#!/usr/bin/python # -*- coding: utf-8 -*- # # Copyright (C) 2009, 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 (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 from webkitpy.common.memoized import memoized # 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>' 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 _rfind_in_lines(regex, lines, start_position, not_found_position): """Does a reverse find starting at start position and going backwards until a match is found. Returns the position where the regex ended. """ # Put the regex in a group and proceed it with a greedy expression that # matches anything to ensure that we get the last possible match in a line. last_in_line_regex = r'.*(' + regex + ')' current_row = start_position.row # Start with the given row and trim off everything past what may be matched. current_line = lines[start_position.row][:start_position.column] while True: found_match = match(last_in_line_regex, current_line) if found_match: return Position(current_row, found_match.end(1)) # A match was not found so continue backward. current_row -= 1 if current_row < 0: return not_found_position current_line = lines[current_row] def _convert_to_lower_with_underscores(text): """Converts all text strings in camelCase or PascalCase to lowers with underscores.""" # First add underscores before any capital letter followed by a lower case letter # as long as it is in a word. # (This put an underscore before Password but not P and A in WPAPassword). text = sub(r'(?<=[A-Za-z0-9])([A-Z])(?=[a-z])', r'_\1', text) # Next add underscores before capitals at the end of words if it was # preceeded by lower case letter or number. # (This puts an underscore before A in isA but not A in CBA). text = sub(r'(?<=[a-z0-9])([A-Z])(?=\b)', r'_\1', text) # Next add underscores when you have a captial letter which is followed by a capital letter # but is not proceeded by one. (This puts an underscore before A in 'WordADay'). text = sub(r'(?<=[a-z0-9])([A-Z][A-Z_])', r'_\1', text) return text.lower() def _create_acronym(text): """Creates an acronym for the given text.""" # Removes all lower case letters except those starting words. text = sub(r'(?= 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: if primary_header_exists: error_message = before_error_message self._section = self._OTHER_SECTION return error_message class Position(object): """Holds the position of something.""" def __init__(self, row, column): self.row = row self.column = column def __str__(self): return '(%s, %s)' % (self.row, self.column) def __cmp__(self, other): return self.row.__cmp__(other.row) or self.column.__cmp__(other.column) class Parameter(object): """Information about one function parameter.""" def __init__(self, parameter, parameter_name_index, row): self.type = parameter[:parameter_name_index].strip() # Remove any initializers from the parameter name (e.g. int i = 5). self.name = sub(r'=.*', '', parameter[parameter_name_index:]).strip() self.row = row @memoized def lower_with_underscores_name(self): """Returns the parameter name in the lower with underscores format.""" return _convert_to_lower_with_underscores(self.name) class SingleLineView(object): """Converts multiple lines into a single line (with line breaks replaced by a space) to allow for easier searching.""" def __init__(self, lines, start_position, end_position): """Create a SingleLineView instance. Args: lines: a list of multiple lines to combine into a single line. start_position: offset within lines of where to start the single line. end_position: just after where to end (like a slice operation). """ # Get the rows of interest. trimmed_lines = lines[start_position.row:end_position.row + 1] # Remove the columns on the last line that aren't included. trimmed_lines[-1] = trimmed_lines[-1][:end_position.column] # Remove the columns on the first line that aren't included. trimmed_lines[0] = trimmed_lines[0][start_position.column:] # Create a single line with all of the parameters. self.single_line = ' '.join(trimmed_lines) # Keep the row lengths, so we can calculate the original row number # given a column in the single line (adding 1 due to the space added # during the join). self._row_lengths = [len(line) + 1 for line in trimmed_lines] self._starting_row = start_position.row def convert_column_to_row(self, single_line_column_number): """Convert the column number from the single line into the original line number. Special cases: * Columns in the added spaces are considered part of the previous line. * Columns beyond the end of the line are consider part the last line in the view.""" total_columns = 0 row_offset = 0 while row_offset < len(self._row_lengths) - 1 and single_line_column_number >= total_columns + self._row_lengths[row_offset]: total_columns += self._row_lengths[row_offset] row_offset += 1 return self._starting_row + row_offset def create_skeleton_parameters(all_parameters): """Converts a parameter list to a skeleton version. The skeleton only has one word for the parameter name, one word for the type, and commas after each parameter and only there. Everything in the skeleton remains in the same columns as the original.""" all_simplifications = ( # Remove template parameters, function declaration parameters, etc. r'(<[^<>]*?>)|(\([^\(\)]*?\))|(\{[^\{\}]*?\})', # Remove all initializers. r'=[^,]*', # Remove :: and everything before it. r'[^,]*::', # Remove modifiers like &, *. r'[&*]', # Remove const modifiers. r'\bconst\s+(?=[A-Za-z])', # Remove numerical modifiers like long. r'\b(unsigned|long|short)\s+(?=unsigned|long|short|int|char|double|float)') skeleton_parameters = all_parameters for simplification in all_simplifications: skeleton_parameters = iteratively_replace_matches_with_char(simplification, ' ', skeleton_parameters) # If there are any parameters, then add a , after the last one to # make a regular pattern of a , following every parameter. if skeleton_parameters.strip(): skeleton_parameters += ',' return skeleton_parameters def find_parameter_name_index(skeleton_parameter): """Determines where the parametere name starts given the skeleton parameter.""" # The first space from the right in the simplified parameter is where the parameter # name starts unless the first space is before any content in the simplified parameter. before_name_index = skeleton_parameter.rstrip().rfind(' ') if before_name_index != -1 and skeleton_parameter[:before_name_index].strip(): return before_name_index + 1 return len(skeleton_parameter) def parameter_list(elided_lines, start_position, end_position): """Generator for a function's parameters.""" # Create new positions that omit the outer parenthesis of the parameters. start_position = Position(row=start_position.row, column=start_position.column + 1) end_position = Position(row=end_position.row, column=end_position.column - 1) single_line_view = SingleLineView(elided_lines, start_position, end_position) skeleton_parameters = create_skeleton_parameters(single_line_view.single_line) end_index = -1 while True: # Find the end of the next parameter. start_index = end_index + 1 end_index = skeleton_parameters.find(',', start_index) # No comma means that all parameters have been parsed. if end_index == -1: return row = single_line_view.convert_column_to_row(end_index) # Parse the parameter into a type and parameter name. skeleton_parameter = skeleton_parameters[start_index:end_index] name_offset = find_parameter_name_index(skeleton_parameter) parameter = single_line_view.single_line[start_index:end_index] yield Parameter(parameter, name_offset, row) 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 positions (even when a # small amount is added to them). self.body_start_position = Position(-1000, 0) self.end_position = Position(-1000, 0) def begin(self, function_name, function_name_start_position, body_start_position, end_position, parameter_start_position, parameter_end_position, clean_lines): """Start analyzing function body. Args: function_name: The name of the function being tracked. function_name_start_position: Position in elided where the function name starts. body_start_position: Position in elided of the { or the ; for a prototype. end_position: Position in elided just after the final } (or ; is. parameter_start_position: Position in elided of the '(' for the parameters. parameter_end_position: Position in elided just after the ')' for the parameters. clean_lines: A CleansedLines instance containing the file. """ self.in_a_function = True self.lines_in_function = -1 # Don't count the open brace line. self.current_function = function_name self.function_name_start_position = function_name_start_position self.body_start_position = body_start_position self.end_position = end_position self.is_declaration = clean_lines.elided[body_start_position.row][body_start_position.column] == ';' self.parameter_start_position = parameter_start_position self.parameter_end_position = parameter_end_position self.is_pure = False if self.is_declaration: characters_after_parameters = SingleLineView(clean_lines.elided, parameter_end_position, body_start_position).single_line self.is_pure = bool(match(r'\s*=\s*0\s*', characters_after_parameters)) self._clean_lines = clean_lines self._parameter_list = None def modifiers_and_return_type(self): """Returns the modifiers and the return type.""" # Go backwards from where the function name is until we encounter one of several things: # ';' or '{' or '}' or 'private:', etc. or '#' or return Position(0, 0) elided = self._clean_lines.elided start_modifiers = _rfind_in_lines(r';|\{|\}|((private|public|protected):)|(#.*)', elided, self.parameter_start_position, Position(0, 0)) return SingleLineView(elided, start_modifiers, self.function_name_start_position).single_line.strip() def parameter_list(self): if not self._parameter_list: # Store the final result as a tuple since that is immutable. self._parameter_list = tuple(parameter_list(self._clean_lines.elided, self.parameter_start_position, self.parameter_end_position)) return self._parameter_list def count(self, line_number): """Count line in current function body.""" if self.in_a_function and line_number >= self.body_start_position.row: 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 , 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(elided, position): """If input points to ( or { or [, finds the position that closes it. If elided[position.row][position.column] points to a '(' or '{' or '[', finds the line_number/pos that correspond to the closing of the expression. Args: elided: A CleansedLines.elided instance containing the file. position: The position of the opening item. Returns: The Position *past* the closing brace, or Position(len(elided), -1) if we never find a close. Note we ignore strings and comments when matching. """ line = elided[position.row] start_character = line[position.column] if start_character == '(': enclosing_character_regex = r'[\(\)]' elif start_character == '[': enclosing_character_regex = r'[\[\]]' elif start_character == '{': enclosing_character_regex = r'[\{\}]' else: return Position(len(elided), -1) current_column = position.column + 1 line_number = position.row net_open = 1 for line in elided[position.row:]: line = line[current_column:] # Search the current line for opening and closing characters. while True: next_enclosing_character = search(enclosing_character_regex, line) # No more on this line. if not next_enclosing_character: break current_column += next_enclosing_character.end(0) line = line[next_enclosing_character.end(0):] if next_enclosing_character.group(0) == start_character: net_open += 1 else: net_open -= 1 if not net_open: return Position(line_number, current_column) # Proceed to the next line. line_number += 1 current_column = 0 # The given item was not closed. return Position(len(elided), -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] "') 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 ?= and )\?=?\s*(\w+|[+-]?\d+)(\.\d*)?', line): error(line_number, 'build/deprecated', 3, '>? and ,:]*>\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'(?|,|~|(operator\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() body_match = search(r'{|;', start_line) if body_match: body_start_position = Position(start_line_number, body_match.start(0)) # 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|:|<|>|,|~|(operator\s*(/|-|=|!|\+)+))*)\(', 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)] function_name_start_position = Position(line_number, match_function.start(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 += '()' parameter_start_position = Position(line_number, match_function.end(1)) parameter_end_position = close_expression(clean_lines.elided, parameter_start_position) if parameter_end_position.row == len(clean_lines.elided): # No end was found. return if start_line[body_start_position.column] == ';': end_position = Position(body_start_position.row, body_start_position.column + 1) else: end_position = close_expression(clean_lines.elided, body_start_position) # Check for nonsensical positions. (This happens in test cases which check code snippets.) if parameter_end_position > body_start_position: return function_state.begin(function, function_name_start_position, body_start_position, end_position, parameter_start_position, parameter_end_position, clean_lines) 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.end_position.row == 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_parameter_name_against_text(parameter, text, error): """Checks to see if the parameter name is contained within the text. Return false if the check failed (i.e. an error was produced). """ # Treat 'lower with underscores' as a canonical form because it is # case insensitive while still retaining word breaks. (This ensures that # 'elate' doesn't look like it is duplicating of 'NateLate'.) canonical_parameter_name = parameter.lower_with_underscores_name() # Appends "object" to all text to catch variables that did the same (but only # do this when the parameter name is more than a single character to avoid # flagging 'b' which may be an ok variable when used in an rgba function). if len(canonical_parameter_name) > 1: text = sub(r'(\w)\b', r'\1Object', text) canonical_text = _convert_to_lower_with_underscores(text) # Used to detect cases like ec for ExceptionCode. acronym = _create_acronym(text).lower() if canonical_text.find(canonical_parameter_name) != -1 or acronym.find(canonical_parameter_name) != -1: error(parameter.row, 'readability/parameter_name', 5, 'The parameter name "%s" adds no information, so it should be removed.' % parameter.name) return False return True def check_function_definition(filename, file_extension, clean_lines, line_number, function_state, error): """Check that function definitions for style issues. Specifically, check that parameter names in declarations add information. Args: filename: Filename of the file that is being processed. 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. function_state: Current function name and lines in body so far. error: The function to call with any errors found. """ if line_number != function_state.body_start_position.row: return modifiers_and_return_type = function_state.modifiers_and_return_type() if filename.find('/chromium/') != -1 and search(r'\bWEBKIT_API\b', modifiers_and_return_type): if filename.find('/chromium/public/') == -1: error(function_state.function_name_start_position.row, 'readability/webkit_api', 5, 'WEBKIT_API should only appear in the chromium public directory.') elif not file_extension == "h": error(function_state.function_name_start_position.row, 'readability/webkit_api', 5, 'WEBKIT_API should only be used in header files.') elif not function_state.is_declaration or search(r'\binline\b', modifiers_and_return_type): error(function_state.function_name_start_position.row, 'readability/webkit_api', 5, 'WEBKIT_API should not be used on a function with a body.') elif function_state.is_pure: error(function_state.function_name_start_position.row, 'readability/webkit_api', 5, 'WEBKIT_API should not be used with a pure virtual function.') # Do checks specific to function declaractions. if not function_state.is_declaration: return parameter_list = function_state.parameter_list() for parameter in parameter_list: if not parameter.name: continue # Check the parameter name against the function name for single parameter set functions. if len(parameter_list) == 1 and match('set[A-Z]', function_state.current_function): trimmed_function_name = function_state.current_function[len('set'):] if not _check_parameter_name_against_text(parameter, trimmed_function_name, error): continue # Since an error was noted for this name, move to the next parameter. # Check the parameter name against the type. if not _check_parameter_name_against_text(parameter, parameter.type, error): continue # Since an error was noted for this name, move to the next parameter. 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_position.row: 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*$|:)', previous_line) or match(r' {4}:', previous_line)) if not exception: error(line_number, 'whitespace/blank_line', 2, 'Blank line at the start of a code block. Is this needed?') # This doesn't ignore whitespace at the end of a namespace block # because that is too hard without pairing open/close braces; # however, a special exception is made for namespace closing # brackets which have a comment containing "namespace". # # Also, ignore blank lines at the end of a block in a long if-else # chain, like this: # if (condition1) { # // Something followed by a blank line # # } else if (condition2) { # // Something else # } if line_number + 1 < clean_lines.num_lines(): next_line = raw[line_number + 1] if (next_line and match(r'\s*}', next_line) and next_line.find('namespace') == -1 and next_line.find('} else ') == -1): error(line_number, 'whitespace/blank_line', 3, 'Blank line at the end of a code block. Is this needed?') # Next, we check for proper spacing with respect to comments. comment_position = line.find('//') if comment_position != -1: # Check if the // may be in quotes. If so, ignore it # Comparisons made explicit for clarity -- pylint: disable-msg=C6403 if (line.count('"', 0, comment_position) - line.count('\\"', 0, comment_position)) % 2 == 0: # not in quotes # Allow one space before end of line comment. if (not match(r'^\s*$', 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') # There should only be one space after punctuation in a comment. if search('[.!?,;:]\s\s', line[comment_position:]): error(line_number, 'whitespace/comments', 5, 'Should only a single space after a punctuation in a 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 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(?Pif|for|foreach|while|switch)\s*\((?P.*)$', line) if matched: statement = matched.group('statement') condition, rest = up_to_unmatched_closing_paren(matched.group('remainder')) if condition is not None: condition_match = search(r'(?P[ ]*)(?P.).*[^ ]+(?P[ ]*)', condition) if condition_match: n_leading = len(condition_match.group('leading')) n_trailing = len(condition_match.group('trailing')) if n_leading != 0: for_exception = statement == 'for' and condition.startswith(' ;') if not for_exception: error(line_number, 'whitespace/parens', 5, 'Extra space after ( in %s' % statement) if n_trailing != 0: for_exception = statement == 'for' and condition.endswith('; ') if not for_exception: error(line_number, 'whitespace/parens', 5, 'Extra space before ) in %s' % statement) # Do not check for more than one command in macros in_preprocessor_directive = match(r'\s*#', line) if not in_preprocessor_directive and not match(r'((\s*{\s*}?)|(\s*;?))\s*\\?$', 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[a-zA-Z0-9_\*&]+)\s\s+(?P[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+(?\*|\&)\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*$', line) and line.count('(') == line.count(')') # Allow do {} while(); and not search(r'}\s*while', line)): error(line_number, 'whitespace/semicolon', 5, 'Semicolon defining empty statement for this loop. Use { } instead.') def get_previous_non_blank_line(clean_lines, line_number): """Return the most recent non-blank line and its line number. Args: clean_lines: A CleansedLines instance containing the file contents. line_number: The number of the line to check. Returns: A tuple with two elements. The first element is the contents of the last non-blank line before the current line, or the empty string if this is the first non-blank line. The second is the line number of that line, or -1 if this is the first non-blank line. """ previous_line_number = line_number - 1 while previous_line_number >= 0: previous_line = clean_lines.elided[previous_line_number] if not is_blank_line(previous_line): # if not a blank line... return (previous_line, previous_line_number) previous_line_number -= 1 return ('', -1) def check_namespace_indentation(clean_lines, line_number, file_extension, file_state, error): """Looks for indentation errors inside of namespaces. Args: clean_lines: A CleansedLines instance containing the file. line_number: The number of the line to check. file_extension: The extension (dot not included) of the file. file_state: A _FileState instance which maintains information about the state of things in the file. error: The function to call with any errors found. """ line = clean_lines.elided[line_number] # Get rid of comments and strings. namespace_match = match(r'(?P\s*)namespace\s+\S+\s*{\s*$', 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\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))\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() 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\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*$', current_line): # Complexity alarm - another switch statement nested inside the one # that we're currently testing. We'll need to track the extent of # that inner switch if the upcoming label tests are still supposed # to work correctly. Let's not do that; instead, we'll finish # checking this line, and then leave it like that. Assuming the # indentation is done consistently (even if incorrectly), this will # still catch all indentation issues in practice. encountered_nested_switch = True current_indentation_match = match(r'(?P\s*)(?P.*)$', 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] last_open_brace = previous_line.rfind('{') if (last_open_brace != -1 and previous_line.find('}', last_open_brace) == -1 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\s*)(\}\s*)?else(\s+if\s*\(|(?P\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*$', current_line): continue # Skip lines with closing braces on the original indentation level. # Even though the styleguide says they should be on the same line as # the "else if" statement, we also want to check for instances where # the current code does not comply with the coding style. Thus, ignore # these lines and proceed to the line before that. if current_line == else_indentation + '}': continue current_indentation_match = match(r'(?P\s*)(?P.*)$', 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 # Don't warn about NULL usage in gtk_widget_style_get(). See Bug 51758. if search(r'\bgtk_widget_style_get\(\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 or null instead of NULL (even in *comments*).') 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