False + self.lines_in_function = 0 + # Make sure these will not be mistaken for real lines (even when a + # small amount is added to them). + self.body_start_line_number = -1000 + self.ending_line_number = -1000 + + def begin(self, function_name, body_start_line_number, ending_line_number, is_declaration): + """Start analyzing function body. + + Args: + function_name: The name of the function being tracked. + body_start_line_number: The line number of the { or the ; for a protoype. + ending_line_number: The line number where the function ends. + is_declaration: True if this is a prototype. + """ + self.in_a_function = True + self.lines_in_function = -1 # Don't count the open brace line. + self.current_function = function_name + self.body_start_line_number = body_start_line_number + self.ending_line_number = ending_line_number + self.is_declaration = is_declaration + + def count(self, line_number): + """Count line in current function body.""" + if self.in_a_function and line_number >= self.body_start_line_number: + self.lines_in_function += 1 + + def check(self, error, line_number): + """Report if too many lines in function body. + + Args: + error: The function to call with any errors found. + line_number: The number of the line to check. + """ + if match(r'T(EST|est)', self.current_function): + base_trigger = self._TEST_TRIGGER + else: + base_trigger = self._NORMAL_TRIGGER + trigger = base_trigger * 2 ** self.min_confidence + + if self.lines_in_function > trigger: + error_level = int(math.log(self.lines_in_function / base_trigger, 2)) + # 50 => 0, 100 => 1, 200 => 2, 400 => 3, 800 => 4, 1600 => 5, ... + if error_level > 5: + error_level = 5 + error(line_number, 'readability/fn_size', error_level, + 'Small and focused functions are preferred:' + ' %s has %d non-comment lines' + ' (error triggered by exceeding %d lines).' % ( + self.current_function, self.lines_in_function, trigger)) + + def end(self): + """Stop analyzing function body.""" + self.in_a_function = False + + +class _IncludeError(Exception): + """Indicates a problem with the include order in a file.""" + pass + + +class FileInfo: + """Provides utility functions for filenames. + + FileInfo provides easy access to the components of a file's path + relative to the project root. + """ + + def __init__(self, filename): + self._filename = filename + + def full_name(self): + """Make Windows paths like Unix.""" + return os.path.abspath(self._filename).replace('\\', '/') + + def repository_name(self): + """Full name after removing the local path to the repository. + + If we have a real absolute path name here we can try to do something smart: + detecting the root of the checkout and truncating /path/to/checkout from + the name so that we get header guards that don't include things like + "C:\Documents and Settings\..." or "/home/username/..." in them and thus + people on different computers who have checked the source out to different + locations won't see bogus errors. + """ + fullname = self.full_name() + + if os.path.exists(fullname): + project_dir = os.path.dirname(fullname) + + if os.path.exists(os.path.join(project_dir, ".svn")): + # If there's a .svn file in the current directory, we + # recursively look up the directory tree for the top + # of the SVN checkout + root_dir = project_dir + one_up_dir = os.path.dirname(root_dir) + while os.path.exists(os.path.join(one_up_dir, ".svn")): + root_dir = os.path.dirname(root_dir) + one_up_dir = os.path.dirname(one_up_dir) + + prefix = os.path.commonprefix([root_dir, project_dir]) + return fullname[len(prefix) + 1:] + + # Not SVN? Try to find a git top level directory by + # searching up from the current path. + root_dir = os.path.dirname(fullname) + while (root_dir != os.path.dirname(root_dir) + and not os.path.exists(os.path.join(root_dir, ".git"))): + root_dir = os.path.dirname(root_dir) + if os.path.exists(os.path.join(root_dir, ".git")): + prefix = os.path.commonprefix([root_dir, project_dir]) + return fullname[len(prefix) + 1:] + + # Don't know what to do; header guard warnings may be wrong... + return fullname + + def split(self): + """Splits the file into the directory, basename, and extension. + + For 'chrome/browser/browser.cpp', Split() would + return ('chrome/browser', 'browser', '.cpp') + + Returns: + A tuple of (directory, basename, extension). + """ + + googlename = self.repository_name() + project, rest = os.path.split(googlename) + return (project,) + os.path.splitext(rest) + + def base_name(self): + """File base name - text after the final slash, before the final period.""" + return self.split()[1] + + def extension(self): + """File extension - text following the final period.""" + return self.split()[2] + + def no_extension(self): + """File has no source file extension.""" + return '/'.join(self.split()[0:2]) + + def is_source(self): + """File has a source file extension.""" + return self.extension()[1:] in ('c', 'cc', 'cpp', 'cxx') + + +# Matches standard C++ escape esequences per 2.13.2.3 of the C++ standard. +_RE_PATTERN_CLEANSE_LINE_ESCAPES = re.compile( + r'\\([abfnrtv?"\\\']|\d+|x[0-9a-fA-F]+)') +# Matches strings. Escape codes should already be removed by ESCAPES. +_RE_PATTERN_CLEANSE_LINE_DOUBLE_QUOTES = re.compile(r'"[^"]*"') +# Matches characters. Escape codes should already be removed by ESCAPES. +_RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES = re.compile(r"'.'") +# Matches multi-line C++ comments. +# This RE is a little bit more complicated than one might expect, because we +# have to take care of space removals tools so we can handle comments inside +# statements better. +# The current rule is: We only clear spaces from both sides when we're at the +# end of the line. Otherwise, we try to remove spaces from the right side, +# if this doesn't work we try on left side but only if there's a non-character +# on the right. +_RE_PATTERN_CLEANSE_LINE_C_COMMENTS = re.compile( + r"""(\s*/\*.*\*/\s*$| + /\*.*\*/\s+| + \s+/\*.*\*/(?=\W)| + /\*.*\*/)""", re.VERBOSE) + + +def is_cpp_string(line): + """Does line terminate so, that the next symbol is in string constant. + + This function does not consider single-line nor multi-line comments. + + Args: + line: is a partial line of code starting from the 0..n. + + Returns: + True, if next character appended to 'line' is inside a + string constant. + """ + + line = line.replace(r'\\', 'XX') # after this, \\" does not match to \" + return ((line.count('"') - line.count(r'\"') - line.count("'\"'")) & 1) == 1 + + +def find_next_multi_line_comment_start(lines, line_index): + """Find the beginning marker for a multiline comment.""" + while line_index < len(lines): + if lines[line_index].strip().startswith('/*'): + # Only return this marker if the comment goes beyond this line + if lines[line_index].strip().find('*/', 2) < 0: + return line_index + line_index += 1 + return len(lines) + + +def find_next_multi_line_comment_end(lines, line_index): + """We are inside a comment, find the end marker.""" + while line_index < len(lines): + if lines[line_index].strip().endswith('*/'): + return line_index + line_index += 1 + return len(lines) + + +def remove_multi_line_comments_from_range(lines, begin, end): + """Clears a range of lines for multi-line comments.""" + # Having // dummy comments makes the lines non-empty, so we will not get + # unnecessary blank line warnings later in the code. + for i in range(begin, end): + lines[i] = '// dummy' + + +def remove_multi_line_comments(lines, error): + """Removes multiline (c-style) comments from lines.""" + line_index = 0 + while line_index < len(lines): + line_index_begin = find_next_multi_line_comment_start(lines, line_index) + if line_index_begin >= len(lines): + return + line_index_end = find_next_multi_line_comment_end(lines, line_index_begin) + if line_index_end >= len(lines): + error(line_index_begin + 1, 'readability/multiline_comment', 5, + 'Could not find end of multi-line comment') + return + remove_multi_line_comments_from_range(lines, line_index_begin, line_index_end + 1) + line_index = line_index_end + 1 + + +def cleanse_comments(line): + """Removes //-comments and single-line C-style /* */ comments. + + Args: + line: A line of C++ source. + + Returns: + The line with single-line comments removed. + """ + comment_position = line.find('//') + if comment_position != -1 and not is_cpp_string(line[:comment_position]): + line = line[:comment_position] + # get rid of /* ... */ + return _RE_PATTERN_CLEANSE_LINE_C_COMMENTS.sub('', line) + + +class CleansedLines(object): + """Holds 3 copies of all lines with different preprocessing applied to them. + + 1) elided member contains lines without strings and comments, + 2) lines member contains lines without comments, and + 3) raw member contains all the lines without processing. + All these three members are of <type 'list'>, and of the same length. + """ + + def __init__(self, lines): + self.elided = [] + self.lines = [] + self.raw_lines = lines + self._num_lines = len(lines) + for line_number in range(len(lines)): + self.lines.append(cleanse_comments(lines[line_number])) + elided = self.collapse_strings(lines[line_number]) + self.elided.append(cleanse_comments(elided)) + + def num_lines(self): + """Returns the number of lines represented.""" + return self._num_lines + + @staticmethod + def collapse_strings(elided): + """Collapses strings and chars on a line to simple "" or '' blocks. + + We nix strings first so we're not fooled by text like '"http://"' + + Args: + elided: The line being processed. + + Returns: + The line with collapsed strings. + """ + if not _RE_PATTERN_INCLUDE.match(elided): + # Remove escaped characters first to make quote/single quote collapsing + # basic. Things that look like escaped characters shouldn't occur + # outside of strings and chars. + elided = _RE_PATTERN_CLEANSE_LINE_ESCAPES.sub('', elided) + elided = _RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES.sub("''", elided) + elided = _RE_PATTERN_CLEANSE_LINE_DOUBLE_QUOTES.sub('""', elided) + return elided + + +def close_expression(clean_lines, line_number, pos): + """If input points to ( or { or [, finds the position that closes it. + + If clean_lines.elided[line_number][pos] points to a '(' or '{' or '[', finds + the line_number/pos that correspond to the closing of the expression. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + pos: A position on the line. + + Returns: + A tuple (line, line_number, pos) pointer *past* the closing brace, or + ('', len(clean_lines.elided), -1) if we never find a close. Note we + ignore strings and comments when matching; and the line we return is the + 'cleansed' line at line_number. + """ + + line = clean_lines.elided[line_number] + start_character = line[pos] + if start_character not in '({[': + return (line, clean_lines.num_lines(), -1) + if start_character == '(': + end_character = ')' + if start_character == '[': + end_character = ']' + if start_character == '{': + end_character = '}' + + num_open = line.count(start_character) - line.count(end_character) + while num_open > 0: + line_number += 1 + if line_number >= clean_lines.num_lines(): + return ('', len(clean_lines.elided), -1) + line = clean_lines.elided[line_number] + num_open += line.count(start_character) - line.count(end_character) + # OK, now find the end_character that actually got us back to even + endpos = len(line) + while num_open >= 0: + endpos = line.rfind(')', 0, endpos) + num_open -= 1 # chopped off another ) + return (line, line_number, endpos + 1) + + +def check_for_copyright(lines, error): + """Logs an error if no Copyright message appears at the top of the file.""" + + # We'll say it should occur by line 10. Don't forget there's a + # dummy line at the front. + for line in xrange(1, min(len(lines), 11)): + if re.search(r'Copyright', lines[line], re.I): + break + else: # means no copyright line was found + error(0, 'legal/copyright', 5, + 'No copyright message found. ' + 'You should have a line: "Copyright [year] <Copyright Owner>"') + + +def get_header_guard_cpp_variable(filename): + """Returns the CPP variable that should be used as a header guard. + + Args: + filename: The name of a C++ header file. + + Returns: + The CPP variable that should be used as a header guard in the + named file. + + """ + + # Restores original filename in case that style checker is invoked from Emacs's + # flymake. + filename = re.sub(r'_flymake\.h$', '.h', filename) + + standard_name = sub(r'[-.\s]', '_', os.path.basename(filename)) + + # Files under WTF typically have header guards that start with WTF_. + if filename.find('/wtf/'): + special_name = "WTF_" + standard_name + else: + special_name = standard_name + return (special_name, standard_name) + + +def check_for_header_guard(filename, lines, error): + """Checks that the file contains a header guard. + + Logs an error if no #ifndef header guard is present. For other + headers, checks that the full pathname is used. + + Args: + filename: The name of the C++ header file. + lines: An array of strings, each representing a line of the file. + error: The function to call with any errors found. + """ + + cppvar = get_header_guard_cpp_variable(filename) + + ifndef = None + ifndef_line_number = 0 + define = None + for line_number, line in enumerate(lines): + line_split = line.split() + if len(line_split) >= 2: + # find the first occurrence of #ifndef and #define, save arg + if not ifndef and line_split[0] == '#ifndef': + # set ifndef to the header guard presented on the #ifndef line. + ifndef = line_split[1] + ifndef_line_number = line_number + if not define and line_split[0] == '#define': + define = line_split[1] + if define and ifndef: + break + + if not ifndef or not define or ifndef != define: + error(0, 'build/header_guard', 5, + 'No #ifndef header guard found, suggested CPP variable is: %s' % + cppvar[0]) + return + + # The guard should be File_h. + if ifndef not in cppvar: + error(ifndef_line_number, 'build/header_guard', 5, + '#ifndef header guard has wrong style, please use: %s' % cppvar[0]) + + +def check_for_unicode_replacement_characters(lines, error): + """Logs an error for each line containing Unicode replacement characters. + + These indicate that either the file contained invalid UTF-8 (likely) + or Unicode replacement characters (which it shouldn't). Note that + it's possible for this to throw off line numbering if the invalid + UTF-8 occurred adjacent to a newline. + + Args: + lines: An array of strings, each representing a line of the file. + error: The function to call with any errors found. + """ + for line_number, line in enumerate(lines): + if u'\ufffd' in line: + error(line_number, 'readability/utf8', 5, + 'Line contains invalid UTF-8 (or Unicode replacement character).') + + +def check_for_new_line_at_eof(lines, error): + """Logs an error if there is no newline char at the end of the file. + + Args: + lines: An array of strings, each representing a line of the file. + error: The function to call with any errors found. + """ + + # The array lines() was created by adding two newlines to the + # original file (go figure), then splitting on \n. + # To verify that the file ends in \n, we just have to make sure the + # last-but-two element of lines() exists and is empty. + if len(lines) < 3 or lines[-2]: + error(len(lines) - 2, 'whitespace/ending_newline', 5, + 'Could not find a newline character at the end of the file.') + + +def check_for_multiline_comments_and_strings(clean_lines, line_number, error): + """Logs an error if we see /* ... */ or "..." that extend past one line. + + /* ... */ comments are legit inside macros, for one line. + Otherwise, we prefer // comments, so it's ok to warn about the + other. Likewise, it's ok for strings to extend across multiple + lines, as long as a line continuation character (backslash) + terminates each line. Although not currently prohibited by the C++ + style guide, it's ugly and unnecessary. We don't do well with either + in this lint program, so we warn about both. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[line_number] + + # Remove all \\ (escaped backslashes) from the line. They are OK, and the + # second (escaped) slash may trigger later \" detection erroneously. + line = line.replace('\\\\', '') + + if line.count('/*') > line.count('*/'): + error(line_number, 'readability/multiline_comment', 5, + 'Complex multi-line /*...*/-style comment found. ' + 'Lint may give bogus warnings. ' + 'Consider replacing these with //-style comments, ' + 'with #if 0...#endif, ' + 'or with more clearly structured multi-line comments.') + + if (line.count('"') - line.count('\\"')) % 2: + error(line_number, 'readability/multiline_string', 5, + 'Multi-line string ("...") found. This lint script doesn\'t ' + 'do well with such strings, and may give bogus warnings. They\'re ' + 'ugly and unnecessary, and you should use concatenation instead".') + + +_THREADING_LIST = ( + ('asctime(', 'asctime_r('), + ('ctime(', 'ctime_r('), + ('getgrgid(', 'getgrgid_r('), + ('getgrnam(', 'getgrnam_r('), + ('getlogin(', 'getlogin_r('), + ('getpwnam(', 'getpwnam_r('), + ('getpwuid(', 'getpwuid_r('), + ('gmtime(', 'gmtime_r('), + ('localtime(', 'localtime_r('), + ('rand(', 'rand_r('), + ('readdir(', 'readdir_r('), + ('strtok(', 'strtok_r('), + ('ttyname(', 'ttyname_r('), + ) + + +def check_posix_threading(clean_lines, line_number, error): + """Checks for calls to thread-unsafe functions. + + Much code has been originally written without consideration of + multi-threading. Also, engineers are relying on their old experience; + they have learned posix before threading extensions were added. These + tests guide the engineers to use thread-safe functions (when using + posix directly). + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[line_number] + for single_thread_function, multithread_safe_function in _THREADING_LIST: + index = line.find(single_thread_function) + # Comparisons made explicit for clarity -- pylint: disable-msg=C6403 + if index >= 0 and (index == 0 or (not line[index - 1].isalnum() + and line[index - 1] not in ('_', '.', '>'))): + error(line_number, 'runtime/threadsafe_fn', 2, + 'Consider using ' + multithread_safe_function + + '...) instead of ' + single_thread_function + + '...) for improved thread safety.') + + +# Matches invalid increment: *count++, which moves pointer instead of +# incrementing a value. +_RE_PATTERN_INVALID_INCREMENT = re.compile( + r'^\s*\*\w+(\+\+|--);') + + +def check_invalid_increment(clean_lines, line_number, error): + """Checks for invalid increment *count++. + + For example following function: + void increment_counter(int* count) { + *count++; + } + is invalid, because it effectively does count++, moving pointer, and should + be replaced with ++*count, (*count)++ or *count += 1. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[line_number] + if _RE_PATTERN_INVALID_INCREMENT.match(line): + error(line_number, 'runtime/invalid_increment', 5, + 'Changing pointer instead of value (or unused value of operator*).') + + +class _ClassInfo(object): + """Stores information about a class.""" + + def __init__(self, name, line_number): + self.name = name + self.line_number = line_number + self.seen_open_brace = False + self.is_derived = False + self.virtual_method_line_number = None + self.has_virtual_destructor = False + self.brace_depth = 0 + + +class _ClassState(object): + """Holds the current state of the parse relating to class declarations. + + It maintains a stack of _ClassInfos representing the parser's guess + as to the current nesting of class declarations. The innermost class + is at the top (back) of the stack. Typically, the stack will either + be empty or have exactly one entry. + """ + + def __init__(self): + self.classinfo_stack = [] + + def check_finished(self, error): + """Checks that all classes have been completely parsed. + + Call this when all lines in a file have been processed. + Args: + error: The function to call with any errors found. + """ + if self.classinfo_stack: + # Note: This test can result in false positives if #ifdef constructs + # get in the way of brace matching. See the testBuildClass test in + # cpp_style_unittest.py for an example of this. + error(self.classinfo_stack[0].line_number, 'build/class', 5, + 'Failed to find complete declaration of class %s' % + self.classinfo_stack[0].name) + + +class _FileState(object): + def __init__(self, clean_lines, file_extension): + self._did_inside_namespace_indent_warning = False + self._clean_lines = clean_lines + if file_extension in ['m', 'mm']: + self._is_objective_c = True + elif file_extension == 'h': + # In the case of header files, it is unknown if the file + # is objective c or not, so set this value to None and then + # if it is requested, use heuristics to guess the value. + self._is_objective_c = None + else: + self._is_objective_c = False + self._is_c = file_extension == 'c' + + def set_did_inside_namespace_indent_warning(self): + self._did_inside_namespace_indent_warning = True + + def did_inside_namespace_indent_warning(self): + return self._did_inside_namespace_indent_warning + + def is_objective_c(self): + if self._is_objective_c is None: + for line in self._clean_lines.elided: + # Starting with @ or #import seem like the best indications + # that we have an Objective C file. + if line.startswith("@") or line.startswith("#import"): + self._is_objective_c = True + break + else: + self._is_objective_c = False + return self._is_objective_c + + def is_c_or_objective_c(self): + """Return whether the file extension corresponds to C or Objective-C.""" + return self._is_c or self.is_objective_c() + + +def check_for_non_standard_constructs(clean_lines, line_number, + class_state, error): + """Logs an error if we see certain non-ANSI constructs ignored by gcc-2. + + Complain about several constructs which gcc-2 accepts, but which are + not standard C++. Warning about these in lint is one way to ease the + transition to new compilers. + - put storage class first (e.g. "static const" instead of "const static"). + - "%lld" instead of %qd" in printf-type functions. + - "%1$d" is non-standard in printf-type functions. + - "\%" is an undefined character escape sequence. + - text after #endif is not allowed. + - invalid inner-style forward declaration. + - >? and <? operators, and their >?= and <?= cousins. + - classes with virtual methods need virtual destructors (compiler warning + available, but not turned on yet.) + + Additionally, check for constructor/destructor style violations as it + is very convenient to do so while checking for gcc-2 compliance. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + class_state: A _ClassState instance which maintains information about + the current stack of nested class declarations being parsed. + error: A callable to which errors are reported, which takes parameters: + line number, error level, and message + """ + + # Remove comments from the line, but leave in strings for now. + line = clean_lines.lines[line_number] + + if search(r'printf\s*\(.*".*%[-+ ]?\d*q', line): + error(line_number, 'runtime/printf_format', 3, + '%q in format strings is deprecated. Use %ll instead.') + + if search(r'printf\s*\(.*".*%\d+\$', line): + error(line_number, 'runtime/printf_format', 2, + '%N$ formats are unconventional. Try rewriting to avoid them.') + + # Remove escaped backslashes before looking for undefined escapes. + line = line.replace('\\\\', '') + + if search(r'("|\').*\\(%|\[|\(|{)', line): + error(line_number, 'build/printf_format', 3, + '%, [, (, and { are undefined character escapes. Unescape them.') + + # For the rest, work with both comments and strings removed. + line = clean_lines.elided[line_number] + + if search(r'\b(const|volatile|void|char|short|int|long' + r'|float|double|signed|unsigned' + r'|schar|u?int8|u?int16|u?int32|u?int64)' + r'\s+(auto|register|static|extern|typedef)\b', + line): + error(line_number, 'build/storage_class', 5, + 'Storage class (static, extern, typedef, etc) should be first.') + + if match(r'\s*#\s*endif\s*[^/\s]+', line): + error(line_number, 'build/endif_comment', 5, + 'Uncommented text after #endif is non-standard. Use a comment.') + + if match(r'\s*class\s+(\w+\s*::\s*)+\w+\s*;', line): + error(line_number, 'build/forward_decl', 5, + 'Inner-style forward declarations are invalid. Remove this line.') + + if search(r'(\w+|[+-]?\d+(\.\d*)?)\s*(<|>)\?=?\s*(\w+|[+-]?\d+)(\.\d*)?', line): + error(line_number, 'build/deprecated', 3, + '>? and <? (max and min) operators are non-standard and deprecated.') + + # Track class entry and exit, and attempt to find cases within the + # class declaration that don't meet the C++ style + # guidelines. Tracking is very dependent on the code matching Google + # style guidelines, but it seems to perform well enough in testing + # to be a worthwhile addition to the checks. + classinfo_stack = class_state.classinfo_stack + # Look for a class declaration + class_decl_match = match( + r'\s*(template\s*<[\w\s<>,:]*>\s*)?(class|struct)\s+(\w+(::\w+)*)', line) + if class_decl_match: + classinfo_stack.append(_ClassInfo(class_decl_match.group(3), line_number)) + + # Everything else in this function uses the top of the stack if it's + # not empty. + if not classinfo_stack: + return + + classinfo = classinfo_stack[-1] + + # If the opening brace hasn't been seen look for it and also + # parent class declarations. + if not classinfo.seen_open_brace: + # If the line has a ';' in it, assume it's a forward declaration or + # a single-line class declaration, which we won't process. + if line.find(';') != -1: + classinfo_stack.pop() + return + classinfo.seen_open_brace = (line.find('{') != -1) + # Look for a bare ':' + if search('(^|[^:]):($|[^:])', line): + classinfo.is_derived = True + if not classinfo.seen_open_brace: + return # Everything else in this function is for after open brace + + # The class may have been declared with namespace or classname qualifiers. + # The constructor and destructor will not have those qualifiers. + base_classname = classinfo.name.split('::')[-1] + + # Look for single-argument constructors that aren't marked explicit. + # Technically a valid construct, but against style. + args = match(r'(?<!explicit)\s+%s\s*\(([^,()]+)\)' + % re.escape(base_classname), + line) + if (args + and args.group(1) != 'void' + and not match(r'(const\s+)?%s\s*&' % re.escape(base_classname), + args.group(1).strip())): + error(line_number, 'runtime/explicit', 5, + 'Single-argument constructors should be marked explicit.') + + # Look for methods declared virtual. + if search(r'\bvirtual\b', line): + classinfo.virtual_method_line_number = line_number + # Only look for a destructor declaration on the same line. It would + # be extremely unlikely for the destructor declaration to occupy + # more than one line. + if search(r'~%s\s*\(' % base_classname, line): + classinfo.has_virtual_destructor = True + + # Look for class end. + brace_depth = classinfo.brace_depth + brace_depth = brace_depth + line.count('{') - line.count('}') + if brace_depth <= 0: + classinfo = classinfo_stack.pop() + # Try to detect missing virtual destructor declarations. + # For now, only warn if a non-derived class with virtual methods lacks + # a virtual destructor. This is to make it less likely that people will + # declare derived virtual destructors without declaring the base + # destructor virtual. + if ((classinfo.virtual_method_line_number is not None) + and (not classinfo.has_virtual_destructor) + and (not classinfo.is_derived)): # Only warn for base classes + error(classinfo.line_number, 'runtime/virtual', 4, + 'The class %s probably needs a virtual destructor due to ' + 'having virtual method(s), one declared at line %d.' + % (classinfo.name, classinfo.virtual_method_line_number)) + else: + classinfo.brace_depth = brace_depth + + +def check_spacing_for_function_call(line, line_number, error): + """Checks for the correctness of various spacing around function calls. + + Args: + line: The text of the line to check. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + + # Since function calls often occur inside if/for/foreach/while/switch + # expressions - which have their own, more liberal conventions - we + # first see if we should be looking inside such an expression for a + # function call, to which we can apply more strict standards. + function_call = line # if there's no control flow construct, look at whole line + for pattern in (r'\bif\s*\((.*)\)\s*{', + r'\bfor\s*\((.*)\)\s*{', + r'\bforeach\s*\((.*)\)\s*{', + r'\bwhile\s*\((.*)\)\s*[{;]', + r'\bswitch\s*\((.*)\)\s*{'): + matched = search(pattern, line) + if matched: + function_call = matched.group(1) # look inside the parens for function calls + break + + # Except in if/for/foreach/while/switch, there should never be space + # immediately inside parens (eg "f( 3, 4 )"). We make an exception + # for nested parens ( (a+b) + c ). Likewise, there should never be + # a space before a ( when it's a function argument. I assume it's a + # function argument when the char before the whitespace is legal in + # a function name (alnum + _) and we're not starting a macro. Also ignore + # pointers and references to arrays and functions coz they're too tricky: + # we use a very simple way to recognize these: + # " (something)(maybe-something)" or + # " (something)(maybe-something," or + # " (something)[something]" + # Note that we assume the contents of [] to be short enough that + # they'll never need to wrap. + if ( # Ignore control structures. + not search(r'\b(if|for|foreach|while|switch|return|new|delete)\b', function_call) + # Ignore pointers/references to functions. + and not search(r' \([^)]+\)\([^)]*(\)|,$)', function_call) + # Ignore pointers/references to arrays. + and not search(r' \([^)]+\)\[[^\]]+\]', function_call)): + if search(r'\w\s*\([ \t](?!\s*\\$)', function_call): # a ( used for a fn call + error(line_number, 'whitespace/parens', 4, + 'Extra space after ( in function call') + elif search(r'\([ \t]+(?!(\s*\\)|\()', function_call): + error(line_number, 'whitespace/parens', 2, + 'Extra space after (') + if (search(r'\w\s+\(', function_call) + and not search(r'#\s*define|typedef', function_call)): + error(line_number, 'whitespace/parens', 4, + 'Extra space before ( in function call') + # If the ) is followed only by a newline or a { + newline, assume it's + # part of a control statement (if/while/etc), and don't complain + if search(r'[^)\s]\s+\)(?!\s*$|{\s*$)', function_call): + error(line_number, 'whitespace/parens', 2, + 'Extra space before )') + + +def is_blank_line(line): + """Returns true if the given line is blank. + + We consider a line to be blank if the line is empty or consists of + only white spaces. + + Args: + line: A line of a string. + + Returns: + True, if the given line is blank. + """ + return not line or line.isspace() + + +def detect_functions(clean_lines, line_number, function_state, error): + """Finds where functions start and end. + + Uses a simplistic algorithm assuming other style guidelines + (especially spacing) are followed. + Trivial bodies are unchecked, so constructors with huge initializer lists + may be missed. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + function_state: Current function name and lines in body so far. + error: The function to call with any errors found. + """ + # Are we now past the end of a function? + if function_state.ending_line_number + 1 == line_number: + function_state.end() + + # If we're in a function, don't try to detect a new one. + if function_state.in_a_function: + return + + lines = clean_lines.lines + line = lines[line_number] + raw = clean_lines.raw_lines + raw_line = raw[line_number] + + regexp = r'\s*(\w(\w|::|\*|\&|\s|<|>|,|~)*)\(' # decls * & space::name( ... + match_result = match(regexp, line) + if not match_result: + return + + # If the name is all caps and underscores, figure it's a macro and + # ignore it, unless it's TEST or TEST_F. + function_name = match_result.group(1).split()[-1] + if function_name != 'TEST' and function_name != 'TEST_F' and match(r'[A-Z_]+$', function_name): + return + + joined_line = '' + for start_line_number in xrange(line_number, clean_lines.num_lines()): + start_line = clean_lines.elided[start_line_number] + joined_line += ' ' + start_line.lstrip() + if search(r'{|;', start_line): + # Replace template constructs with _ so that no spaces remain in the function name, + # while keeping the column numbers of other characters the same as "line". + line_with_no_templates = iteratively_replace_matches_with_char(r'<[^<>]*>', '_', line) + match_function = search(r'((\w|:|<|>|,|~)*)\(', line_with_no_templates) + if not match_function: + return # The '(' must have been inside of a template. + + # Use the column numbers from the modified line to find the + # function name in the original line. + function = line[match_function.start(1):match_function.end(1)] + + if match(r'TEST', function): # Handle TEST... macros + parameter_regexp = search(r'(\(.*\))', joined_line) + if parameter_regexp: # Ignore bad syntax + function += parameter_regexp.group(1) + else: + function += '()' + is_declaration = bool(search(r'^[^{]*;', start_line)) + if is_declaration: + ending_line_number = start_line_number + else: + open_brace_index = start_line.find('{') + ending_line_number = close_expression(clean_lines, start_line_number, open_brace_index)[1] + function_state.begin(function, start_line_number, ending_line_number, is_declaration) + return + + # No body for the function (or evidence of a non-function) was found. + error(line_number, 'readability/fn_size', 5, + 'Lint failed to find start of function body.') + + +def check_for_function_lengths(clean_lines, line_number, function_state, error): + """Reports for long function bodies. + + For an overview why this is done, see: + http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions + + Blank/comment lines are not counted so as to avoid encouraging the removal + of vertical space and commments just to get through a lint check. + NOLINT *on the last line of a function* disables this check. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + function_state: Current function name and lines in body so far. + error: The function to call with any errors found. + """ + lines = clean_lines.lines + line = lines[line_number] + raw = clean_lines.raw_lines + raw_line = raw[line_number] + + if function_state.ending_line_number == line_number: # last line + if not search(r'\bNOLINT\b', raw_line): + function_state.check(error, line_number) + elif not match(r'^\s*$', line): + function_state.count(line_number) # Count non-blank/non-comment lines. + + +def check_pass_ptr_usage(clean_lines, line_number, function_state, error): + """Check for proper usage of Pass*Ptr. + + Currently this is limited to detecting declarations of Pass*Ptr + variables inside of functions. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + function_state: Current function name and lines in body so far. + error: The function to call with any errors found. + """ + if not function_state.in_a_function: + return + + lines = clean_lines.lines + line = lines[line_number] + if line_number > function_state.body_start_line_number: + matched_pass_ptr = match(r'^\s*Pass([A-Z][A-Za-z]*)Ptr<', line) + if matched_pass_ptr: + type_name = 'Pass%sPtr' % matched_pass_ptr.group(1) + error(line_number, 'readability/pass_ptr', 5, + 'Local variables should never be %s (see ' + 'http://webkit.org/coding/RefPtr.html).' % type_name) + + +def check_spacing(file_extension, clean_lines, line_number, error): + """Checks for the correctness of various spacing issues in the code. + + Things we check for: spaces around operators, spaces after + if/for/while/switch, no spaces around parens in function calls, two + spaces between code and comment, don't start a block with a blank + line, don't end a function with a blank line, don't have too many + blank lines in a row. + + Args: + file_extension: The current file extension, without the leading dot. + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + + raw = clean_lines.raw_lines + line = raw[line_number] + + # Before nixing comments, check if the line is blank for no good + # reason. This includes the first line after a block is opened, and + # blank lines at the end of a function (ie, right before a line like '}'). + if is_blank_line(line): + elided = clean_lines.elided + previous_line = elided[line_number - 1] + previous_brace = previous_line.rfind('{') + # FIXME: Don't complain if line before blank line, and line after, + # both start with alnums and are indented the same amount. + # This ignores whitespace at the start of a namespace block + # because those are not usually indented. + if (previous_brace != -1 and previous_line[previous_brace:].find('}') == -1 + and previous_line[:previous_brace].find('namespace') == -1): + # OK, we have a blank line at the start of a code block. Before we + # complain, we check if it is an exception to the rule: The previous + # non-empty line has the parameters of a function header that are indented + # 4 spaces (because they did not fit in a 80 column line when placed on + # the same line as the function name). We also check for the case where + # the previous line is indented 6 spaces, which may happen when the + # initializers of a constructor do not fit into a 80 column line. + exception = False + if match(r' {6}\w', previous_line): # Initializer list? + # We are looking for the opening column of initializer list, which + # should be indented 4 spaces to cause 6 space indentation afterwards. + search_position = line_number - 2 + while (search_position >= 0 + and match(r' {6}\w', elided[search_position])): + search_position -= 1 + exception = (search_position >= 0 + and elided[search_position][:5] == ' :') + else: + # Search for the function arguments or an initializer list. We use a + # simple heuristic here: If the line is indented 4 spaces; and we have a + # closing paren, without the opening paren, followed by an opening brace + # or colon (for initializer lists) we assume that it is the last line of + # a function header. If we have a colon indented 4 spaces, it is an + # initializer list. + exception = (match(r' {4}\w[^\(]*\)\s*(const\s*)?(\{\s*$|:)', + previous_line) + or match(r' {4}:', previous_line)) + + if not exception: + error(line_number, 'whitespace/blank_line', 2, + 'Blank line at the start of a code block. Is this needed?') + # This doesn't ignore whitespace at the end of a namespace block + # because that is too hard without pairing open/close braces; + # however, a special exception is made for namespace closing + # brackets which have a comment containing "namespace". + # + # Also, ignore blank lines at the end of a block in a long if-else + # chain, like this: + # if (condition1) { + # // Something followed by a blank line + # + # } else if (condition2) { + # // Something else + # } + if line_number + 1 < clean_lines.num_lines(): + next_line = raw[line_number + 1] + if (next_line + and match(r'\s*}', next_line) + and next_line.find('namespace') == -1 + and next_line.find('} else ') == -1): + error(line_number, 'whitespace/blank_line', 3, + 'Blank line at the end of a code block. Is this needed?') + + # Next, we complain if there's a comment too near the text + comment_position = line.find('//') + if comment_position != -1: + # Check if the // may be in quotes. If so, ignore it + # Comparisons made explicit for clarity -- pylint: disable-msg=C6403 + if (line.count('"', 0, comment_position) - line.count('\\"', 0, comment_position)) % 2 == 0: # not in quotes + # Allow one space before end of line comment. + if (not match(r'^\s*$', line[:comment_position]) + and (comment_position >= 1 + and ((line[comment_position - 1] not in string.whitespace) + or (comment_position >= 2 + and line[comment_position - 2] in string.whitespace)))): + error(line_number, 'whitespace/comments', 5, + 'One space before end of line comments') + # There should always be a space between the // and the comment + commentend = comment_position + 2 + if commentend < len(line) and not line[commentend] == ' ': + # but some lines are exceptions -- e.g. if they're big + # comment delimiters like: + # //---------------------------------------------------------- + # or they begin with multiple slashes followed by a space: + # //////// Header comment + matched = (search(r'[=/-]{4,}\s*$', line[commentend:]) + or search(r'^/+ ', line[commentend:])) + if not matched: + error(line_number, 'whitespace/comments', 4, + 'Should have a space between // and comment') + + line = clean_lines.elided[line_number] # get rid of comments and strings + + # Don't try to do spacing checks for operator methods + line = sub(r'operator(==|!=|<|<<|<=|>=|>>|>|\+=|-=|\*=|/=|%=|&=|\|=|^=|<<=|>>=)\(', 'operator\(', line) + # Don't try to do spacing checks for #include or #import statements at + # minimum because it messes up checks for spacing around / + if match(r'\s*#\s*(?:include|import)', line): + return + if search(r'[\w.]=[\w.]', line): + error(line_number, 'whitespace/operators', 4, + 'Missing spaces around =') + + # FIXME: It's not ok to have spaces around binary operators like . + + # You should always have whitespace around binary operators. + # Alas, we can't test < or > because they're legitimately used sans spaces + # (a->b, vector<int> a). The only time we can tell is a < with no >, and + # only if it's not template params list spilling into the next line. + matched = search(r'[^<>=!\s](==|!=|\+=|-=|\*=|/=|/|\|=|&=|<<=|>>=|<=|>=|\|\||\||&&|>>|<<)[^<>=!\s]', line) + if not matched: + # Note that while it seems that the '<[^<]*' term in the following + # regexp could be simplified to '<.*', which would indeed match + # the same class of strings, the [^<] means that searching for the + # regexp takes linear rather than quadratic time. + if not search(r'<[^<]*,\s*$', line): # template params spill + matched = search(r'[^<>=!\s](<)[^<>=!\s]([^>]|->)*$', line) + if matched: + error(line_number, 'whitespace/operators', 3, + 'Missing spaces around %s' % matched.group(1)) + + # There shouldn't be space around unary operators + matched = search(r'(!\s|~\s|[\s]--[\s;]|[\s]\+\+[\s;])', line) + if matched: + error(line_number, 'whitespace/operators', 4, + 'Extra space for operator %s' % matched.group(1)) + + # A pet peeve of mine: no spaces after an if, while, switch, or for + matched = search(r' (if\(|for\(|foreach\(|while\(|switch\()', line) + if matched: + error(line_number, 'whitespace/parens', 5, + 'Missing space before ( in %s' % matched.group(1)) + + # For if/for/foreach/while/switch, the left and right parens should be + # consistent about how many spaces are inside the parens, and + # there should either be zero or one spaces inside the parens. + # We don't want: "if ( foo)" or "if ( foo )". + # Exception: "for ( ; foo; bar)" and "for (foo; bar; )" are allowed. + matched = search(r'\b(?P<statement>if|for|foreach|while|switch)\s*\((?P<remainder>.*)$', line) + if matched: + statement = matched.group('statement') + condition, rest = up_to_unmatched_closing_paren(matched.group('remainder')) + if condition is not None: + condition_match = search(r'(?P<leading>[ ]*)(?P<separator>.).*[^ ]+(?P<trailing>[ ]*)', condition) + if condition_match: + n_leading = len(condition_match.group('leading')) + n_trailing = len(condition_match.group('trailing')) + if n_leading != 0: + for_exception = statement == 'for' and condition.startswith(' ;') + if not for_exception: + error(line_number, 'whitespace/parens', 5, + 'Extra space after ( in %s' % statement) + if n_trailing != 0: + for_exception = statement == 'for' and condition.endswith('; ') + if not for_exception: + error(line_number, 'whitespace/parens', 5, + 'Extra space before ) in %s' % statement) + + # Do not check for more than one command in macros + in_preprocessor_directive = match(r'\s*#', line) + if not in_preprocessor_directive and not match(r'((\s*{\s*}?)|(\s*;?))\s*\\?$', rest): + error(line_number, 'whitespace/parens', 4, + 'More than one command on the same line in %s' % statement) + + # You should always have a space after a comma (either as fn arg or operator) + if search(r',[^\s]', line): + error(line_number, 'whitespace/comma', 3, + 'Missing space after ,') + + matched = search(r'^\s*(?P<token1>[a-zA-Z0-9_\*&]+)\s\s+(?P<token2>[a-zA-Z0-9_\*&]+)', line) + if matched: + error(line_number, 'whitespace/declaration', 3, + 'Extra space between %s and %s' % (matched.group('token1'), matched.group('token2'))) + + if file_extension == 'cpp': + # C++ should have the & or * beside the type not the variable name. + matched = match(r'\s*\w+(?<!\breturn|\bdelete)\s+(?P<pointer_operator>\*|\&)\w+', line) + if matched: + error(line_number, 'whitespace/declaration', 3, + 'Declaration has space between type name and %s in %s' % (matched.group('pointer_operator'), matched.group(0).strip())) + + elif file_extension == 'c': + # C Pointer declaration should have the * beside the variable not the type name. + matched = search(r'^\s*\w+\*\s+\w+', line) + if matched: + error(line_number, 'whitespace/declaration', 3, + 'Declaration has space between * and variable name in %s' % matched.group(0).strip()) + + # Next we will look for issues with function calls. + check_spacing_for_function_call(line, line_number, error) + + # Except after an opening paren, you should have spaces before your braces. + # And since you should never have braces at the beginning of a line, this is + # an easy test. + if search(r'[^ ({]{', line): + error(line_number, 'whitespace/braces', 5, + 'Missing space before {') + + # Make sure '} else {' has spaces. + if search(r'}else', line): + error(line_number, 'whitespace/braces', 5, + 'Missing space before else') + + # You shouldn't have spaces before your brackets, except maybe after + # 'delete []' or 'new char * []'. + if search(r'\w\s+\[', line) and not search(r'delete\s+\[', line): + error(line_number, 'whitespace/braces', 5, + 'Extra space before [') + + # You shouldn't have a space before a semicolon at the end of the line. + # There's a special case for "for" since the style guide allows space before + # the semicolon there. + if search(r':\s*;\s*$', line): + error(line_number, 'whitespace/semicolon', 5, + 'Semicolon defining empty statement. Use { } instead.') + elif search(r'^\s*;\s*$', line): + error(line_number, 'whitespace/semicolon', 5, + 'Line contains only semicolon. If this should be an empty statement, ' + 'use { } instead.') + elif (search(r'\s+;\s*$', line) and not search(r'\bfor\b', line)): + error(line_number, 'whitespace/semicolon', 5, + 'Extra space before last semicolon. If this should be an empty ' + 'statement, use { } instead.') + elif (search(r'\b(for|while)\s*\(.*\)\s*;\s*$', line) + and line.count('(') == line.count(')') + # Allow do {} while(); + and not search(r'}\s*while', line)): + error(line_number, 'whitespace/semicolon', 5, + 'Semicolon defining empty statement for this loop. Use { } instead.') + + +def get_previous_non_blank_line(clean_lines, line_number): + """Return the most recent non-blank line and its line number. + + Args: + clean_lines: A CleansedLines instance containing the file contents. + line_number: The number of the line to check. + + Returns: + A tuple with two elements. The first element is the contents of the last + non-blank line before the current line, or the empty string if this is the + first non-blank line. The second is the line number of that line, or -1 + if this is the first non-blank line. + """ + + previous_line_number = line_number - 1 + while previous_line_number >= 0: + previous_line = clean_lines.elided[previous_line_number] + if not is_blank_line(previous_line): # if not a blank line... + return (previous_line, previous_line_number) + previous_line_number -= 1 + return ('', -1) + + +def check_namespace_indentation(clean_lines, line_number, file_extension, file_state, error): + """Looks for indentation errors inside of namespaces. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + file_extension: The extension (dot not included) of the file. + file_state: A _FileState instance which maintains information about + the state of things in the file. + error: The function to call with any errors found. + """ + + line = clean_lines.elided[line_number] # Get rid of comments and strings. + + namespace_match = match(r'(?P<namespace_indentation>\s*)namespace\s+\S+\s*{\s*$', line) + if not namespace_match: + return + + current_indentation_level = len(namespace_match.group('namespace_indentation')) + if current_indentation_level > 0: + # Don't warn about an indented namespace if we already warned about indented code. + if not file_state.did_inside_namespace_indent_warning(): + error(line_number, 'whitespace/indent', 4, + 'namespace should never be indented.') + return + looking_for_semicolon = False; + line_offset = 0 + in_preprocessor_directive = False; + for current_line in clean_lines.elided[line_number + 1:]: + line_offset += 1 + if not current_line.strip(): + continue + if not current_indentation_level: + if not (in_preprocessor_directive or looking_for_semicolon): + if not match(r'\S', current_line) and not file_state.did_inside_namespace_indent_warning(): + file_state.set_did_inside_namespace_indent_warning() + error(line_number + line_offset, 'whitespace/indent', 4, + 'Code inside a namespace should not be indented.') + if in_preprocessor_directive or (current_line.strip()[0] == '#'): # This takes care of preprocessor directive syntax. + in_preprocessor_directive = current_line[-1] == '\\' + else: + looking_for_semicolon = ((current_line.find(';') == -1) and (current_line.strip()[-1] != '}')) or (current_line[-1] == '\\') + else: + looking_for_semicolon = False; # If we have a brace we may not need a semicolon. + current_indentation_level += current_line.count('{') - current_line.count('}') + if current_indentation_level < 0: + break; + + +def check_using_std(clean_lines, line_number, file_state, error): + """Looks for 'using std::foo;' statements which should be replaced with 'using namespace std;'. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + file_state: A _FileState instance which maintains information about + the state of things in the file. + error: The function to call with any errors found. + """ + + # This check doesn't apply to C or Objective-C implementation files. + if file_state.is_c_or_objective_c(): + return + + line = clean_lines.elided[line_number] # Get rid of comments and strings. + + using_std_match = match(r'\s*using\s+std::(?P<method_name>\S+)\s*;\s*$', line) + if not using_std_match: + return + + method_name = using_std_match.group('method_name') + error(line_number, 'build/using_std', 4, + "Use 'using namespace std;' instead of 'using std::%s;'." % method_name) + + +def check_max_min_macros(clean_lines, line_number, file_state, error): + """Looks use of MAX() and MIN() macros that should be replaced with std::max() and std::min(). + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + file_state: A _FileState instance which maintains information about + the state of things in the file. + error: The function to call with any errors found. + """ + + # This check doesn't apply to C or Objective-C implementation files. + if file_state.is_c_or_objective_c(): + return + + line = clean_lines.elided[line_number] # Get rid of comments and strings. + + max_min_macros_search = search(r'\b(?P<max_min_macro>(MAX|MIN))\s*\(', line) + if not max_min_macros_search: + return + + max_min_macro = max_min_macros_search.group('max_min_macro') + max_min_macro_lower = max_min_macro.lower() + error(line_number, 'runtime/max_min_macros', 4, + 'Use std::%s() or std::%s<type>() instead of the %s() macro.' + % (max_min_macro_lower, max_min_macro_lower, max_min_macro)) + + +def check_switch_indentation(clean_lines, line_number, error): + """Looks for indentation errors inside of switch statements. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + + line = clean_lines.elided[line_number] # Get rid of comments and strings. + + switch_match = match(r'(?P<switch_indentation>\s*)switch\s*\(.+\)\s*{\s*$', line) + if not switch_match: + return + + switch_indentation = switch_match.group('switch_indentation') + inner_indentation = switch_indentation + ' ' * 4 + line_offset = 0 + encountered_nested_switch = False + + for current_line in clean_lines.elided[line_number + 1:]: + line_offset += 1 + + # Skip not only empty lines but also those with preprocessor directives. + if current_line.strip() == '' or current_line.startswith('#'): + continue + + if match(r'\s*switch\s*\(.+\)\s*{\s*$', current_line): + # Complexity alarm - another switch statement nested inside the one + # that we're currently testing. We'll need to track the extent of + # that inner switch if the upcoming label tests are still supposed + # to work correctly. Let's not do that; instead, we'll finish + # checking this line, and then leave it like that. Assuming the + # indentation is done consistently (even if incorrectly), this will + # still catch all indentation issues in practice. + encountered_nested_switch = True + + current_indentation_match = match(r'(?P<indentation>\s*)(?P<remaining_line>.*)$', current_line); + current_indentation = current_indentation_match.group('indentation') + remaining_line = current_indentation_match.group('remaining_line') + + # End the check at the end of the switch statement. + if remaining_line.startswith('}') and current_indentation == switch_indentation: + break + # Case and default branches should not be indented. The regexp also + # catches single-line cases like "default: break;" but does not trigger + # on stuff like "Document::Foo();". + elif match(r'(default|case\s+.*)\s*:([^:].*)?$', remaining_line): + if current_indentation != switch_indentation: + error(line_number + line_offset, 'whitespace/indent', 4, + 'A case label should not be indented, but line up with its switch statement.') + # Don't throw an error for multiple badly indented labels, + # one should be enough to figure out the problem. + break + # We ignore goto labels at the very beginning of a line. + elif match(r'\w+\s*:\s*$', remaining_line): + continue + # It's not a goto label, so check if it's indented at least as far as + # the switch statement plus one more level of indentation. + elif not current_indentation.startswith(inner_indentation): + error(line_number + line_offset, 'whitespace/indent', 4, + 'Non-label code inside switch statements should be indented.') + # Don't throw an error for multiple badly indented statements, + # one should be enough to figure out the problem. + break + + if encountered_nested_switch: + break + + +def check_braces(clean_lines, line_number, error): + """Looks for misplaced braces (e.g. at the end of line). + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + + line = clean_lines.elided[line_number] # Get rid of comments and strings. + + if match(r'\s*{\s*$', line): + # We allow an open brace to start a line in the case where someone + # is using braces for function definition or in a block to + # explicitly create a new scope, which is commonly used to control + # the lifetime of stack-allocated variables. We don't detect this + # perfectly: we just don't complain if the last non-whitespace + # character on the previous non-blank line is ';', ':', '{', '}', + # ')', or ') const' and doesn't begin with 'if|for|while|switch|else'. + # We also allow '#' for #endif and '=' for array initialization. + previous_line = get_previous_non_blank_line(clean_lines, line_number)[0] + if ((not search(r'[;:}{)=]\s*$|\)\s*const\s*$', previous_line) + or search(r'\b(if|for|foreach|while|switch|else)\b', previous_line)) + and previous_line.find('#') < 0): + error(line_number, 'whitespace/braces', 4, + 'This { should be at the end of the previous line') + elif (search(r'\)\s*(const\s*)?{\s*$', line) + and line.count('(') == line.count(')') + and not search(r'\b(if|for|foreach|while|switch)\b', line) + and not match(r'\s+[A-Z_][A-Z_0-9]+\b', line)): + error(line_number, 'whitespace/braces', 4, + 'Place brace on its own line for function definitions.') + + if (match(r'\s*}\s*(else\s*({\s*)?)?$', line) and line_number > 1): + # We check if a closed brace has started a line to see if a + # one line control statement was previous. + previous_line = clean_lines.elided[line_number - 2] + if (previous_line.find('{') > 0 and previous_line.find('}') < 0 + and search(r'\b(if|for|foreach|while|else)\b', previous_line)): + error(line_number, 'whitespace/braces', 4, + 'One line control clauses should not use braces.') + + # An else clause should be on the same line as the preceding closing brace. + if match(r'\s*else\s*', line): + previous_line = get_previous_non_blank_line(clean_lines, line_number)[0] + if match(r'\s*}\s*$', previous_line): + error(line_number, 'whitespace/newline', 4, + 'An else should appear on the same line as the preceding }') + + # Likewise, an else should never have the else clause on the same line + if search(r'\belse [^\s{]', line) and not search(r'\belse if\b', line): + error(line_number, 'whitespace/newline', 4, + 'Else clause should never be on same line as else (use 2 lines)') + + # In the same way, a do/while should never be on one line + if match(r'\s*do [^\s{]', line): + error(line_number, 'whitespace/newline', 4, + 'do/while clauses should not be on a single line') + + # Braces shouldn't be followed by a ; unless they're defining a struct + # or initializing an array. + # We can't tell in general, but we can for some common cases. + previous_line_number = line_number + while True: + (previous_line, previous_line_number) = get_previous_non_blank_line(clean_lines, previous_line_number) + if match(r'\s+{.*}\s*;', line) and not previous_line.count(';'): + line = previous_line + line + else: + break + if (search(r'{.*}\s*;', line) + and line.count('{') == line.count('}') + and not search(r'struct|class|enum|\s*=\s*{', line)): + error(line_number, 'readability/braces', 4, + "You don't need a ; after a }") + + +def check_exit_statement_simplifications(clean_lines, line_number, error): + """Looks for else or else-if statements that should be written as an + if statement when the prior if concludes with a return, break, continue or + goto statement. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + + line = clean_lines.elided[line_number] # Get rid of comments and strings. + + else_match = match(r'(?P<else_indentation>\s*)(\}\s*)?else(\s+if\s*\(|(?P<else>\s*(\{\s*)?\Z))', line) + if not else_match: + return + + else_indentation = else_match.group('else_indentation') + inner_indentation = else_indentation + ' ' * 4 + + previous_lines = clean_lines.elided[:line_number] + previous_lines.reverse() + line_offset = 0 + encountered_exit_statement = False + + for current_line in previous_lines: + line_offset -= 1 + + # Skip not only empty lines but also those with preprocessor directives + # and goto labels. + if current_line.strip() == '' or current_line.startswith('#') or match(r'\w+\s*:\s*$', current_line): + continue + + # Skip lines with closing braces on the original indentation level. + # Even though the styleguide says they should be on the same line as + # the "else if" statement, we also want to check for instances where + # the current code does not comply with the coding style. Thus, ignore + # these lines and proceed to the line before that. + if current_line == else_indentation + '}': + continue + + current_indentation_match = match(r'(?P<indentation>\s*)(?P<remaining_line>.*)$', current_line); + current_indentation = current_indentation_match.group('indentation') + remaining_line = current_indentation_match.group('remaining_line') + + # As we're going up the lines, the first real statement to encounter + # has to be an exit statement (return, break, continue or goto) - + # otherwise, this check doesn't apply. + if not encountered_exit_statement: + # We only want to find exit statements if they are on exactly + # the same level of indentation as expected from the code inside + # the block. If the indentation doesn't strictly match then we + # might have a nested if or something, which must be ignored. + if current_indentation != inner_indentation: + break + if match(r'(return(\W+.*)|(break|continue)\s*;|goto\s*\w+;)$', remaining_line): + encountered_exit_statement = True + continue + break + + # When code execution reaches this point, we've found an exit statement + # as last statement of the previous block. Now we only need to make + # sure that the block belongs to an "if", then we can throw an error. + + # Skip lines with opening braces on the original indentation level, + # similar to the closing braces check above. ("if (condition)\n{") + if current_line == else_indentation + '{': + continue + + # Skip everything that's further indented than our "else" or "else if". + if current_indentation.startswith(else_indentation) and current_indentation != else_indentation: + continue + + # So we've got a line with same (or less) indentation. Is it an "if"? + # If yes: throw an error. If no: don't throw an error. + # Whatever the outcome, this is the end of our loop. + if match(r'if\s*\(', remaining_line): + if else_match.start('else') != -1: + error(line_number + line_offset, 'readability/control_flow', 4, + 'An else statement can be removed when the prior "if" ' + 'concludes with a return, break, continue or goto statement.') + else: + error(line_number + line_offset, 'readability/control_flow', 4, + 'An else if statement should be written as an if statement ' + 'when the prior "if" concludes with a return, break, ' + 'continue or goto statement.') + break + + +def replaceable_check(operator, macro, line): + """Determine whether a basic CHECK can be replaced with a more specific one. + + For example suggest using CHECK_EQ instead of CHECK(a == b) and + similarly for CHECK_GE, CHECK_GT, CHECK_LE, CHECK_LT, CHECK_NE. + + Args: + operator: The C++ operator used in the CHECK. + macro: The CHECK or EXPECT macro being called. + line: The current source line. + + Returns: + True if the CHECK can be replaced with a more specific one. + """ + + # This matches decimal and hex integers, strings, and chars (in that order). + match_constant = r'([-+]?(\d+|0[xX][0-9a-fA-F]+)[lLuU]{0,3}|".*"|\'.*\')' + + # Expression to match two sides of the operator with something that + # looks like a literal, since CHECK(x == iterator) won't compile. + # This means we can't catch all the cases where a more specific + # CHECK is possible, but it's less annoying than dealing with + # extraneous warnings. + match_this = (r'\s*' + macro + r'\((\s*' + + match_constant + r'\s*' + operator + r'[^<>].*|' + r'.*[^<>]' + operator + r'\s*' + match_constant + + r'\s*\))') + + # Don't complain about CHECK(x == NULL) or similar because + # CHECK_EQ(x, NULL) won't compile (requires a cast). + # Also, don't complain about more complex boolean expressions + # involving && or || such as CHECK(a == b || c == d). + return match(match_this, line) and not search(r'NULL|&&|\|\|', line) + + +def check_check(clean_lines, line_number, error): + """Checks the use of CHECK and EXPECT macros. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + error: The function to call with any errors found. + """ + + # Decide the set of replacement macros that should be suggested + raw_lines = clean_lines.raw_lines + current_macro = '' + for macro in _CHECK_MACROS: + if raw_lines[line_number].find(macro) >= 0: + current_macro = macro + break + if not current_macro: + # Don't waste time here if line doesn't contain 'CHECK' or 'EXPECT' + return + + line = clean_lines.elided[line_number] # get rid of comments and strings + + # Encourage replacing plain CHECKs with CHECK_EQ/CHECK_NE/etc. + for operator in ['==', '!=', '>=', '>', '<=', '<']: + if replaceable_check(operator, current_macro, line): + error(line_number, 'readability/check', 2, + 'Consider using %s instead of %s(a %s b)' % ( + _CHECK_REPLACEMENT[current_macro][operator], + current_macro, operator)) + break + + +def check_for_comparisons_to_zero(clean_lines, line_number, error): + # Get the line without comments and strings. + line = clean_lines.elided[line_number] + + # Include NULL here so that users don't have to convert NULL to 0 first and then get this error. + if search(r'[=!]=\s*(NULL|0|true|false)\W', line) or search(r'\W(NULL|0|true|false)\s*[=!]=', line): + error(line_number, 'readability/comparison_to_zero', 5, + 'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.') + + +def check_for_null(clean_lines, line_number, file_state, error): + # This check doesn't apply to C or Objective-C implementation files. + if file_state.is_c_or_objective_c(): + return + + line = clean_lines.elided[line_number] + + # Don't warn about NULL usage in g_*(). See Bug 32858 and 39372. + if search(r'\bg(_[a-z]+)+\b', line): + return + + # Don't warn about NULL usage in gst_*_many(). See Bug 39740 + if search(r'\bgst_\w+_many\b', line): + return + + # Don't warn about NULL usage in g_str{join,concat}(). See Bug 34834 + if search(r'\bg_str(join|concat)\b', line): + return + + # Don't warn about NULL usage in gdk_pixbuf_save_to_*{join,concat}(). See Bug 43090. + if search(r'\bgdk_pixbuf_save_to\w+\b', line): + return + + if search(r'\bNULL\b', line): + error(line_number, 'readability/null', 5, 'Use 0 instead of NULL.') + return + + line = clean_lines.raw_lines[line_number] + # See if NULL occurs in any comments in the line. If the search for NULL using the raw line + # matches, then do the check with strings collapsed to avoid giving errors for + # NULLs occurring in strings. + if search(r'\bNULL\b', line) and search(r'\bNULL\b', CleansedLines.collapse_strings(line)): + error(line_number, 'readability/null', 4, 'Use 0 instead of NULL.') + +def get_line_width(line): + """Determines the width of the line in column positions. + + Args: + line: A string, which may be a Unicode string. + + Returns: + The width of the line in column positions, accounting for Unicode + combining characters and wide characters. + """ + if isinstance(line, unicode): + width = 0 + for c in unicodedata.normalize('NFC', line): + if unicodedata.east_asian_width(c) in ('W', 'F'): + width += 2 + elif not unicodedata.combining(c): + width += 1 + return width + return len(line) + + +def check_style(clean_lines, line_number, file_extension, class_state, file_state, error): + """Checks rules from the 'C++ style rules' section of cppguide.html. + + Most of these rules are hard to test (naming, comment style), but we + do what we can. In particular we check for 4-space indents, line lengths, + tab usage, spaces inside code, etc. + + Args: + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + file_extension: The extension (without the dot) of the filename. + class_state: A _ClassState instance which maintains information about + the current stack of nested class declarations being parsed. + file_state: A _FileState instance which maintains information about + the state of things in the file. + error: The function to call with any errors found. + """ + + raw_lines = clean_lines.raw_lines + line = raw_lines[line_number] + + if line.find('\t') != -1: + error(line_number, 'whitespace/tab', 1, + 'Tab found; better to use spaces') + + # One or three blank spaces at the beginning of the line is weird; it's + # hard to reconcile that with 4-space indents. + # NOTE: here are the conditions rob pike used for his tests. Mine aren't + # as sophisticated, but it may be worth becoming so: RLENGTH==initial_spaces + # if(RLENGTH > 20) complain = 0; + # if(match($0, " +(error|private|public|protected):")) complain = 0; + # if(match(prev, "&& *$")) complain = 0; + # if(match(prev, "\\|\\| *$")) complain = 0; + # if(match(prev, "[\",=><] *$")) complain = 0; + # if(match($0, " <<")) complain = 0; + # if(match(prev, " +for \\(")) complain = 0; + # if(prevodd && match(prevprev, " +for \\(")) complain = 0; + initial_spaces = 0 + cleansed_line = clean_lines.elided[line_number] + while initial_spaces < len(line) and line[initial_spaces] == ' ': + initial_spaces += 1 + if line and line[-1].isspace(): + error(line_number, 'whitespace/end_of_line', 4, + 'Line ends in whitespace. Consider deleting these extra spaces.') + # There are certain situations we allow one space, notably for labels + elif ((initial_spaces >= 1 and initial_spaces <= 3) + and not match(r'\s*\w+\s*:\s*$', cleansed_line)): + error(line_number, 'whitespace/indent', 3, + 'Weird number of spaces at line-start. ' + 'Are you using a 4-space indent?') + # Labels should always be indented at least one space. + elif not initial_spaces and line[:2] != '//': + label_match = match(r'(?P<label>[^:]+):\s*$', line) + + if label_match: + label = label_match.group('label') + # Only throw errors for stuff that is definitely not a goto label, + # because goto labels can in fact occur at the start of the line. + if label in ['public', 'private', 'protected'] or label.find(' ') != -1: + error(line_number, 'whitespace/labels', 4, + 'Labels should always be indented at least one space. ' + 'If this is a member-initializer list in a constructor, ' + 'the colon should be on the line after the definition header.') + + if (cleansed_line.count(';') > 1 + # for loops are allowed two ;'s (and may run over two lines). + and cleansed_line.find('for') == -1 + and (get_previous_non_blank_line(clean_lines, line_number)[0].find('for') == -1 + or get_previous_non_blank_line(clean_lines, line_number)[0].find(';') != -1) + # It's ok to have many commands in a switch case that fits in 1 line + and not ((cleansed_line.find('case ') != -1 + or cleansed_line.find('default:') != -1) + and cleansed_line.find('break;') != -1) + # Also it's ok to have many commands in trivial single-line accessors in class definitions. + and not (match(r'.*\(.*\).*{.*.}', line) + and class_state.classinfo_stack + and line.count('{') == line.count('}')) + and not cleansed_line.startswith('#define ')): + error(line_number, 'whitespace/newline', 4, + 'More than one command on the same line') + + if cleansed_line.strip().endswith('||') or cleansed_line.strip().endswith('&&'): + error(line_number, 'whitespace/operators', 4, + 'Boolean expressions that span multiple lines should have their ' + 'operators on the left side of the line instead of the right side.') + + # Some more style checks + check_namespace_indentation(clean_lines, line_number, file_extension, file_state, error) + check_using_std(clean_lines, line_number, file_state, error) + check_max_min_macros(clean_lines, line_number, file_state, error) + check_switch_indentation(clean_lines, line_number, error) + check_braces(clean_lines, line_number, error) + check_exit_statement_simplifications(clean_lines, line_number, error) + check_spacing(file_extension, clean_lines, line_number, error) + check_check(clean_lines, line_number, error) + check_for_comparisons_to_zero(clean_lines, line_number, error) + check_for_null(clean_lines, line_number, file_state, error) + + +_RE_PATTERN_INCLUDE_NEW_STYLE = re.compile(r'#include +"[^/]+\.h"') +_RE_PATTERN_INCLUDE = re.compile(r'^\s*#\s*include\s*([<"])([^>"]*)[>"].*$') +# Matches the first component of a filename delimited by -s and _s. That is: +# _RE_FIRST_COMPONENT.match('foo').group(0) == 'foo' +# _RE_FIRST_COMPONENT.match('foo.cpp').group(0) == 'foo' +# _RE_FIRST_COMPONENT.match('foo-bar_baz.cpp').group(0) == 'foo' +# _RE_FIRST_COMPONENT.match('foo_bar-baz.cpp').group(0) == 'foo' +_RE_FIRST_COMPONENT = re.compile(r'^[^-_.]+') + + +def _drop_common_suffixes(filename): + """Drops common suffixes like _test.cpp or -inl.h from filename. + + For example: + >>> _drop_common_suffixes('foo/foo-inl.h') + 'foo/foo' + >>> _drop_common_suffixes('foo/bar/foo.cpp') + 'foo/bar/foo' + >>> _drop_common_suffixes('foo/foo_internal.h') + 'foo/foo' + >>> _drop_common_suffixes('foo/foo_unusualinternal.h') + 'foo/foo_unusualinternal' + + Args: + filename: The input filename. + + Returns: + The filename with the common suffix removed. + """ + for suffix in ('test.cpp', 'regtest.cpp', 'unittest.cpp', + 'inl.h', 'impl.h', 'internal.h'): + if (filename.endswith(suffix) and len(filename) > len(suffix) + and filename[-len(suffix) - 1] in ('-', '_')): + return filename[:-len(suffix) - 1] + return os.path.splitext(filename)[0] + + +def _classify_include(filename, include, is_system, include_state): + """Figures out what kind of header 'include' is. + + Args: + filename: The current file cpp_style is running over. + include: The path to a #included file. + is_system: True if the #include used <> rather than "". + include_state: An _IncludeState instance in which the headers are inserted. + + Returns: + One of the _XXX_HEADER constants. + + For example: + >>> _classify_include('foo.cpp', 'config.h', False) + _CONFIG_HEADER + >>> _classify_include('foo.cpp', 'foo.h', False) + _PRIMARY_HEADER + >>> _classify_include('foo.cpp', 'bar.h', False) + _OTHER_HEADER + """ + + # If it is a system header we know it is classified as _OTHER_HEADER. + if is_system: + return _OTHER_HEADER + + # If the include is named config.h then this is WebCore/config.h. + if include == "config.h": + return _CONFIG_HEADER + + # There cannot be primary includes in header files themselves. Only an + # include exactly matches the header filename will be is flagged as + # primary, so that it triggers the "don't include yourself" check. + if filename.endswith('.h') and filename != include: + return _OTHER_HEADER; + + # Qt's moc files do not follow the naming and ordering rules, so they should be skipped + if include.startswith('moc_') and include.endswith('.cpp'): + return _MOC_HEADER + + if include.endswith('.moc'): + return _MOC_HEADER + + # If the target file basename starts with the include we're checking + # then we consider it the primary header. + target_base = FileInfo(filename).base_name() + include_base = FileInfo(include).base_name() + + # If we haven't encountered a primary header, then be lenient in checking. + if not include_state.visited_primary_section() and target_base.find(include_base) != -1: + return _PRIMARY_HEADER + # If we already encountered a primary header, perform a strict comparison. + # In case the two filename bases are the same then the above lenient check + # probably was a false positive. + elif include_state.visited_primary_section() and target_base == include_base: + if include == "ResourceHandleWin.h": + # FIXME: Thus far, we've only seen one example of these, but if we + # start to see more, please consider generalizing this check + # somehow. + return _OTHER_HEADER + return _PRIMARY_HEADER + + return _OTHER_HEADER + + +def check_include_line(filename, file_extension, clean_lines, line_number, include_state, error): + """Check rules that are applicable to #include lines. + + Strings on #include lines are NOT removed from elided line, to make + certain tasks easier. However, to prevent false positives, checks + applicable to #include lines in CheckLanguage must be put here. + + Args: + filename: The name of the current file. + file_extension: The current file extension, without the leading dot. + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + include_state: An _IncludeState instance in which the headers are inserted. + error: The function to call with any errors found. + """ + # FIXME: For readability or as a possible optimization, consider + # exiting early here by checking whether the "build/include" + # category should be checked for the given filename. This + # may involve having the error handler classes expose a + # should_check() method, in addition to the usual __call__ + # method. + line = clean_lines.lines[line_number] + + matched = _RE_PATTERN_INCLUDE.search(line) + if not matched: + return + + include = matched.group(2) + is_system = (matched.group(1) == '<') + + # Look for any of the stream classes that are part of standard C++. + if match(r'(f|ind|io|i|o|parse|pf|stdio|str|)?stream$', include): + error(line_number, 'readability/streams', 3, + 'Streams are highly discouraged.') + + # Look for specific includes to fix. + if include.startswith('wtf/') and not is_system: + error(line_number, 'build/include', 4, + 'wtf includes should be <wtf/file.h> instead of "wtf/file.h".') + + duplicate_header = include in include_state + if duplicate_header: + error(line_number, 'build/include', 4, + '"%s" already included at %s:%s' % + (include, filename, include_state[include])) + else: + include_state[include] = line_number + + header_type = _classify_include(filename, include, is_system, include_state) + include_state.header_types[line_number] = header_type + + # Only proceed if this isn't a duplicate header. + if duplicate_header: + return + + # We want to ensure that headers appear in the right order: + # 1) for implementation files: config.h, primary header, blank line, alphabetically sorted + # 2) for header files: alphabetically sorted + # The include_state object keeps track of the last type seen + # and complains if the header types are out of order or missing. + error_message = include_state.check_next_include_order(header_type, file_extension == "h") + + # Check to make sure we have a blank line after primary header. + if not error_message and header_type == _PRIMARY_HEADER: + next_line = clean_lines.raw_lines[line_number + 1] + if not is_blank_line(next_line): + error(line_number, 'build/include_order', 4, + 'You should add a blank line after implementation file\'s own header.') + + # Check to make sure all headers besides config.h and the primary header are + # alphabetically sorted. Skip Qt's moc files. + if not error_message and header_type == _OTHER_HEADER: + previous_line_number = line_number - 1; + previous_line = clean_lines.lines[previous_line_number] + previous_match = _RE_PATTERN_INCLUDE.search(previous_line) + while (not previous_match and previous_line_number > 0 + and not search(r'\A(#if|#ifdef|#ifndef|#else|#elif|#endif)', previous_line)): + previous_line_number -= 1; + previous_line = clean_lines.lines[previous_line_number] + previous_match = _RE_PATTERN_INCLUDE.search(previous_line) + if previous_match: + previous_header_type = include_state.header_types[previous_line_number] + if previous_header_type == _OTHER_HEADER and previous_line.strip() > line.strip(): + error(line_number, 'build/include_order', 4, + 'Alphabetical sorting problem.') + + if error_message: + if file_extension == 'h': + error(line_number, 'build/include_order', 4, + '%s Should be: alphabetically sorted.' % + error_message) + else: + error(line_number, 'build/include_order', 4, + '%s Should be: config.h, primary header, blank line, and then alphabetically sorted.' % + error_message) + + +def check_language(filename, clean_lines, line_number, file_extension, include_state, + file_state, error): + """Checks rules from the 'C++ language rules' section of cppguide.html. + + Some of these rules are hard to test (function overloading, using + uint32 inappropriately), but we do the best we can. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + line_number: The number of the line to check. + file_extension: The extension (without the dot) of the filename. + include_state: An _IncludeState instance in which the headers are inserted. + file_state: A _FileState instance which maintains information about + the state of things in the file. + error: The function to call with any errors found. + """ + # If the line is empty or consists of entirely a comment, no need to + # check it. + line = clean_lines.elided[line_number] + if not line: + return + + matched = _RE_PATTERN_INCLUDE.search(line) + if matched: + check_include_line(filename, file_extension, clean_lines, line_number, include_state, error) + return + + # FIXME: figure out if they're using default arguments in fn proto. + + # Check to see if they're using an conversion function cast. + # I just try to capture the most common basic types, though there are more. + # Parameterless conversion functions, such as bool(), are allowed as they are + # probably a member operator declaration or default constructor. + matched = search( + r'\b(int|float|double|bool|char|int32|uint32|int64|uint64)\([^)]', line) + if matched: + # gMock methods are defined using some variant of MOCK_METHODx(name, type) + # where type may be float(), int(string), etc. Without context they are + # virtually indistinguishable from int(x) casts. + if not match(r'^\s*MOCK_(CONST_)?METHOD\d+(_T)?\(', line): + error(line_number, 'readability/casting', 4, + 'Using deprecated casting style. ' + 'Use static_cast<%s>(...) instead' % + matched.group(1)) + + check_c_style_cast(line_number, line, clean_lines.raw_lines[line_number], + 'static_cast', + r'\((int|float|double|bool|char|u?int(16|32|64))\)', + error) + # This doesn't catch all cases. Consider (const char * const)"hello". + check_c_style_cast(line_number, line, clean_lines.raw_lines[line_number], + 'reinterpret_cast', r'\((\w+\s?\*+\s?)\)', error) + + # In addition, we look for people taking the address of a cast. This + # is dangerous -- casts can assign to temporaries, so the pointer doesn't + # point where you think. + if search( + r'(&\([^)]+\)[\w(])|(&(static|dynamic|reinterpret)_cast\b)', line): + error(line_number, 'runtime/casting', 4, + ('Are you taking an address of a cast? ' + 'This is dangerous: could be a temp var. ' + 'Take the address before doing the cast, rather than after')) + + # Check for people declaring static/global STL strings at the top level. + # This is dangerous because the C++ language does not guarantee that + # globals with constructors are initialized before the first access. + matched = match( + r'((?:|static +)(?:|const +))string +([a-zA-Z0-9_:]+)\b(.*)', + line) + # Make sure it's not a function. + # Function template specialization looks like: "string foo<Type>(...". + # Class template definitions look like: "string Foo<Type>::Method(...". + if matched and not match(r'\s*(<.*>)?(::[a-zA-Z0-9_]+)?\s*\(([^"]|$)', + matched.group(3)): + error(line_number, 'runtime/string', 4, + 'For a static/global string constant, use a C style string instead: ' + '"%schar %s[]".' % + (matched.group(1), matched.group(2))) + + # Check that we're not using RTTI outside of testing code. + if search(r'\bdynamic_cast<', line): + error(line_number, 'runtime/rtti', 5, + 'Do not use dynamic_cast<>. If you need to cast within a class ' + "hierarchy, use static_cast<> to upcast. Google doesn't support " + 'RTTI.') + + if search(r'\b([A-Za-z0-9_]*_)\(\1\)', line): + error(line_number, 'runtime/init', 4, + 'You seem to be initializing a member variable with itself.') + + if file_extension == 'h': + # FIXME: check that 1-arg constructors are explicit. + # How to tell it's a constructor? + # (handled in check_for_non_standard_constructs for now) + pass + + # Check if people are using the verboten C basic types. The only exception + # we regularly allow is "unsigned short port" for port. + if search(r'\bshort port\b', line): + if not search(r'\bunsigned short port\b', line): + error(line_number, 'runtime/int', 4, + 'Use "unsigned short" for ports, not "short"') + + # When snprintf is used, the second argument shouldn't be a literal. + matched = search(r'snprintf\s*\(([^,]*),\s*([0-9]*)\s*,', line) + if matched: + error(line_number, 'runtime/printf', 3, + 'If you can, use sizeof(%s) instead of %s as the 2nd arg ' + 'to snprintf.' % (matched.group(1), matched.group(2))) + + # Check if some verboten C functions are being used. + if search(r'\bsprintf\b', line): + error(line_number, 'runtime/printf', 5, + 'Never use sprintf. Use snprintf instead.') + matched = search(r'\b(strcpy|strcat)\b', line) + if matched: + error(line_number, 'runtime/printf', 4, + 'Almost always, snprintf is better than %s' % matched.group(1)) + + if search(r'\bsscanf\b', line): + error(line_number, 'runtime/printf', 1, + 'sscanf can be ok, but is slow and can overflow buffers.') + + # Check for suspicious usage of "if" like + # } if (a == b) { + if search(r'\}\s*if\s*\(', line): + error(line_number, 'readability/braces', 4, + 'Did you mean "else if"? If not, start a new line for "if".') + + # Check for potential format string bugs like printf(foo). + # We constrain the pattern not to pick things like DocidForPrintf(foo). + # Not perfect but it can catch printf(foo.c_str()) and printf(foo->c_str()) + matched = re.search(r'\b((?:string)?printf)\s*\(([\w.\->()]+)\)', line, re.I) + if matched: + error(line_number, 'runtime/printf', 4, + 'Potential format string bug. Do %s("%%s", %s) instead.' + % (matched.group(1), matched.group(2))) + + # Check for potential memset bugs like memset(buf, sizeof(buf), 0). + matched = search(r'memset\s*\(([^,]*),\s*([^,]*),\s*0\s*\)', line) + if matched and not match(r"^''|-?[0-9]+|0x[0-9A-Fa-f]$", matched.group(2)): + error(line_number, 'runtime/memset', 4, + 'Did you mean "memset(%s, 0, %s)"?' + % (matched.group(1), matched.group(2))) + + # Detect variable-length arrays. + matched = match(r'\s*(.+::)?(\w+) [a-z]\w*\[(.+)];', line) + if (matched and matched.group(2) != 'return' and matched.group(2) != 'delete' and + matched.group(3).find(']') == -1): + # Split the size using space and arithmetic operators as delimiters. + # If any of the resulting tokens are not compile time constants then + # report the error. + tokens = re.split(r'\s|\+|\-|\*|\/|<<|>>]', matched.group(3)) + is_const = True + skip_next = False + for tok in tokens: + if skip_next: + skip_next = False + continue + + if search(r'sizeof\(.+\)', tok): + continue + if search(r'arraysize\(\w+\)', tok): + continue + + tok = tok.lstrip('(') + tok = tok.rstrip(')') + if not tok: + continue + if match(r'\d+', tok): + continue + if match(r'0[xX][0-9a-fA-F]+', tok): + continue + if match(r'k[A-Z0-9]\w*', tok): + continue + if match(r'(.+::)?k[A-Z0-9]\w*', tok): + continue + if match(r'(.+::)?[A-Z][A-Z0-9_]*', tok): + continue + # A catch all for tricky sizeof cases, including 'sizeof expression', + # 'sizeof(*type)', 'sizeof(const type)', 'sizeof(struct StructName)' + # requires skipping the next token becasue we split on ' ' and '*'. + if tok.startswith('sizeof'): + skip_next = True + continue + is_const = False + break + if not is_const: + error(line_number, 'runtime/arrays', 1, + 'Do not use variable-length arrays. Use an appropriately named ' + "('k' followed by CamelCase) compile-time constant for the size.") + + # Check for use of unnamed namespaces in header files. Registration + # macros are typically OK, so we allow use of "namespace {" on lines + # that end with backslashes. + if (file_extension == 'h' + and search(r'\bnamespace\s*{', line) + and line[-1] != '\\'): + error(line_number, 'build/namespaces', 4, + 'Do not use unnamed namespaces in header files. See ' + 'http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Namespaces' + ' for more information.') + + check_identifier_name_in_declaration(filename, line_number, line, file_state, error) + + +def check_identifier_name_in_declaration(filename, line_number, line, file_state, error): + """Checks if identifier names contain any underscores. + + As identifiers in libraries we are using have a bunch of + underscores, we only warn about the declarations of identifiers + and don't check use of identifiers. + + Args: + filename: The name of the current file. + line_number: The number of the line to check. + line: The line of code to check. + file_state: A _FileState instance which maintains information about + the state of things in the file. + error: The function to call with any errors found. + """ + # We don't check a return statement. + if match(r'\s*(return|delete)\b', line): + return + + # Basically, a declaration is a type name followed by whitespaces + # followed by an identifier. The type name can be complicated + # due to type adjectives and templates. We remove them first to + # simplify the process to find declarations of identifiers. + + # Convert "long long", "long double", and "long long int" to + # simple types, but don't remove simple "long". + line = sub(r'long (long )?(?=long|double|int)', '', line) + # Convert unsigned/signed types to simple types, too. + line = sub(r'(unsigned|signed) (?=char|short|int|long)', '', line) + line = sub(r'\b(inline|using|static|const|volatile|auto|register|extern|typedef|restrict|struct|class|virtual)(?=\W)', '', line) + + # Remove "new" and "new (expr)" to simplify, too. + line = sub(r'new\s*(\([^)]*\))?', '', line) + + # Remove all template parameters by removing matching < and >. + # Loop until no templates are removed to remove nested templates. + while True: + line, number_of_replacements = subn(r'<([\w\s:]|::)+\s*[*&]*\s*>', '', line) + if not number_of_replacements: + break + + # Declarations of local variables can be in condition expressions + # of control flow statements (e.g., "if (RenderObject* p = o->parent())"). + # We remove the keywords and the first parenthesis. + # + # Declarations in "while", "if", and "switch" are different from + # other declarations in two aspects: + # + # - There can be only one declaration between the parentheses. + # (i.e., you cannot write "if (int i = 0, j = 1) {}") + # - The variable must be initialized. + # (i.e., you cannot write "if (int i) {}") + # + # and we will need different treatments for them. + line = sub(r'^\s*for\s*\(', '', line) + line, control_statement = subn(r'^\s*(while|else if|if|switch)\s*\(', '', line) + + # Detect variable and functions. + type_regexp = r'\w([\w]|\s*[*&]\s*|::)+' + identifier_regexp = r'(?P<identifier>[\w:]+)' + maybe_bitfield_regexp = r'(:\s*\d+\s*)?' + character_after_identifier_regexp = r'(?P<character_after_identifier>[[;()=,])(?!=)' + declaration_without_type_regexp = r'\s*' + identifier_regexp + r'\s*' + maybe_bitfield_regexp + character_after_identifier_regexp + declaration_with_type_regexp = r'\s*' + type_regexp + r'\s' + declaration_without_type_regexp + is_function_arguments = False + number_of_identifiers = 0 + while True: + # If we are seeing the first identifier or arguments of a + # function, there should be a type name before an identifier. + if not number_of_identifiers or is_function_arguments: + declaration_regexp = declaration_with_type_regexp + else: + declaration_regexp = declaration_without_type_regexp + + matched = match(declaration_regexp, line) + if not matched: + return + identifier = matched.group('identifier') + character_after_identifier = matched.group('character_after_identifier') + + # If we removed a non-for-control statement, the character after + # the identifier should be '='. With this rule, we can avoid + # warning for cases like "if (val & INT_MAX) {". + if control_statement and character_after_identifier != '=': + return + + is_function_arguments = is_function_arguments or character_after_identifier == '(' + + # Remove "m_" and "s_" to allow them. + modified_identifier = sub(r'(^|(?<=::))[ms]_', '', identifier) + if not file_state.is_objective_c() and modified_identifier.find('_') >= 0: + # Various exceptions to the rule: JavaScript op codes functions, const_iterator. + if (not (filename.find('JavaScriptCore') >= 0 and modified_identifier.find('op_') >= 0) + and not modified_identifier.startswith('tst_') + and not modified_identifier.startswith('webkit_dom_object_') + and not modified_identifier.startswith('NPN_') + and not modified_identifier.startswith('NPP_') + and not modified_identifier.startswith('NP_') + and not modified_identifier.startswith('qt_') + and not modified_identifier.startswith('cairo_') + and not modified_identifier.find('::qt_') >= 0 + and not modified_identifier == "const_iterator" + and not modified_identifier == "vm_throw"): + error(line_number, 'readability/naming', 4, identifier + " is incorrectly named. Don't use underscores in your identifier names.") + + # Check for variables named 'l', these are too easy to confuse with '1' in some fonts + if modified_identifier == 'l': + error(line_number, 'readability/naming', 4, identifier + " is incorrectly named. Don't use the single letter 'l' as an identifier name.") + + # There can be only one declaration in non-for-control statements. + if control_statement: + return + # We should continue checking if this is a function + # declaration because we need to check its arguments. + # Also, we need to check multiple declarations. + if character_after_identifier != '(' and character_after_identifier != ',': + return + + number_of_identifiers += 1 + line = line[matched.end():] + +def check_c_style_cast(line_number, line, raw_line, cast_type, pattern, + error): + """Checks for a C-style cast by looking for the pattern. + + This also handles sizeof(type) warnings, due to similarity of content. + + Args: + line_number: The number of the line to check. + line: The line of code to check. + raw_line: The raw line of code to check, with comments. + cast_type: The string for the C++ cast to recommend. This is either + reinterpret_cast or static_cast, depending. + pattern: The regular expression used to find C-style casts. + error: The function to call with any errors found. + """ + matched = search(pattern, line) + if not matched: + return + + # e.g., sizeof(int) + sizeof_match = match(r'.*sizeof\s*$', line[0:matched.start(1) - 1]) + if sizeof_match: + error(line_number, 'runtime/sizeof', 1, + 'Using sizeof(type). Use sizeof(varname) instead if possible') + return + + remainder = line[matched.end(0):] + + # The close paren is for function pointers as arguments to a function. + # eg, void foo(void (*bar)(int)); + # The semicolon check is a more basic function check; also possibly a + # function pointer typedef. + # eg, void foo(int); or void foo(int) const; + # The equals check is for function pointer assignment. + # eg, void *(*foo)(int) = ... + # + # Right now, this will only catch cases where there's a single argument, and + # it's unnamed. It should probably be expanded to check for multiple + # arguments with some unnamed. + function_match = match(r'\s*(\)|=|(const)?\s*(;|\{|throw\(\)))', remainder) + if function_match: + if (not function_match.group(3) + or function_match.group(3) == ';' + or raw_line.find('/*') < 0): + error(line_number, 'readability/function', 3, + 'All parameters should be named in a function') + return + + # At this point, all that should be left is actual casts. + error(line_number, 'readability/casting', 4, + 'Using C-style cast. Use %s<%s>(...) instead' % + (cast_type, matched.group(1))) + + +_HEADERS_CONTAINING_TEMPLATES = ( + ('<deque>', ('deque',)), + ('<functional>', ('unary_function', 'binary_function', + 'plus', 'minus', 'multiplies', 'divides', 'modulus', + 'negate', + 'equal_to', 'not_equal_to', 'greater', 'less', + 'greater_equal', 'less_equal', + 'logical_and', 'logical_or', 'logical_not', + 'unary_negate', 'not1', 'binary_negate', 'not2', + 'bind1st', 'bind2nd', + 'pointer_to_unary_function', + 'pointer_to_binary_function', + 'ptr_fun', + 'mem_fun_t', 'mem_fun', 'mem_fun1_t', 'mem_fun1_ref_t', + 'mem_fun_ref_t', + 'const_mem_fun_t', 'const_mem_fun1_t', + 'const_mem_fun_ref_t', 'const_mem_fun1_ref_t', + 'mem_fun_ref', + )), + ('<limits>', ('numeric_limits',)), + ('<list>', ('list',)), + ('<map>', ('map', 'multimap',)), + ('<memory>', ('allocator',)), + ('<queue>', ('queue', 'priority_queue',)), + ('<set>', ('set', 'multiset',)), + ('<stack>', ('stack',)), + ('<string>', ('char_traits', 'basic_string',)), + ('<utility>', ('pair',)), + ('<vector>', ('vector',)), + + # gcc extensions. + # Note: std::hash is their hash, ::hash is our hash + ('<hash_map>', ('hash_map', 'hash_multimap',)), + ('<hash_set>', ('hash_set', 'hash_multiset',)), + ('<slist>', ('slist',)), + ) + +_HEADERS_ACCEPTED_BUT_NOT_PROMOTED = { + # We can trust with reasonable confidence that map gives us pair<>, too. + 'pair<>': ('map', 'multimap', 'hash_map', 'hash_multimap') +} + +_RE_PATTERN_STRING = re.compile(r'\bstring\b') + +_re_pattern_algorithm_header = [] +for _template in ('copy', 'max', 'min', 'min_element', 'sort', 'swap', + 'transform'): + # Match max<type>(..., ...), max(..., ...), but not foo->max, foo.max or + # type::max(). + _re_pattern_algorithm_header.append( + (re.compile(r'[^>.]\b' + _template + r'(<.*?>)?\([^\)]'), + _template, + '<algorithm>')) + +_re_pattern_templates = [] +for _header, _templates in _HEADERS_CONTAINING_TEMPLATES: + for _template in _templates: + _re_pattern_templates.append( + (re.compile(r'(\<|\b)' + _template + r'\s*\<'), + _template + '<>', + _header)) + + +def files_belong_to_same_module(filename_cpp, filename_h): + """Check if these two filenames belong to the same module. + + The concept of a 'module' here is a as follows: + foo.h, foo-inl.h, foo.cpp, foo_test.cpp and foo_unittest.cpp belong to the + same 'module' if they are in the same directory. + some/path/public/xyzzy and some/path/internal/xyzzy are also considered + to belong to the same module here. + + If the filename_cpp contains a longer path than the filename_h, for example, + '/absolute/path/to/base/sysinfo.cpp', and this file would include + 'base/sysinfo.h', this function also produces the prefix needed to open the + header. This is used by the caller of this function to more robustly open the + header file. We don't have access to the real include paths in this context, + so we need this guesswork here. + + Known bugs: tools/base/bar.cpp and base/bar.h belong to the same module + according to this implementation. Because of this, this function gives + some false positives. This should be sufficiently rare in practice. + + Args: + filename_cpp: is the path for the .cpp file + filename_h: is the path for the header path + + Returns: + Tuple with a bool and a string: + bool: True if filename_cpp and filename_h belong to the same module. + string: the additional prefix needed to open the header file. + """ + + if not filename_cpp.endswith('.cpp'): + return (False, '') + filename_cpp = filename_cpp[:-len('.cpp')] + if filename_cpp.endswith('_unittest'): + filename_cpp = filename_cpp[:-len('_unittest')] + elif filename_cpp.endswith('_test'): + filename_cpp = filename_cpp[:-len('_test')] + filename_cpp = filename_cpp.replace('/public/', '/') + filename_cpp = filename_cpp.replace('/internal/', '/') + + if not filename_h.endswith('.h'): + return (False, '') + filename_h = filename_h[:-len('.h')] + if filename_h.endswith('-inl'): + filename_h = filename_h[:-len('-inl')] + filename_h = filename_h.replace('/public/', '/') + filename_h = filename_h.replace('/internal/', '/') + + files_belong_to_same_module = filename_cpp.endswith(filename_h) + common_path = '' + if files_belong_to_same_module: + common_path = filename_cpp[:-len(filename_h)] + return files_belong_to_same_module, common_path + + +def update_include_state(filename, include_state, io=codecs): + """Fill up the include_state with new includes found from the file. + + Args: + filename: the name of the header to read. + include_state: an _IncludeState instance in which the headers are inserted. + io: The io factory to use to read the file. Provided for testability. + + Returns: + True if a header was succesfully added. False otherwise. + """ + io = _unit_test_config.get(INCLUDE_IO_INJECTION_KEY, codecs) + header_file = None + try: + header_file = io.open(filename, 'r', 'utf8', 'replace') + except IOError: + return False + line_number = 0 + for line in header_file: + line_number += 1 + clean_line = cleanse_comments(line) + matched = _RE_PATTERN_INCLUDE.search(clean_line) + if matched: + include = matched.group(2) + # The value formatting is cute, but not really used right now. + # What matters here is that the key is in include_state. + include_state.setdefault(include, '%s:%d' % (filename, line_number)) + return True + + +def check_for_include_what_you_use(filename, clean_lines, include_state, error): + """Reports for missing stl includes. + + This function will output warnings to make sure you are including the headers + necessary for the stl containers and functions that you use. We only give one + reason to include a header. For example, if you use both equal_to<> and + less<> in a .h file, only one (the latter in the file) of these will be + reported as a reason to include the <functional>. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + include_state: An _IncludeState instance. + error: The function to call with any errors found. + """ + required = {} # A map of header name to line_number and the template entity. + # Example of required: { '<functional>': (1219, 'less<>') } + + for line_number in xrange(clean_lines.num_lines()): + line = clean_lines.elided[line_number] + if not line or line[0] == '#': + continue + + # String is special -- it is a non-templatized type in STL. + if _RE_PATTERN_STRING.search(line): + required['<string>'] = (line_number, 'string') + + for pattern, template, header in _re_pattern_algorithm_header: + if pattern.search(line): + required[header] = (line_number, template) + + # The following function is just a speed up, no semantics are changed. + if not '<' in line: # Reduces the cpu time usage by skipping lines. + continue + + for pattern, template, header in _re_pattern_templates: + if pattern.search(line): + required[header] = (line_number, template) + + # The policy is that if you #include something in foo.h you don't need to + # include it again in foo.cpp. Here, we will look at possible includes. + # Let's copy the include_state so it is only messed up within this function. + include_state = include_state.copy() + + # Did we find the header for this file (if any) and succesfully load it? + header_found = False + + # Use the absolute path so that matching works properly. + abs_filename = os.path.abspath(filename) + + # For Emacs's flymake. + # If cpp_style is invoked from Emacs's flymake, a temporary file is generated + # by flymake and that file name might end with '_flymake.cpp'. In that case, + # restore original file name here so that the corresponding header file can be + # found. + # e.g. If the file name is 'foo_flymake.cpp', we should search for 'foo.h' + # instead of 'foo_flymake.h' + abs_filename = re.sub(r'_flymake\.cpp$', '.cpp', abs_filename) + + # include_state is modified during iteration, so we iterate over a copy of + # the keys. + for header in include_state.keys(): #NOLINT + (same_module, common_path) = files_belong_to_same_module(abs_filename, header) + fullpath = common_path + header + if same_module and update_include_state(fullpath, include_state): + header_found = True + + # If we can't find the header file for a .cpp, assume it's because we don't + # know where to look. In that case we'll give up as we're not sure they + # didn't include it in the .h file. + # FIXME: Do a better job of finding .h files so we are confident that + # not having the .h file means there isn't one. + if filename.endswith('.cpp') and not header_found: + return + + # All the lines have been processed, report the errors found. + for required_header_unstripped in required: + template = required[required_header_unstripped][1] + if template in _HEADERS_ACCEPTED_BUT_NOT_PROMOTED: + headers = _HEADERS_ACCEPTED_BUT_NOT_PROMOTED[template] + if [True for header in headers if header in include_state]: + continue + if required_header_unstripped.strip('<>"') not in include_state: + error(required[required_header_unstripped][0], + 'build/include_what_you_use', 4, + 'Add #include ' + required_header_unstripped + ' for ' + template) + + +def process_line(filename, file_extension, + clean_lines, line, include_state, function_state, + class_state, file_state, error): + """Processes a single line in the file. + + Args: + filename: Filename of the file that is being processed. + file_extension: The extension (dot not included) of the file. + clean_lines: An array of strings, each representing a line of the file, + with comments stripped. + line: Number of line being processed. + include_state: An _IncludeState instance in which the headers are inserted. + function_state: A _FunctionState instance which counts function lines, etc. + class_state: A _ClassState instance which maintains information about + the current stack of nested class declarations being parsed. + file_state: A _FileState instance which maintains information about + the state of things in the file. + error: A callable to which errors are reported, which takes arguments: + line number, error level, and message + + """ + raw_lines = clean_lines.raw_lines + detect_functions(clean_lines, line, function_state, error) + check_for_function_lengths(clean_lines, line, function_state, error) + if search(r'\bNOLINT\b', raw_lines[line]): # ignore nolint lines + return + check_pass_ptr_usage(clean_lines, line, function_state, error) + check_for_multiline_comments_and_strings(clean_lines, line, error) + check_style(clean_lines, line, file_extension, class_state, file_state, error) + check_language(filename, clean_lines, line, file_extension, include_state, + file_state, error) + check_for_non_standard_constructs(clean_lines, line, class_state, error) + check_posix_threading(clean_lines, line, error) + check_invalid_increment(clean_lines, line, error) + + +def _process_lines(filename, file_extension, lines, error, min_confidence): + """Performs lint checks and reports any errors to the given error function. + + Args: + filename: Filename of the file that is being processed. + file_extension: The extension (dot not included) of the file. + lines: An array of strings, each representing a line of the file, with the + last element being empty if the file is termined with a newline. + error: A callable to which errors are reported, which takes 4 arguments: + """ + lines = (['// marker so line numbers and indices both start at 1'] + lines + + ['// marker so line numbers end in a known way']) + + include_state = _IncludeState() + function_state = _FunctionState(min_confidence) + class_state = _ClassState() + + check_for_copyright(lines, error) + + if file_extension == 'h': + check_for_header_guard(filename, lines, error) + + remove_multi_line_comments(lines, error) + clean_lines = CleansedLines(lines) + file_state = _FileState(clean_lines, file_extension) + for line in xrange(clean_lines.num_lines()): + process_line(filename, file_extension, clean_lines, line, + include_state, function_state, class_state, file_state, error) + class_state.check_finished(error) + + check_for_include_what_you_use(filename, clean_lines, include_state, error) + + # We check here rather than inside process_line so that we see raw + # lines rather than "cleaned" lines. + check_for_unicode_replacement_characters(lines, error) + + check_for_new_line_at_eof(lines, error) + + +class CppChecker(object): + + """Processes C++ lines for checking style.""" + + # This list is used to-- + # + # (1) generate an explicit list of all possible categories, + # (2) unit test that all checked categories have valid names, and + # (3) unit test that all categories are getting unit tested. + # + categories = set([ + 'build/class', + 'build/deprecated', + 'build/endif_comment', + 'build/forward_decl', + 'build/header_guard', + 'build/include', + 'build/include_order', + 'build/include_what_you_use', + 'build/namespaces', + 'build/printf_format', + 'build/storage_class', + 'build/using_std', + 'legal/copyright', + 'readability/braces', + 'readability/casting', + 'readability/check', + 'readability/comparison_to_zero', + 'readability/constructors', + 'readability/control_flow', + 'readability/fn_size', + 'readability/function', + 'readability/multiline_comment', + 'readability/multiline_string', + 'readability/naming', + 'readability/null', + 'readability/pass_ptr', + 'readability/streams', + 'readability/todo', + 'readability/utf8', + 'runtime/arrays', + 'runtime/casting', + 'runtime/explicit', + 'runtime/init', + 'runtime/int', + 'runtime/invalid_increment', + 'runtime/max_min_macros', + 'runtime/memset', + 'runtime/printf', + 'runtime/printf_format', + 'runtime/references', + 'runtime/rtti', + 'runtime/sizeof', + 'runtime/string', + 'runtime/threadsafe_fn', + 'runtime/virtual', + 'whitespace/blank_line', + 'whitespace/braces', + 'whitespace/comma', + 'whitespace/comments', + 'whitespace/declaration', + 'whitespace/end_of_line', + 'whitespace/ending_newline', + 'whitespace/indent', + 'whitespace/labels', + 'whitespace/line_length', + 'whitespace/newline', + 'whitespace/operators', + 'whitespace/parens', + 'whitespace/semicolon', + 'whitespace/tab', + 'whitespace/todo', + ]) + + def __init__(self, file_path, file_extension, handle_style_error, + min_confidence): + """Create a CppChecker instance. + + Args: + file_extension: A string that is the file extension, without + the leading dot. + + """ + self.file_extension = file_extension + self.file_path = file_path + self.handle_style_error = handle_style_error + self.min_confidence = min_confidence + + # Useful for unit testing. + def __eq__(self, other): + """Return whether this CppChecker instance is equal to another.""" + if self.file_extension != other.file_extension: + return False + if self.file_path != other.file_path: + return False + if self.handle_style_error != other.handle_style_error: + return False + if self.min_confidence != other.min_confidence: + return False + + return True + + # Useful for unit testing. + def __ne__(self, other): + # Python does not automatically deduce __ne__() from __eq__(). + return not self.__eq__(other) + + def check(self, lines): + _process_lines(self.file_path, self.file_extension, lines, + self.handle_style_error, self.min_confidence) + + +# FIXME: Remove this function (requires refactoring unit tests). +def process_file_data(filename, file_extension, lines, error, min_confidence, unit_test_config): + global _unit_test_config + _unit_test_config = unit_test_config + checker = CppChecker(filename, file_extension, error, min_confidence) + checker.check(lines) + _unit_test_config = {} diff --git a/Tools/Scripts/webkitpy/style/checkers/cpp_unittest.py b/Tools/Scripts/webkitpy/style/checkers/cpp_unittest.py new file mode 100644 index 0000000..70df1ea --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/cpp_unittest.py @@ -0,0 +1,3998 @@ +#!/usr/bin/python +# -*- coding: utf-8; -*- +# +# Copyright (C) 2009 Google Inc. All rights reserved. +# Copyright (C) 2009 Torch Mobile Inc. +# Copyright (C) 2009 Apple Inc. All rights reserved. +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit test for cpp_style.py.""" + +# FIXME: Add a good test that tests UpdateIncludeState. + +import codecs +import os +import random +import re +import unittest +import cpp as cpp_style +from cpp import CppChecker +from ..filter import FilterConfiguration + +# This class works as an error collector and replaces cpp_style.Error +# function for the unit tests. We also verify each category we see +# is in STYLE_CATEGORIES, to help keep that list up to date. +class ErrorCollector: + _all_style_categories = CppChecker.categories + # This is a list including all categories seen in any unit test. + _seen_style_categories = {} + + def __init__(self, assert_fn, filter=None): + """assert_fn: a function to call when we notice a problem. + filter: filters the errors that we are concerned about.""" + self._assert_fn = assert_fn + self._errors = [] + if not filter: + filter = FilterConfiguration() + self._filter = filter + + def __call__(self, unused_linenum, category, confidence, message): + self._assert_fn(category in self._all_style_categories, + 'Message "%s" has category "%s",' + ' which is not in STYLE_CATEGORIES' % (message, category)) + if self._filter.should_check(category, ""): + self._seen_style_categories[category] = 1 + self._errors.append('%s [%s] [%d]' % (message, category, confidence)) + + def results(self): + if len(self._errors) < 2: + return ''.join(self._errors) # Most tests expect to have a string. + else: + return self._errors # Let's give a list if there is more than one. + + def result_list(self): + return self._errors + + def verify_all_categories_are_seen(self): + """Fails if there's a category in _all_style_categories - _seen_style_categories. + + This should only be called after all tests are run, so + _seen_style_categories has had a chance to fully populate. Since + this isn't called from within the normal unittest framework, we + can't use the normal unittest assert macros. Instead we just exit + when we see an error. Good thing this test is always run last! + """ + for category in self._all_style_categories: + if category not in self._seen_style_categories: + import sys + sys.exit('FATAL ERROR: There are no tests for category "%s"' % category) + + +# This class is a lame mock of codecs. We do not verify filename, mode, or +# encoding, but for the current use case it is not needed. +class MockIo: + def __init__(self, mock_file): + self.mock_file = mock_file + + def open(self, unused_filename, unused_mode, unused_encoding, _): # NOLINT + # (lint doesn't like open as a method name) + return self.mock_file + + +class CppFunctionsTest(unittest.TestCase): + + """Supports testing functions that do not need CppStyleTestBase.""" + + def test_is_c_or_objective_c(self): + clean_lines = cpp_style.CleansedLines(['']) + clean_objc_lines = cpp_style.CleansedLines(['#import "header.h"']) + self.assertTrue(cpp_style._FileState(clean_lines, 'c').is_c_or_objective_c()) + self.assertTrue(cpp_style._FileState(clean_lines, 'm').is_c_or_objective_c()) + self.assertFalse(cpp_style._FileState(clean_lines, 'cpp').is_c_or_objective_c()) + self.assertFalse(cpp_style._FileState(clean_lines, 'cc').is_c_or_objective_c()) + self.assertFalse(cpp_style._FileState(clean_lines, 'h').is_c_or_objective_c()) + self.assertTrue(cpp_style._FileState(clean_objc_lines, 'h').is_c_or_objective_c()) + + +class CppStyleTestBase(unittest.TestCase): + """Provides some useful helper functions for cpp_style tests. + + Attributes: + min_confidence: An integer that is the current minimum confidence + level for the tests. + + """ + + # FIXME: Refactor the unit tests so the confidence level is passed + # explicitly, just like it is in the real code. + min_confidence = 1; + + # Helper function to avoid needing to explicitly pass confidence + # in all the unit test calls to cpp_style.process_file_data(). + def process_file_data(self, filename, file_extension, lines, error, unit_test_config={}): + """Call cpp_style.process_file_data() with the min_confidence.""" + return cpp_style.process_file_data(filename, file_extension, lines, + error, self.min_confidence, unit_test_config) + + def perform_lint(self, code, filename, basic_error_rules, unit_test_config={}): + error_collector = ErrorCollector(self.assert_, FilterConfiguration(basic_error_rules)) + lines = code.split('\n') + extension = filename.split('.')[1] + self.process_file_data(filename, extension, lines, error_collector, unit_test_config) + return error_collector.results() + + # Perform lint on single line of input and return the error message. + def perform_single_line_lint(self, code, filename): + basic_error_rules = ('-build/header_guard', + '-legal/copyright', + '-readability/fn_size', + '-whitespace/ending_newline') + return self.perform_lint(code, filename, basic_error_rules) + + # Perform lint over multiple lines and return the error message. + def perform_multi_line_lint(self, code, file_extension): + basic_error_rules = ('-build/header_guard', + '-legal/copyright', + '-multi_line_filter', + '-whitespace/ending_newline') + return self.perform_lint(code, 'test.' + file_extension, basic_error_rules) + + # Only keep some errors related to includes, namespaces and rtti. + def perform_language_rules_check(self, filename, code): + basic_error_rules = ('-', + '+build/include', + '+build/include_order', + '+build/namespaces', + '+runtime/rtti') + return self.perform_lint(code, filename, basic_error_rules) + + # Only keep function length errors. + def perform_function_lengths_check(self, code): + basic_error_rules = ('-', + '+readability/fn_size') + return self.perform_lint(code, 'test.cpp', basic_error_rules) + + # Only keep pass ptr errors. + def perform_pass_ptr_check(self, code): + basic_error_rules = ('-', + '+readability/pass_ptr') + return self.perform_lint(code, 'test.cpp', basic_error_rules) + + # Only include what you use errors. + def perform_include_what_you_use(self, code, filename='foo.h', io=codecs): + basic_error_rules = ('-', + '+build/include_what_you_use') + unit_test_config = {cpp_style.INCLUDE_IO_INJECTION_KEY: io} + return self.perform_lint(code, filename, basic_error_rules, unit_test_config) + + # Perform lint and compare the error message with "expected_message". + def assert_lint(self, code, expected_message, file_name='foo.cpp'): + self.assertEquals(expected_message, self.perform_single_line_lint(code, file_name)) + + def assert_lint_one_of_many_errors_re(self, code, expected_message_re, file_name='foo.cpp'): + messages = self.perform_single_line_lint(code, file_name) + for message in messages: + if re.search(expected_message_re, message): + return + + self.assertEquals(expected_message_re, messages) + + def assert_multi_line_lint(self, code, expected_message, file_name='foo.h'): + file_extension = file_name[file_name.rfind('.') + 1:] + self.assertEquals(expected_message, self.perform_multi_line_lint(code, file_extension)) + + def assert_multi_line_lint_re(self, code, expected_message_re, file_name='foo.h'): + file_extension = file_name[file_name.rfind('.') + 1:] + message = self.perform_multi_line_lint(code, file_extension) + if not re.search(expected_message_re, message): + self.fail('Message was:\n' + message + 'Expected match to "' + expected_message_re + '"') + + def assert_language_rules_check(self, file_name, code, expected_message): + self.assertEquals(expected_message, + self.perform_language_rules_check(file_name, code)) + + def assert_include_what_you_use(self, code, expected_message): + self.assertEquals(expected_message, + self.perform_include_what_you_use(code)) + + def assert_blank_lines_check(self, lines, start_errors, end_errors): + error_collector = ErrorCollector(self.assert_) + self.process_file_data('foo.cpp', 'cpp', lines, error_collector) + self.assertEquals( + start_errors, + error_collector.results().count( + 'Blank line at the start of a code block. Is this needed?' + ' [whitespace/blank_line] [2]')) + self.assertEquals( + end_errors, + error_collector.results().count( + 'Blank line at the end of a code block. Is this needed?' + ' [whitespace/blank_line] [3]')) + + +class FunctionDetectionTest(CppStyleTestBase): + def perform_function_detection(self, lines, function_information): + clean_lines = cpp_style.CleansedLines(lines) + function_state = cpp_style._FunctionState(5) + error_collector = ErrorCollector(self.assert_) + cpp_style.detect_functions(clean_lines, 0, function_state, error_collector) + if not function_information: + self.assertEquals(function_state.in_a_function, False) + return + self.assertEquals(function_state.in_a_function, True) + self.assertEquals(function_state.current_function, function_information['name'] + '()') + self.assertEquals(function_state.body_start_line_number, function_information['body_start_line_number']) + self.assertEquals(function_state.ending_line_number, function_information['ending_line_number']) + self.assertEquals(function_state.is_declaration, function_information['is_declaration']) + + def test_basic_function_detection(self): + self.perform_function_detection( + ['void theTestFunctionName(int) {', + '}'], + {'name': 'theTestFunctionName', + 'body_start_line_number': 0, + 'ending_line_number': 1, + 'is_declaration': False}) + + def test_function_declaration_detection(self): + self.perform_function_detection( + ['void aFunctionName(int);'], + {'name': 'aFunctionName', + 'body_start_line_number': 0, + 'ending_line_number': 0, + 'is_declaration': True}) + + def test_non_functions(self): + # This case exposed an error because the open brace was in quotes. + self.perform_function_detection( + ['asm(', + ' "stmdb sp!, {r1-r3}" "\n"', + ');'], + # This isn't a function but it looks like one to our simple + # algorithm and that is ok. + {'name': 'asm', + 'body_start_line_number': 2, + 'ending_line_number': 2, + 'is_declaration': True}) + + # Simple test case with something that is not a function. + self.perform_function_detection(['class Stuff;'], None) + +class CppStyleTest(CppStyleTestBase): + + # Test get line width. + def test_get_line_width(self): + self.assertEquals(0, cpp_style.get_line_width('')) + self.assertEquals(10, cpp_style.get_line_width(u'x' * 10)) + self.assertEquals(16, cpp_style.get_line_width(u'都|道|府|県|支庁')) + + def test_find_next_multi_line_comment_start(self): + self.assertEquals(1, cpp_style.find_next_multi_line_comment_start([''], 0)) + + lines = ['a', 'b', '/* c'] + self.assertEquals(2, cpp_style.find_next_multi_line_comment_start(lines, 0)) + + lines = ['char a[] = "/*";'] # not recognized as comment. + self.assertEquals(1, cpp_style.find_next_multi_line_comment_start(lines, 0)) + + def test_find_next_multi_line_comment_end(self): + self.assertEquals(1, cpp_style.find_next_multi_line_comment_end([''], 0)) + lines = ['a', 'b', ' c */'] + self.assertEquals(2, cpp_style.find_next_multi_line_comment_end(lines, 0)) + + def test_remove_multi_line_comments_from_range(self): + lines = ['a', ' /* comment ', ' * still comment', ' comment */ ', 'b'] + cpp_style.remove_multi_line_comments_from_range(lines, 1, 4) + self.assertEquals(['a', '// dummy', '// dummy', '// dummy', 'b'], lines) + + def test_spaces_at_end_of_line(self): + self.assert_lint( + '// Hello there ', + 'Line ends in whitespace. Consider deleting these extra spaces.' + ' [whitespace/end_of_line] [4]') + + # Test C-style cast cases. + def test_cstyle_cast(self): + self.assert_lint( + 'int a = (int)1.0;', + 'Using C-style cast. Use static_cast<int>(...) instead' + ' [readability/casting] [4]') + self.assert_lint( + 'int *a = (int *)DEFINED_VALUE;', + 'Using C-style cast. Use reinterpret_cast<int *>(...) instead' + ' [readability/casting] [4]', 'foo.c') + self.assert_lint( + 'uint16 a = (uint16)1.0;', + 'Using C-style cast. Use static_cast<uint16>(...) instead' + ' [readability/casting] [4]') + self.assert_lint( + 'int32 a = (int32)1.0;', + 'Using C-style cast. Use static_cast<int32>(...) instead' + ' [readability/casting] [4]') + self.assert_lint( + 'uint64 a = (uint64)1.0;', + 'Using C-style cast. Use static_cast<uint64>(...) instead' + ' [readability/casting] [4]') + + # Test taking address of casts (runtime/casting) + def test_runtime_casting(self): + self.assert_lint( + 'int* x = &static_cast<int*>(foo);', + 'Are you taking an address of a cast? ' + 'This is dangerous: could be a temp var. ' + 'Take the address before doing the cast, rather than after' + ' [runtime/casting] [4]') + + self.assert_lint( + 'int* x = &dynamic_cast<int *>(foo);', + ['Are you taking an address of a cast? ' + 'This is dangerous: could be a temp var. ' + 'Take the address before doing the cast, rather than after' + ' [runtime/casting] [4]', + 'Do not use dynamic_cast<>. If you need to cast within a class ' + 'hierarchy, use static_cast<> to upcast. Google doesn\'t support ' + 'RTTI. [runtime/rtti] [5]']) + + self.assert_lint( + 'int* x = &reinterpret_cast<int *>(foo);', + 'Are you taking an address of a cast? ' + 'This is dangerous: could be a temp var. ' + 'Take the address before doing the cast, rather than after' + ' [runtime/casting] [4]') + + # It's OK to cast an address. + self.assert_lint( + 'int* x = reinterpret_cast<int *>(&foo);', + '') + + def test_runtime_selfinit(self): + self.assert_lint( + 'Foo::Foo(Bar r, Bel l) : r_(r_), l_(l_) { }', + 'You seem to be initializing a member variable with itself.' + ' [runtime/init] [4]') + self.assert_lint( + 'Foo::Foo(Bar r, Bel l) : r_(r), l_(l) { }', + '') + self.assert_lint( + 'Foo::Foo(Bar r) : r_(r), l_(r_), ll_(l_) { }', + '') + + def test_runtime_rtti(self): + statement = 'int* x = dynamic_cast<int*>(&foo);' + error_message = ( + 'Do not use dynamic_cast<>. If you need to cast within a class ' + 'hierarchy, use static_cast<> to upcast. Google doesn\'t support ' + 'RTTI. [runtime/rtti] [5]') + # dynamic_cast is disallowed in most files. + self.assert_language_rules_check('foo.cpp', statement, error_message) + self.assert_language_rules_check('foo.h', statement, error_message) + + # We cannot test this functionality because of difference of + # function definitions. Anyway, we may never enable this. + # + # # Test for unnamed arguments in a method. + # def test_check_for_unnamed_params(self): + # message = ('All parameters should be named in a function' + # ' [readability/function] [3]') + # self.assert_lint('virtual void A(int*) const;', message) + # self.assert_lint('virtual void B(void (*fn)(int*));', message) + # self.assert_lint('virtual void C(int*);', message) + # self.assert_lint('void *(*f)(void *) = x;', message) + # self.assert_lint('void Method(char*) {', message) + # self.assert_lint('void Method(char*);', message) + # self.assert_lint('void Method(char* /*x*/);', message) + # self.assert_lint('typedef void (*Method)(int32);', message) + # self.assert_lint('static void operator delete[](void*) throw();', message) + # + # self.assert_lint('virtual void D(int* p);', '') + # self.assert_lint('void operator delete(void* x) throw();', '') + # self.assert_lint('void Method(char* x)\n{', '') + # self.assert_lint('void Method(char* /*x*/)\n{', '') + # self.assert_lint('void Method(char* x);', '') + # self.assert_lint('typedef void (*Method)(int32 x);', '') + # self.assert_lint('static void operator delete[](void* x) throw();', '') + # self.assert_lint('static void operator delete[](void* /*x*/) throw();', '') + # + # # This one should technically warn, but doesn't because the function + # # pointer is confusing. + # self.assert_lint('virtual void E(void (*fn)(int* p));', '') + + # Test deprecated casts such as int(d) + def test_deprecated_cast(self): + self.assert_lint( + 'int a = int(2.2);', + 'Using deprecated casting style. ' + 'Use static_cast<int>(...) instead' + ' [readability/casting] [4]') + # Checks for false positives... + self.assert_lint( + 'int a = int(); // Constructor, o.k.', + '') + self.assert_lint( + 'X::X() : a(int()) {} // default Constructor, o.k.', + '') + self.assert_lint( + 'operator bool(); // Conversion operator, o.k.', + '') + + # The second parameter to a gMock method definition is a function signature + # that often looks like a bad cast but should not picked up by lint. + def test_mock_method(self): + self.assert_lint( + 'MOCK_METHOD0(method, int());', + '') + self.assert_lint( + 'MOCK_CONST_METHOD1(method, float(string));', + '') + self.assert_lint( + 'MOCK_CONST_METHOD2_T(method, double(float, float));', + '') + + # Test sizeof(type) cases. + def test_sizeof_type(self): + self.assert_lint( + 'sizeof(int);', + 'Using sizeof(type). Use sizeof(varname) instead if possible' + ' [runtime/sizeof] [1]') + self.assert_lint( + 'sizeof(int *);', + 'Using sizeof(type). Use sizeof(varname) instead if possible' + ' [runtime/sizeof] [1]') + + # Test typedef cases. There was a bug that cpp_style misidentified + # typedef for pointer to function as C-style cast and produced + # false-positive error messages. + def test_typedef_for_pointer_to_function(self): + self.assert_lint( + 'typedef void (*Func)(int x);', + '') + self.assert_lint( + 'typedef void (*Func)(int *x);', + '') + self.assert_lint( + 'typedef void Func(int x);', + '') + self.assert_lint( + 'typedef void Func(int *x);', + '') + + def test_include_what_you_use_no_implementation_files(self): + code = 'std::vector<int> foo;' + self.assertEquals('Add #include <vector> for vector<>' + ' [build/include_what_you_use] [4]', + self.perform_include_what_you_use(code, 'foo.h')) + self.assertEquals('', + self.perform_include_what_you_use(code, 'foo.cpp')) + + def test_include_what_you_use(self): + self.assert_include_what_you_use( + '''#include <vector> + std::vector<int> foo; + ''', + '') + self.assert_include_what_you_use( + '''#include <map> + std::pair<int,int> foo; + ''', + '') + self.assert_include_what_you_use( + '''#include <multimap> + std::pair<int,int> foo; + ''', + '') + self.assert_include_what_you_use( + '''#include <hash_map> + std::pair<int,int> foo; + ''', + '') + self.assert_include_what_you_use( + '''#include <utility> + std::pair<int,int> foo; + ''', + '') + self.assert_include_what_you_use( + '''#include <vector> + DECLARE_string(foobar); + ''', + '') + self.assert_include_what_you_use( + '''#include <vector> + DEFINE_string(foobar, "", ""); + ''', + '') + self.assert_include_what_you_use( + '''#include <vector> + std::pair<int,int> foo; + ''', + 'Add #include <utility> for pair<>' + ' [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include "base/foobar.h" + std::vector<int> foo; + ''', + 'Add #include <vector> for vector<>' + ' [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include <vector> + std::set<int> foo; + ''', + 'Add #include <set> for set<>' + ' [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include "base/foobar.h" + hash_map<int, int> foobar; + ''', + 'Add #include <hash_map> for hash_map<>' + ' [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include "base/foobar.h" + bool foobar = std::less<int>(0,1); + ''', + 'Add #include <functional> for less<>' + ' [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include "base/foobar.h" + bool foobar = min<int>(0,1); + ''', + 'Add #include <algorithm> for min [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + 'void a(const string &foobar);', + 'Add #include <string> for string [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include "base/foobar.h" + bool foobar = swap(0,1); + ''', + 'Add #include <algorithm> for swap [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include "base/foobar.h" + bool foobar = transform(a.begin(), a.end(), b.start(), Foo); + ''', + 'Add #include <algorithm> for transform ' + '[build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include "base/foobar.h" + bool foobar = min_element(a.begin(), a.end()); + ''', + 'Add #include <algorithm> for min_element ' + '[build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''foo->swap(0,1); + foo.swap(0,1); + ''', + '') + self.assert_include_what_you_use( + '''#include <string> + void a(const std::multimap<int,string> &foobar); + ''', + 'Add #include <map> for multimap<>' + ' [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include <queue> + void a(const std::priority_queue<int> &foobar); + ''', + '') + self.assert_include_what_you_use( + '''#include "base/basictypes.h" + #include "base/port.h" + #include <assert.h> + #include <string> + #include <vector> + vector<string> hajoa;''', '') + self.assert_include_what_you_use( + '''#include <string> + int i = numeric_limits<int>::max() + ''', + 'Add #include <limits> for numeric_limits<>' + ' [build/include_what_you_use] [4]') + self.assert_include_what_you_use( + '''#include <limits> + int i = numeric_limits<int>::max() + ''', + '') + + # Test the UpdateIncludeState code path. + mock_header_contents = ['#include "blah/foo.h"', '#include "blah/bar.h"'] + message = self.perform_include_what_you_use( + '#include "config.h"\n' + '#include "blah/a.h"\n', + filename='blah/a.cpp', + io=MockIo(mock_header_contents)) + self.assertEquals(message, '') + + mock_header_contents = ['#include <set>'] + message = self.perform_include_what_you_use( + '''#include "config.h" + #include "blah/a.h" + + std::set<int> foo;''', + filename='blah/a.cpp', + io=MockIo(mock_header_contents)) + self.assertEquals(message, '') + + # If there's just a .cpp and the header can't be found then it's ok. + message = self.perform_include_what_you_use( + '''#include "config.h" + #include "blah/a.h" + + std::set<int> foo;''', + filename='blah/a.cpp') + self.assertEquals(message, '') + + # Make sure we find the headers with relative paths. + mock_header_contents = [''] + message = self.perform_include_what_you_use( + '''#include "config.h" + #include "%s/a.h" + + std::set<int> foo;''' % os.path.basename(os.getcwd()), + filename='a.cpp', + io=MockIo(mock_header_contents)) + self.assertEquals(message, 'Add #include <set> for set<> ' + '[build/include_what_you_use] [4]') + + def test_files_belong_to_same_module(self): + f = cpp_style.files_belong_to_same_module + self.assertEquals((True, ''), f('a.cpp', 'a.h')) + self.assertEquals((True, ''), f('base/google.cpp', 'base/google.h')) + self.assertEquals((True, ''), f('base/google_test.cpp', 'base/google.h')) + self.assertEquals((True, ''), + f('base/google_unittest.cpp', 'base/google.h')) + self.assertEquals((True, ''), + f('base/internal/google_unittest.cpp', + 'base/public/google.h')) + self.assertEquals((True, 'xxx/yyy/'), + f('xxx/yyy/base/internal/google_unittest.cpp', + 'base/public/google.h')) + self.assertEquals((True, 'xxx/yyy/'), + f('xxx/yyy/base/google_unittest.cpp', + 'base/public/google.h')) + self.assertEquals((True, ''), + f('base/google_unittest.cpp', 'base/google-inl.h')) + self.assertEquals((True, '/home/build/google3/'), + f('/home/build/google3/base/google.cpp', 'base/google.h')) + + self.assertEquals((False, ''), + f('/home/build/google3/base/google.cpp', 'basu/google.h')) + self.assertEquals((False, ''), f('a.cpp', 'b.h')) + + def test_cleanse_line(self): + self.assertEquals('int foo = 0; ', + cpp_style.cleanse_comments('int foo = 0; // danger!')) + self.assertEquals('int o = 0;', + cpp_style.cleanse_comments('int /* foo */ o = 0;')) + self.assertEquals('foo(int a, int b);', + cpp_style.cleanse_comments('foo(int a /* abc */, int b);')) + self.assertEqual('f(a, b);', + cpp_style.cleanse_comments('f(a, /* name */ b);')) + self.assertEqual('f(a, b);', + cpp_style.cleanse_comments('f(a /* name */, b);')) + self.assertEqual('f(a, b);', + cpp_style.cleanse_comments('f(a, /* name */b);')) + + def test_multi_line_comments(self): + # missing explicit is bad + self.assert_multi_line_lint( + r'''int a = 0; + /* multi-liner + class Foo { + Foo(int f); // should cause a lint warning in code + } + */ ''', + '') + self.assert_multi_line_lint( + r'''/* int a = 0; multi-liner + static const int b = 0;''', + ['Could not find end of multi-line comment' + ' [readability/multiline_comment] [5]', + 'Complex multi-line /*...*/-style comment found. ' + 'Lint may give bogus warnings. Consider replacing these with ' + '//-style comments, with #if 0...#endif, or with more clearly ' + 'structured multi-line comments. [readability/multiline_comment] [5]']) + self.assert_multi_line_lint(r''' /* multi-line comment''', + ['Could not find end of multi-line comment' + ' [readability/multiline_comment] [5]', + 'Complex multi-line /*...*/-style comment found. ' + 'Lint may give bogus warnings. Consider replacing these with ' + '//-style comments, with #if 0...#endif, or with more clearly ' + 'structured multi-line comments. [readability/multiline_comment] [5]']) + self.assert_multi_line_lint(r''' // /* comment, but not multi-line''', '') + + def test_multiline_strings(self): + multiline_string_error_message = ( + 'Multi-line string ("...") found. This lint script doesn\'t ' + 'do well with such strings, and may give bogus warnings. They\'re ' + 'ugly and unnecessary, and you should use concatenation instead".' + ' [readability/multiline_string] [5]') + + file_path = 'mydir/foo.cpp' + + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'cpp', + ['const char* str = "This is a\\', + ' multiline string.";'], + error_collector) + self.assertEquals( + 2, # One per line. + error_collector.result_list().count(multiline_string_error_message)) + + # Test non-explicit single-argument constructors + def test_explicit_single_argument_constructors(self): + # missing explicit is bad + self.assert_multi_line_lint( + '''class Foo { + Foo(int f); + };''', + 'Single-argument constructors should be marked explicit.' + ' [runtime/explicit] [5]') + # missing explicit is bad, even with whitespace + self.assert_multi_line_lint( + '''class Foo { + Foo (int f); + };''', + ['Extra space before ( in function call [whitespace/parens] [4]', + 'Single-argument constructors should be marked explicit.' + ' [runtime/explicit] [5]']) + # missing explicit, with distracting comment, is still bad + self.assert_multi_line_lint( + '''class Foo { + Foo(int f); // simpler than Foo(blargh, blarg) + };''', + 'Single-argument constructors should be marked explicit.' + ' [runtime/explicit] [5]') + # missing explicit, with qualified classname + self.assert_multi_line_lint( + '''class Qualifier::AnotherOne::Foo { + Foo(int f); + };''', + 'Single-argument constructors should be marked explicit.' + ' [runtime/explicit] [5]') + # structs are caught as well. + self.assert_multi_line_lint( + '''struct Foo { + Foo(int f); + };''', + 'Single-argument constructors should be marked explicit.' + ' [runtime/explicit] [5]') + # Templatized classes are caught as well. + self.assert_multi_line_lint( + '''template<typename T> class Foo { + Foo(int f); + };''', + 'Single-argument constructors should be marked explicit.' + ' [runtime/explicit] [5]') + # proper style is okay + self.assert_multi_line_lint( + '''class Foo { + explicit Foo(int f); + };''', + '') + # two argument constructor is okay + self.assert_multi_line_lint( + '''class Foo { + Foo(int f, int b); + };''', + '') + # two argument constructor, across two lines, is okay + self.assert_multi_line_lint( + '''class Foo { + Foo(int f, + int b); + };''', + '') + # non-constructor (but similar name), is okay + self.assert_multi_line_lint( + '''class Foo { + aFoo(int f); + };''', + '') + # constructor with void argument is okay + self.assert_multi_line_lint( + '''class Foo { + Foo(void); + };''', + '') + # single argument method is okay + self.assert_multi_line_lint( + '''class Foo { + Bar(int b); + };''', + '') + # comments should be ignored + self.assert_multi_line_lint( + '''class Foo { + // Foo(int f); + };''', + '') + # single argument function following class definition is okay + # (okay, it's not actually valid, but we don't want a false positive) + self.assert_multi_line_lint( + '''class Foo { + Foo(int f, int b); + }; + Foo(int f);''', + '') + # single argument function is okay + self.assert_multi_line_lint( + '''static Foo(int f);''', + '') + # single argument copy constructor is okay. + self.assert_multi_line_lint( + '''class Foo { + Foo(const Foo&); + };''', + '') + self.assert_multi_line_lint( + '''class Foo { + Foo(Foo&); + };''', + '') + + def test_slash_star_comment_on_single_line(self): + self.assert_multi_line_lint( + '''/* static */ Foo(int f);''', + '') + self.assert_multi_line_lint( + '''/*/ static */ Foo(int f);''', + '') + self.assert_multi_line_lint( + '''/*/ static Foo(int f);''', + 'Could not find end of multi-line comment' + ' [readability/multiline_comment] [5]') + self.assert_multi_line_lint( + ''' /*/ static Foo(int f);''', + 'Could not find end of multi-line comment' + ' [readability/multiline_comment] [5]') + self.assert_multi_line_lint( + ''' /**/ static Foo(int f);''', + '') + + # Test suspicious usage of "if" like this: + # if (a == b) { + # DoSomething(); + # } if (a == c) { // Should be "else if". + # DoSomething(); // This gets called twice if a == b && a == c. + # } + def test_suspicious_usage_of_if(self): + self.assert_lint( + ' if (a == b) {', + '') + self.assert_lint( + ' } if (a == b) {', + 'Did you mean "else if"? If not, start a new line for "if".' + ' [readability/braces] [4]') + + # Test suspicious usage of memset. Specifically, a 0 + # as the final argument is almost certainly an error. + def test_suspicious_usage_of_memset(self): + # Normal use is okay. + self.assert_lint( + ' memset(buf, 0, sizeof(buf))', + '') + + # A 0 as the final argument is almost certainly an error. + self.assert_lint( + ' memset(buf, sizeof(buf), 0)', + 'Did you mean "memset(buf, 0, sizeof(buf))"?' + ' [runtime/memset] [4]') + self.assert_lint( + ' memset(buf, xsize * ysize, 0)', + 'Did you mean "memset(buf, 0, xsize * ysize)"?' + ' [runtime/memset] [4]') + + # There is legitimate test code that uses this form. + # This is okay since the second argument is a literal. + self.assert_lint( + " memset(buf, 'y', 0)", + '') + self.assert_lint( + ' memset(buf, 4, 0)', + '') + self.assert_lint( + ' memset(buf, -1, 0)', + '') + self.assert_lint( + ' memset(buf, 0xF1, 0)', + '') + self.assert_lint( + ' memset(buf, 0xcd, 0)', + '') + + def test_check_posix_threading(self): + self.assert_lint('sctime_r()', '') + self.assert_lint('strtok_r()', '') + self.assert_lint(' strtok_r(foo, ba, r)', '') + self.assert_lint('brand()', '') + self.assert_lint('_rand()', '') + self.assert_lint('.rand()', '') + self.assert_lint('>rand()', '') + self.assert_lint('rand()', + 'Consider using rand_r(...) instead of rand(...)' + ' for improved thread safety.' + ' [runtime/threadsafe_fn] [2]') + self.assert_lint('strtok()', + 'Consider using strtok_r(...) ' + 'instead of strtok(...)' + ' for improved thread safety.' + ' [runtime/threadsafe_fn] [2]') + + # Test potential format string bugs like printf(foo). + def test_format_strings(self): + self.assert_lint('printf("foo")', '') + self.assert_lint('printf("foo: %s", foo)', '') + self.assert_lint('DocidForPrintf(docid)', '') # Should not trigger. + self.assert_lint( + 'printf(foo)', + 'Potential format string bug. Do printf("%s", foo) instead.' + ' [runtime/printf] [4]') + self.assert_lint( + 'printf(foo.c_str())', + 'Potential format string bug. ' + 'Do printf("%s", foo.c_str()) instead.' + ' [runtime/printf] [4]') + self.assert_lint( + 'printf(foo->c_str())', + 'Potential format string bug. ' + 'Do printf("%s", foo->c_str()) instead.' + ' [runtime/printf] [4]') + self.assert_lint( + 'StringPrintf(foo)', + 'Potential format string bug. Do StringPrintf("%s", foo) instead.' + '' + ' [runtime/printf] [4]') + + # Variable-length arrays are not permitted. + def test_variable_length_array_detection(self): + errmsg = ('Do not use variable-length arrays. Use an appropriately named ' + "('k' followed by CamelCase) compile-time constant for the size." + ' [runtime/arrays] [1]') + + self.assert_lint('int a[any_old_variable];', errmsg) + self.assert_lint('int doublesize[some_var * 2];', errmsg) + self.assert_lint('int a[afunction()];', errmsg) + self.assert_lint('int a[function(kMaxFooBars)];', errmsg) + self.assert_lint('bool aList[items_->size()];', errmsg) + self.assert_lint('namespace::Type buffer[len+1];', errmsg) + + self.assert_lint('int a[64];', '') + self.assert_lint('int a[0xFF];', '') + self.assert_lint('int first[256], second[256];', '') + self.assert_lint('int arrayName[kCompileTimeConstant];', '') + self.assert_lint('char buf[somenamespace::kBufSize];', '') + self.assert_lint('int arrayName[ALL_CAPS];', '') + self.assert_lint('AClass array1[foo::bar::ALL_CAPS];', '') + self.assert_lint('int a[kMaxStrLen + 1];', '') + self.assert_lint('int a[sizeof(foo)];', '') + self.assert_lint('int a[sizeof(*foo)];', '') + self.assert_lint('int a[sizeof foo];', '') + self.assert_lint('int a[sizeof(struct Foo)];', '') + self.assert_lint('int a[128 - sizeof(const bar)];', '') + self.assert_lint('int a[(sizeof(foo) * 4)];', '') + self.assert_lint('int a[(arraysize(fixed_size_array)/2) << 1];', 'Missing spaces around / [whitespace/operators] [3]') + self.assert_lint('delete a[some_var];', '') + self.assert_lint('return a[some_var];', '') + + # Brace usage + def test_braces(self): + # Braces shouldn't be followed by a ; unless they're defining a struct + # or initializing an array + self.assert_lint('int a[3] = { 1, 2, 3 };', '') + self.assert_lint( + '''const int foo[] = + {1, 2, 3 };''', + '') + # For single line, unmatched '}' with a ';' is ignored (not enough context) + self.assert_multi_line_lint( + '''int a[3] = { 1, + 2, + 3 };''', + '') + self.assert_multi_line_lint( + '''int a[2][3] = { { 1, 2 }, + { 3, 4 } };''', + '') + self.assert_multi_line_lint( + '''int a[2][3] = + { { 1, 2 }, + { 3, 4 } };''', + '') + + # CHECK/EXPECT_TRUE/EXPECT_FALSE replacements + def test_check_check(self): + self.assert_lint('CHECK(x == 42)', + 'Consider using CHECK_EQ instead of CHECK(a == b)' + ' [readability/check] [2]') + self.assert_lint('CHECK(x != 42)', + 'Consider using CHECK_NE instead of CHECK(a != b)' + ' [readability/check] [2]') + self.assert_lint('CHECK(x >= 42)', + 'Consider using CHECK_GE instead of CHECK(a >= b)' + ' [readability/check] [2]') + self.assert_lint('CHECK(x > 42)', + 'Consider using CHECK_GT instead of CHECK(a > b)' + ' [readability/check] [2]') + self.assert_lint('CHECK(x <= 42)', + 'Consider using CHECK_LE instead of CHECK(a <= b)' + ' [readability/check] [2]') + self.assert_lint('CHECK(x < 42)', + 'Consider using CHECK_LT instead of CHECK(a < b)' + ' [readability/check] [2]') + + self.assert_lint('DCHECK(x == 42)', + 'Consider using DCHECK_EQ instead of DCHECK(a == b)' + ' [readability/check] [2]') + self.assert_lint('DCHECK(x != 42)', + 'Consider using DCHECK_NE instead of DCHECK(a != b)' + ' [readability/check] [2]') + self.assert_lint('DCHECK(x >= 42)', + 'Consider using DCHECK_GE instead of DCHECK(a >= b)' + ' [readability/check] [2]') + self.assert_lint('DCHECK(x > 42)', + 'Consider using DCHECK_GT instead of DCHECK(a > b)' + ' [readability/check] [2]') + self.assert_lint('DCHECK(x <= 42)', + 'Consider using DCHECK_LE instead of DCHECK(a <= b)' + ' [readability/check] [2]') + self.assert_lint('DCHECK(x < 42)', + 'Consider using DCHECK_LT instead of DCHECK(a < b)' + ' [readability/check] [2]') + + self.assert_lint( + 'EXPECT_TRUE("42" == x)', + 'Consider using EXPECT_EQ instead of EXPECT_TRUE(a == b)' + ' [readability/check] [2]') + self.assert_lint( + 'EXPECT_TRUE("42" != x)', + 'Consider using EXPECT_NE instead of EXPECT_TRUE(a != b)' + ' [readability/check] [2]') + self.assert_lint( + 'EXPECT_TRUE(+42 >= x)', + 'Consider using EXPECT_GE instead of EXPECT_TRUE(a >= b)' + ' [readability/check] [2]') + self.assert_lint( + 'EXPECT_TRUE_M(-42 > x)', + 'Consider using EXPECT_GT_M instead of EXPECT_TRUE_M(a > b)' + ' [readability/check] [2]') + self.assert_lint( + 'EXPECT_TRUE_M(42U <= x)', + 'Consider using EXPECT_LE_M instead of EXPECT_TRUE_M(a <= b)' + ' [readability/check] [2]') + self.assert_lint( + 'EXPECT_TRUE_M(42L < x)', + 'Consider using EXPECT_LT_M instead of EXPECT_TRUE_M(a < b)' + ' [readability/check] [2]') + + self.assert_lint( + 'EXPECT_FALSE(x == 42)', + 'Consider using EXPECT_NE instead of EXPECT_FALSE(a == b)' + ' [readability/check] [2]') + self.assert_lint( + 'EXPECT_FALSE(x != 42)', + 'Consider using EXPECT_EQ instead of EXPECT_FALSE(a != b)' + ' [readability/check] [2]') + self.assert_lint( + 'EXPECT_FALSE(x >= 42)', + 'Consider using EXPECT_LT instead of EXPECT_FALSE(a >= b)' + ' [readability/check] [2]') + self.assert_lint( + 'ASSERT_FALSE(x > 42)', + 'Consider using ASSERT_LE instead of ASSERT_FALSE(a > b)' + ' [readability/check] [2]') + self.assert_lint( + 'ASSERT_FALSE(x <= 42)', + 'Consider using ASSERT_GT instead of ASSERT_FALSE(a <= b)' + ' [readability/check] [2]') + self.assert_lint( + 'ASSERT_FALSE_M(x < 42)', + 'Consider using ASSERT_GE_M instead of ASSERT_FALSE_M(a < b)' + ' [readability/check] [2]') + + self.assert_lint('CHECK(some_iterator == obj.end())', '') + self.assert_lint('EXPECT_TRUE(some_iterator == obj.end())', '') + self.assert_lint('EXPECT_FALSE(some_iterator == obj.end())', '') + + self.assert_lint('CHECK(CreateTestFile(dir, (1 << 20)));', '') + self.assert_lint('CHECK(CreateTestFile(dir, (1 >> 20)));', '') + + self.assert_lint('CHECK(x<42)', + ['Missing spaces around <' + ' [whitespace/operators] [3]', + 'Consider using CHECK_LT instead of CHECK(a < b)' + ' [readability/check] [2]']) + self.assert_lint('CHECK(x>42)', + 'Consider using CHECK_GT instead of CHECK(a > b)' + ' [readability/check] [2]') + + self.assert_lint( + ' EXPECT_TRUE(42 < x) // Random comment.', + 'Consider using EXPECT_LT instead of EXPECT_TRUE(a < b)' + ' [readability/check] [2]') + self.assert_lint( + 'EXPECT_TRUE( 42 < x )', + ['Extra space after ( in function call' + ' [whitespace/parens] [4]', + 'Consider using EXPECT_LT instead of EXPECT_TRUE(a < b)' + ' [readability/check] [2]']) + self.assert_lint( + 'CHECK("foo" == "foo")', + 'Consider using CHECK_EQ instead of CHECK(a == b)' + ' [readability/check] [2]') + + self.assert_lint('CHECK_EQ("foo", "foo")', '') + + def test_brace_at_begin_of_line(self): + self.assert_lint('{', + 'This { should be at the end of the previous line' + ' [whitespace/braces] [4]') + self.assert_multi_line_lint( + '#endif\n' + '{\n' + '}\n', + '') + self.assert_multi_line_lint( + 'if (condition) {', + '') + self.assert_multi_line_lint( + ' MACRO1(macroArg) {', + '') + self.assert_multi_line_lint( + 'ACCESSOR_GETTER(MessageEventPorts) {', + 'Place brace on its own line for function definitions. [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'int foo() {', + 'Place brace on its own line for function definitions. [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'int foo() const {', + 'Place brace on its own line for function definitions. [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'int foo() const\n' + '{\n' + '}\n', + '') + self.assert_multi_line_lint( + 'if (condition\n' + ' && condition2\n' + ' && condition3) {\n' + '}\n', + '') + + def test_mismatching_spaces_in_parens(self): + self.assert_lint('if (foo ) {', 'Extra space before ) in if' + ' [whitespace/parens] [5]') + self.assert_lint('switch ( foo) {', 'Extra space after ( in switch' + ' [whitespace/parens] [5]') + self.assert_lint('for (foo; ba; bar ) {', 'Extra space before ) in for' + ' [whitespace/parens] [5]') + self.assert_lint('for ((foo); (ba); (bar) ) {', 'Extra space before ) in for' + ' [whitespace/parens] [5]') + self.assert_lint('for (; foo; bar) {', '') + self.assert_lint('for (; (foo); (bar)) {', '') + self.assert_lint('for ( ; foo; bar) {', '') + self.assert_lint('for ( ; (foo); (bar)) {', '') + self.assert_lint('for ( ; foo; bar ) {', 'Extra space before ) in for' + ' [whitespace/parens] [5]') + self.assert_lint('for ( ; (foo); (bar) ) {', 'Extra space before ) in for' + ' [whitespace/parens] [5]') + self.assert_lint('for (foo; bar; ) {', '') + self.assert_lint('for ((foo); (bar); ) {', '') + self.assert_lint('foreach (foo, foos ) {', 'Extra space before ) in foreach' + ' [whitespace/parens] [5]') + self.assert_lint('foreach ( foo, foos) {', 'Extra space after ( in foreach' + ' [whitespace/parens] [5]') + self.assert_lint('while ( foo) {', 'Extra space after ( in while' + ' [whitespace/parens] [5]') + + def test_spacing_for_fncall(self): + self.assert_lint('if (foo) {', '') + self.assert_lint('for (foo;bar;baz) {', '') + self.assert_lint('foreach (foo, foos) {', '') + self.assert_lint('while (foo) {', '') + self.assert_lint('switch (foo) {', '') + self.assert_lint('new (RenderArena()) RenderInline(document())', '') + self.assert_lint('foo( bar)', 'Extra space after ( in function call' + ' [whitespace/parens] [4]') + self.assert_lint('foobar( \\', '') + self.assert_lint('foobar( \\', '') + self.assert_lint('( a + b)', 'Extra space after (' + ' [whitespace/parens] [2]') + self.assert_lint('((a+b))', '') + self.assert_lint('foo (foo)', 'Extra space before ( in function call' + ' [whitespace/parens] [4]') + self.assert_lint('typedef foo (*foo)(foo)', '') + self.assert_lint('typedef foo (*foo12bar_)(foo)', '') + self.assert_lint('typedef foo (Foo::*bar)(foo)', '') + self.assert_lint('foo (Foo::*bar)(', + 'Extra space before ( in function call' + ' [whitespace/parens] [4]') + self.assert_lint('typedef foo (Foo::*bar)(', '') + self.assert_lint('(foo)(bar)', '') + self.assert_lint('Foo (*foo)(bar)', '') + self.assert_lint('Foo (*foo)(Bar bar,', '') + self.assert_lint('char (*p)[sizeof(foo)] = &foo', '') + self.assert_lint('char (&ref)[sizeof(foo)] = &foo', '') + self.assert_lint('const char32 (*table[])[6];', '') + + def test_spacing_before_braces(self): + self.assert_lint('if (foo){', 'Missing space before {' + ' [whitespace/braces] [5]') + self.assert_lint('for{', 'Missing space before {' + ' [whitespace/braces] [5]') + self.assert_lint('for {', '') + self.assert_lint('EXPECT_DEBUG_DEATH({', '') + + def test_spacing_around_else(self): + self.assert_lint('}else {', 'Missing space before else' + ' [whitespace/braces] [5]') + self.assert_lint('} else{', 'Missing space before {' + ' [whitespace/braces] [5]') + self.assert_lint('} else {', '') + self.assert_lint('} else if', '') + + def test_spacing_for_binary_ops(self): + self.assert_lint('if (foo<=bar) {', 'Missing spaces around <=' + ' [whitespace/operators] [3]') + self.assert_lint('if (foo<bar) {', 'Missing spaces around <' + ' [whitespace/operators] [3]') + self.assert_lint('if (foo<bar->baz) {', 'Missing spaces around <' + ' [whitespace/operators] [3]') + self.assert_lint('if (foo<bar->bar) {', 'Missing spaces around <' + ' [whitespace/operators] [3]') + self.assert_lint('typedef hash_map<Foo, Bar', 'Missing spaces around <' + ' [whitespace/operators] [3]') + self.assert_lint('typedef hash_map<FoooooType, BaaaaarType,', '') + self.assert_lint('a<Foo> t+=b;', 'Missing spaces around +=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo> t-=b;', 'Missing spaces around -=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t*=b;', 'Missing spaces around *=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t/=b;', 'Missing spaces around /=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t|=b;', 'Missing spaces around |=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t&=b;', 'Missing spaces around &=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t<<=b;', 'Missing spaces around <<=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t>>=b;', 'Missing spaces around >>=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t>>=&b|c;', 'Missing spaces around >>=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t<<=*b/c;', 'Missing spaces around <<=' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo> t -= b;', '') + self.assert_lint('a<Foo> t += b;', '') + self.assert_lint('a<Foo*> t *= b;', '') + self.assert_lint('a<Foo*> t /= b;', '') + self.assert_lint('a<Foo*> t |= b;', '') + self.assert_lint('a<Foo*> t &= b;', '') + self.assert_lint('a<Foo*> t <<= b;', '') + self.assert_lint('a<Foo*> t >>= b;', '') + self.assert_lint('a<Foo*> t >>= &b|c;', 'Missing spaces around |' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t <<= *b/c;', 'Missing spaces around /' + ' [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t <<= b/c; //Test', [ + 'Should have a space between // and comment ' + '[whitespace/comments] [4]', 'Missing' + ' spaces around / [whitespace/operators] [3]']) + self.assert_lint('a<Foo*> t <<= b||c; //Test', ['One space before end' + ' of line comments [whitespace/comments] [5]', + 'Should have a space between // and comment ' + '[whitespace/comments] [4]', + 'Missing spaces around || [whitespace/operators] [3]']) + self.assert_lint('a<Foo*> t <<= b&&c; // Test', 'Missing spaces around' + ' && [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t <<= b&&&c; // Test', 'Missing spaces around' + ' && [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t <<= b&&*c; // Test', 'Missing spaces around' + ' && [whitespace/operators] [3]') + self.assert_lint('a<Foo*> t <<= b && *c; // Test', '') + self.assert_lint('a<Foo*> t <<= b && &c; // Test', '') + self.assert_lint('a<Foo*> t <<= b || &c; /*Test', 'Complex multi-line ' + '/*...*/-style comment found. Lint may give bogus ' + 'warnings. Consider replacing these with //-style' + ' comments, with #if 0...#endif, or with more clearly' + ' structured multi-line comments. [readability/multiline_comment] [5]') + self.assert_lint('a<Foo&> t <<= &b | &c;', '') + self.assert_lint('a<Foo*> t <<= &b & &c; // Test', '') + self.assert_lint('a<Foo*> t <<= *b / &c; // Test', '') + self.assert_lint('if (a=b == 1)', 'Missing spaces around = [whitespace/operators] [4]') + self.assert_lint('a = 1<<20', 'Missing spaces around << [whitespace/operators] [3]') + self.assert_lint('if (a = b == 1)', '') + self.assert_lint('a = 1 << 20', '') + self.assert_multi_line_lint('#include <sys/io.h>\n', '') + self.assert_multi_line_lint('#import <foo/bar.h>\n', '') + + def test_operator_methods(self): + self.assert_lint('String operator+(const String&, const String&);', '') + self.assert_lint('bool operator==(const String&, const String&);', '') + self.assert_lint('String& operator-=(const String&, const String&);', '') + self.assert_lint('String& operator+=(const String&, const String&);', '') + self.assert_lint('String& operator*=(const String&, const String&);', '') + self.assert_lint('String& operator%=(const String&, const String&);', '') + self.assert_lint('String& operator&=(const String&, const String&);', '') + self.assert_lint('String& operator<<=(const String&, const String&);', '') + self.assert_lint('String& operator>>=(const String&, const String&);', '') + self.assert_lint('String& operator|=(const String&, const String&);', '') + self.assert_lint('String& operator^=(const String&, const String&);', '') + + def test_spacing_before_last_semicolon(self): + self.assert_lint('call_function() ;', + 'Extra space before last semicolon. If this should be an ' + 'empty statement, use { } instead.' + ' [whitespace/semicolon] [5]') + self.assert_lint('while (true) ;', + 'Extra space before last semicolon. If this should be an ' + 'empty statement, use { } instead.' + ' [whitespace/semicolon] [5]') + self.assert_lint('default:;', + 'Semicolon defining empty statement. Use { } instead.' + ' [whitespace/semicolon] [5]') + self.assert_lint(' ;', + 'Line contains only semicolon. If this should be an empty ' + 'statement, use { } instead.' + ' [whitespace/semicolon] [5]') + self.assert_lint('for (int i = 0; ;', '') + + # Static or global STL strings. + def test_static_or_global_stlstrings(self): + self.assert_lint('string foo;', + 'For a static/global string constant, use a C style ' + 'string instead: "char foo[]".' + ' [runtime/string] [4]') + self.assert_lint('string kFoo = "hello"; // English', + 'For a static/global string constant, use a C style ' + 'string instead: "char kFoo[]".' + ' [runtime/string] [4]') + self.assert_lint('static string foo;', + 'For a static/global string constant, use a C style ' + 'string instead: "static char foo[]".' + ' [runtime/string] [4]') + self.assert_lint('static const string foo;', + 'For a static/global string constant, use a C style ' + 'string instead: "static const char foo[]".' + ' [runtime/string] [4]') + self.assert_lint('string Foo::bar;', + 'For a static/global string constant, use a C style ' + 'string instead: "char Foo::bar[]".' + ' [runtime/string] [4]') + # Rare case. + self.assert_lint('string foo("foobar");', + 'For a static/global string constant, use a C style ' + 'string instead: "char foo[]".' + ' [runtime/string] [4]') + # Should not catch local or member variables. + self.assert_lint(' string foo', '') + # Should not catch functions. + self.assert_lint('string EmptyString() { return ""; }', '') + self.assert_lint('string EmptyString () { return ""; }', '') + self.assert_lint('string VeryLongNameFunctionSometimesEndsWith(\n' + ' VeryLongNameType veryLongNameVariable) {}', '') + self.assert_lint('template<>\n' + 'string FunctionTemplateSpecialization<SomeType>(\n' + ' int x) { return ""; }', '') + self.assert_lint('template<>\n' + 'string FunctionTemplateSpecialization<vector<A::B>* >(\n' + ' int x) { return ""; }', '') + + # should not catch methods of template classes. + self.assert_lint('string Class<Type>::Method() const\n' + '{\n' + ' return "";\n' + '}\n', '') + self.assert_lint('string Class<Type>::Method(\n' + ' int arg) const\n' + '{\n' + ' return "";\n' + '}\n', '') + + def test_no_spaces_in_function_calls(self): + self.assert_lint('TellStory(1, 3);', + '') + self.assert_lint('TellStory(1, 3 );', + 'Extra space before )' + ' [whitespace/parens] [2]') + self.assert_lint('TellStory(1 /* wolf */, 3 /* pigs */);', + '') + self.assert_multi_line_lint('#endif\n );', + '') + + def test_two_spaces_between_code_and_comments(self): + self.assert_lint('} // namespace foo', + '') + self.assert_lint('}// namespace foo', + 'One space before end of line comments' + ' [whitespace/comments] [5]') + self.assert_lint('printf("foo"); // Outside quotes.', + '') + self.assert_lint('int i = 0; // Having one space is fine.','') + self.assert_lint('int i = 0; // Having two spaces is bad.', + 'One space before end of line comments' + ' [whitespace/comments] [5]') + self.assert_lint('int i = 0; // Having three spaces is bad.', + 'One space before end of line comments' + ' [whitespace/comments] [5]') + self.assert_lint('// Top level comment', '') + self.assert_lint(' // Line starts with four spaces.', '') + self.assert_lint('foo();\n' + '{ // A scope is opening.', '') + self.assert_lint(' foo();\n' + ' { // An indented scope is opening.', '') + self.assert_lint('if (foo) { // not a pure scope', + '') + self.assert_lint('printf("// In quotes.")', '') + self.assert_lint('printf("\\"%s // In quotes.")', '') + self.assert_lint('printf("%s", "// In quotes.")', '') + + def test_space_after_comment_marker(self): + self.assert_lint('//', '') + self.assert_lint('//x', 'Should have a space between // and comment' + ' [whitespace/comments] [4]') + self.assert_lint('// x', '') + self.assert_lint('//----', '') + self.assert_lint('//====', '') + self.assert_lint('//////', '') + self.assert_lint('////// x', '') + self.assert_lint('/// x', '') + self.assert_lint('////x', 'Should have a space between // and comment' + ' [whitespace/comments] [4]') + + def test_newline_at_eof(self): + def do_test(self, data, is_missing_eof): + error_collector = ErrorCollector(self.assert_) + self.process_file_data('foo.cpp', 'cpp', data.split('\n'), + error_collector) + # The warning appears only once. + self.assertEquals( + int(is_missing_eof), + error_collector.results().count( + 'Could not find a newline character at the end of the file.' + ' [whitespace/ending_newline] [5]')) + + do_test(self, '// Newline\n// at EOF\n', False) + do_test(self, '// No newline\n// at EOF', True) + + def test_invalid_utf8(self): + def do_test(self, raw_bytes, has_invalid_utf8): + error_collector = ErrorCollector(self.assert_) + self.process_file_data('foo.cpp', 'cpp', + unicode(raw_bytes, 'utf8', 'replace').split('\n'), + error_collector) + # The warning appears only once. + self.assertEquals( + int(has_invalid_utf8), + error_collector.results().count( + 'Line contains invalid UTF-8' + ' (or Unicode replacement character).' + ' [readability/utf8] [5]')) + + do_test(self, 'Hello world\n', False) + do_test(self, '\xe9\x8e\xbd\n', False) + do_test(self, '\xe9x\x8e\xbd\n', True) + # This is the encoding of the replacement character itself (which + # you can see by evaluating codecs.getencoder('utf8')(u'\ufffd')). + do_test(self, '\xef\xbf\xbd\n', True) + + def test_is_blank_line(self): + self.assert_(cpp_style.is_blank_line('')) + self.assert_(cpp_style.is_blank_line(' ')) + self.assert_(cpp_style.is_blank_line(' \t\r\n')) + self.assert_(not cpp_style.is_blank_line('int a;')) + self.assert_(not cpp_style.is_blank_line('{')) + + def test_blank_lines_check(self): + self.assert_blank_lines_check(['{\n', '\n', '\n', '}\n'], 1, 1) + self.assert_blank_lines_check([' if (foo) {\n', '\n', ' }\n'], 1, 1) + self.assert_blank_lines_check( + ['\n', '// {\n', '\n', '\n', '// Comment\n', '{\n', '}\n'], 0, 0) + self.assert_blank_lines_check(['\n', 'run("{");\n', '\n'], 0, 0) + self.assert_blank_lines_check(['\n', ' if (foo) { return 0; }\n', '\n'], 0, 0) + + def test_allow_blank_line_before_closing_namespace(self): + error_collector = ErrorCollector(self.assert_) + self.process_file_data('foo.cpp', 'cpp', + ['namespace {', '', '} // namespace'], + error_collector) + self.assertEquals(0, error_collector.results().count( + 'Blank line at the end of a code block. Is this needed?' + ' [whitespace/blank_line] [3]')) + + def test_allow_blank_line_before_if_else_chain(self): + error_collector = ErrorCollector(self.assert_) + self.process_file_data('foo.cpp', 'cpp', + ['if (hoge) {', + '', # No warning + '} else if (piyo) {', + '', # No warning + '} else if (piyopiyo) {', + ' hoge = true;', # No warning + '} else {', + '', # Warning on this line + '}'], + error_collector) + self.assertEquals(1, error_collector.results().count( + 'Blank line at the end of a code block. Is this needed?' + ' [whitespace/blank_line] [3]')) + + def test_else_on_same_line_as_closing_braces(self): + error_collector = ErrorCollector(self.assert_) + self.process_file_data('foo.cpp', 'cpp', + ['if (hoge) {', + '', + '}', + ' else {' # Warning on this line + '', + '}'], + error_collector) + self.assertEquals(1, error_collector.results().count( + 'An else should appear on the same line as the preceding }' + ' [whitespace/newline] [4]')) + + def test_else_clause_not_on_same_line_as_else(self): + self.assert_lint(' else DoSomethingElse();', + 'Else clause should never be on same line as else ' + '(use 2 lines) [whitespace/newline] [4]') + self.assert_lint(' else ifDoSomethingElse();', + 'Else clause should never be on same line as else ' + '(use 2 lines) [whitespace/newline] [4]') + self.assert_lint(' else if (blah) {', '') + self.assert_lint(' variable_ends_in_else = true;', '') + + def test_comma(self): + self.assert_lint('a = f(1,2);', + 'Missing space after , [whitespace/comma] [3]') + self.assert_lint('int tmp=a,a=b,b=tmp;', + ['Missing spaces around = [whitespace/operators] [4]', + 'Missing space after , [whitespace/comma] [3]']) + self.assert_lint('f(a, /* name */ b);', '') + self.assert_lint('f(a, /* name */b);', '') + + def test_declaration(self): + self.assert_lint('int a;', '') + self.assert_lint('int a;', 'Extra space between int and a [whitespace/declaration] [3]') + self.assert_lint('int* a;', 'Extra space between int* and a [whitespace/declaration] [3]') + self.assert_lint('else if { }', '') + self.assert_lint('else if { }', 'Extra space between else and if [whitespace/declaration] [3]') + + def test_pointer_reference_marker_location(self): + self.assert_lint('int* b;', '', 'foo.cpp') + self.assert_lint('int *b;', + 'Declaration has space between type name and * in int *b [whitespace/declaration] [3]', + 'foo.cpp') + self.assert_lint('return *b;', '', 'foo.cpp') + self.assert_lint('delete *b;', '', 'foo.cpp') + self.assert_lint('int *b;', '', 'foo.c') + self.assert_lint('int* b;', + 'Declaration has space between * and variable name in int* b [whitespace/declaration] [3]', + 'foo.c') + self.assert_lint('int& b;', '', 'foo.cpp') + self.assert_lint('int &b;', + 'Declaration has space between type name and & in int &b [whitespace/declaration] [3]', + 'foo.cpp') + self.assert_lint('return &b;', '', 'foo.cpp') + + def test_indent(self): + self.assert_lint('static int noindent;', '') + self.assert_lint(' int fourSpaceIndent;', '') + self.assert_lint(' int oneSpaceIndent;', + 'Weird number of spaces at line-start. ' + 'Are you using a 4-space indent? [whitespace/indent] [3]') + self.assert_lint(' int threeSpaceIndent;', + 'Weird number of spaces at line-start. ' + 'Are you using a 4-space indent? [whitespace/indent] [3]') + self.assert_lint(' char* oneSpaceIndent = "public:";', + 'Weird number of spaces at line-start. ' + 'Are you using a 4-space indent? [whitespace/indent] [3]') + self.assert_lint(' public:', '') + self.assert_lint(' public:', '') + self.assert_lint(' public:', '') + + def test_label(self): + self.assert_lint('public:', + 'Labels should always be indented at least one space. ' + 'If this is a member-initializer list in a constructor, ' + 'the colon should be on the line after the definition ' + 'header. [whitespace/labels] [4]') + self.assert_lint(' public:', '') + self.assert_lint(' public:', '') + self.assert_lint(' public:', '') + self.assert_lint(' public:', '') + self.assert_lint(' public:', '') + + def test_not_alabel(self): + self.assert_lint('MyVeryLongNamespace::MyVeryLongClassName::', '') + + def test_tab(self): + self.assert_lint('\tint a;', + 'Tab found; better to use spaces [whitespace/tab] [1]') + self.assert_lint('int a = 5;\t// set a to 5', + 'Tab found; better to use spaces [whitespace/tab] [1]') + + def test_unnamed_namespaces_in_headers(self): + self.assert_language_rules_check( + 'foo.h', 'namespace {', + 'Do not use unnamed namespaces in header files. See' + ' http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Namespaces' + ' for more information. [build/namespaces] [4]') + # namespace registration macros are OK. + self.assert_language_rules_check('foo.h', 'namespace { \\', '') + # named namespaces are OK. + self.assert_language_rules_check('foo.h', 'namespace foo {', '') + self.assert_language_rules_check('foo.h', 'namespace foonamespace {', '') + self.assert_language_rules_check('foo.cpp', 'namespace {', '') + self.assert_language_rules_check('foo.cpp', 'namespace foo {', '') + + def test_build_class(self): + # Test that the linter can parse to the end of class definitions, + # and that it will report when it can't. + # Use multi-line linter because it performs the ClassState check. + self.assert_multi_line_lint( + 'class Foo {', + 'Failed to find complete declaration of class Foo' + ' [build/class] [5]') + # Don't warn on forward declarations of various types. + self.assert_multi_line_lint( + 'class Foo;', + '') + self.assert_multi_line_lint( + '''struct Foo* + foo = NewFoo();''', + '') + # Here is an example where the linter gets confused, even though + # the code doesn't violate the style guide. + self.assert_multi_line_lint( + '''class Foo + #ifdef DERIVE_FROM_GOO + : public Goo { + #else + : public Hoo { + #endif + };''', + 'Failed to find complete declaration of class Foo' + ' [build/class] [5]') + + def test_build_end_comment(self): + # The crosstool compiler we currently use will fail to compile the + # code in this test, so we might consider removing the lint check. + self.assert_lint('#endif Not a comment', + 'Uncommented text after #endif is non-standard.' + ' Use a comment.' + ' [build/endif_comment] [5]') + + def test_build_forward_decl(self): + # The crosstool compiler we currently use will fail to compile the + # code in this test, so we might consider removing the lint check. + self.assert_lint('class Foo::Goo;', + 'Inner-style forward declarations are invalid.' + ' Remove this line.' + ' [build/forward_decl] [5]') + + def test_build_header_guard(self): + file_path = 'mydir/Foo.h' + + # We can't rely on our internal stuff to get a sane path on the open source + # side of things, so just parse out the suggested header guard. This + # doesn't allow us to test the suggested header guard, but it does let us + # test all the other header tests. + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'h', [], error_collector) + expected_guard = '' + matcher = re.compile( + 'No \#ifndef header guard found\, suggested CPP variable is\: ([A-Za-z_0-9]+) ') + for error in error_collector.result_list(): + matches = matcher.match(error) + if matches: + expected_guard = matches.group(1) + break + + # Make sure we extracted something for our header guard. + self.assertNotEqual(expected_guard, '') + + # Wrong guard + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'h', + ['#ifndef FOO_H', '#define FOO_H'], error_collector) + self.assertEquals( + 1, + error_collector.result_list().count( + '#ifndef header guard has wrong style, please use: %s' + ' [build/header_guard] [5]' % expected_guard), + error_collector.result_list()) + + # No define + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'h', + ['#ifndef %s' % expected_guard], error_collector) + self.assertEquals( + 1, + error_collector.result_list().count( + 'No #ifndef header guard found, suggested CPP variable is: %s' + ' [build/header_guard] [5]' % expected_guard), + error_collector.result_list()) + + # Mismatched define + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'h', + ['#ifndef %s' % expected_guard, + '#define FOO_H'], + error_collector) + self.assertEquals( + 1, + error_collector.result_list().count( + 'No #ifndef header guard found, suggested CPP variable is: %s' + ' [build/header_guard] [5]' % expected_guard), + error_collector.result_list()) + + # No header guard errors + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'h', + ['#ifndef %s' % expected_guard, + '#define %s' % expected_guard, + '#endif // %s' % expected_guard], + error_collector) + for line in error_collector.result_list(): + if line.find('build/header_guard') != -1: + self.fail('Unexpected error: %s' % line) + + # Completely incorrect header guard + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'h', + ['#ifndef FOO', + '#define FOO', + '#endif // FOO'], + error_collector) + self.assertEquals( + 1, + error_collector.result_list().count( + '#ifndef header guard has wrong style, please use: %s' + ' [build/header_guard] [5]' % expected_guard), + error_collector.result_list()) + + # Special case for flymake + error_collector = ErrorCollector(self.assert_) + self.process_file_data('mydir/Foo_flymake.h', 'h', + ['#ifndef %s' % expected_guard, + '#define %s' % expected_guard, + '#endif // %s' % expected_guard], + error_collector) + for line in error_collector.result_list(): + if line.find('build/header_guard') != -1: + self.fail('Unexpected error: %s' % line) + + error_collector = ErrorCollector(self.assert_) + self.process_file_data('mydir/Foo_flymake.h', 'h', [], error_collector) + self.assertEquals( + 1, + error_collector.result_list().count( + 'No #ifndef header guard found, suggested CPP variable is: %s' + ' [build/header_guard] [5]' % expected_guard), + error_collector.result_list()) + + # Allow the WTF_ prefix for files in that directory. + header_guard_filter = FilterConfiguration(('-', '+build/header_guard')) + error_collector = ErrorCollector(self.assert_, header_guard_filter) + self.process_file_data('JavaScriptCore/wtf/TestName.h', 'h', + ['#ifndef WTF_TestName_h', '#define WTF_TestName_h'], + error_collector) + self.assertEquals(0, len(error_collector.result_list()), + error_collector.result_list()) + + # Also allow the non WTF_ prefix for files in that directory. + error_collector = ErrorCollector(self.assert_, header_guard_filter) + self.process_file_data('JavaScriptCore/wtf/TestName.h', 'h', + ['#ifndef TestName_h', '#define TestName_h'], + error_collector) + self.assertEquals(0, len(error_collector.result_list()), + error_collector.result_list()) + + # Verify that we suggest the WTF prefix version. + error_collector = ErrorCollector(self.assert_, header_guard_filter) + self.process_file_data('JavaScriptCore/wtf/TestName.h', 'h', + ['#ifndef BAD_TestName_h', '#define BAD_TestName_h'], + error_collector) + self.assertEquals( + 1, + error_collector.result_list().count( + '#ifndef header guard has wrong style, please use: WTF_TestName_h' + ' [build/header_guard] [5]'), + error_collector.result_list()) + + def test_build_printf_format(self): + self.assert_lint( + r'printf("\%%d", value);', + '%, [, (, and { are undefined character escapes. Unescape them.' + ' [build/printf_format] [3]') + + self.assert_lint( + r'snprintf(buffer, sizeof(buffer), "\[%d", value);', + '%, [, (, and { are undefined character escapes. Unescape them.' + ' [build/printf_format] [3]') + + self.assert_lint( + r'fprintf(file, "\(%d", value);', + '%, [, (, and { are undefined character escapes. Unescape them.' + ' [build/printf_format] [3]') + + self.assert_lint( + r'vsnprintf(buffer, sizeof(buffer), "\\\{%d", ap);', + '%, [, (, and { are undefined character escapes. Unescape them.' + ' [build/printf_format] [3]') + + # Don't warn if double-slash precedes the symbol + self.assert_lint(r'printf("\\%%%d", value);', + '') + + def test_runtime_printf_format(self): + self.assert_lint( + r'fprintf(file, "%q", value);', + '%q in format strings is deprecated. Use %ll instead.' + ' [runtime/printf_format] [3]') + + self.assert_lint( + r'aprintf(file, "The number is %12q", value);', + '%q in format strings is deprecated. Use %ll instead.' + ' [runtime/printf_format] [3]') + + self.assert_lint( + r'printf(file, "The number is" "%-12q", value);', + '%q in format strings is deprecated. Use %ll instead.' + ' [runtime/printf_format] [3]') + + self.assert_lint( + r'printf(file, "The number is" "%+12q", value);', + '%q in format strings is deprecated. Use %ll instead.' + ' [runtime/printf_format] [3]') + + self.assert_lint( + r'printf(file, "The number is" "% 12q", value);', + '%q in format strings is deprecated. Use %ll instead.' + ' [runtime/printf_format] [3]') + + self.assert_lint( + r'snprintf(file, "Never mix %d and %1$d parmaeters!", value);', + '%N$ formats are unconventional. Try rewriting to avoid them.' + ' [runtime/printf_format] [2]') + + def assert_lintLogCodeOnError(self, code, expected_message): + # Special assert_lint which logs the input code on error. + result = self.perform_single_line_lint(code, 'foo.cpp') + if result != expected_message: + self.fail('For code: "%s"\nGot: "%s"\nExpected: "%s"' + % (code, result, expected_message)) + + def test_build_storage_class(self): + qualifiers = [None, 'const', 'volatile'] + signs = [None, 'signed', 'unsigned'] + types = ['void', 'char', 'int', 'float', 'double', + 'schar', 'int8', 'uint8', 'int16', 'uint16', + 'int32', 'uint32', 'int64', 'uint64'] + storage_classes = ['auto', 'extern', 'register', 'static', 'typedef'] + + build_storage_class_error_message = ( + 'Storage class (static, extern, typedef, etc) should be first.' + ' [build/storage_class] [5]') + + # Some explicit cases. Legal in C++, deprecated in C99. + self.assert_lint('const int static foo = 5;', + build_storage_class_error_message) + + self.assert_lint('char static foo;', + build_storage_class_error_message) + + self.assert_lint('double const static foo = 2.0;', + build_storage_class_error_message) + + self.assert_lint('uint64 typedef unsignedLongLong;', + build_storage_class_error_message) + + self.assert_lint('int register foo = 0;', + build_storage_class_error_message) + + # Since there are a very large number of possibilities, randomly + # construct declarations. + # Make sure that the declaration is logged if there's an error. + # Seed generator with an integer for absolute reproducibility. + random.seed(25) + for unused_i in range(10): + # Build up random list of non-storage-class declaration specs. + other_decl_specs = [random.choice(qualifiers), random.choice(signs), + random.choice(types)] + # remove None + other_decl_specs = filter(lambda x: x is not None, other_decl_specs) + + # shuffle + random.shuffle(other_decl_specs) + + # insert storage class after the first + storage_class = random.choice(storage_classes) + insertion_point = random.randint(1, len(other_decl_specs)) + decl_specs = (other_decl_specs[0:insertion_point] + + [storage_class] + + other_decl_specs[insertion_point:]) + + self.assert_lintLogCodeOnError( + ' '.join(decl_specs) + ';', + build_storage_class_error_message) + + # but no error if storage class is first + self.assert_lintLogCodeOnError( + storage_class + ' ' + ' '.join(other_decl_specs), + '') + + def test_legal_copyright(self): + legal_copyright_message = ( + 'No copyright message found. ' + 'You should have a line: "Copyright [year] <Copyright Owner>"' + ' [legal/copyright] [5]') + + copyright_line = '// Copyright 2008 Google Inc. All Rights Reserved.' + + file_path = 'mydir/googleclient/foo.cpp' + + # There should be a copyright message in the first 10 lines + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'cpp', [], error_collector) + self.assertEquals( + 1, + error_collector.result_list().count(legal_copyright_message)) + + error_collector = ErrorCollector(self.assert_) + self.process_file_data( + file_path, 'cpp', + ['' for unused_i in range(10)] + [copyright_line], + error_collector) + self.assertEquals( + 1, + error_collector.result_list().count(legal_copyright_message)) + + # Test that warning isn't issued if Copyright line appears early enough. + error_collector = ErrorCollector(self.assert_) + self.process_file_data(file_path, 'cpp', [copyright_line], error_collector) + for message in error_collector.result_list(): + if message.find('legal/copyright') != -1: + self.fail('Unexpected error: %s' % message) + + error_collector = ErrorCollector(self.assert_) + self.process_file_data( + file_path, 'cpp', + ['' for unused_i in range(9)] + [copyright_line], + error_collector) + for message in error_collector.result_list(): + if message.find('legal/copyright') != -1: + self.fail('Unexpected error: %s' % message) + + def test_invalid_increment(self): + self.assert_lint('*count++;', + 'Changing pointer instead of value (or unused value of ' + 'operator*). [runtime/invalid_increment] [5]') + + +class CleansedLinesTest(unittest.TestCase): + def test_init(self): + lines = ['Line 1', + 'Line 2', + 'Line 3 // Comment test', + 'Line 4 "foo"'] + + clean_lines = cpp_style.CleansedLines(lines) + self.assertEquals(lines, clean_lines.raw_lines) + self.assertEquals(4, clean_lines.num_lines()) + + self.assertEquals(['Line 1', + 'Line 2', + 'Line 3 ', + 'Line 4 "foo"'], + clean_lines.lines) + + self.assertEquals(['Line 1', + 'Line 2', + 'Line 3 ', + 'Line 4 ""'], + clean_lines.elided) + + def test_init_empty(self): + clean_lines = cpp_style.CleansedLines([]) + self.assertEquals([], clean_lines.raw_lines) + self.assertEquals(0, clean_lines.num_lines()) + + def test_collapse_strings(self): + collapse = cpp_style.CleansedLines.collapse_strings + self.assertEquals('""', collapse('""')) # "" (empty) + self.assertEquals('"""', collapse('"""')) # """ (bad) + self.assertEquals('""', collapse('"xyz"')) # "xyz" (string) + self.assertEquals('""', collapse('"\\\""')) # "\"" (string) + self.assertEquals('""', collapse('"\'"')) # "'" (string) + self.assertEquals('"\"', collapse('"\"')) # "\" (bad) + self.assertEquals('""', collapse('"\\\\"')) # "\\" (string) + self.assertEquals('"', collapse('"\\\\\\"')) # "\\\" (bad) + self.assertEquals('""', collapse('"\\\\\\\\"')) # "\\\\" (string) + + self.assertEquals('\'\'', collapse('\'\'')) # '' (empty) + self.assertEquals('\'\'', collapse('\'a\'')) # 'a' (char) + self.assertEquals('\'\'', collapse('\'\\\'\'')) # '\'' (char) + self.assertEquals('\'', collapse('\'\\\'')) # '\' (bad) + self.assertEquals('', collapse('\\012')) # '\012' (char) + self.assertEquals('', collapse('\\xfF0')) # '\xfF0' (char) + self.assertEquals('', collapse('\\n')) # '\n' (char) + self.assertEquals('\#', collapse('\\#')) # '\#' (bad) + + self.assertEquals('StringReplace(body, "", "");', + collapse('StringReplace(body, "\\\\", "\\\\\\\\");')) + self.assertEquals('\'\' ""', + collapse('\'"\' "foo"')) + + +class OrderOfIncludesTest(CppStyleTestBase): + def setUp(self): + self.include_state = cpp_style._IncludeState() + + # Cheat os.path.abspath called in FileInfo class. + self.os_path_abspath_orig = os.path.abspath + os.path.abspath = lambda value: value + + def tearDown(self): + os.path.abspath = self.os_path_abspath_orig + + def test_try_drop_common_suffixes(self): + self.assertEqual('foo/foo', cpp_style._drop_common_suffixes('foo/foo-inl.h')) + self.assertEqual('foo/bar/foo', + cpp_style._drop_common_suffixes('foo/bar/foo_inl.h')) + self.assertEqual('foo/foo', cpp_style._drop_common_suffixes('foo/foo.cpp')) + self.assertEqual('foo/foo_unusualinternal', + cpp_style._drop_common_suffixes('foo/foo_unusualinternal.h')) + self.assertEqual('', + cpp_style._drop_common_suffixes('_test.cpp')) + self.assertEqual('test', + cpp_style._drop_common_suffixes('test.cpp')) + + +class OrderOfIncludesTest(CppStyleTestBase): + def setUp(self): + self.include_state = cpp_style._IncludeState() + + # Cheat os.path.abspath called in FileInfo class. + self.os_path_abspath_orig = os.path.abspath + os.path.abspath = lambda value: value + + def tearDown(self): + os.path.abspath = self.os_path_abspath_orig + + def test_check_next_include_order__no_config(self): + self.assertEqual('Header file should not contain WebCore config.h.', + self.include_state.check_next_include_order(cpp_style._CONFIG_HEADER, True)) + + def test_check_next_include_order__no_self(self): + self.assertEqual('Header file should not contain itself.', + self.include_state.check_next_include_order(cpp_style._PRIMARY_HEADER, True)) + # Test actual code to make sure that header types are correctly assigned. + self.assert_language_rules_check('Foo.h', + '#include "Foo.h"\n', + 'Header file should not contain itself. Should be: alphabetically sorted.' + ' [build/include_order] [4]') + self.assert_language_rules_check('FooBar.h', + '#include "Foo.h"\n', + '') + + def test_check_next_include_order__likely_then_config(self): + self.assertEqual('Found header this file implements before WebCore config.h.', + self.include_state.check_next_include_order(cpp_style._PRIMARY_HEADER, False)) + self.assertEqual('Found WebCore config.h after a header this file implements.', + self.include_state.check_next_include_order(cpp_style._CONFIG_HEADER, False)) + + def test_check_next_include_order__other_then_config(self): + self.assertEqual('Found other header before WebCore config.h.', + self.include_state.check_next_include_order(cpp_style._OTHER_HEADER, False)) + self.assertEqual('Found WebCore config.h after other header.', + self.include_state.check_next_include_order(cpp_style._CONFIG_HEADER, False)) + + def test_check_next_include_order__config_then_other_then_likely(self): + self.assertEqual('', self.include_state.check_next_include_order(cpp_style._CONFIG_HEADER, False)) + self.assertEqual('Found other header before a header this file implements.', + self.include_state.check_next_include_order(cpp_style._OTHER_HEADER, False)) + self.assertEqual('Found header this file implements after other header.', + self.include_state.check_next_include_order(cpp_style._PRIMARY_HEADER, False)) + + def test_check_alphabetical_include_order(self): + self.assert_language_rules_check('foo.h', + '#include "a.h"\n' + '#include "c.h"\n' + '#include "b.h"\n', + 'Alphabetical sorting problem. [build/include_order] [4]') + + self.assert_language_rules_check('foo.h', + '#include "a.h"\n' + '#include "b.h"\n' + '#include "c.h"\n', + '') + + self.assert_language_rules_check('foo.h', + '#include <assert.h>\n' + '#include "bar.h"\n', + 'Alphabetical sorting problem. [build/include_order] [4]') + + self.assert_language_rules_check('foo.h', + '#include "bar.h"\n' + '#include <assert.h>\n', + '') + + def test_check_line_break_after_own_header(self): + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '#include "bar.h"\n', + 'You should add a blank line after implementation file\'s own header. [build/include_order] [4]') + + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '\n' + '#include "bar.h"\n', + '') + + def test_check_preprocessor_in_include_section(self): + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '\n' + '#ifdef BAZ\n' + '#include "baz.h"\n' + '#else\n' + '#include "foobar.h"\n' + '#endif"\n' + '#include "bar.h"\n', # No flag because previous is in preprocessor section + '') + + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '\n' + '#ifdef BAZ\n' + '#include "baz.h"\n' + '#endif"\n' + '#include "bar.h"\n' + '#include "a.h"\n', # Should still flag this. + 'Alphabetical sorting problem. [build/include_order] [4]') + + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '\n' + '#ifdef BAZ\n' + '#include "baz.h"\n' + '#include "bar.h"\n' #Should still flag this + '#endif"\n', + 'Alphabetical sorting problem. [build/include_order] [4]') + + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '\n' + '#ifdef BAZ\n' + '#include "baz.h"\n' + '#endif"\n' + '#ifdef FOOBAR\n' + '#include "foobar.h"\n' + '#endif"\n' + '#include "bar.h"\n' + '#include "a.h"\n', # Should still flag this. + 'Alphabetical sorting problem. [build/include_order] [4]') + + # Check that after an already included error, the sorting rules still work. + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '\n' + '#include "foo.h"\n' + '#include "g.h"\n', + '"foo.h" already included at foo.cpp:2 [build/include] [4]') + + def test_check_wtf_includes(self): + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '\n' + '#include <wtf/Assertions.h>\n', + '') + self.assert_language_rules_check('foo.cpp', + '#include "config.h"\n' + '#include "foo.h"\n' + '\n' + '#include "wtf/Assertions.h"\n', + 'wtf includes should be <wtf/file.h> instead of "wtf/file.h".' + ' [build/include] [4]') + + def test_classify_include(self): + classify_include = cpp_style._classify_include + include_state = cpp_style._IncludeState() + self.assertEqual(cpp_style._CONFIG_HEADER, + classify_include('foo/foo.cpp', + 'config.h', + False, include_state)) + self.assertEqual(cpp_style._PRIMARY_HEADER, + classify_include('foo/internal/foo.cpp', + 'foo/public/foo.h', + False, include_state)) + self.assertEqual(cpp_style._PRIMARY_HEADER, + classify_include('foo/internal/foo.cpp', + 'foo/other/public/foo.h', + False, include_state)) + self.assertEqual(cpp_style._OTHER_HEADER, + classify_include('foo/internal/foo.cpp', + 'foo/other/public/foop.h', + False, include_state)) + self.assertEqual(cpp_style._OTHER_HEADER, + classify_include('foo/foo.cpp', + 'string', + True, include_state)) + self.assertEqual(cpp_style._PRIMARY_HEADER, + classify_include('fooCustom.cpp', + 'foo.h', + False, include_state)) + self.assertEqual(cpp_style._PRIMARY_HEADER, + classify_include('PrefixFooCustom.cpp', + 'Foo.h', + False, include_state)) + self.assertEqual(cpp_style._MOC_HEADER, + classify_include('foo.cpp', + 'foo.moc', + False, include_state)) + self.assertEqual(cpp_style._MOC_HEADER, + classify_include('foo.cpp', + 'moc_foo.cpp', + False, include_state)) + # Tricky example where both includes might be classified as primary. + self.assert_language_rules_check('ScrollbarThemeWince.cpp', + '#include "config.h"\n' + '#include "ScrollbarThemeWince.h"\n' + '\n' + '#include "Scrollbar.h"\n', + '') + self.assert_language_rules_check('ScrollbarThemeWince.cpp', + '#include "config.h"\n' + '#include "Scrollbar.h"\n' + '\n' + '#include "ScrollbarThemeWince.h"\n', + 'Found header this file implements after a header this file implements.' + ' Should be: config.h, primary header, blank line, and then alphabetically sorted.' + ' [build/include_order] [4]') + self.assert_language_rules_check('ResourceHandleWin.cpp', + '#include "config.h"\n' + '#include "ResourceHandle.h"\n' + '\n' + '#include "ResourceHandleWin.h"\n', + '') + + def test_try_drop_common_suffixes(self): + self.assertEqual('foo/foo', cpp_style._drop_common_suffixes('foo/foo-inl.h')) + self.assertEqual('foo/bar/foo', + cpp_style._drop_common_suffixes('foo/bar/foo_inl.h')) + self.assertEqual('foo/foo', cpp_style._drop_common_suffixes('foo/foo.cpp')) + self.assertEqual('foo/foo_unusualinternal', + cpp_style._drop_common_suffixes('foo/foo_unusualinternal.h')) + self.assertEqual('', + cpp_style._drop_common_suffixes('_test.cpp')) + self.assertEqual('test', + cpp_style._drop_common_suffixes('test.cpp')) + self.assertEqual('test', + cpp_style._drop_common_suffixes('test.cpp')) + +class CheckForFunctionLengthsTest(CppStyleTestBase): + def setUp(self): + # Reducing these thresholds for the tests speeds up tests significantly. + self.old_normal_trigger = cpp_style._FunctionState._NORMAL_TRIGGER + self.old_test_trigger = cpp_style._FunctionState._TEST_TRIGGER + + cpp_style._FunctionState._NORMAL_TRIGGER = 10 + cpp_style._FunctionState._TEST_TRIGGER = 25 + + def tearDown(self): + cpp_style._FunctionState._NORMAL_TRIGGER = self.old_normal_trigger + cpp_style._FunctionState._TEST_TRIGGER = self.old_test_trigger + + # FIXME: Eliminate the need for this function. + def set_min_confidence(self, min_confidence): + """Set new test confidence and return old test confidence.""" + old_min_confidence = self.min_confidence + self.min_confidence = min_confidence + return old_min_confidence + + def assert_function_lengths_check(self, code, expected_message): + """Check warnings for long function bodies are as expected. + + Args: + code: C++ source code expected to generate a warning message. + expected_message: Message expected to be generated by the C++ code. + """ + self.assertEquals(expected_message, + self.perform_function_lengths_check(code)) + + def trigger_lines(self, error_level): + """Return number of lines needed to trigger a function length warning. + + Args: + error_level: --v setting for cpp_style. + + Returns: + Number of lines needed to trigger a function length warning. + """ + return cpp_style._FunctionState._NORMAL_TRIGGER * 2 ** error_level + + def trigger_test_lines(self, error_level): + """Return number of lines needed to trigger a test function length warning. + + Args: + error_level: --v setting for cpp_style. + + Returns: + Number of lines needed to trigger a test function length warning. + """ + return cpp_style._FunctionState._TEST_TRIGGER * 2 ** error_level + + def assert_function_length_check_definition(self, lines, error_level): + """Generate long function definition and check warnings are as expected. + + Args: + lines: Number of lines to generate. + error_level: --v setting for cpp_style. + """ + trigger_level = self.trigger_lines(self.min_confidence) + self.assert_function_lengths_check( + 'void test(int x)' + self.function_body(lines), + ('Small and focused functions are preferred: ' + 'test() has %d non-comment lines ' + '(error triggered by exceeding %d lines).' + ' [readability/fn_size] [%d]' + % (lines, trigger_level, error_level))) + + def assert_function_length_check_definition_ok(self, lines): + """Generate shorter function definition and check no warning is produced. + + Args: + lines: Number of lines to generate. + """ + self.assert_function_lengths_check( + 'void test(int x)' + self.function_body(lines), + '') + + def assert_function_length_check_at_error_level(self, error_level): + """Generate and check function at the trigger level for --v setting. + + Args: + error_level: --v setting for cpp_style. + """ + self.assert_function_length_check_definition(self.trigger_lines(error_level), + error_level) + + def assert_function_length_check_below_error_level(self, error_level): + """Generate and check function just below the trigger level for --v setting. + + Args: + error_level: --v setting for cpp_style. + """ + self.assert_function_length_check_definition(self.trigger_lines(error_level) - 1, + error_level - 1) + + def assert_function_length_check_above_error_level(self, error_level): + """Generate and check function just above the trigger level for --v setting. + + Args: + error_level: --v setting for cpp_style. + """ + self.assert_function_length_check_definition(self.trigger_lines(error_level) + 1, + error_level) + + def function_body(self, number_of_lines): + return ' {\n' + ' this_is_just_a_test();\n' * number_of_lines + '}' + + def function_body_with_blank_lines(self, number_of_lines): + return ' {\n' + ' this_is_just_a_test();\n\n' * number_of_lines + '}' + + def function_body_with_no_lints(self, number_of_lines): + return ' {\n' + ' this_is_just_a_test(); // NOLINT\n' * number_of_lines + '}' + + # Test line length checks. + def test_function_length_check_declaration(self): + self.assert_function_lengths_check( + 'void test();', # Not a function definition + '') + + def test_function_length_check_declaration_with_block_following(self): + self.assert_function_lengths_check( + ('void test();\n' + + self.function_body(66)), # Not a function definition + '') + + def test_function_length_check_class_definition(self): + self.assert_function_lengths_check( # Not a function definition + 'class Test' + self.function_body(66) + ';', + '') + + def test_function_length_check_trivial(self): + self.assert_function_lengths_check( + 'void test() {}', # Not counted + '') + + def test_function_length_check_empty(self): + self.assert_function_lengths_check( + 'void test() {\n}', + '') + + def test_function_length_check_definition_below_severity0(self): + old_min_confidence = self.set_min_confidence(0) + self.assert_function_length_check_definition_ok(self.trigger_lines(0) - 1) + self.set_min_confidence(old_min_confidence) + + def test_function_length_check_definition_at_severity0(self): + old_min_confidence = self.set_min_confidence(0) + self.assert_function_length_check_definition_ok(self.trigger_lines(0)) + self.set_min_confidence(old_min_confidence) + + def test_function_length_check_definition_above_severity0(self): + old_min_confidence = self.set_min_confidence(0) + self.assert_function_length_check_above_error_level(0) + self.set_min_confidence(old_min_confidence) + + def test_function_length_check_definition_below_severity1v0(self): + old_min_confidence = self.set_min_confidence(0) + self.assert_function_length_check_below_error_level(1) + self.set_min_confidence(old_min_confidence) + + def test_function_length_check_definition_at_severity1v0(self): + old_min_confidence = self.set_min_confidence(0) + self.assert_function_length_check_at_error_level(1) + self.set_min_confidence(old_min_confidence) + + def test_function_length_check_definition_below_severity1(self): + self.assert_function_length_check_definition_ok(self.trigger_lines(1) - 1) + + def test_function_length_check_definition_at_severity1(self): + self.assert_function_length_check_definition_ok(self.trigger_lines(1)) + + def test_function_length_check_definition_above_severity1(self): + self.assert_function_length_check_above_error_level(1) + + def test_function_length_check_definition_severity1_plus_indented(self): + error_level = 1 + error_lines = self.trigger_lines(error_level) + 1 + trigger_level = self.trigger_lines(self.min_confidence) + indent_spaces = ' ' + self.assert_function_lengths_check( + re.sub(r'(?m)^(.)', indent_spaces + r'\1', + 'void test_indent(int x)\n' + self.function_body(error_lines)), + ('Small and focused functions are preferred: ' + 'test_indent() has %d non-comment lines ' + '(error triggered by exceeding %d lines).' + ' [readability/fn_size] [%d]') + % (error_lines, trigger_level, error_level)) + + def test_function_length_check_definition_severity1_plus_blanks(self): + error_level = 1 + error_lines = self.trigger_lines(error_level) + 1 + trigger_level = self.trigger_lines(self.min_confidence) + self.assert_function_lengths_check( + 'void test_blanks(int x)' + self.function_body(error_lines), + ('Small and focused functions are preferred: ' + 'test_blanks() has %d non-comment lines ' + '(error triggered by exceeding %d lines).' + ' [readability/fn_size] [%d]') + % (error_lines, trigger_level, error_level)) + + def test_function_length_check_complex_definition_severity1(self): + error_level = 1 + error_lines = self.trigger_lines(error_level) + 1 + trigger_level = self.trigger_lines(self.min_confidence) + self.assert_function_lengths_check( + ('my_namespace::my_other_namespace::MyVeryLongTypeName<Type1, bool func(const Element*)>*\n' + 'my_namespace::my_other_namespace<Type3, Type4>::~MyFunction<Type5<Type6, Type7> >(int arg1, char* arg2)' + + self.function_body(error_lines)), + ('Small and focused functions are preferred: ' + 'my_namespace::my_other_namespace<Type3, Type4>::~MyFunction<Type5<Type6, Type7> >()' + ' has %d non-comment lines ' + '(error triggered by exceeding %d lines).' + ' [readability/fn_size] [%d]') + % (error_lines, trigger_level, error_level)) + + def test_function_length_check_definition_severity1_for_test(self): + error_level = 1 + error_lines = self.trigger_test_lines(error_level) + 1 + trigger_level = self.trigger_test_lines(self.min_confidence) + self.assert_function_lengths_check( + 'TEST_F(Test, Mutator)' + self.function_body(error_lines), + ('Small and focused functions are preferred: ' + 'TEST_F(Test, Mutator) has %d non-comment lines ' + '(error triggered by exceeding %d lines).' + ' [readability/fn_size] [%d]') + % (error_lines, trigger_level, error_level)) + + def test_function_length_check_definition_severity1_for_split_line_test(self): + error_level = 1 + error_lines = self.trigger_test_lines(error_level) + 1 + trigger_level = self.trigger_test_lines(self.min_confidence) + self.assert_function_lengths_check( + ('TEST_F(GoogleUpdateRecoveryRegistryProtectedTest,\n' + ' FixGoogleUpdate_AllValues_MachineApp)' # note: 4 spaces + + self.function_body(error_lines)), + ('Small and focused functions are preferred: ' + 'TEST_F(GoogleUpdateRecoveryRegistryProtectedTest, ' # 1 space + 'FixGoogleUpdate_AllValues_MachineApp) has %d non-comment lines ' + '(error triggered by exceeding %d lines).' + ' [readability/fn_size] [%d]') + % (error_lines, trigger_level, error_level)) + + def test_function_length_check_definition_severity1_for_bad_test_doesnt_break(self): + error_level = 1 + error_lines = self.trigger_test_lines(error_level) + 1 + trigger_level = self.trigger_test_lines(self.min_confidence) + self.assert_function_lengths_check( + ('TEST_F(' + + self.function_body(error_lines)), + ('Small and focused functions are preferred: ' + 'TEST_F has %d non-comment lines ' + '(error triggered by exceeding %d lines).' + ' [readability/fn_size] [%d]') + % (error_lines, trigger_level, error_level)) + + def test_function_length_check_definition_severity1_with_embedded_no_lints(self): + error_level = 1 + error_lines = self.trigger_lines(error_level) + 1 + trigger_level = self.trigger_lines(self.min_confidence) + self.assert_function_lengths_check( + 'void test(int x)' + self.function_body_with_no_lints(error_lines), + ('Small and focused functions are preferred: ' + 'test() has %d non-comment lines ' + '(error triggered by exceeding %d lines).' + ' [readability/fn_size] [%d]') + % (error_lines, trigger_level, error_level)) + + def test_function_length_check_definition_severity1_with_no_lint(self): + self.assert_function_lengths_check( + ('void test(int x)' + self.function_body(self.trigger_lines(1)) + + ' // NOLINT -- long function'), + '') + + def test_function_length_check_definition_below_severity2(self): + self.assert_function_length_check_below_error_level(2) + + def test_function_length_check_definition_severity2(self): + self.assert_function_length_check_at_error_level(2) + + def test_function_length_check_definition_above_severity2(self): + self.assert_function_length_check_above_error_level(2) + + def test_function_length_check_definition_below_severity3(self): + self.assert_function_length_check_below_error_level(3) + + def test_function_length_check_definition_severity3(self): + self.assert_function_length_check_at_error_level(3) + + def test_function_length_check_definition_above_severity3(self): + self.assert_function_length_check_above_error_level(3) + + def test_function_length_check_definition_below_severity4(self): + self.assert_function_length_check_below_error_level(4) + + def test_function_length_check_definition_severity4(self): + self.assert_function_length_check_at_error_level(4) + + def test_function_length_check_definition_above_severity4(self): + self.assert_function_length_check_above_error_level(4) + + def test_function_length_check_definition_below_severity5(self): + self.assert_function_length_check_below_error_level(5) + + def test_function_length_check_definition_at_severity5(self): + self.assert_function_length_check_at_error_level(5) + + def test_function_length_check_definition_above_severity5(self): + self.assert_function_length_check_above_error_level(5) + + def test_function_length_check_definition_huge_lines(self): + # 5 is the limit + self.assert_function_length_check_definition(self.trigger_lines(10), 5) + + def test_function_length_not_determinable(self): + # Macro invocation without terminating semicolon. + self.assert_function_lengths_check( + 'MACRO(arg)', + '') + + # Macro with underscores + self.assert_function_lengths_check( + 'MACRO_WITH_UNDERSCORES(arg1, arg2, arg3)', + '') + + self.assert_function_lengths_check( + 'NonMacro(arg)', + 'Lint failed to find start of function body.' + ' [readability/fn_size] [5]') + + +class NoNonVirtualDestructorsTest(CppStyleTestBase): + + def test_no_error(self): + self.assert_multi_line_lint( + '''class Foo { + virtual ~Foo(); + virtual void foo(); + };''', + '') + + self.assert_multi_line_lint( + '''class Foo { + virtual inline ~Foo(); + virtual void foo(); + };''', + '') + + self.assert_multi_line_lint( + '''class Foo { + inline virtual ~Foo(); + virtual void foo(); + };''', + '') + + self.assert_multi_line_lint( + '''class Foo::Goo { + virtual ~Goo(); + virtual void goo(); + };''', + '') + self.assert_multi_line_lint( + 'class Foo { void foo(); };', + 'More than one command on the same line [whitespace/newline] [4]') + self.assert_multi_line_lint( + 'class MyClass {\n' + ' int getIntValue() { ASSERT(m_ptr); return *m_ptr; }\n' + '};\n', + '') + self.assert_multi_line_lint( + 'class MyClass {\n' + ' int getIntValue()\n' + ' {\n' + ' ASSERT(m_ptr); return *m_ptr;\n' + ' }\n' + '};\n', + 'More than one command on the same line [whitespace/newline] [4]') + + self.assert_multi_line_lint( + '''class Qualified::Goo : public Foo { + virtual void goo(); + };''', + '') + + self.assert_multi_line_lint( + # Line-ending : + '''class Goo : + public Foo { + virtual void goo(); + };''', + 'Labels should always be indented at least one space. If this is a ' + 'member-initializer list in a constructor, the colon should be on the ' + 'line after the definition header. [whitespace/labels] [4]') + + def test_no_destructor_when_virtual_needed(self): + self.assert_multi_line_lint_re( + '''class Foo { + virtual void foo(); + };''', + 'The class Foo probably needs a virtual destructor') + + def test_destructor_non_virtual_when_virtual_needed(self): + self.assert_multi_line_lint_re( + '''class Foo { + ~Foo(); + virtual void foo(); + };''', + 'The class Foo probably needs a virtual destructor') + + def test_no_warn_when_derived(self): + self.assert_multi_line_lint( + '''class Foo : public Goo { + virtual void foo(); + };''', + '') + + def test_internal_braces(self): + self.assert_multi_line_lint_re( + '''class Foo { + enum Goo { + GOO + }; + virtual void foo(); + };''', + 'The class Foo probably needs a virtual destructor') + + def test_inner_class_needs_virtual_destructor(self): + self.assert_multi_line_lint_re( + '''class Foo { + class Goo { + virtual void goo(); + }; + };''', + 'The class Goo probably needs a virtual destructor') + + def test_outer_class_needs_virtual_destructor(self): + self.assert_multi_line_lint_re( + '''class Foo { + class Goo { + }; + virtual void foo(); + };''', + 'The class Foo probably needs a virtual destructor') + + def test_qualified_class_needs_virtual_destructor(self): + self.assert_multi_line_lint_re( + '''class Qualified::Foo { + virtual void foo(); + };''', + 'The class Qualified::Foo probably needs a virtual destructor') + + def test_multi_line_declaration_no_error(self): + self.assert_multi_line_lint_re( + '''class Foo + : public Goo { + virtual void foo(); + };''', + '') + + def test_multi_line_declaration_with_error(self): + self.assert_multi_line_lint( + '''class Foo + { + virtual void foo(); + };''', + ['This { should be at the end of the previous line ' + '[whitespace/braces] [4]', + 'The class Foo probably needs a virtual destructor due to having ' + 'virtual method(s), one declared at line 3. [runtime/virtual] [4]']) + + +class PassPtrTest(CppStyleTestBase): + # For http://webkit.org/coding/RefPtr.html + + def assert_pass_ptr_check(self, code, expected_message): + """Check warnings for Pass*Ptr are as expected. + + Args: + code: C++ source code expected to generate a warning message. + expected_message: Message expected to be generated by the C++ code. + """ + self.assertEquals(expected_message, + self.perform_pass_ptr_check(code)) + + def test_pass_ref_ptr_in_function(self): + # Local variables should never be PassRefPtr. + self.assert_pass_ptr_check( + 'int myFunction()\n' + '{\n' + ' PassRefPtr<Type1> variable = variable2;\n' + '}', + 'Local variables should never be PassRefPtr (see ' + 'http://webkit.org/coding/RefPtr.html). [readability/pass_ptr] [5]') + + def test_pass_own_ptr_in_function(self): + # Local variables should never be PassRefPtr. + self.assert_pass_ptr_check( + 'int myFunction()\n' + '{\n' + ' PassOwnPtr<Type1> variable = variable2;\n' + '}', + 'Local variables should never be PassOwnPtr (see ' + 'http://webkit.org/coding/RefPtr.html). [readability/pass_ptr] [5]') + + def test_pass_other_type_ptr_in_function(self): + # Local variables should never be PassRefPtr. + self.assert_pass_ptr_check( + 'int myFunction()\n' + '{\n' + ' PassOtherTypePtr<Type1> variable;\n' + '}', + 'Local variables should never be PassOtherTypePtr (see ' + 'http://webkit.org/coding/RefPtr.html). [readability/pass_ptr] [5]') + + def test_pass_ref_ptr_return_value(self): + self.assert_pass_ptr_check( + 'PassRefPtr<Type1>\n' + 'myFunction(int)\n' + '{\n' + '}', + '') + self.assert_pass_ptr_check( + 'PassRefPtr<Type1> myFunction(int)\n' + '{\n' + '}', + '') + self.assert_pass_ptr_check( + 'PassRefPtr<Type1> myFunction();\n', + '') + + def test_pass_ref_ptr_parameter_value(self): + self.assert_pass_ptr_check( + 'int myFunction(PassRefPtr<Type1>)\n' + '{\n' + '}', + '') + + def test_ref_ptr_member_variable(self): + self.assert_pass_ptr_check( + 'class Foo {' + ' RefPtr<Type1> m_other;\n' + '};\n', + '') + + +class WebKitStyleTest(CppStyleTestBase): + + # for http://webkit.org/coding/coding-style.html + def test_indentation(self): + # 1. Use spaces, not tabs. Tabs should only appear in files that + # require them for semantic meaning, like Makefiles. + self.assert_multi_line_lint( + 'class Foo {\n' + ' int goo;\n' + '};', + '') + self.assert_multi_line_lint( + 'class Foo {\n' + '\tint goo;\n' + '};', + 'Tab found; better to use spaces [whitespace/tab] [1]') + + # 2. The indent size is 4 spaces. + self.assert_multi_line_lint( + 'class Foo {\n' + ' int goo;\n' + '};', + '') + self.assert_multi_line_lint( + 'class Foo {\n' + ' int goo;\n' + '};', + 'Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3]') + # FIXME: No tests for 8-spaces. + + # 3. In a header, code inside a namespace should not be indented. + self.assert_multi_line_lint( + 'namespace WebCore {\n\n' + 'class Document {\n' + ' int myVariable;\n' + '};\n' + '}', + '', + 'foo.h') + self.assert_multi_line_lint( + 'namespace OuterNamespace {\n' + ' namespace InnerNamespace {\n' + ' class Document {\n' + '};\n' + '};\n' + '}', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.h') + self.assert_multi_line_lint( + 'namespace OuterNamespace {\n' + ' class Document {\n' + ' namespace InnerNamespace {\n' + '};\n' + '};\n' + '}', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.h') + self.assert_multi_line_lint( + 'namespace WebCore {\n' + '#if 0\n' + ' class Document {\n' + '};\n' + '#endif\n' + '}', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.h') + self.assert_multi_line_lint( + 'namespace WebCore {\n' + 'class Document {\n' + '};\n' + '}', + '', + 'foo.h') + + # 4. In an implementation file (files with the extension .cpp, .c + # or .mm), code inside a namespace should not be indented. + self.assert_multi_line_lint( + 'namespace WebCore {\n\n' + 'Document::Foo()\n' + ' : foo(bar)\n' + ' , boo(far)\n' + '{\n' + ' stuff();\n' + '}', + '', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace OuterNamespace {\n' + 'namespace InnerNamespace {\n' + 'Document::Foo() { }\n' + ' void* p;\n' + '}\n' + '}\n', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace OuterNamespace {\n' + 'namespace InnerNamespace {\n' + 'Document::Foo() { }\n' + '}\n' + ' void* p;\n' + '}\n', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace WebCore {\n\n' + ' const char* foo = "start:;"\n' + ' "dfsfsfs";\n' + '}\n', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace WebCore {\n\n' + 'const char* foo(void* a = ";", // ;\n' + ' void* b);\n' + ' void* p;\n' + '}\n', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace WebCore {\n\n' + 'const char* foo[] = {\n' + ' "void* b);", // ;\n' + ' "asfdf",\n' + ' }\n' + ' void* p;\n' + '}\n', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace WebCore {\n\n' + 'const char* foo[] = {\n' + ' "void* b);", // }\n' + ' "asfdf",\n' + ' }\n' + '}\n', + '', + 'foo.cpp') + self.assert_multi_line_lint( + ' namespace WebCore {\n\n' + ' void Document::Foo()\n' + ' {\n' + 'start: // infinite loops are fun!\n' + ' goto start;\n' + ' }', + 'namespace should never be indented. [whitespace/indent] [4]', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace WebCore {\n' + ' Document::Foo() { }\n' + '}', + 'Code inside a namespace should not be indented.' + ' [whitespace/indent] [4]', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace WebCore {\n' + '#define abc(x) x; \\\n' + ' x\n' + '}', + '', + 'foo.cpp') + self.assert_multi_line_lint( + 'namespace WebCore {\n' + '#define abc(x) x; \\\n' + ' x\n' + ' void* x;' + '}', + 'Code inside a namespace should not be indented. [whitespace/indent] [4]', + 'foo.cpp') + + # 5. A case label should line up with its switch statement. The + # case statement is indented. + self.assert_multi_line_lint( + ' switch (condition) {\n' + ' case fooCondition:\n' + ' case barCondition:\n' + ' i++;\n' + ' break;\n' + ' default:\n' + ' i--;\n' + ' }\n', + '') + self.assert_multi_line_lint( + ' switch (condition) {\n' + ' case fooCondition:\n' + ' switch (otherCondition) {\n' + ' default:\n' + ' return;\n' + ' }\n' + ' default:\n' + ' i--;\n' + ' }\n', + '') + self.assert_multi_line_lint( + ' switch (condition) {\n' + ' case fooCondition: break;\n' + ' default: return;\n' + ' }\n', + '') + self.assert_multi_line_lint( + ' switch (condition) {\n' + ' case fooCondition:\n' + ' case barCondition:\n' + ' i++;\n' + ' break;\n' + ' default:\n' + ' i--;\n' + ' }\n', + 'A case label should not be indented, but line up with its switch statement.' + ' [whitespace/indent] [4]') + self.assert_multi_line_lint( + ' switch (condition) {\n' + ' case fooCondition:\n' + ' break;\n' + ' default:\n' + ' i--;\n' + ' }\n', + 'A case label should not be indented, but line up with its switch statement.' + ' [whitespace/indent] [4]') + self.assert_multi_line_lint( + ' switch (condition) {\n' + ' case fooCondition:\n' + ' case barCondition:\n' + ' switch (otherCondition) {\n' + ' default:\n' + ' return;\n' + ' }\n' + ' default:\n' + ' i--;\n' + ' }\n', + 'A case label should not be indented, but line up with its switch statement.' + ' [whitespace/indent] [4]') + self.assert_multi_line_lint( + ' switch (condition) {\n' + ' case fooCondition:\n' + ' case barCondition:\n' + ' i++;\n' + ' break;\n\n' + ' default:\n' + ' i--;\n' + ' }\n', + 'Non-label code inside switch statements should be indented.' + ' [whitespace/indent] [4]') + self.assert_multi_line_lint( + ' switch (condition) {\n' + ' case fooCondition:\n' + ' case barCondition:\n' + ' switch (otherCondition) {\n' + ' default:\n' + ' return;\n' + ' }\n' + ' default:\n' + ' i--;\n' + ' }\n', + 'Non-label code inside switch statements should be indented.' + ' [whitespace/indent] [4]') + + # 6. Boolean expressions at the same nesting level that span + # multiple lines should have their operators on the left side of + # the line instead of the right side. + self.assert_multi_line_lint( + ' return attr->name() == srcAttr\n' + ' || attr->name() == lowsrcAttr;\n', + '') + self.assert_multi_line_lint( + ' return attr->name() == srcAttr ||\n' + ' attr->name() == lowsrcAttr;\n', + 'Boolean expressions that span multiple lines should have their ' + 'operators on the left side of the line instead of the right side.' + ' [whitespace/operators] [4]') + + def test_spacing(self): + # 1. Do not place spaces around unary operators. + self.assert_multi_line_lint( + 'i++;', + '') + self.assert_multi_line_lint( + 'i ++;', + 'Extra space for operator ++; [whitespace/operators] [4]') + + # 2. Do place spaces around binary and ternary operators. + self.assert_multi_line_lint( + 'y = m * x + b;', + '') + self.assert_multi_line_lint( + 'f(a, b);', + '') + self.assert_multi_line_lint( + 'c = a | b;', + '') + self.assert_multi_line_lint( + 'return condition ? 1 : 0;', + '') + self.assert_multi_line_lint( + 'y=m*x+b;', + 'Missing spaces around = [whitespace/operators] [4]') + self.assert_multi_line_lint( + 'f(a,b);', + 'Missing space after , [whitespace/comma] [3]') + self.assert_multi_line_lint( + 'c = a|b;', + 'Missing spaces around | [whitespace/operators] [3]') + # FIXME: We cannot catch this lint error. + # self.assert_multi_line_lint( + # 'return condition ? 1:0;', + # '') + + # 3. Place spaces between control statements and their parentheses. + self.assert_multi_line_lint( + ' if (condition)\n' + ' doIt();\n', + '') + self.assert_multi_line_lint( + ' if(condition)\n' + ' doIt();\n', + 'Missing space before ( in if( [whitespace/parens] [5]') + + # 4. Do not place spaces between a function and its parentheses, + # or between a parenthesis and its content. + self.assert_multi_line_lint( + 'f(a, b);', + '') + self.assert_multi_line_lint( + 'f (a, b);', + 'Extra space before ( in function call [whitespace/parens] [4]') + self.assert_multi_line_lint( + 'f( a, b );', + ['Extra space after ( in function call [whitespace/parens] [4]', + 'Extra space before ) [whitespace/parens] [2]']) + + def test_line_breaking(self): + # 1. Each statement should get its own line. + self.assert_multi_line_lint( + ' x++;\n' + ' y++;\n' + ' if (condition);\n' + ' doIt();\n', + '') + self.assert_multi_line_lint( + ' if (condition) \\\n' + ' doIt();\n', + '') + self.assert_multi_line_lint( + ' x++; y++;', + 'More than one command on the same line [whitespace/newline] [4]') + self.assert_multi_line_lint( + ' if (condition) doIt();\n', + 'More than one command on the same line in if [whitespace/parens] [4]') + # Ensure that having a # in the line doesn't hide the error. + self.assert_multi_line_lint( + ' x++; char a[] = "#";', + 'More than one command on the same line [whitespace/newline] [4]') + # Ignore preprocessor if's. + self.assert_multi_line_lint( + ' #if (condition) || (condition2)\n', + '') + + # 2. An else statement should go on the same line as a preceding + # close brace if one is present, else it should line up with the + # if statement. + self.assert_multi_line_lint( + 'if (condition) {\n' + ' doSomething();\n' + ' doSomethingAgain();\n' + '} else {\n' + ' doSomethingElse();\n' + ' doSomethingElseAgain();\n' + '}\n', + '') + self.assert_multi_line_lint( + 'if (condition)\n' + ' doSomething();\n' + 'else\n' + ' doSomethingElse();\n', + '') + self.assert_multi_line_lint( + 'if (condition)\n' + ' doSomething();\n' + 'else {\n' + ' doSomethingElse();\n' + ' doSomethingElseAgain();\n' + '}\n', + '') + self.assert_multi_line_lint( + '#define TEST_ASSERT(expression) do { if (!(expression)) { TestsController::shared().testFailed(__FILE__, __LINE__, #expression); return; } } while (0)\n', + '') + self.assert_multi_line_lint( + '#define TEST_ASSERT(expression) do { if ( !(expression)) { TestsController::shared().testFailed(__FILE__, __LINE__, #expression); return; } } while (0)\n', + 'Extra space after ( in if [whitespace/parens] [5]') + # FIXME: currently we only check first conditional, so we cannot detect errors in next ones. + # self.assert_multi_line_lint( + # '#define TEST_ASSERT(expression) do { if (!(expression)) { TestsController::shared().testFailed(__FILE__, __LINE__, #expression); return; } } while (0 )\n', + # 'Mismatching spaces inside () in if [whitespace/parens] [5]') + self.assert_multi_line_lint( + 'if (condition) {\n' + ' doSomething();\n' + ' doSomethingAgain();\n' + '}\n' + 'else {\n' + ' doSomethingElse();\n' + ' doSomethingElseAgain();\n' + '}\n', + 'An else should appear on the same line as the preceding } [whitespace/newline] [4]') + self.assert_multi_line_lint( + 'if (condition) doSomething(); else doSomethingElse();\n', + ['More than one command on the same line [whitespace/newline] [4]', + 'Else clause should never be on same line as else (use 2 lines) [whitespace/newline] [4]', + 'More than one command on the same line in if [whitespace/parens] [4]']) + self.assert_multi_line_lint( + 'if (condition) doSomething(); else {\n' + ' doSomethingElse();\n' + '}\n', + ['More than one command on the same line in if [whitespace/parens] [4]', + 'One line control clauses should not use braces. [whitespace/braces] [4]']) + self.assert_multi_line_lint( + 'void func()\n' + '{\n' + ' while (condition) { }\n' + ' return 0;\n' + '}\n', + '') + self.assert_multi_line_lint( + 'void func()\n' + '{\n' + ' for (i = 0; i < 42; i++) { foobar(); }\n' + ' return 0;\n' + '}\n', + 'More than one command on the same line in for [whitespace/parens] [4]') + + # 3. An else if statement should be written as an if statement + # when the prior if concludes with a return statement. + self.assert_multi_line_lint( + 'if (motivated) {\n' + ' if (liquid)\n' + ' return money;\n' + '} else if (tired)\n' + ' break;\n', + '') + self.assert_multi_line_lint( + 'if (condition)\n' + ' doSomething();\n' + 'else if (otherCondition)\n' + ' doSomethingElse();\n', + '') + self.assert_multi_line_lint( + 'if (condition)\n' + ' doSomething();\n' + 'else\n' + ' doSomethingElse();\n', + '') + self.assert_multi_line_lint( + 'if (condition)\n' + ' returnValue = foo;\n' + 'else if (otherCondition)\n' + ' returnValue = bar;\n', + '') + self.assert_multi_line_lint( + 'if (condition)\n' + ' returnValue = foo;\n' + 'else\n' + ' returnValue = bar;\n', + '') + self.assert_multi_line_lint( + 'if (condition)\n' + ' doSomething();\n' + 'else if (liquid)\n' + ' return money;\n' + 'else if (broke)\n' + ' return favor;\n' + 'else\n' + ' sleep(28800);\n', + '') + self.assert_multi_line_lint( + 'if (liquid) {\n' + ' prepare();\n' + ' return money;\n' + '} else if (greedy) {\n' + ' keep();\n' + ' return nothing;\n' + '}\n', + 'An else if statement should be written as an if statement when the ' + 'prior "if" concludes with a return, break, continue or goto statement.' + ' [readability/control_flow] [4]') + self.assert_multi_line_lint( + ' if (stupid) {\n' + 'infiniteLoop:\n' + ' goto infiniteLoop;\n' + ' } else if (evil)\n' + ' goto hell;\n', + 'An else if statement should be written as an if statement when the ' + 'prior "if" concludes with a return, break, continue or goto statement.' + ' [readability/control_flow] [4]') + self.assert_multi_line_lint( + 'if (liquid)\n' + '{\n' + ' prepare();\n' + ' return money;\n' + '}\n' + 'else if (greedy)\n' + ' keep();\n', + ['This { should be at the end of the previous line [whitespace/braces] [4]', + 'An else should appear on the same line as the preceding } [whitespace/newline] [4]', + 'An else if statement should be written as an if statement when the ' + 'prior "if" concludes with a return, break, continue or goto statement.' + ' [readability/control_flow] [4]']) + self.assert_multi_line_lint( + 'if (gone)\n' + ' return;\n' + 'else if (here)\n' + ' go();\n', + 'An else if statement should be written as an if statement when the ' + 'prior "if" concludes with a return, break, continue or goto statement.' + ' [readability/control_flow] [4]') + self.assert_multi_line_lint( + 'if (gone)\n' + ' return;\n' + 'else\n' + ' go();\n', + 'An else statement can be removed when the prior "if" concludes ' + 'with a return, break, continue or goto statement.' + ' [readability/control_flow] [4]') + self.assert_multi_line_lint( + 'if (motivated) {\n' + ' prepare();\n' + ' continue;\n' + '} else {\n' + ' cleanUp();\n' + ' break;\n' + '}\n', + 'An else statement can be removed when the prior "if" concludes ' + 'with a return, break, continue or goto statement.' + ' [readability/control_flow] [4]') + self.assert_multi_line_lint( + 'if (tired)\n' + ' break;\n' + 'else {\n' + ' prepare();\n' + ' continue;\n' + '}\n', + 'An else statement can be removed when the prior "if" concludes ' + 'with a return, break, continue or goto statement.' + ' [readability/control_flow] [4]') + + def test_braces(self): + # 1. Function definitions: place each brace on its own line. + self.assert_multi_line_lint( + 'int main()\n' + '{\n' + ' doSomething();\n' + '}\n', + '') + self.assert_multi_line_lint( + 'int main() {\n' + ' doSomething();\n' + '}\n', + 'Place brace on its own line for function definitions. [whitespace/braces] [4]') + + # 2. Other braces: place the open brace on the line preceding the + # code block; place the close brace on its own line. + self.assert_multi_line_lint( + 'class MyClass {\n' + ' int foo;\n' + '};\n', + '') + self.assert_multi_line_lint( + 'namespace WebCore {\n' + 'int foo;\n' + '};\n', + '') + self.assert_multi_line_lint( + 'for (int i = 0; i < 10; i++) {\n' + ' DoSomething();\n' + '};\n', + '') + self.assert_multi_line_lint( + 'class MyClass\n' + '{\n' + ' int foo;\n' + '};\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'if (condition)\n' + '{\n' + ' int foo;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'for (int i = 0; i < 10; i++)\n' + '{\n' + ' int foo;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'while (true)\n' + '{\n' + ' int foo;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'foreach (Foo* foo, foos)\n' + '{\n' + ' int bar;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'switch (type)\n' + '{\n' + 'case foo: return;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'if (condition)\n' + '{\n' + ' int foo;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'for (int i = 0; i < 10; i++)\n' + '{\n' + ' int foo;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'while (true)\n' + '{\n' + ' int foo;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'switch (type)\n' + '{\n' + 'case foo: return;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + self.assert_multi_line_lint( + 'else if (type)\n' + '{\n' + 'case foo: return;\n' + '}\n', + 'This { should be at the end of the previous line [whitespace/braces] [4]') + + # 3. One-line control clauses should not use braces unless + # comments are included or a single statement spans multiple + # lines. + self.assert_multi_line_lint( + 'if (true) {\n' + ' int foo;\n' + '}\n', + 'One line control clauses should not use braces. [whitespace/braces] [4]') + + self.assert_multi_line_lint( + 'for (; foo; bar) {\n' + ' int foo;\n' + '}\n', + 'One line control clauses should not use braces. [whitespace/braces] [4]') + + self.assert_multi_line_lint( + 'foreach (foo, foos) {\n' + ' int bar;\n' + '}\n', + 'One line control clauses should not use braces. [whitespace/braces] [4]') + + self.assert_multi_line_lint( + 'while (true) {\n' + ' int foo;\n' + '}\n', + 'One line control clauses should not use braces. [whitespace/braces] [4]') + + self.assert_multi_line_lint( + 'if (true)\n' + ' int foo;\n' + 'else {\n' + ' int foo;\n' + '}\n', + 'One line control clauses should not use braces. [whitespace/braces] [4]') + + self.assert_multi_line_lint( + 'if (true) {\n' + ' int foo;\n' + '} else\n' + ' int foo;\n', + 'One line control clauses should not use braces. [whitespace/braces] [4]') + + self.assert_multi_line_lint( + 'if (true) {\n' + ' // Some comment\n' + ' int foo;\n' + '}\n', + '') + + self.assert_multi_line_lint( + 'if (true) {\n' + ' myFunction(reallyLongParam1, reallyLongParam2,\n' + ' reallyLongParam3);\n' + '}\n', + '') + + # 4. Control clauses without a body should use empty braces. + self.assert_multi_line_lint( + 'for ( ; current; current = current->next) { }\n', + '') + self.assert_multi_line_lint( + 'for ( ; current;\n' + ' current = current->next) {}\n', + '') + self.assert_multi_line_lint( + 'for ( ; current; current = current->next);\n', + 'Semicolon defining empty statement for this loop. Use { } instead. [whitespace/semicolon] [5]') + self.assert_multi_line_lint( + 'while (true);\n', + 'Semicolon defining empty statement for this loop. Use { } instead. [whitespace/semicolon] [5]') + self.assert_multi_line_lint( + '} while (true);\n', + '') + + def test_null_false_zero(self): + # 1. In C++, the null pointer value should be written as 0. In C, + # it should be written as NULL. In Objective-C and Objective-C++, + # follow the guideline for C or C++, respectively, but use nil to + # represent a null Objective-C object. + self.assert_lint( + 'functionCall(NULL)', + 'Use 0 instead of NULL.' + ' [readability/null] [5]', + 'foo.cpp') + self.assert_lint( + "// Don't use NULL in comments since it isn't in code.", + 'Use 0 instead of NULL.' + ' [readability/null] [4]', + 'foo.cpp') + self.assert_lint( + '"A string with NULL" // and a comment with NULL is tricky to flag correctly in cpp_style.', + 'Use 0 instead of NULL.' + ' [readability/null] [4]', + 'foo.cpp') + self.assert_lint( + '"A string containing NULL is ok"', + '', + 'foo.cpp') + self.assert_lint( + 'if (aboutNULL)', + '', + 'foo.cpp') + self.assert_lint( + 'myVariable = NULLify', + '', + 'foo.cpp') + # Make sure that the NULL check does not apply to C and Objective-C files. + self.assert_lint( + 'functionCall(NULL)', + '', + 'foo.c') + self.assert_lint( + 'functionCall(NULL)', + '', + 'foo.m') + + # Make sure that the NULL check does not apply to g_object_{set,get} and + # g_str{join,concat} + self.assert_lint( + 'g_object_get(foo, "prop", &bar, NULL);', + '') + self.assert_lint( + 'g_object_set(foo, "prop", bar, NULL);', + '') + self.assert_lint( + 'g_build_filename(foo, bar, NULL);', + '') + self.assert_lint( + 'gst_bin_add_many(foo, bar, boo, NULL);', + '') + self.assert_lint( + 'gst_bin_remove_many(foo, bar, boo, NULL);', + '') + self.assert_lint( + 'gst_element_link_many(foo, bar, boo, NULL);', + '') + self.assert_lint( + 'gst_element_unlink_many(foo, bar, boo, NULL);', + '') + self.assert_lint( + 'gchar* result = g_strconcat("part1", "part2", "part3", NULL);', + '') + self.assert_lint( + 'gchar* result = g_strconcat("part1", NULL);', + '') + self.assert_lint( + 'gchar* result = g_strjoin(",", "part1", "part2", "part3", NULL);', + '') + self.assert_lint( + 'gchar* result = g_strjoin(",", "part1", NULL);', + '') + self.assert_lint( + 'gchar* result = gdk_pixbuf_save_to_callback(pixbuf, function, data, type, error, NULL);', + '') + self.assert_lint( + 'gchar* result = gdk_pixbuf_save_to_buffer(pixbuf, function, data, type, error, NULL);', + '') + self.assert_lint( + 'gchar* result = gdk_pixbuf_save_to_stream(pixbuf, function, data, type, error, NULL);', + '') + + # 2. C++ and C bool values should be written as true and + # false. Objective-C BOOL values should be written as YES and NO. + # FIXME: Implement this. + + # 3. Tests for true/false, null/non-null, and zero/non-zero should + # all be done without equality comparisons. + self.assert_lint( + 'if (count == 0)', + 'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.' + ' [readability/comparison_to_zero] [5]') + self.assert_lint_one_of_many_errors_re( + 'if (string != NULL)', + r'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons\.') + self.assert_lint( + 'if (condition == true)', + 'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.' + ' [readability/comparison_to_zero] [5]') + self.assert_lint( + 'if (myVariable != /* Why would anyone put a comment here? */ false)', + 'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.' + ' [readability/comparison_to_zero] [5]') + + self.assert_lint( + 'if (0 /* This comment also looks odd to me. */ != aLongerVariableName)', + 'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.' + ' [readability/comparison_to_zero] [5]') + self.assert_lint_one_of_many_errors_re( + 'if (NULL == thisMayBeNull)', + r'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons\.') + self.assert_lint( + 'if (true != anotherCondition)', + 'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.' + ' [readability/comparison_to_zero] [5]') + self.assert_lint( + 'if (false == myBoolValue)', + 'Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.' + ' [readability/comparison_to_zero] [5]') + + self.assert_lint( + 'if (fontType == trueType)', + '') + self.assert_lint( + 'if (othertrue == fontType)', + '') + + def test_using_std(self): + self.assert_lint( + 'using std::min;', + "Use 'using namespace std;' instead of 'using std::min;'." + " [build/using_std] [4]", + 'foo.cpp') + + def test_max_macro(self): + self.assert_lint( + 'int i = MAX(0, 1);', + '', + 'foo.c') + + self.assert_lint( + 'int i = MAX(0, 1);', + 'Use std::max() or std::max<type>() instead of the MAX() macro.' + ' [runtime/max_min_macros] [4]', + 'foo.cpp') + + self.assert_lint( + 'inline int foo() { return MAX(0, 1); }', + 'Use std::max() or std::max<type>() instead of the MAX() macro.' + ' [runtime/max_min_macros] [4]', + 'foo.h') + + def test_min_macro(self): + self.assert_lint( + 'int i = MIN(0, 1);', + '', + 'foo.c') + + self.assert_lint( + 'int i = MIN(0, 1);', + 'Use std::min() or std::min<type>() instead of the MIN() macro.' + ' [runtime/max_min_macros] [4]', + 'foo.cpp') + + self.assert_lint( + 'inline int foo() { return MIN(0, 1); }', + 'Use std::min() or std::min<type>() instead of the MIN() macro.' + ' [runtime/max_min_macros] [4]', + 'foo.h') + + def test_names(self): + name_underscore_error_message = " is incorrectly named. Don't use underscores in your identifier names. [readability/naming] [4]" + name_tooshort_error_message = " is incorrectly named. Don't use the single letter 'l' as an identifier name. [readability/naming] [4]" + + # Basic cases from WebKit style guide. + self.assert_lint('struct Data;', '') + self.assert_lint('size_t bufferSize;', '') + self.assert_lint('class HTMLDocument;', '') + self.assert_lint('String mimeType();', '') + self.assert_lint('size_t buffer_size;', + 'buffer_size' + name_underscore_error_message) + self.assert_lint('short m_length;', '') + self.assert_lint('short _length;', + '_length' + name_underscore_error_message) + self.assert_lint('short length_;', + 'length_' + name_underscore_error_message) + self.assert_lint('unsigned _length;', + '_length' + name_underscore_error_message) + self.assert_lint('unsigned int _length;', + '_length' + name_underscore_error_message) + self.assert_lint('unsigned long long _length;', + '_length' + name_underscore_error_message) + + # Allow underscores in Objective C files. + self.assert_lint('unsigned long long _length;', + '', + 'foo.m') + self.assert_lint('unsigned long long _length;', + '', + 'foo.mm') + self.assert_lint('#import "header_file.h"\n' + 'unsigned long long _length;', + '', + 'foo.h') + self.assert_lint('unsigned long long _length;\n' + '@interface WebFullscreenWindow;', + '', + 'foo.h') + self.assert_lint('unsigned long long _length;\n' + '@implementation WebFullscreenWindow;', + '', + 'foo.h') + self.assert_lint('unsigned long long _length;\n' + '@class WebWindowFadeAnimation;', + '', + 'foo.h') + + # Variable name 'l' is easy to confuse with '1' + self.assert_lint('int l;', 'l' + name_tooshort_error_message) + self.assert_lint('size_t l;', 'l' + name_tooshort_error_message) + self.assert_lint('long long l;', 'l' + name_tooshort_error_message) + + # Pointers, references, functions, templates, and adjectives. + self.assert_lint('char* under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('const int UNDER_SCORE;', + 'UNDER_SCORE' + name_underscore_error_message) + self.assert_lint('static inline const char const& const under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('WebCore::RenderObject* under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('int func_name();', + 'func_name' + name_underscore_error_message) + self.assert_lint('RefPtr<RenderObject*> under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('WTF::Vector<WTF::RefPtr<const RenderObject* const> > under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('int under_score[];', + 'under_score' + name_underscore_error_message) + self.assert_lint('struct dirent* under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('long under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('long long under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('long double under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('long long int under_score;', + 'under_score' + name_underscore_error_message) + + # Declarations in control statement. + self.assert_lint('if (int under_score = 42) {', + 'under_score' + name_underscore_error_message) + self.assert_lint('else if (int under_score = 42) {', + 'under_score' + name_underscore_error_message) + self.assert_lint('for (int under_score = 42; cond; i++) {', + 'under_score' + name_underscore_error_message) + self.assert_lint('while (foo & under_score = bar) {', + 'under_score' + name_underscore_error_message) + self.assert_lint('for (foo * under_score = p; cond; i++) {', + 'under_score' + name_underscore_error_message) + self.assert_lint('for (foo * under_score; cond; i++) {', + 'under_score' + name_underscore_error_message) + self.assert_lint('while (foo & value_in_thirdparty_library) {', '') + self.assert_lint('while (foo * value_in_thirdparty_library) {', '') + self.assert_lint('if (mli && S_OK == mli->foo()) {', '') + + # More member variables and functions. + self.assert_lint('int SomeClass::s_validName', '') + self.assert_lint('int m_under_score;', + 'm_under_score' + name_underscore_error_message) + self.assert_lint('int SomeClass::s_under_score = 0;', + 'SomeClass::s_under_score' + name_underscore_error_message) + self.assert_lint('int SomeClass::under_score = 0;', + 'SomeClass::under_score' + name_underscore_error_message) + + # Other statements. + self.assert_lint('return INT_MAX;', '') + self.assert_lint('return_t under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('goto under_score;', + 'under_score' + name_underscore_error_message) + self.assert_lint('delete static_cast<Foo*>(p);', '') + + # Multiple variables in one line. + self.assert_lint('void myFunction(int variable1, int another_variable);', + 'another_variable' + name_underscore_error_message) + self.assert_lint('int variable1, another_variable;', + 'another_variable' + name_underscore_error_message) + self.assert_lint('int first_variable, secondVariable;', + 'first_variable' + name_underscore_error_message) + self.assert_lint('void my_function(int variable_1, int variable_2);', + ['my_function' + name_underscore_error_message, + 'variable_1' + name_underscore_error_message, + 'variable_2' + name_underscore_error_message]) + self.assert_lint('for (int variable_1, variable_2;;) {', + ['variable_1' + name_underscore_error_message, + 'variable_2' + name_underscore_error_message]) + + # There is an exception for op code functions but only in the JavaScriptCore directory. + self.assert_lint('void this_op_code(int var1, int var2)', '', 'JavaScriptCore/foo.cpp') + self.assert_lint('void op_code(int var1, int var2)', '', 'JavaScriptCore/foo.cpp') + self.assert_lint('void this_op_code(int var1, int var2)', 'this_op_code' + name_underscore_error_message) + + # GObject requires certain magical names in class declarations. + self.assert_lint('void webkit_dom_object_init();', '') + self.assert_lint('void webkit_dom_object_class_init();', '') + + # There is an exception for some unit tests that begin with "tst_". + self.assert_lint('void tst_QWebFrame::arrayObjectEnumerable(int var1, int var2)', '') + + # The Qt API uses names that begin with "qt_". + self.assert_lint('void QTFrame::qt_drt_is_awesome(int var1, int var2)', '') + self.assert_lint('void qt_drt_is_awesome(int var1, int var2);', '') + + # Cairo forward-declarations should not be a failure. + self.assert_lint('typedef struct _cairo cairo_t;', '') + self.assert_lint('typedef struct _cairo_surface cairo_surface_t;', '') + self.assert_lint('typedef struct _cairo_scaled_font cairo_scaled_font_t;', '') + + # NPAPI functions that start with NPN_, NPP_ or NP_ are allowed. + self.assert_lint('void NPN_Status(NPP, const char*)', '') + self.assert_lint('NPError NPP_SetWindow(NPP instance, NPWindow *window)', '') + self.assert_lint('NPObject* NP_Allocate(NPP, NPClass*)', '') + + # const_iterator is allowed as well. + self.assert_lint('typedef VectorType::const_iterator const_iterator;', '') + + # vm_throw is allowed as well. + self.assert_lint('int vm_throw;', '') + + # Bitfields. + self.assert_lint('unsigned _fillRule : 1;', + '_fillRule' + name_underscore_error_message) + + # new operators in initialization. + self.assert_lint('OwnPtr<uint32_t> variable(new uint32_t);', '') + self.assert_lint('OwnPtr<uint32_t> variable(new (expr) uint32_t);', '') + self.assert_lint('OwnPtr<uint32_t> under_score(new uint32_t);', + 'under_score' + name_underscore_error_message) + + + def test_comments(self): + # A comment at the beginning of a line is ok. + self.assert_lint('// comment', '') + self.assert_lint(' // comment', '') + + self.assert_lint('} // namespace WebCore', + 'One space before end of line comments' + ' [whitespace/comments] [5]') + + def test_other(self): + # FIXME: Implement this. + pass + + +class CppCheckerTest(unittest.TestCase): + + """Tests CppChecker class.""" + + def mock_handle_style_error(self): + pass + + def _checker(self): + return CppChecker("foo", "h", self.mock_handle_style_error, 3) + + def test_init(self): + """Test __init__ constructor.""" + checker = self._checker() + self.assertEquals(checker.file_extension, "h") + self.assertEquals(checker.file_path, "foo") + self.assertEquals(checker.handle_style_error, self.mock_handle_style_error) + self.assertEquals(checker.min_confidence, 3) + + def test_eq(self): + """Test __eq__ equality function.""" + checker1 = self._checker() + checker2 = self._checker() + + # == calls __eq__. + self.assertTrue(checker1 == checker2) + + def mock_handle_style_error2(self): + pass + + # Verify that a difference in any argument cause equality to fail. + checker = CppChecker("foo", "h", self.mock_handle_style_error, 3) + self.assertFalse(checker == CppChecker("bar", "h", self.mock_handle_style_error, 3)) + self.assertFalse(checker == CppChecker("foo", "c", self.mock_handle_style_error, 3)) + self.assertFalse(checker == CppChecker("foo", "h", mock_handle_style_error2, 3)) + self.assertFalse(checker == CppChecker("foo", "h", self.mock_handle_style_error, 4)) + + def test_ne(self): + """Test __ne__ inequality function.""" + checker1 = self._checker() + checker2 = self._checker() + + # != calls __ne__. + # By default, __ne__ always returns true on different objects. + # Thus, just check the distinguishing case to verify that the + # code defines __ne__. + self.assertFalse(checker1 != checker2) + + +def tearDown(): + """A global check to make sure all error-categories have been tested. + + The main tearDown() routine is the only code we can guarantee will be + run after all other tests have been executed. + """ + try: + if _run_verifyallcategoriesseen: + ErrorCollector(None).verify_all_categories_are_seen() + except NameError: + # If nobody set the global _run_verifyallcategoriesseen, then + # we assume we shouldn't run the test + pass + +if __name__ == '__main__': + import sys + # We don't want to run the verify_all_categories_are_seen() test unless + # we're running the full test suite: if we only run one test, + # obviously we're not going to see all the error categories. So we + # only run verify_all_categories_are_seen() when no commandline flags + # are passed in. + global _run_verifyallcategoriesseen + _run_verifyallcategoriesseen = (len(sys.argv) == 1) + + unittest.main() diff --git a/Tools/Scripts/webkitpy/style/checkers/python.py b/Tools/Scripts/webkitpy/style/checkers/python.py new file mode 100644 index 0000000..70d4450 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/python.py @@ -0,0 +1,56 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND +# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR +# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Supports checking WebKit style in Python files.""" + +from ...style_references import pep8 + + +class PythonChecker(object): + + """Processes text lines for checking style.""" + + def __init__(self, file_path, handle_style_error): + self._file_path = file_path + self._handle_style_error = handle_style_error + + def check(self, lines): + # Initialize pep8.options, which is necessary for + # Checker.check_all() to execute. + pep8.process_options(arglist=[self._file_path]) + + checker = pep8.Checker(self._file_path) + + def _pep8_handle_error(line_number, offset, text, check): + # FIXME: Incorporate the character offset into the error output. + # This will require updating the error handler __call__ + # signature to include an optional "offset" parameter. + pep8_code = text[:4] + pep8_message = text[5:] + + category = "pep8/" + pep8_code + + self._handle_style_error(line_number, category, 5, pep8_message) + + checker.report_error = _pep8_handle_error + + errors = checker.check_all() diff --git a/Tools/Scripts/webkitpy/style/checkers/python_unittest.py b/Tools/Scripts/webkitpy/style/checkers/python_unittest.py new file mode 100644 index 0000000..e003eb8 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/python_unittest.py @@ -0,0 +1,62 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND +# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR +# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit tests for python.py.""" + +import os +import unittest + +from python import PythonChecker + + +class PythonCheckerTest(unittest.TestCase): + + """Tests the PythonChecker class.""" + + def test_init(self): + """Test __init__() method.""" + def _mock_handle_style_error(self): + pass + + checker = PythonChecker("foo.txt", _mock_handle_style_error) + self.assertEquals(checker._file_path, "foo.txt") + self.assertEquals(checker._handle_style_error, + _mock_handle_style_error) + + def test_check(self): + """Test check() method.""" + errors = [] + + def _mock_handle_style_error(line_number, category, confidence, + message): + error = (line_number, category, confidence, message) + errors.append(error) + + current_dir = os.path.dirname(__file__) + file_path = os.path.join(current_dir, "python_unittest_input.py") + + checker = PythonChecker(file_path, _mock_handle_style_error) + checker.check(lines=[]) + + self.assertEquals(len(errors), 1) + self.assertEquals(errors[0], + (2, "pep8/W291", 5, "trailing whitespace")) diff --git a/Tools/Scripts/webkitpy/style/checkers/python_unittest_input.py b/Tools/Scripts/webkitpy/style/checkers/python_unittest_input.py new file mode 100644 index 0000000..9f1d118 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/python_unittest_input.py @@ -0,0 +1,2 @@ +# This file is sample input for python_unittest.py and includes a single +# error which is an extra space at the end of this line. diff --git a/Tools/Scripts/webkitpy/style/checkers/test_expectations.py b/Tools/Scripts/webkitpy/style/checkers/test_expectations.py new file mode 100644 index 0000000..c86b32c --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/test_expectations.py @@ -0,0 +1,120 @@ +# Copyright (C) 2010 Google Inc. All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Checks WebKit style for test_expectations files.""" + +import logging +import os +import re +import sys + +from common import TabChecker +from webkitpy.style_references import port +from webkitpy.style_references import test_expectations + +_log = logging.getLogger("webkitpy.style.checkers.test_expectations") + + +class ChromiumOptions(object): + """A mock object for creating chromium port object. + + port.get() requires an options object which has 'chromium' attribute to create + chromium port object for each platform. This class mocks such object. + """ + def __init__(self): + self.chromium = True + + +class TestExpectationsChecker(object): + """Processes test_expectations.txt lines for validating the syntax.""" + + categories = set(['test/expectations']) + + def __init__(self, file_path, handle_style_error): + self._file_path = file_path + self._handle_style_error = handle_style_error + self._tab_checker = TabChecker(file_path, handle_style_error) + self._output_regex = re.compile('Line:(?P<line>\d+)\s*(?P<message>.+)') + # Determining the port of this expectations. + try: + port_name = self._file_path.split(os.sep)[-2] + if port_name == "chromium": + options = ChromiumOptions() + self._port_obj = port.get(port_name=None, options=options) + else: + self._port_obj = port.get(port_name=port_name) + except: + # Using 'test' port when we couldn't determine the port for this + # expectations. + _log.warn("Could not determine the port for %s. " + "Using 'test' port, but platform-specific expectations " + "will fail the check." % self._file_path) + self._port_obj = port.get('test') + self._port_to_check = self._port_obj.test_platform_name() + # Suppress error messages of test_expectations module since they will be + # reported later. + log = logging.getLogger("webkitpy.layout_tests.layout_package." + "test_expectations") + log.setLevel(logging.CRITICAL) + + def _handle_error_message(self, lineno, message, confidence): + pass + + def check_test_expectations(self, expectations_str, tests=None, overrides=None): + err = None + expectations = None + try: + expectations = test_expectations.TestExpectationsFile( + port=self._port_obj, expectations=expectations_str, full_test_list=tests, + test_platform_name=self._port_to_check, is_debug_mode=False, + is_lint_mode=True, overrides=overrides) + except test_expectations.ParseError, error: + err = error + + if err: + level = 2 + if err.fatal: + level = 5 + for error in err.errors: + matched = self._output_regex.match(error) + if matched: + lineno, message = matched.group('line', 'message') + self._handle_style_error(int(lineno), 'test/expectations', level, message) + + + def check_tabs(self, lines): + self._tab_checker.check(lines) + + def check(self, lines): + overrides = self._port_obj.test_expectations_overrides() + expectations = '\n'.join(lines) + self.check_test_expectations(expectations_str=expectations, + tests=None, + overrides=overrides) + # Warn tabs in lines as well + self.check_tabs(lines) diff --git a/Tools/Scripts/webkitpy/style/checkers/test_expectations_unittest.py b/Tools/Scripts/webkitpy/style/checkers/test_expectations_unittest.py new file mode 100644 index 0000000..9817c5d --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/test_expectations_unittest.py @@ -0,0 +1,173 @@ +#!/usr/bin/python +# Copyright (C) 2010 Google Inc. All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit tests for test_expectations.py.""" + +import os +import sys +import unittest + +# We need following workaround hack to run this unit tests in stand-alone. +try: + d = os.path.dirname(__file__) +except NameError: + d = os.path.dirname(sys.argv[0]) +sys.path.append(os.path.abspath(os.path.join(d, '../../../'))) + +from test_expectations import TestExpectationsChecker +from webkitpy.style_references import port +from webkitpy.style_references import test_expectations as test_expectations_style + + +class ErrorCollector(object): + """An error handler class for unit tests.""" + + def __init__(self): + self._errors = [] + + def __call__(self, lineno, category, confidence, message): + self._errors.append('%s [%s] [%d]' % (message, category, confidence)) + + def get_errors(self): + return ''.join(self._errors) + + def reset_errors(self): + self._errors = [] + + +class TestExpectationsTestCase(unittest.TestCase): + """TestCase for test_expectations.py""" + + def setUp(self): + self._error_collector = ErrorCollector() + port_obj = port.get('test') + self._test_file = os.path.join(port_obj.layout_tests_dir(), 'passes/text.html') + + def process_expectations(self, expectations, overrides=None): + self._checker = TestExpectationsChecker() + + def assert_lines_lint(self, lines, expected): + self._error_collector.reset_errors() + checker = TestExpectationsChecker('test/test_expectations.txt', + self._error_collector) + checker.check_test_expectations(expectations_str='\n'.join(lines), + tests=[self._test_file], + overrides=None) + checker.check_tabs(lines) + self.assertEqual(expected, self._error_collector.get_errors()) + + def test_valid_expectations(self): + self.assert_lines_lint( + ["passes/text.html = PASS"], + "") + self.assert_lines_lint( + ["passes/text.html = FAIL PASS"], + "") + self.assert_lines_lint( + ["passes/text.html = CRASH TIMEOUT FAIL PASS"], + "") + self.assert_lines_lint( + ["BUGCR1234 MAC : passes/text.html = PASS FAIL"], + "") + self.assert_lines_lint( + ["SKIP BUGCR1234 : passes/text.html = TIMEOUT PASS"], + "") + self.assert_lines_lint( + ["BUGCR1234 DEBUG : passes/text.html = TIMEOUT PASS"], + "") + self.assert_lines_lint( + ["BUGCR1234 DEBUG SKIP : passes/text.html = TIMEOUT PASS"], + "") + self.assert_lines_lint( + ["BUGCR1234 MAC DEBUG SKIP : passes/text.html = TIMEOUT PASS"], + "") + self.assert_lines_lint( + ["BUGCR1234 DEBUG MAC : passes/text.html = TIMEOUT PASS"], + "") + self.assert_lines_lint( + ["SLOW BUGCR1234 : passes/text.html = PASS"], + "") + self.assert_lines_lint( + ["WONTFIX SKIP : passes/text.html = TIMEOUT"], + "") + + def test_modifier_errors(self): + self.assert_lines_lint( + ["BUG1234 : passes/text.html = FAIL"], + 'Bug must be either BUGCR, BUGWK, or BUGV8_ for test: bug1234 passes/text.html [test/expectations] [5]') + + def test_valid_modifiers(self): + self.assert_lines_lint( + ["INVALID-MODIFIER : passes/text.html = PASS"], + "Invalid modifier for test: invalid-modifier " + "passes/text.html [test/expectations] [5]") + self.assert_lines_lint( + ["SKIP : passes/text.html = PASS"], + "Test lacks BUG modifier. " + "passes/text.html [test/expectations] [2]") + + def test_expectation_errors(self): + self.assert_lines_lint( + ["missing expectations"], + "Missing expectations. ['missing expectations'] [test/expectations] [5]") + self.assert_lines_lint( + ["SLOW : passes/text.html = TIMEOUT"], + "A test can not be both slow and timeout. " + "If it times out indefinitely, then it should be just timeout. " + "passes/text.html [test/expectations] [5]") + self.assert_lines_lint( + ["does/not/exist.html = FAIL"], + "Path does not exist. does/not/exist.html [test/expectations] [2]") + + def test_parse_expectations(self): + self.assert_lines_lint( + ["passes/text.html = PASS"], + "") + self.assert_lines_lint( + ["passes/text.html = UNSUPPORTED"], + "Unsupported expectation: unsupported " + "passes/text.html [test/expectations] [5]") + self.assert_lines_lint( + ["passes/text.html = PASS UNSUPPORTED"], + "Unsupported expectation: unsupported " + "passes/text.html [test/expectations] [5]") + + def test_already_seen_test(self): + self.assert_lines_lint( + ["passes/text.html = PASS", + "passes/text.html = TIMEOUT"], + "Duplicate expectation. %s [test/expectations] [5]" % self._test_file) + + def test_tab(self): + self.assert_lines_lint( + ["\tpasses/text.html = PASS"], + "Line contains tab character. [whitespace/tab] [5]") + +if __name__ == '__main__': + unittest.main() diff --git a/Tools/Scripts/webkitpy/style/checkers/text.py b/Tools/Scripts/webkitpy/style/checkers/text.py new file mode 100644 index 0000000..1147658 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/text.py @@ -0,0 +1,51 @@ +# Copyright (C) 2009 Google Inc. All rights reserved. +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Checks WebKit style for text files.""" + +from common import TabChecker + +class TextChecker(object): + + """Processes text lines for checking style.""" + + def __init__(self, file_path, handle_style_error): + self.file_path = file_path + self.handle_style_error = handle_style_error + self._tab_checker = TabChecker(file_path, handle_style_error) + + def check(self, lines): + self._tab_checker.check(lines) + + +# FIXME: Remove this function (requires refactoring unit tests). +def process_file_data(filename, lines, error): + checker = TextChecker(filename, error) + checker.check(lines) + diff --git a/Tools/Scripts/webkitpy/style/checkers/text_unittest.py b/Tools/Scripts/webkitpy/style/checkers/text_unittest.py new file mode 100644 index 0000000..ced49a9 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/text_unittest.py @@ -0,0 +1,94 @@ +#!/usr/bin/python +# Copyright (C) 2009 Google Inc. All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit test for text_style.py.""" + +import unittest + +import text as text_style +from text import TextChecker + +class TextStyleTestCase(unittest.TestCase): + """TestCase for text_style.py""" + + def assertNoError(self, lines): + """Asserts that the specified lines has no errors.""" + self.had_error = False + + def error_for_test(line_number, category, confidence, message): + """Records if an error occurs.""" + self.had_error = True + + text_style.process_file_data('', lines, error_for_test) + self.assert_(not self.had_error, '%s should not have any errors.' % lines) + + def assertError(self, lines, expected_line_number): + """Asserts that the specified lines has an error.""" + self.had_error = False + + def error_for_test(line_number, category, confidence, message): + """Checks if the expected error occurs.""" + self.assertEquals(expected_line_number, line_number) + self.assertEquals('whitespace/tab', category) + self.had_error = True + + text_style.process_file_data('', lines, error_for_test) + self.assert_(self.had_error, '%s should have an error [whitespace/tab].' % lines) + + + def test_no_error(self): + """Tests for no error cases.""" + self.assertNoError(['']) + self.assertNoError(['abc def', 'ggg']) + + + def test_error(self): + """Tests for error cases.""" + self.assertError(['2009-12-16\tKent Tamura\t<tkent@chromium.org>'], 1) + self.assertError(['2009-12-16 Kent Tamura <tkent@chromium.org>', + '', + '\tReviewed by NOBODY.'], 3) + + +class TextCheckerTest(unittest.TestCase): + + """Tests TextChecker class.""" + + def mock_handle_style_error(self): + pass + + def test_init(self): + """Test __init__ constructor.""" + checker = TextChecker("foo.txt", self.mock_handle_style_error) + self.assertEquals(checker.file_path, "foo.txt") + self.assertEquals(checker.handle_style_error, self.mock_handle_style_error) + + +if __name__ == '__main__': + unittest.main() diff --git a/Tools/Scripts/webkitpy/style/checkers/xml.py b/Tools/Scripts/webkitpy/style/checkers/xml.py new file mode 100644 index 0000000..2f7c0ce --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/xml.py @@ -0,0 +1,45 @@ +# Copyright (C) 2010 Apple Inc. All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND +# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR +# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Checks WebKit style for XML files.""" + +from __future__ import absolute_import + +from xml.parsers import expat + + +class XMLChecker(object): + """Processes XML lines for checking style.""" + + def __init__(self, file_path, handle_style_error): + self.file_path = file_path + self.handle_style_error = handle_style_error + + def check(self, lines): + parser = expat.ParserCreate() + try: + for line in lines: + parser.Parse(line) + parser.Parse('\n') + parser.Parse('', True) + except expat.ExpatError, error: + self.handle_style_error(error.lineno, 'xml/syntax', 5, expat.ErrorString(error.code)) diff --git a/Tools/Scripts/webkitpy/style/checkers/xml_unittest.py b/Tools/Scripts/webkitpy/style/checkers/xml_unittest.py new file mode 100644 index 0000000..3825660 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/checkers/xml_unittest.py @@ -0,0 +1,72 @@ +#!/usr/bin/env python +# +# Copyright (C) 2010 Apple Inc. All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND +# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR +# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit test for xml.py.""" + +import unittest + +import xml + + +class XMLCheckerTest(unittest.TestCase): + """Tests XMLChecker class.""" + + def assert_no_error(self, xml_data): + def handle_style_error(line_number, category, confidence, message): + self.fail('Unexpected error: %d %s %d %s' % (line_number, category, confidence, message)) + checker = xml.XMLChecker('foo.xml', handle_style_error) + checker.check(xml_data.split('\n')) + + def assert_error(self, expected_line_number, expected_category, xml_data): + def handle_style_error(line_number, category, confidence, message): + self.had_error = True + self.assertEquals(expected_line_number, line_number) + self.assertEquals(expected_category, category) + checker = xml.XMLChecker('foo.xml', handle_style_error) + checker.check(xml_data.split('\n')) + self.assertTrue(self.had_error) + + def mock_handle_style_error(self): + pass + + def test_conflict_marker(self): + self.assert_error(1, 'xml/syntax', '<<<<<<< HEAD\n<foo>\n</foo>\n') + + def test_extra_closing_tag(self): + self.assert_error(3, 'xml/syntax', '<foo>\n</foo>\n</foo>\n') + + def test_init(self): + checker = xml.XMLChecker('foo.xml', self.mock_handle_style_error) + self.assertEquals(checker.file_path, 'foo.xml') + self.assertEquals(checker.handle_style_error, self.mock_handle_style_error) + + def test_missing_closing_tag(self): + self.assert_error(3, 'xml/syntax', '<foo>\n<bar>\n</foo>\n') + + def test_no_error(self): + checker = xml.XMLChecker('foo.xml', self.assert_no_error) + checker.check(['<foo>', '</foo>']) + +if __name__ == '__main__': + unittest.main() diff --git a/Tools/Scripts/webkitpy/style/error_handlers.py b/Tools/Scripts/webkitpy/style/error_handlers.py new file mode 100644 index 0000000..0bede24 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/error_handlers.py @@ -0,0 +1,159 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY +# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Defines style error handler classes. + +A style error handler is a function to call when a style error is +found. Style error handlers can also have state. A class that represents +a style error handler should implement the following methods. + +Methods: + + __call__(self, line_number, category, confidence, message): + + Handle the occurrence of a style error. + + Check whether the error is reportable. If so, increment the total + error count and report the details. Note that error reporting can + be suppressed after reaching a certain number of reports. + + Args: + line_number: The integer line number of the line containing the error. + category: The name of the category of the error, for example + "whitespace/newline". + confidence: An integer between 1 and 5 inclusive that represents the + application's level of confidence in the error. The value + 5 means that we are certain of the problem, and the + value 1 means that it could be a legitimate construct. + message: The error message to report. + +""" + + +import sys + + +class DefaultStyleErrorHandler(object): + + """The default style error handler.""" + + def __init__(self, file_path, configuration, increment_error_count, + line_numbers=None): + """Create a default style error handler. + + Args: + file_path: The path to the file containing the error. This + is used for reporting to the user. + configuration: A StyleProcessorConfiguration instance. + increment_error_count: A function that takes no arguments and + increments the total count of reportable + errors. + line_numbers: An array of line numbers of the lines for which + style errors should be reported, or None if errors + for all lines should be reported. When it is not + None, this array normally contains the line numbers + corresponding to the modified lines of a patch. + + """ + if line_numbers is not None: + line_numbers = set(line_numbers) + + self._file_path = file_path + self._configuration = configuration + self._increment_error_count = increment_error_count + self._line_numbers = line_numbers + + # A string to integer dictionary cache of the number of reportable + # errors per category passed to this instance. + self._category_totals = {} + + # Useful for unit testing. + def __eq__(self, other): + """Return whether this instance is equal to another.""" + if self._configuration != other._configuration: + return False + if self._file_path != other._file_path: + return False + if self._increment_error_count != other._increment_error_count: + return False + if self._line_numbers != other._line_numbers: + return False + + return True + + # Useful for unit testing. + def __ne__(self, other): + # Python does not automatically deduce __ne__ from __eq__. + return not self.__eq__(other) + + def _add_reportable_error(self, category): + """Increment the error count and return the new category total.""" + self._increment_error_count() # Increment the total. + + # Increment the category total. + if not category in self._category_totals: + self._category_totals[category] = 1 + else: + self._category_totals[category] += 1 + + return self._category_totals[category] + + def _max_reports(self, category): + """Return the maximum number of errors to report.""" + if not category in self._configuration.max_reports_per_category: + return None + return self._configuration.max_reports_per_category[category] + + def __call__(self, line_number, category, confidence, message): + """Handle the occurrence of a style error. + + See the docstring of this module for more information. + + """ + if (self._line_numbers is not None and + line_number not in self._line_numbers): + # Then the error occurred in a line that was not modified, so + # the error is not reportable. + return + + if not self._configuration.is_reportable(category=category, + confidence_in_error=confidence, + file_path=self._file_path): + return + + category_total = self._add_reportable_error(category) + + max_reports = self._max_reports(category) + + if (max_reports is not None) and (category_total > max_reports): + # Then suppress displaying the error. + return + + self._configuration.write_style_error(category=category, + confidence_in_error=confidence, + file_path=self._file_path, + line_number=line_number, + message=message) + + if category_total == max_reports: + self._configuration.stderr_write("Suppressing further [%s] reports " + "for this file.\n" % category) diff --git a/Tools/Scripts/webkitpy/style/error_handlers_unittest.py b/Tools/Scripts/webkitpy/style/error_handlers_unittest.py new file mode 100644 index 0000000..23619cc --- /dev/null +++ b/Tools/Scripts/webkitpy/style/error_handlers_unittest.py @@ -0,0 +1,187 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY +# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit tests for error_handlers.py.""" + + +import unittest + +from checker import StyleProcessorConfiguration +from error_handlers import DefaultStyleErrorHandler +from filter import FilterConfiguration + + +class DefaultStyleErrorHandlerTest(unittest.TestCase): + + """Tests the DefaultStyleErrorHandler class.""" + + def setUp(self): + self._error_messages = [] + self._error_count = 0 + + _category = "whitespace/tab" + """The category name for the tests in this class.""" + + _file_path = "foo.h" + """The file path for the tests in this class.""" + + def _mock_increment_error_count(self): + self._error_count += 1 + + def _mock_stderr_write(self, message): + self._error_messages.append(message) + + def _style_checker_configuration(self): + """Return a StyleProcessorConfiguration instance for testing.""" + base_rules = ["-whitespace", "+whitespace/tab"] + filter_configuration = FilterConfiguration(base_rules=base_rules) + + return StyleProcessorConfiguration( + filter_configuration=filter_configuration, + max_reports_per_category={"whitespace/tab": 2}, + min_confidence=3, + output_format="vs7", + stderr_write=self._mock_stderr_write) + + def _error_handler(self, configuration, line_numbers=None): + return DefaultStyleErrorHandler(configuration=configuration, + file_path=self._file_path, + increment_error_count=self._mock_increment_error_count, + line_numbers=line_numbers) + + def _check_initialized(self): + """Check that count and error messages are initialized.""" + self.assertEquals(0, self._error_count) + self.assertEquals(0, len(self._error_messages)) + + def _call_error_handler(self, handle_error, confidence, line_number=100): + """Call the given error handler with a test error.""" + handle_error(line_number=line_number, + category=self._category, + confidence=confidence, + message="message") + + def test_eq__true_return_value(self): + """Test the __eq__() method for the return value of True.""" + handler1 = self._error_handler(configuration=None) + handler2 = self._error_handler(configuration=None) + + self.assertTrue(handler1.__eq__(handler2)) + + def test_eq__false_return_value(self): + """Test the __eq__() method for the return value of False.""" + def make_handler(configuration=self._style_checker_configuration(), + file_path='foo.txt', increment_error_count=lambda: True, + line_numbers=[100]): + return DefaultStyleErrorHandler(configuration=configuration, + file_path=file_path, + increment_error_count=increment_error_count, + line_numbers=line_numbers) + + handler = make_handler() + + # Establish a baseline for our comparisons below. + self.assertTrue(handler.__eq__(make_handler())) + + # Verify that a difference in any argument causes equality to fail. + self.assertFalse(handler.__eq__(make_handler(configuration=None))) + self.assertFalse(handler.__eq__(make_handler(file_path='bar.txt'))) + self.assertFalse(handler.__eq__(make_handler(increment_error_count=None))) + self.assertFalse(handler.__eq__(make_handler(line_numbers=[50]))) + + def test_ne(self): + """Test the __ne__() method.""" + # By default, __ne__ always returns true on different objects. + # Thus, check just the distinguishing case to verify that the + # code defines __ne__. + handler1 = self._error_handler(configuration=None) + handler2 = self._error_handler(configuration=None) + + self.assertFalse(handler1.__ne__(handler2)) + + def test_non_reportable_error(self): + """Test __call__() with a non-reportable error.""" + self._check_initialized() + configuration = self._style_checker_configuration() + + confidence = 1 + # Confirm the error is not reportable. + self.assertFalse(configuration.is_reportable(self._category, + confidence, + self._file_path)) + error_handler = self._error_handler(configuration) + self._call_error_handler(error_handler, confidence) + + self.assertEquals(0, self._error_count) + self.assertEquals([], self._error_messages) + + # Also serves as a reportable error test. + def test_max_reports_per_category(self): + """Test error report suppression in __call__() method.""" + self._check_initialized() + configuration = self._style_checker_configuration() + error_handler = self._error_handler(configuration) + + confidence = 5 + + # First call: usual reporting. + self._call_error_handler(error_handler, confidence) + self.assertEquals(1, self._error_count) + self.assertEquals(1, len(self._error_messages)) + self.assertEquals(self._error_messages, + ["foo.h(100): message [whitespace/tab] [5]\n"]) + + # Second call: suppression message reported. + self._call_error_handler(error_handler, confidence) + # The "Suppressing further..." message counts as an additional + # message (but not as an addition to the error count). + self.assertEquals(2, self._error_count) + self.assertEquals(3, len(self._error_messages)) + self.assertEquals(self._error_messages[-2], + "foo.h(100): message [whitespace/tab] [5]\n") + self.assertEquals(self._error_messages[-1], + "Suppressing further [whitespace/tab] reports " + "for this file.\n") + + # Third call: no report. + self._call_error_handler(error_handler, confidence) + self.assertEquals(3, self._error_count) + self.assertEquals(3, len(self._error_messages)) + + def test_line_numbers(self): + """Test the line_numbers parameter.""" + self._check_initialized() + configuration = self._style_checker_configuration() + error_handler = self._error_handler(configuration, + line_numbers=[50]) + confidence = 5 + + # Error on non-modified line: no error. + self._call_error_handler(error_handler, confidence, line_number=60) + self.assertEquals(0, self._error_count) + self.assertEquals([], self._error_messages) + + # Error on modified line: error. + self._call_error_handler(error_handler, confidence, line_number=50) + self.assertEquals(1, self._error_count) + self.assertEquals(self._error_messages, + ["foo.h(50): message [whitespace/tab] [5]\n"]) diff --git a/Tools/Scripts/webkitpy/style/filereader.py b/Tools/Scripts/webkitpy/style/filereader.py new file mode 100644 index 0000000..1a24cb5 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/filereader.py @@ -0,0 +1,162 @@ +# Copyright (C) 2009 Google Inc. All rights reserved. +# Copyright (C) 2010 Chris Jerdonek (chris.jerdonek@gmail.com) +# Copyright (C) 2010 ProFUSION embedded systems +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Supports reading and processing text files.""" + +import codecs +import logging +import os +import sys + + +_log = logging.getLogger(__name__) + + +class TextFileReader(object): + + """Supports reading and processing text files. + + Attributes: + file_count: The total number of files passed to this instance + for processing, including non-text files and files + that should be skipped. + delete_only_file_count: The total number of files that are not + processed this instance actually because + the files don't have any modified lines + but should be treated as processed. + + """ + + def __init__(self, processor): + """Create an instance. + + Arguments: + processor: A ProcessorBase instance. + + """ + self._processor = processor + self.file_count = 0 + self.delete_only_file_count = 0 + + def _read_lines(self, file_path): + """Read the file at a path, and return its lines. + + Raises: + IOError: If the file does not exist or cannot be read. + + """ + # Support the UNIX convention of using "-" for stdin. + if file_path == '-': + file = codecs.StreamReaderWriter(sys.stdin, + codecs.getreader('utf8'), + codecs.getwriter('utf8'), + 'replace') + else: + # We do not open the file with universal newline support + # (codecs does not support it anyway), so the resulting + # lines contain trailing "\r" characters if we are reading + # a file with CRLF endings. + file = codecs.open(file_path, 'r', 'utf8', 'replace') + + try: + contents = file.read() + finally: + file.close() + + lines = contents.split('\n') + return lines + + def process_file(self, file_path, **kwargs): + """Process the given file by calling the processor's process() method. + + Args: + file_path: The path of the file to process. + **kwargs: Any additional keyword parameters that should be passed + to the processor's process() method. The process() + method should support these keyword arguments. + + Raises: + SystemExit: If no file at file_path exists. + + """ + self.file_count += 1 + + if not os.path.exists(file_path) and file_path != "-": + _log.error("File does not exist: '%s'" % file_path) + sys.exit(1) + + if not self._processor.should_process(file_path): + _log.debug("Skipping file: '%s'" % file_path) + return + _log.debug("Processing file: '%s'" % file_path) + + try: + lines = self._read_lines(file_path) + except IOError, err: + message = ("Could not read file. Skipping: '%s'\n %s" + % (file_path, err)) + _log.warn(message) + return + + self._processor.process(lines, file_path, **kwargs) + + def _process_directory(self, directory): + """Process all files in the given directory, recursively. + + Args: + directory: A directory path. + + """ + for dir_path, dir_names, file_names in os.walk(directory): + for file_name in file_names: + file_path = os.path.join(dir_path, file_name) + self.process_file(file_path) + + def process_paths(self, paths): + """Process the given file and directory paths. + + Args: + paths: A list of file and directory paths. + + """ + for path in paths: + if os.path.isdir(path): + self._process_directory(directory=path) + else: + self.process_file(path) + + def count_delete_only_file(self): + """Count up files that contains only deleted lines. + + Files which has no modified or newly-added lines don't need + to check style, but should be treated as checked. For that + purpose, we just count up the number of such files. + """ + self.delete_only_file_count += 1 diff --git a/Tools/Scripts/webkitpy/style/filereader_unittest.py b/Tools/Scripts/webkitpy/style/filereader_unittest.py new file mode 100644 index 0000000..6328337 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/filereader_unittest.py @@ -0,0 +1,166 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND +# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR +# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Contains unit tests for filereader.py.""" + +from __future__ import with_statement + +import codecs +import os +import shutil +import tempfile +import unittest + +from webkitpy.common.system.logtesting import LoggingTestCase +from webkitpy.style.checker import ProcessorBase +from webkitpy.style.filereader import TextFileReader + + +class TextFileReaderTest(LoggingTestCase): + + class MockProcessor(ProcessorBase): + + """A processor for test purposes. + + This processor simply records the parameters passed to its process() + method for later checking by the unittest test methods. + + """ + + def __init__(self): + self.processed = [] + """The parameters passed for all calls to the process() method.""" + + def should_process(self, file_path): + return not file_path.endswith('should_not_process.txt') + + def process(self, lines, file_path, test_kwarg=None): + self.processed.append((lines, file_path, test_kwarg)) + + def setUp(self): + LoggingTestCase.setUp(self) + processor = TextFileReaderTest.MockProcessor() + + temp_dir = tempfile.mkdtemp() + + self._file_reader = TextFileReader(processor) + self._processor = processor + self._temp_dir = temp_dir + + def tearDown(self): + LoggingTestCase.tearDown(self) + shutil.rmtree(self._temp_dir) + + def _create_file(self, rel_path, text, encoding="utf-8"): + """Create a file with given text and return the path to the file.""" + # FIXME: There are better/more secure APIs for creatin tmp file paths. + file_path = os.path.join(self._temp_dir, rel_path) + with codecs.open(file_path, "w", encoding) as file: + file.write(text) + return file_path + + def _passed_to_processor(self): + """Return the parameters passed to MockProcessor.process().""" + return self._processor.processed + + def _assert_file_reader(self, passed_to_processor, file_count): + """Assert the state of the file reader.""" + self.assertEquals(passed_to_processor, self._passed_to_processor()) + self.assertEquals(file_count, self._file_reader.file_count) + + def test_process_file__does_not_exist(self): + try: + self._file_reader.process_file('does_not_exist.txt') + except SystemExit, err: + self.assertEquals(str(err), '1') + else: + self.fail('No Exception raised.') + self._assert_file_reader([], 1) + self.assertLog(["ERROR: File does not exist: 'does_not_exist.txt'\n"]) + + def test_process_file__is_dir(self): + temp_dir = os.path.join(self._temp_dir, 'test_dir') + os.mkdir(temp_dir) + + self._file_reader.process_file(temp_dir) + + # Because the log message below contains exception text, it is + # possible that the text varies across platforms. For this reason, + # we check only the portion of the log message that we control, + # namely the text at the beginning. + log_messages = self.logMessages() + # We remove the message we are looking at to prevent the tearDown() + # from raising an exception when it asserts that no log messages + # remain. + message = log_messages.pop() + + self.assertTrue(message.startswith('WARNING: Could not read file. ' + "Skipping: '%s'\n " % temp_dir)) + + self._assert_file_reader([], 1) + + def test_process_file__should_not_process(self): + file_path = self._create_file('should_not_process.txt', 'contents') + + self._file_reader.process_file(file_path) + self._assert_file_reader([], 1) + + def test_process_file__multiple_lines(self): + file_path = self._create_file('foo.txt', 'line one\r\nline two\n') + + self._file_reader.process_file(file_path) + processed = [(['line one\r', 'line two', ''], file_path, None)] + self._assert_file_reader(processed, 1) + + def test_process_file__file_stdin(self): + file_path = self._create_file('-', 'file contents') + + self._file_reader.process_file(file_path=file_path, test_kwarg='foo') + processed = [(['file contents'], file_path, 'foo')] + self._assert_file_reader(processed, 1) + + def test_process_file__with_kwarg(self): + file_path = self._create_file('foo.txt', 'file contents') + + self._file_reader.process_file(file_path=file_path, test_kwarg='foo') + processed = [(['file contents'], file_path, 'foo')] + self._assert_file_reader(processed, 1) + + def test_process_paths(self): + # We test a list of paths that contains both a file and a directory. + dir = os.path.join(self._temp_dir, 'foo_dir') + os.mkdir(dir) + + file_path1 = self._create_file('file1.txt', 'foo') + + rel_path = os.path.join('foo_dir', 'file2.txt') + file_path2 = self._create_file(rel_path, 'bar') + + self._file_reader.process_paths([dir, file_path1]) + processed = [(['bar'], file_path2, None), + (['foo'], file_path1, None)] + self._assert_file_reader(processed, 2) + + def test_count_delete_only_file(self): + self._file_reader.count_delete_only_file() + delete_only_file_count = self._file_reader.delete_only_file_count + self.assertEquals(delete_only_file_count, 1) diff --git a/Tools/Scripts/webkitpy/style/filter.py b/Tools/Scripts/webkitpy/style/filter.py new file mode 100644 index 0000000..608a9e6 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/filter.py @@ -0,0 +1,278 @@ +# Copyright (C) 2010 Chris Jerdonek (chris.jerdonek@gmail.com) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY +# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Contains filter-related code.""" + + +def validate_filter_rules(filter_rules, all_categories): + """Validate the given filter rules, and raise a ValueError if not valid. + + Args: + filter_rules: A list of boolean filter rules, for example-- + ["-whitespace", "+whitespace/braces"] + all_categories: A list of all available category names, for example-- + ["whitespace/tabs", "whitespace/braces"] + + Raises: + ValueError: An error occurs if a filter rule does not begin + with "+" or "-" or if a filter rule does not match + the beginning of some category name in the list + of all available categories. + + """ + for rule in filter_rules: + if not (rule.startswith('+') or rule.startswith('-')): + raise ValueError('Invalid filter rule "%s": every rule ' + "must start with + or -." % rule) + + for category in all_categories: + if category.startswith(rule[1:]): + break + else: + raise ValueError('Suspected incorrect filter rule "%s": ' + "the rule does not match the beginning " + "of any category name." % rule) + + +class _CategoryFilter(object): + + """Filters whether to check style categories.""" + + def __init__(self, filter_rules=None): + """Create a category filter. + + Args: + filter_rules: A list of strings that are filter rules, which + are strings beginning with the plus or minus + symbol (+/-). The list should include any + default filter rules at the beginning. + Defaults to the empty list. + + Raises: + ValueError: Invalid filter rule if a rule does not start with + plus ("+") or minus ("-"). + + """ + if filter_rules is None: + filter_rules = [] + + self._filter_rules = filter_rules + self._should_check_category = {} # Cached dictionary of category to True/False + + def __str__(self): + return ",".join(self._filter_rules) + + # Useful for unit testing. + def __eq__(self, other): + """Return whether this CategoryFilter instance is equal to another.""" + return self._filter_rules == other._filter_rules + + # Useful for unit testing. + def __ne__(self, other): + # Python does not automatically deduce from __eq__(). + return not (self == other) + + def should_check(self, category): + """Return whether the category should be checked. + + The rules for determining whether a category should be checked + are as follows. By default all categories should be checked. + Then apply the filter rules in order from first to last, with + later flags taking precedence. + + A filter rule applies to a category if the string after the + leading plus/minus (+/-) matches the beginning of the category + name. A plus (+) means the category should be checked, while a + minus (-) means the category should not be checked. + + """ + if category in self._should_check_category: + return self._should_check_category[category] + + should_check = True # All categories checked by default. + for rule in self._filter_rules: + if not category.startswith(rule[1:]): + continue + should_check = rule.startswith('+') + self._should_check_category[category] = should_check # Update cache. + return should_check + + +class FilterConfiguration(object): + + """Supports filtering with path-specific and user-specified rules.""" + + def __init__(self, base_rules=None, path_specific=None, user_rules=None): + """Create a FilterConfiguration instance. + + Args: + base_rules: The starting list of filter rules to use for + processing. The default is the empty list, which + by itself would mean that all categories should be + checked. + + path_specific: A list of (sub_paths, path_rules) pairs + that stores the path-specific filter rules for + appending to the base rules. + The "sub_paths" value is a list of path + substrings. If a file path contains one of the + substrings, then the corresponding path rules + are appended. The first substring match takes + precedence, i.e. only the first match triggers + an append. + The "path_rules" value is a list of filter + rules that can be appended to the base rules. + + user_rules: A list of filter rules that is always appended + to the base rules and any path rules. In other + words, the user rules take precedence over the + everything. In practice, the user rules are + provided by the user from the command line. + + """ + if base_rules is None: + base_rules = [] + if path_specific is None: + path_specific = [] + if user_rules is None: + user_rules = [] + + self._base_rules = base_rules + self._path_specific = path_specific + self._path_specific_lower = None + """The backing store for self._get_path_specific_lower().""" + + self._user_rules = user_rules + + self._path_rules_to_filter = {} + """Cached dictionary of path rules to CategoryFilter instance.""" + + # The same CategoryFilter instance can be shared across + # multiple keys in this dictionary. This allows us to take + # greater advantage of the caching done by + # CategoryFilter.should_check(). + self._path_to_filter = {} + """Cached dictionary of file path to CategoryFilter instance.""" + + # Useful for unit testing. + def __eq__(self, other): + """Return whether this FilterConfiguration is equal to another.""" + if self._base_rules != other._base_rules: + return False + if self._path_specific != other._path_specific: + return False + if self._user_rules != other._user_rules: + return False + + return True + + # Useful for unit testing. + def __ne__(self, other): + # Python does not automatically deduce this from __eq__(). + return not self.__eq__(other) + + # We use the prefix "_get" since the name "_path_specific_lower" + # is already taken up by the data attribute backing store. + def _get_path_specific_lower(self): + """Return a copy of self._path_specific with the paths lower-cased.""" + if self._path_specific_lower is None: + self._path_specific_lower = [] + for (sub_paths, path_rules) in self._path_specific: + sub_paths = map(str.lower, sub_paths) + self._path_specific_lower.append((sub_paths, path_rules)) + return self._path_specific_lower + + def _path_rules_from_path(self, path): + """Determine the path-specific rules to use, and return as a tuple. + + This method returns a tuple rather than a list so the return + value can be passed to _filter_from_path_rules() without change. + + """ + path = path.lower() + for (sub_paths, path_rules) in self._get_path_specific_lower(): + for sub_path in sub_paths: + if path.find(sub_path) > -1: + return tuple(path_rules) + return () # Default to the empty tuple. + + def _filter_from_path_rules(self, path_rules): + """Return the CategoryFilter associated to the given path rules. + + Args: + path_rules: A tuple of path rules. We require a tuple rather + than a list so the value can be used as a dictionary + key in self._path_rules_to_filter. + + """ + # We reuse the same CategoryFilter where possible to take + # advantage of the caching they do. + if path_rules not in self._path_rules_to_filter: + rules = list(self._base_rules) # Make a copy + rules.extend(path_rules) + rules.extend(self._user_rules) + self._path_rules_to_filter[path_rules] = _CategoryFilter(rules) + + return self._path_rules_to_filter[path_rules] + + def _filter_from_path(self, path): + """Return the CategoryFilter associated to a path.""" + if path not in self._path_to_filter: + path_rules = self._path_rules_from_path(path) + filter = self._filter_from_path_rules(path_rules) + self._path_to_filter[path] = filter + + return self._path_to_filter[path] + + def should_check(self, category, path): + """Return whether the given category should be checked. + + This method determines whether a category should be checked + by checking the category name against the filter rules for + the given path. + + For a given path, the filter rules are the combination of + the base rules, the path-specific rules, and the user-provided + rules -- in that order. As we will describe below, later rules + in the list take precedence. The path-specific rules are the + rules corresponding to the first element of the "path_specific" + parameter that contains a string case-insensitively matching + some substring of the path. If there is no such element, + there are no path-specific rules for that path. + + Given a list of filter rules, the logic for determining whether + a category should be checked is as follows. By default all + categories should be checked. Then apply the filter rules in + order from first to last, with later flags taking precedence. + + A filter rule applies to a category if the string after the + leading plus/minus (+/-) matches the beginning of the category + name. A plus (+) means the category should be checked, while a + minus (-) means the category should not be checked. + + Args: + category: The category name. + path: The path of the file being checked. + + """ + return self._filter_from_path(path).should_check(category) + diff --git a/Tools/Scripts/webkitpy/style/filter_unittest.py b/Tools/Scripts/webkitpy/style/filter_unittest.py new file mode 100644 index 0000000..7b8a5402 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/filter_unittest.py @@ -0,0 +1,256 @@ +# Copyright (C) 2010 Chris Jerdonek (chris.jerdonek@gmail.com) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY +# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit tests for filter.py.""" + +import unittest + +from filter import _CategoryFilter as CategoryFilter +from filter import validate_filter_rules +from filter import FilterConfiguration + +# On Testing __eq__() and __ne__(): +# +# In the tests below, we deliberately do not use assertEquals() or +# assertNotEquals() to test __eq__() or __ne__(). We do this to be +# very explicit about what we are testing, especially in the case +# of assertNotEquals(). +# +# Part of the reason is that it is not immediately clear what +# expression the unittest module uses to assert "not equals" -- the +# negation of __eq__() or __ne__(), which are not necessarily +# equivalent expresions in Python. For example, from Python's "Data +# Model" documentation-- +# +# "There are no implied relationships among the comparison +# operators. The truth of x==y does not imply that x!=y is +# false. Accordingly, when defining __eq__(), one should +# also define __ne__() so that the operators will behave as +# expected." +# +# (from http://docs.python.org/reference/datamodel.html#object.__ne__ ) + +class ValidateFilterRulesTest(unittest.TestCase): + + """Tests validate_filter_rules() function.""" + + def test_validate_filter_rules(self): + all_categories = ["tabs", "whitespace", "build/include"] + + bad_rules = [ + "tabs", + "*tabs", + " tabs", + " +tabs", + "+whitespace/newline", + "+xxx", + ] + + good_rules = [ + "+tabs", + "-tabs", + "+build" + ] + + for rule in bad_rules: + self.assertRaises(ValueError, validate_filter_rules, + [rule], all_categories) + + for rule in good_rules: + # This works: no error. + validate_filter_rules([rule], all_categories) + + +class CategoryFilterTest(unittest.TestCase): + + """Tests CategoryFilter class.""" + + def test_init(self): + """Test __init__ method.""" + # Test that the attributes are getting set correctly. + filter = CategoryFilter(["+"]) + self.assertEquals(["+"], filter._filter_rules) + + def test_init_default_arguments(self): + """Test __init__ method default arguments.""" + filter = CategoryFilter() + self.assertEquals([], filter._filter_rules) + + def test_str(self): + """Test __str__ "to string" operator.""" + filter = CategoryFilter(["+a", "-b"]) + self.assertEquals(str(filter), "+a,-b") + + def test_eq(self): + """Test __eq__ equality function.""" + filter1 = CategoryFilter(["+a", "+b"]) + filter2 = CategoryFilter(["+a", "+b"]) + filter3 = CategoryFilter(["+b", "+a"]) + + # See the notes at the top of this module about testing + # __eq__() and __ne__(). + self.assertTrue(filter1.__eq__(filter2)) + self.assertFalse(filter1.__eq__(filter3)) + + def test_ne(self): + """Test __ne__ inequality function.""" + # By default, __ne__ always returns true on different objects. + # Thus, just check the distinguishing case to verify that the + # code defines __ne__. + # + # Also, see the notes at the top of this module about testing + # __eq__() and __ne__(). + self.assertFalse(CategoryFilter().__ne__(CategoryFilter())) + + def test_should_check(self): + """Test should_check() method.""" + filter = CategoryFilter() + self.assertTrue(filter.should_check("everything")) + # Check a second time to exercise cache. + self.assertTrue(filter.should_check("everything")) + + filter = CategoryFilter(["-"]) + self.assertFalse(filter.should_check("anything")) + # Check a second time to exercise cache. + self.assertFalse(filter.should_check("anything")) + + filter = CategoryFilter(["-", "+ab"]) + self.assertTrue(filter.should_check("abc")) + self.assertFalse(filter.should_check("a")) + + filter = CategoryFilter(["+", "-ab"]) + self.assertFalse(filter.should_check("abc")) + self.assertTrue(filter.should_check("a")) + + +class FilterConfigurationTest(unittest.TestCase): + + """Tests FilterConfiguration class.""" + + def _config(self, base_rules, path_specific, user_rules): + """Return a FilterConfiguration instance.""" + return FilterConfiguration(base_rules=base_rules, + path_specific=path_specific, + user_rules=user_rules) + + def test_init(self): + """Test __init__ method.""" + # Test that the attributes are getting set correctly. + # We use parameter values that are different from the defaults. + base_rules = ["-"] + path_specific = [(["path"], ["+a"])] + user_rules = ["+"] + + config = self._config(base_rules, path_specific, user_rules) + + self.assertEquals(base_rules, config._base_rules) + self.assertEquals(path_specific, config._path_specific) + self.assertEquals(user_rules, config._user_rules) + + def test_default_arguments(self): + # Test that the attributes are getting set correctly to the defaults. + config = FilterConfiguration() + + self.assertEquals([], config._base_rules) + self.assertEquals([], config._path_specific) + self.assertEquals([], config._user_rules) + + def test_eq(self): + """Test __eq__ method.""" + # See the notes at the top of this module about testing + # __eq__() and __ne__(). + self.assertTrue(FilterConfiguration().__eq__(FilterConfiguration())) + + # Verify that a difference in any argument causes equality to fail. + config = FilterConfiguration() + + # These parameter values are different from the defaults. + base_rules = ["-"] + path_specific = [(["path"], ["+a"])] + user_rules = ["+"] + + self.assertFalse(config.__eq__(FilterConfiguration( + base_rules=base_rules))) + self.assertFalse(config.__eq__(FilterConfiguration( + path_specific=path_specific))) + self.assertFalse(config.__eq__(FilterConfiguration( + user_rules=user_rules))) + + def test_ne(self): + """Test __ne__ method.""" + # By default, __ne__ always returns true on different objects. + # Thus, just check the distinguishing case to verify that the + # code defines __ne__. + # + # Also, see the notes at the top of this module about testing + # __eq__() and __ne__(). + self.assertFalse(FilterConfiguration().__ne__(FilterConfiguration())) + + def test_base_rules(self): + """Test effect of base_rules on should_check().""" + base_rules = ["-b"] + path_specific = [] + user_rules = [] + + config = self._config(base_rules, path_specific, user_rules) + + self.assertTrue(config.should_check("a", "path")) + self.assertFalse(config.should_check("b", "path")) + + def test_path_specific(self): + """Test effect of path_rules_specifier on should_check().""" + base_rules = ["-"] + path_specific = [(["path1"], ["+b"]), + (["path2"], ["+c"])] + user_rules = [] + + config = self._config(base_rules, path_specific, user_rules) + + self.assertFalse(config.should_check("c", "path1")) + self.assertTrue(config.should_check("c", "path2")) + # Test that first match takes precedence. + self.assertFalse(config.should_check("c", "path2/path1")) + + def test_path_with_different_case(self): + """Test a path that differs only in case.""" + base_rules = ["-"] + path_specific = [(["Foo/"], ["+whitespace"])] + user_rules = [] + + config = self._config(base_rules, path_specific, user_rules) + + self.assertFalse(config.should_check("whitespace", "Fooo/bar.txt")) + self.assertTrue(config.should_check("whitespace", "Foo/bar.txt")) + # Test different case. + self.assertTrue(config.should_check("whitespace", "FOO/bar.txt")) + + def test_user_rules(self): + """Test effect of user_rules on should_check().""" + base_rules = ["-"] + path_specific = [] + user_rules = ["+b"] + + config = self._config(base_rules, path_specific, user_rules) + + self.assertFalse(config.should_check("a", "path")) + self.assertTrue(config.should_check("b", "path")) + diff --git a/Tools/Scripts/webkitpy/style/main.py b/Tools/Scripts/webkitpy/style/main.py new file mode 100644 index 0000000..83c0323 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/main.py @@ -0,0 +1,130 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND +# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR +# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +import logging +import os +import sys + +from webkitpy.common.system.ospath import relpath as _relpath + + +_log = logging.getLogger(__name__) + + +def change_directory(checkout_root, paths, mock_os=None): + """Change the working directory to the WebKit checkout root, if possible. + + If every path in the paths parameter is below the checkout root (or if + the paths parameter is empty or None), this method changes the current + working directory to the checkout root and converts the paths parameter + as described below. + This allows the paths being checked to be displayed relative to the + checkout root, and for path-specific style checks to work as expected. + Path-specific checks include whether files should be skipped, whether + custom style rules should apply to certain files, etc. + If the checkout root is None or the empty string, this method returns + the paths parameter unchanged. + + Returns: + paths: A copy of the paths parameter -- possibly converted, as follows. + If this method changed the current working directory to the + checkout root, then the list is the paths parameter converted to + normalized paths relative to the checkout root. Otherwise, the + paths are not converted. + + Args: + paths: A list of paths to the files that should be checked for style. + This argument can be None or the empty list if a git commit + or all changes under the checkout root should be checked. + checkout_root: The path to the root of the WebKit checkout, or None or + the empty string if no checkout could be detected. + mock_os: A replacement module for unit testing. Defaults to os. + + """ + os_module = os if mock_os is None else mock_os + + if paths is not None: + paths = list(paths) + + if not checkout_root: + if not paths: + raise Exception("The paths parameter must be non-empty if " + "there is no checkout root.") + + # FIXME: Consider trying to detect the checkout root for each file + # being checked rather than only trying to detect the checkout + # root for the current working directory. This would allow + # files to be checked correctly even if the script is being + # run from outside any WebKit checkout. + # + # Moreover, try to find the "source root" for each file + # using path-based heuristics rather than using only the + # presence of a WebKit checkout. For example, we could + # examine parent directories until a directory is found + # containing JavaScriptCore, WebCore, WebKit, Websites, + # and Tools. + # Then log an INFO message saying that a source root not + # in a WebKit checkout was found. This will allow us to check + # the style of non-scm copies of the source tree (e.g. + # nightlies). + _log.warn("WebKit checkout root not found:\n" + " Path-dependent style checks may not work correctly.\n" + " See the help documentation for more info.") + + return paths + + if paths: + # Then try converting all of the paths to paths relative to + # the checkout root. + rel_paths = [] + for path in paths: + rel_path = _relpath(path, checkout_root) + if rel_path is None: + # Then the path is not below the checkout root. Since all + # paths should be interpreted relative to the same root, + # do not interpret any of the paths as relative to the + # checkout root. Interpret all of them relative to the + # current working directory, and do not change the current + # working directory. + _log.warn( +"""Path-dependent style checks may not work correctly: + + One of the given paths is outside the WebKit checkout of the current + working directory: + + Path: %s + Checkout root: %s + + Pass only files below the checkout root to ensure correct results. + See the help documentation for more info. +""" + % (path, checkout_root)) + + return paths + rel_paths.append(rel_path) + # If we got here, the conversion was successful. + paths = rel_paths + + _log.debug("Changing to checkout root: " + checkout_root) + os_module.chdir(checkout_root) + + return paths diff --git a/Tools/Scripts/webkitpy/style/main_unittest.py b/Tools/Scripts/webkitpy/style/main_unittest.py new file mode 100644 index 0000000..fe448f5 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/main_unittest.py @@ -0,0 +1,124 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND +# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR +# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit tests for main.py.""" + +import os +import unittest + +from main import change_directory +from webkitpy.style_references import LogTesting + + +class ChangeDirectoryTest(unittest.TestCase): + + """Tests change_directory().""" + + _original_directory = "/original" + _checkout_root = "/WebKit" + + class _MockOs(object): + + """A mock os module for unit testing.""" + + def __init__(self, test_case): + self._test_case = test_case + self._current_directory = \ + ChangeDirectoryTest._original_directory + + def chdir(self, current_directory): + self._current_directory = current_directory + + def assertCurrentDirectory(self, expected_directory): + self._test_case.assertEquals(expected_directory, + self._current_directory) + + def setUp(self): + self._log = LogTesting.setUp(self) + self._mock_os = self._MockOs(self) + + def tearDown(self): + self._log.tearDown() + + # This method is a convenient wrapper for change_working_directory() that + # passes the mock_os for this unit testing class. + def _change_directory(self, paths, checkout_root): + return change_directory(paths=paths, + checkout_root=checkout_root, + mock_os=self._mock_os) + + def _assert_result(self, actual_return_value, expected_return_value, + expected_log_messages, expected_current_directory): + self.assertEquals(actual_return_value, expected_return_value) + self._log.assertMessages(expected_log_messages) + self._mock_os.assertCurrentDirectory(expected_current_directory) + + def test_checkout_root_none_paths_none(self): + self.assertRaises(Exception, self._change_directory, + checkout_root=None, paths=None) + self._log.assertMessages([]) + self._mock_os.assertCurrentDirectory(self._original_directory) + + def test_checkout_root_none(self): + paths = self._change_directory(checkout_root=None, + paths=["path1"]) + log_messages = [ +"""WARNING: WebKit checkout root not found: + Path-dependent style checks may not work correctly. + See the help documentation for more info. +"""] + self._assert_result(paths, ["path1"], log_messages, + self._original_directory) + + def test_paths_none(self): + paths = self._change_directory(checkout_root=self._checkout_root, + paths=None) + self._assert_result(paths, None, [], self._checkout_root) + + def test_paths_convertible(self): + paths=["/WebKit/foo1.txt", + "/WebKit/foo2.txt"] + paths = self._change_directory(checkout_root=self._checkout_root, + paths=paths) + self._assert_result(paths, ["foo1.txt", "foo2.txt"], [], + self._checkout_root) + + def test_with_scm_paths_unconvertible(self): + paths=["/WebKit/foo1.txt", + "/outside/foo2.txt"] + paths = self._change_directory(checkout_root=self._checkout_root, + paths=paths) + log_messages = [ +"""WARNING: Path-dependent style checks may not work correctly: + + One of the given paths is outside the WebKit checkout of the current + working directory: + + Path: /outside/foo2.txt + Checkout root: /WebKit + + Pass only files below the checkout root to ensure correct results. + See the help documentation for more info. + +"""] + self._assert_result(paths, paths, log_messages, + self._original_directory) diff --git a/Tools/Scripts/webkitpy/style/optparser.py b/Tools/Scripts/webkitpy/style/optparser.py new file mode 100644 index 0000000..f4e9923 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/optparser.py @@ -0,0 +1,457 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY +# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Supports the parsing of command-line options for check-webkit-style.""" + +import logging +from optparse import OptionParser +import os.path +import sys + +from filter import validate_filter_rules +# This module should not import anything from checker.py. + +_log = logging.getLogger(__name__) + +_USAGE = """usage: %prog [--help] [options] [path1] [path2] ... + +Overview: + Check coding style according to WebKit style guidelines: + + http://webkit.org/coding/coding-style.html + + Path arguments can be files and directories. If neither a git commit nor + paths are passed, then all changes in your source control working directory + are checked. + +Style errors: + This script assigns to every style error a confidence score from 1-5 and + a category name. A confidence score of 5 means the error is certainly + a problem, and 1 means it could be fine. + + Category names appear in error messages in brackets, for example + [whitespace/indent]. See the options section below for an option that + displays all available categories and which are reported by default. + +Filters: + Use filters to configure what errors to report. Filters are specified using + a comma-separated list of boolean filter rules. The script reports errors + in a category if the category passes the filter, as described below. + + All categories start out passing. Boolean filter rules are then evaluated + from left to right, with later rules taking precedence. For example, the + rule "+foo" passes any category that starts with "foo", and "-foo" fails + any such category. The filter input "-whitespace,+whitespace/braces" fails + the category "whitespace/tab" and passes "whitespace/braces". + + Examples: --filter=-whitespace,+whitespace/braces + --filter=-whitespace,-runtime/printf,+runtime/printf_format + --filter=-,+build/include_what_you_use + +Paths: + Certain style-checking behavior depends on the paths relative to + the WebKit source root of the files being checked. For example, + certain types of errors may be handled differently for files in + WebKit/gtk/webkit/ (e.g. by suppressing "readability/naming" errors + for files in this directory). + + Consequently, if the path relative to the source root cannot be + determined for a file being checked, then style checking may not + work correctly for that file. This can occur, for example, if no + WebKit checkout can be found, or if the source root can be detected, + but one of the files being checked lies outside the source tree. + + If a WebKit checkout can be detected and all files being checked + are in the source tree, then all paths will automatically be + converted to paths relative to the source root prior to checking. + This is also useful for display purposes. + + Currently, this command can detect the source root only if the + command is run from within a WebKit checkout (i.e. if the current + working directory is below the root of a checkout). In particular, + it is not recommended to run this script from a directory outside + a checkout. + + Running this script from a top-level WebKit source directory and + checking only files in the source tree will ensure that all style + checking behaves correctly -- whether or not a checkout can be + detected. This is because all file paths will already be relative + to the source root and so will not need to be converted.""" + +_EPILOG = ("This script can miss errors and does not substitute for " + "code review.") + + +# This class should not have knowledge of the flag key names. +class DefaultCommandOptionValues(object): + + """Stores the default check-webkit-style command-line options. + + Attributes: + output_format: A string that is the default output format. + min_confidence: An integer that is the default minimum confidence level. + + """ + + def __init__(self, min_confidence, output_format): + self.min_confidence = min_confidence + self.output_format = output_format + + +# This class should not have knowledge of the flag key names. +class CommandOptionValues(object): + + """Stores the option values passed by the user via the command line. + + Attributes: + is_verbose: A boolean value of whether verbose logging is enabled. + + filter_rules: The list of filter rules provided by the user. + These rules are appended to the base rules and + path-specific rules and so take precedence over + the base filter rules, etc. + + git_commit: A string representing the git commit to check. + The default is None. + + min_confidence: An integer between 1 and 5 inclusive that is the + minimum confidence level of style errors to report. + The default is 1, which reports all errors. + + output_format: A string that is the output format. The supported + output formats are "emacs" which emacs can parse + and "vs7" which Microsoft Visual Studio 7 can parse. + + """ + def __init__(self, + filter_rules=None, + git_commit=None, + diff_files=None, + is_verbose=False, + min_confidence=1, + output_format="emacs"): + if filter_rules is None: + filter_rules = [] + + if (min_confidence < 1) or (min_confidence > 5): + raise ValueError('Invalid "min_confidence" parameter: value ' + "must be an integer between 1 and 5 inclusive. " + 'Value given: "%s".' % min_confidence) + + if output_format not in ("emacs", "vs7"): + raise ValueError('Invalid "output_format" parameter: ' + 'value must be "emacs" or "vs7". ' + 'Value given: "%s".' % output_format) + + self.filter_rules = filter_rules + self.git_commit = git_commit + self.diff_files = diff_files + self.is_verbose = is_verbose + self.min_confidence = min_confidence + self.output_format = output_format + + # Useful for unit testing. + def __eq__(self, other): + """Return whether this instance is equal to another.""" + if self.filter_rules != other.filter_rules: + return False + if self.git_commit != other.git_commit: + return False + if self.diff_files != other.diff_files: + return False + if self.is_verbose != other.is_verbose: + return False + if self.min_confidence != other.min_confidence: + return False + if self.output_format != other.output_format: + return False + + return True + + # Useful for unit testing. + def __ne__(self, other): + # Python does not automatically deduce this from __eq__(). + return not self.__eq__(other) + + +class ArgumentPrinter(object): + + """Supports the printing of check-webkit-style command arguments.""" + + def _flag_pair_to_string(self, flag_key, flag_value): + return '--%(key)s=%(val)s' % {'key': flag_key, 'val': flag_value } + + def to_flag_string(self, options): + """Return a flag string of the given CommandOptionValues instance. + + This method orders the flag values alphabetically by the flag key. + + Args: + options: A CommandOptionValues instance. + + """ + flags = {} + flags['min-confidence'] = options.min_confidence + flags['output'] = options.output_format + # Only include the filter flag if user-provided rules are present. + filter_rules = options.filter_rules + if filter_rules: + flags['filter'] = ",".join(filter_rules) + if options.git_commit: + flags['git-commit'] = options.git_commit + if options.diff_files: + flags['diff_files'] = options.diff_files + + flag_string = '' + # Alphabetizing lets us unit test this method. + for key in sorted(flags.keys()): + flag_string += self._flag_pair_to_string(key, flags[key]) + ' ' + + return flag_string.strip() + + +class ArgumentParser(object): + + # FIXME: Move the documentation of the attributes to the __init__ + # docstring after making the attributes internal. + """Supports the parsing of check-webkit-style command arguments. + + Attributes: + create_usage: A function that accepts a DefaultCommandOptionValues + instance and returns a string of usage instructions. + Defaults to the function that generates the usage + string for check-webkit-style. + default_options: A DefaultCommandOptionValues instance that provides + the default values for options not explicitly + provided by the user. + stderr_write: A function that takes a string as a parameter and + serves as stderr.write. Defaults to sys.stderr.write. + This parameter should be specified only for unit tests. + + """ + + def __init__(self, + all_categories, + default_options, + base_filter_rules=None, + mock_stderr=None, + usage=None): + """Create an ArgumentParser instance. + + Args: + all_categories: The set of all available style categories. + default_options: See the corresponding attribute in the class + docstring. + Keyword Args: + base_filter_rules: The list of filter rules at the beginning of + the list of rules used to check style. This + list has the least precedence when checking + style and precedes any user-provided rules. + The class uses this parameter only for display + purposes to the user. Defaults to the empty list. + create_usage: See the documentation of the corresponding + attribute in the class docstring. + stderr_write: See the documentation of the corresponding + attribute in the class docstring. + + """ + if base_filter_rules is None: + base_filter_rules = [] + stderr = sys.stderr if mock_stderr is None else mock_stderr + if usage is None: + usage = _USAGE + + self._all_categories = all_categories + self._base_filter_rules = base_filter_rules + + # FIXME: Rename these to reflect that they are internal. + self.default_options = default_options + self.stderr_write = stderr.write + + self._parser = self._create_option_parser(stderr=stderr, + usage=usage, + default_min_confidence=self.default_options.min_confidence, + default_output_format=self.default_options.output_format) + + def _create_option_parser(self, stderr, usage, + default_min_confidence, default_output_format): + # Since the epilog string is short, it is not necessary to replace + # the epilog string with a mock epilog string when testing. + # For this reason, we use _EPILOG directly rather than passing it + # as an argument like we do for the usage string. + parser = OptionParser(usage=usage, epilog=_EPILOG) + + filter_help = ('set a filter to control what categories of style ' + 'errors to report. Specify a filter using a comma-' + 'delimited list of boolean filter rules, for example ' + '"--filter -whitespace,+whitespace/braces". To display ' + 'all categories and which are enabled by default, pass ' + """no value (e.g. '-f ""' or '--filter=').""") + parser.add_option("-f", "--filter-rules", metavar="RULES", + dest="filter_value", help=filter_help) + + git_commit_help = ("check all changes in the given commit. " + "Use 'commit_id..' to check all changes after commmit_id") + parser.add_option("-g", "--git-diff", "--git-commit", + metavar="COMMIT", dest="git_commit", help=git_commit_help,) + + diff_files_help = "diff the files passed on the command line rather than checking the style of every line" + parser.add_option("--diff-files", action="store_true", dest="diff_files", default=False, help=diff_files_help) + + min_confidence_help = ("set the minimum confidence of style errors " + "to report. Can be an integer 1-5, with 1 " + "displaying all errors. Defaults to %default.") + parser.add_option("-m", "--min-confidence", metavar="INT", + type="int", dest="min_confidence", + default=default_min_confidence, + help=min_confidence_help) + + output_format_help = ('set the output format, which can be "emacs" ' + 'or "vs7" (for Visual Studio). ' + 'Defaults to "%default".') + parser.add_option("-o", "--output-format", metavar="FORMAT", + choices=["emacs", "vs7"], + dest="output_format", default=default_output_format, + help=output_format_help) + + verbose_help = "enable verbose logging." + parser.add_option("-v", "--verbose", dest="is_verbose", default=False, + action="store_true", help=verbose_help) + + # Override OptionParser's error() method so that option help will + # also display when an error occurs. Normally, just the usage + # string displays and not option help. + parser.error = self._parse_error + + # Override OptionParser's print_help() method so that help output + # does not render to the screen while running unit tests. + print_help = parser.print_help + parser.print_help = lambda: print_help(file=stderr) + + return parser + + def _parse_error(self, error_message): + """Print the help string and an error message, and exit.""" + # The method format_help() includes both the usage string and + # the flag options. + help = self._parser.format_help() + # Separate help from the error message with a single blank line. + self.stderr_write(help + "\n") + if error_message: + _log.error(error_message) + + # Since we are using this method to replace/override the Python + # module optparse's OptionParser.error() method, we match its + # behavior and exit with status code 2. + # + # As additional background, Python documentation says-- + # + # "Unix programs generally use 2 for command line syntax errors + # and 1 for all other kind of errors." + # + # (from http://docs.python.org/library/sys.html#sys.exit ) + sys.exit(2) + + def _exit_with_categories(self): + """Exit and print the style categories and default filter rules.""" + self.stderr_write('\nAll categories:\n') + for category in sorted(self._all_categories): + self.stderr_write(' ' + category + '\n') + + self.stderr_write('\nDefault filter rules**:\n') + for filter_rule in sorted(self._base_filter_rules): + self.stderr_write(' ' + filter_rule + '\n') + self.stderr_write('\n**The command always evaluates the above rules, ' + 'and before any --filter flag.\n\n') + + sys.exit(0) + + def _parse_filter_flag(self, flag_value): + """Parse the --filter flag, and return a list of filter rules. + + Args: + flag_value: A string of comma-separated filter rules, for + example "-whitespace,+whitespace/indent". + + """ + filters = [] + for uncleaned_filter in flag_value.split(','): + filter = uncleaned_filter.strip() + if not filter: + continue + filters.append(filter) + return filters + + def parse(self, args): + """Parse the command line arguments to check-webkit-style. + + Args: + args: A list of command-line arguments as returned by sys.argv[1:]. + + Returns: + A tuple of (paths, options) + + paths: The list of paths to check. + options: A CommandOptionValues instance. + + """ + (options, paths) = self._parser.parse_args(args=args) + + filter_value = options.filter_value + git_commit = options.git_commit + diff_files = options.diff_files + is_verbose = options.is_verbose + min_confidence = options.min_confidence + output_format = options.output_format + + if filter_value is not None and not filter_value: + # Then the user explicitly passed no filter, for + # example "-f ''" or "--filter=". + self._exit_with_categories() + + # Validate user-provided values. + + min_confidence = int(min_confidence) + if (min_confidence < 1) or (min_confidence > 5): + self._parse_error('option --min-confidence: invalid integer: ' + '%s: value must be between 1 and 5' + % min_confidence) + + if filter_value: + filter_rules = self._parse_filter_flag(filter_value) + else: + filter_rules = [] + + try: + validate_filter_rules(filter_rules, self._all_categories) + except ValueError, err: + self._parse_error(err) + + options = CommandOptionValues(filter_rules=filter_rules, + git_commit=git_commit, + diff_files=diff_files, + is_verbose=is_verbose, + min_confidence=min_confidence, + output_format=output_format) + + return (paths, options) + diff --git a/Tools/Scripts/webkitpy/style/optparser_unittest.py b/Tools/Scripts/webkitpy/style/optparser_unittest.py new file mode 100644 index 0000000..a6b64da --- /dev/null +++ b/Tools/Scripts/webkitpy/style/optparser_unittest.py @@ -0,0 +1,258 @@ +# Copyright (C) 2010 Chris Jerdonek (cjerdonek@webkit.org) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# 1. Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# 2. Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# +# THIS SOFTWARE IS PROVIDED BY APPLE INC. AND ITS CONTRIBUTORS ``AS IS'' AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR ITS CONTRIBUTORS BE LIABLE FOR ANY +# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Unit tests for parser.py.""" + +import unittest + +from webkitpy.common.system.logtesting import LoggingTestCase +from webkitpy.style.optparser import ArgumentParser +from webkitpy.style.optparser import ArgumentPrinter +from webkitpy.style.optparser import CommandOptionValues as ProcessorOptions +from webkitpy.style.optparser import DefaultCommandOptionValues + + +class ArgumentPrinterTest(unittest.TestCase): + + """Tests the ArgumentPrinter class.""" + + _printer = ArgumentPrinter() + + def _create_options(self, + output_format='emacs', + min_confidence=3, + filter_rules=[], + git_commit=None): + return ProcessorOptions(filter_rules=filter_rules, + git_commit=git_commit, + min_confidence=min_confidence, + output_format=output_format) + + def test_to_flag_string(self): + options = self._create_options('vs7', 5, ['+foo', '-bar'], 'git') + self.assertEquals('--filter=+foo,-bar --git-commit=git ' + '--min-confidence=5 --output=vs7', + self._printer.to_flag_string(options)) + + # This is to check that --filter and --git-commit do not + # show up when not user-specified. + options = self._create_options() + self.assertEquals('--min-confidence=3 --output=emacs', + self._printer.to_flag_string(options)) + + +class ArgumentParserTest(LoggingTestCase): + + """Test the ArgumentParser class.""" + + class _MockStdErr(object): + + def write(self, message): + # We do not want the usage string or style categories + # to print during unit tests, so print nothing. + return + + def _parse(self, args): + """Call a test parser.parse().""" + parser = self._create_parser() + return parser.parse(args) + + def _create_defaults(self): + """Return a DefaultCommandOptionValues instance for testing.""" + base_filter_rules = ["-", "+whitespace"] + return DefaultCommandOptionValues(min_confidence=3, + output_format="vs7") + + def _create_parser(self): + """Return an ArgumentParser instance for testing.""" + default_options = self._create_defaults() + + all_categories = ["build" ,"whitespace"] + + mock_stderr = self._MockStdErr() + + return ArgumentParser(all_categories=all_categories, + base_filter_rules=[], + default_options=default_options, + mock_stderr=mock_stderr, + usage="test usage") + + def test_parse_documentation(self): + parse = self._parse + + # FIXME: Test both the printing of the usage string and the + # filter categories help. + + # Request the usage string. + self.assertRaises(SystemExit, parse, ['--help']) + # Request default filter rules and available style categories. + self.assertRaises(SystemExit, parse, ['--filter=']) + + def test_parse_bad_values(self): + parse = self._parse + + # Pass an unsupported argument. + self.assertRaises(SystemExit, parse, ['--bad']) + self.assertLog(['ERROR: no such option: --bad\n']) + + self.assertRaises(SystemExit, parse, ['--min-confidence=bad']) + self.assertLog(['ERROR: option --min-confidence: ' + "invalid integer value: 'bad'\n"]) + self.assertRaises(SystemExit, parse, ['--min-confidence=0']) + self.assertLog(['ERROR: option --min-confidence: invalid integer: 0: ' + 'value must be between 1 and 5\n']) + self.assertRaises(SystemExit, parse, ['--min-confidence=6']) + self.assertLog(['ERROR: option --min-confidence: invalid integer: 6: ' + 'value must be between 1 and 5\n']) + parse(['--min-confidence=1']) # works + parse(['--min-confidence=5']) # works + + self.assertRaises(SystemExit, parse, ['--output=bad']) + self.assertLog(['ERROR: option --output-format: invalid choice: ' + "'bad' (choose from 'emacs', 'vs7')\n"]) + parse(['--output=vs7']) # works + + # Pass a filter rule not beginning with + or -. + self.assertRaises(SystemExit, parse, ['--filter=build']) + self.assertLog(['ERROR: Invalid filter rule "build": ' + 'every rule must start with + or -.\n']) + parse(['--filter=+build']) # works + + def test_parse_default_arguments(self): + parse = self._parse + + (files, options) = parse([]) + + self.assertEquals(files, []) + + self.assertEquals(options.filter_rules, []) + self.assertEquals(options.git_commit, None) + self.assertEquals(options.diff_files, False) + self.assertEquals(options.is_verbose, False) + self.assertEquals(options.min_confidence, 3) + self.assertEquals(options.output_format, 'vs7') + + def test_parse_explicit_arguments(self): + parse = self._parse + + # Pass non-default explicit values. + (files, options) = parse(['--min-confidence=4']) + self.assertEquals(options.min_confidence, 4) + (files, options) = parse(['--output=emacs']) + self.assertEquals(options.output_format, 'emacs') + (files, options) = parse(['-g', 'commit']) + self.assertEquals(options.git_commit, 'commit') + (files, options) = parse(['--git-commit=commit']) + self.assertEquals(options.git_commit, 'commit') + (files, options) = parse(['--git-diff=commit']) + self.assertEquals(options.git_commit, 'commit') + (files, options) = parse(['--verbose']) + self.assertEquals(options.is_verbose, True) + (files, options) = parse(['--diff-files', 'file.txt']) + self.assertEquals(options.diff_files, True) + + # Pass user_rules. + (files, options) = parse(['--filter=+build,-whitespace']) + self.assertEquals(options.filter_rules, + ["+build", "-whitespace"]) + + # Pass spurious white space in user rules. + (files, options) = parse(['--filter=+build, -whitespace']) + self.assertEquals(options.filter_rules, + ["+build", "-whitespace"]) + + def test_parse_files(self): + parse = self._parse + + (files, options) = parse(['foo.cpp']) + self.assertEquals(files, ['foo.cpp']) + + # Pass multiple files. + (files, options) = parse(['--output=emacs', 'foo.cpp', 'bar.cpp']) + self.assertEquals(files, ['foo.cpp', 'bar.cpp']) + + +class CommandOptionValuesTest(unittest.TestCase): + + """Tests CommandOptionValues class.""" + + def test_init(self): + """Test __init__ constructor.""" + # Check default parameters. + options = ProcessorOptions() + self.assertEquals(options.filter_rules, []) + self.assertEquals(options.git_commit, None) + self.assertEquals(options.is_verbose, False) + self.assertEquals(options.min_confidence, 1) + self.assertEquals(options.output_format, "emacs") + + # Check argument validation. + self.assertRaises(ValueError, ProcessorOptions, output_format="bad") + ProcessorOptions(output_format="emacs") # No ValueError: works + ProcessorOptions(output_format="vs7") # works + self.assertRaises(ValueError, ProcessorOptions, min_confidence=0) + self.assertRaises(ValueError, ProcessorOptions, min_confidence=6) + ProcessorOptions(min_confidence=1) # works + ProcessorOptions(min_confidence=5) # works + + # Check attributes. + options = ProcessorOptions(filter_rules=["+"], + git_commit="commit", + is_verbose=True, + min_confidence=3, + output_format="vs7") + self.assertEquals(options.filter_rules, ["+"]) + self.assertEquals(options.git_commit, "commit") + self.assertEquals(options.is_verbose, True) + self.assertEquals(options.min_confidence, 3) + self.assertEquals(options.output_format, "vs7") + + def test_eq(self): + """Test __eq__ equality function.""" + self.assertTrue(ProcessorOptions().__eq__(ProcessorOptions())) + + # Also verify that a difference in any argument causes equality to fail. + + # Explicitly create a ProcessorOptions instance with all default + # values. We do this to be sure we are assuming the right default + # values in our self.assertFalse() calls below. + options = ProcessorOptions(filter_rules=[], + git_commit=None, + is_verbose=False, + min_confidence=1, + output_format="emacs") + # Verify that we created options correctly. + self.assertTrue(options.__eq__(ProcessorOptions())) + + self.assertFalse(options.__eq__(ProcessorOptions(filter_rules=["+"]))) + self.assertFalse(options.__eq__(ProcessorOptions(git_commit="commit"))) + self.assertFalse(options.__eq__(ProcessorOptions(is_verbose=True))) + self.assertFalse(options.__eq__(ProcessorOptions(min_confidence=2))) + self.assertFalse(options.__eq__(ProcessorOptions(output_format="vs7"))) + + def test_ne(self): + """Test __ne__ inequality function.""" + # By default, __ne__ always returns true on different objects. + # Thus, just check the distinguishing case to verify that the + # code defines __ne__. + self.assertFalse(ProcessorOptions().__ne__(ProcessorOptions())) + diff --git a/Tools/Scripts/webkitpy/style/patchreader.py b/Tools/Scripts/webkitpy/style/patchreader.py new file mode 100644 index 0000000..f44839d --- /dev/null +++ b/Tools/Scripts/webkitpy/style/patchreader.py @@ -0,0 +1,66 @@ +# Copyright (C) 2010 Google Inc. All rights reserved. +# Copyright (C) 2010 Chris Jerdonek (chris.jerdonek@gmail.com) +# Copyright (C) 2010 ProFUSION embedded systems +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +import logging + +from webkitpy.common.checkout.diff_parser import DiffParser + + +_log = logging.getLogger("webkitpy.style.patchreader") + + +class PatchReader(object): + """Supports checking style in patches.""" + + def __init__(self, text_file_reader): + """Create a PatchReader instance. + + Args: + text_file_reader: A TextFileReader instance. + + """ + self._text_file_reader = text_file_reader + + def check(self, patch_string): + """Check style in the given patch.""" + patch_files = DiffParser(patch_string.splitlines()).files + + for path, diff_file in patch_files.iteritems(): + line_numbers = diff_file.added_or_modified_line_numbers() + _log.debug('Found %s new or modified lines in: %s' % (len(line_numbers), path)) + + if not line_numbers: + # Don't check files which contain only deleted lines + # as they can never add style errors. However, mark them as + # processed so that we count up number of such files. + self._text_file_reader.count_delete_only_file() + continue + + self._text_file_reader.process_file(file_path=path, line_numbers=line_numbers) diff --git a/Tools/Scripts/webkitpy/style/patchreader_unittest.py b/Tools/Scripts/webkitpy/style/patchreader_unittest.py new file mode 100644 index 0000000..b121082 --- /dev/null +++ b/Tools/Scripts/webkitpy/style/patchreader_unittest.py @@ -0,0 +1,92 @@ +#!/usr/bin/python +# +# Copyright (C) 2010 Google Inc. All rights reserved. +# Copyright (C) 2009 Torch Mobile Inc. +# Copyright (C) 2009 Apple Inc. All rights reserved. +# Copyright (C) 2010 Chris Jerdonek (chris.jerdonek@gmail.com) +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +import unittest + +from webkitpy.style.patchreader import PatchReader + + +class PatchReaderTest(unittest.TestCase): + + """Test the PatchReader class.""" + + class MockTextFileReader(object): + + def __init__(self): + self.passed_to_process_file = [] + """A list of (file_path, line_numbers) pairs.""" + self.delete_only_file_count = 0 + """A number of times count_delete_only_file() called""" + + def process_file(self, file_path, line_numbers): + self.passed_to_process_file.append((file_path, line_numbers)) + + def count_delete_only_file(self): + self.delete_only_file_count += 1 + + def setUp(self): + file_reader = self.MockTextFileReader() + self._file_reader = file_reader + self._patch_checker = PatchReader(file_reader) + + def _call_check_patch(self, patch_string): + self._patch_checker.check(patch_string) + + def _assert_checked(self, passed_to_process_file, delete_only_file_count): + self.assertEquals(self._file_reader.passed_to_process_file, + passed_to_process_file) + self.assertEquals(self._file_reader.delete_only_file_count, + delete_only_file_count) + + def test_check_patch(self): + # The modified line_numbers array for this patch is: [2]. + self._call_check_patch("""diff --git a/__init__.py b/__init__.py +index ef65bee..e3db70e 100644 +--- a/__init__.py ++++ b/__init__.py +@@ -1,1 +1,2 @@ + # Required for Python to search this directory for module files ++# New line +""") + self._assert_checked([("__init__.py", [2])], 0) + + def test_check_patch_with_deletion(self): + self._call_check_patch("""Index: __init__.py +=================================================================== +--- __init__.py (revision 3593) ++++ __init__.py (working copy) +@@ -1 +0,0 @@ +-foobar +""") + # _mock_check_file should not be called for the deletion patch. + self._assert_checked([], 1) |