Repository: mesos Updated Branches: refs/heads/master 44856413f -> e88fe8c31
http://git-wip-us.apache.org/repos/asf/mesos/blob/7454420c/support/cpplint.py ---------------------------------------------------------------------- diff --git a/support/cpplint.py b/support/cpplint.py index 1e5d96f..7a020a4 100644 --- a/support/cpplint.py +++ b/support/cpplint.py @@ -1,4 +1,4 @@ -#!/usr/bin/python +#!/usr/bin/env python # # Copyright (c) 2009 Google Inc. All rights reserved. # @@ -63,11 +63,12 @@ import unicodedata _USAGE = """ Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...] [--counting=total|toplevel|detailed] [--root=subdir] - [--linelength=digits] + [--linelength=digits] [--headers=x,y,...] + [--quiet] <file> [file] ... The style guidelines this tries to follow are those in - http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml + https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml Every problem is given a confidence score from 1-5, with 5 meaning we are certain of the problem, and 1 meaning it could be a legitimate construct. @@ -78,7 +79,7 @@ Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...] suppresses errors of all categories on that line. The files passed in will be linted; at least one file must be provided. - Default linted extensions are .cc, .cpp, .cu, .cuh, .hpp and .h. Change the + Default linted extensions are .cc, .cpp, .cu, .cuh and .h. Change the extensions with the --extensions flag. Flags: @@ -90,6 +91,9 @@ Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...] verbose=# Specify a number 0-5 to restrict errors to certain verbosity levels. + quiet + Don't print anything if no errors are found. + filter=-x,+y,... Specify a comma-separated list of category-filters to apply: only error messages whose category names pass the filters will be printed. @@ -121,12 +125,13 @@ Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...] ignored. Examples: - Assuing that src/.git exists, the header guard CPP variables for - src/chrome/browser/ui/browser.h are: + Assuming that top/src/.git exists (and cwd=top/src), the header guard + CPP variables for top/src/chrome/browser/ui/browser.h are: No flag => CHROME_BROWSER_UI_BROWSER_H_ --root=chrome => BROWSER_UI_BROWSER_H_ --root=chrome/browser => UI_BROWSER_H_ + --root=.. => SRC_CHROME_BROWSER_UI_BROWSER_H_ linelength=digits This is the allowed line length for the project. The default value is @@ -140,6 +145,57 @@ Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...] Examples: --extensions=hpp,cpp + + headers=x,y,... + The header extensions that cpplint will treat as .h in checks. Values are + automatically added to --extensions list. + + Examples: + --headers=hpp,hxx + --headers=hpp + + cpplint.py supports per-directory configurations specified in CPPLINT.cfg + files. CPPLINT.cfg file can contain a number of key=value pairs. + Currently the following options are supported: + + set noparent + filter=+filter1,-filter2,... + exclude_files=regex + linelength=80 + root=subdir + headers=x,y,... + + "set noparent" option prevents cpplint from traversing directory tree + upwards looking for more .cfg files in parent directories. This option + is usually placed in the top-level project directory. + + The "filter" option is similar in function to --filter flag. It specifies + message filters in addition to the |_DEFAULT_FILTERS| and those specified + through --filter command-line flag. + + "exclude_files" allows to specify a regular expression to be matched against + a file name. If the expression matches, the file is skipped and not run + through liner. + + "linelength" allows to specify the allowed line length for the project. + + The "root" option is similar in function to the --root flag (see example + above). Paths are relative to the directory of the CPPLINT.cfg. + + The "headers" option is similar in function to the --headers flag + (see example above). + + CPPLINT.cfg has an effect on files in the same directory and all + sub-directories, unless overridden by a nested configuration file. + + Example file: + filter=-build/include_order,+build/include_alpha + exclude_files=.*\.cc + + The above example disables build/include_order warning and enables + build/include_alpha as well as excludes all .cc from being + processed by linter, in the current directory (where the .cfg + file is located) and all sub-directories. """ # We categorize each error message we print. Here are the categories. @@ -147,81 +203,102 @@ Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...] # If you add a new error message with a new category, add it to the list # here! cpplint_unittest.py should tell you if you forget to do this. _ERROR_CATEGORIES = [ - 'build/class', - 'build/deprecated', - 'build/endif_comment', - 'build/explicit_make_pair', - 'build/forward_decl', - 'build/header_guard', - 'build/include', - 'build/include_alpha', - 'build/include_order', - 'build/include_what_you_use', - 'build/namespaces', - 'build/nullptr', - 'build/printf_format', - 'build/storage_class', - 'legal/copyright', - 'readability/alt_tokens', - 'readability/braces', - 'readability/casting', - 'readability/check', - 'readability/constructors', - 'readability/fn_size', - 'readability/function', - 'readability/multiline_comment', - 'readability/multiline_string', - 'readability/namespace', - 'readability/nolint', - 'readability/nul', - 'readability/streams', - 'readability/todo', - 'readability/utf8', - 'runtime/arrays', - 'runtime/casting', - 'runtime/explicit', - 'runtime/int', - 'runtime/init', - 'runtime/invalid_increment', - 'runtime/member_string_references', - 'runtime/memset', - 'runtime/operator', - 'runtime/printf', - 'runtime/printf_format', - 'runtime/references', - 'runtime/string', - 'runtime/threadsafe_fn', - 'runtime/vlog', - 'whitespace/blank_line', - 'whitespace/braces', - 'whitespace/comma', - 'whitespace/comments', - 'whitespace/empty_conditional_body', - 'whitespace/empty_loop_body', - 'whitespace/end_of_line', - 'whitespace/ending_newline', - 'whitespace/forcolon', - 'whitespace/indent', - 'whitespace/line_length', - 'whitespace/newline', - 'whitespace/operators', - 'whitespace/parens', - 'whitespace/semicolon', - 'whitespace/tab', - 'whitespace/todo' - ] - -# The default state of the category filter. This is overrided by the --filter= + 'build/class', + 'build/c++11', + 'build/c++14', + 'build/c++tr1', + 'build/deprecated', + 'build/endif_comment', + 'build/explicit_make_pair', + 'build/forward_decl', + 'build/header_guard', + 'build/include', + 'build/include_alpha', + 'build/include_order', + 'build/include_what_you_use', + 'build/namespaces', + 'build/nullptr', + 'build/printf_format', + 'build/storage_class', + 'legal/copyright', + 'readability/alt_tokens', + 'readability/braces', + 'readability/casting', + 'readability/check', + 'readability/constructors', + 'readability/fn_size', + 'readability/inheritance', + 'readability/multiline_comment', + 'readability/multiline_string', + 'readability/namespace', + 'readability/nolint', + 'readability/nul', + 'readability/strings', + 'readability/todo', + 'readability/utf8', + 'runtime/arrays', + 'runtime/casting', + 'runtime/explicit', + 'runtime/int', + 'runtime/init', + 'runtime/invalid_increment', + 'runtime/member_string_references', + 'runtime/memset', + 'runtime/indentation_namespace', + 'runtime/operator', + 'runtime/printf', + 'runtime/printf_format', + 'runtime/references', + 'runtime/string', + 'runtime/threadsafe_fn', + 'runtime/vlog', + 'whitespace/blank_line', + 'whitespace/braces', + 'whitespace/comma', + 'whitespace/comments', + 'whitespace/empty_conditional_body', + 'whitespace/empty_if_body', + 'whitespace/empty_loop_body', + 'whitespace/end_of_line', + 'whitespace/ending_newline', + 'whitespace/forcolon', + 'whitespace/indent', + 'whitespace/line_length', + 'whitespace/newline', + 'whitespace/operators', + 'whitespace/parens', + 'whitespace/semicolon', + 'whitespace/tab', + 'whitespace/todo', + ] + +# These error categories are no longer enforced by cpplint, but for backwards- +# compatibility they may still appear in NOLINT comments. +_LEGACY_ERROR_CATEGORIES = [ + 'readability/streams', + 'readability/function', + ] + +# The default state of the category filter. This is overridden by the --filter= # flag. By default all errors are on, so only add here categories that should be # off by default (i.e., categories that must be enabled by the --filter= flags). # All entries here should start with a '-' or '+', as in the --filter= flag. _DEFAULT_FILTERS = ['-build/include_alpha'] +# The default list of categories suppressed for C (not C++) files. +_DEFAULT_C_SUPPRESSED_CATEGORIES = [ + 'readability/casting', + ] + +# The default list of categories suppressed for Linux Kernel files. +_DEFAULT_KERNEL_SUPPRESSED_CATEGORIES = [ + 'whitespace/tab', + ] + # We used to check for high-bit characters, but after much discussion we # decided those were OK, as long as they were in UTF-8 and didn't represent # hard-coded international strings, which belong in a separate i18n file. - # C++ headers _CPP_HEADERS = frozenset([ # Legacy @@ -311,6 +388,7 @@ _CPP_HEADERS = frozenset([ 'random', 'ratio', 'regex', + 'scoped_allocator', 'set', 'sstream', 'stack', @@ -358,15 +436,40 @@ _CPP_HEADERS = frozenset([ 'cwctype', ]) +# Type names +_TYPES = re.compile( + r'^(?:' + # [dcl.type.simple] + r'(char(16_t|32_t)?)|wchar_t|' + r'bool|short|int|long|signed|unsigned|float|double|' + # [support.types] + r'(ptrdiff_t|size_t|max_align_t|nullptr_t)|' + # [cstdint.syn] + r'(u?int(_fast|_least)?(8|16|32|64)_t)|' + r'(u?int(max|ptr)_t)|' + r')$') + + +# These headers are excluded from [build/include] and [build/include_order] +# checks: +# - Anything not following google file name conventions (containing an +# uppercase character, such as Python.h or nsStringAPI.h, for example). +# - Lua headers. +_THIRD_PARTY_HEADERS_PATTERN = re.compile( + r'^(?:[^/]*[A-Z][^/]*\.h|lua\.h|lauxlib\.h|lualib\.h)$') + +# Pattern for matching FileInfo.BaseName() against test file name +_TEST_FILE_SUFFIX = r'(_test|_unittest|_regtest)$' + +# Pattern that matches only complete whitespace, possibly across multiple lines. +_EMPTY_CONDITIONAL_BODY_PATTERN = re.compile(r'^\s*$', re.DOTALL) + # 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. +# testing/base/public/gunit.h. _CHECK_MACROS = [ 'DCHECK', 'CHECK', - 'EXPECT_TRUE_M', 'EXPECT_TRUE', - 'ASSERT_TRUE_M', 'ASSERT_TRUE', - 'EXPECT_FALSE_M', 'EXPECT_FALSE', - 'ASSERT_FALSE_M', 'ASSERT_FALSE', + 'EXPECT_TRUE', 'ASSERT_TRUE', + 'EXPECT_FALSE', 'ASSERT_FALSE', ] # Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE @@ -379,16 +482,12 @@ for op, replacement in [('==', 'EQ'), ('!=', 'NE'), _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 # Alternative tokens and their replacements. For full list, see section 2.5 # Alternative tokens [lex.digraph] in the C++ standard. @@ -437,11 +536,14 @@ _MATCH_ASM = re.compile(r'^\s*(?:asm|_asm|__asm|__asm__)' r'(?:\s+(volatile|__volatile__))?' r'\s*[{(]') +# Match strings that indicate we're working on a C (not C++) file. +_SEARCH_C_FILE = re.compile(r'\b(?:LINT_C_FILE|' + r'vim?:\s*.*(\s*|:)filetype=c(\s*|:|$))') -_regexp_compile_cache = {} +# Match string that indicates we're working on a Linux Kernel file. +_SEARCH_KERNEL_FILE = re.compile(r'\b(?:LINT_KERNEL_FILE)') -# Finds occurrences of NOLINT or NOLINT(...). -_RE_SUPPRESSION = re.compile(r'\bNOLINT\b(\([^)]*\))?') +_regexp_compile_cache = {} # {str, set(int)}: a map from error categories to sets of linenumbers # on which those errors are expected and should be suppressed. @@ -450,6 +552,7 @@ _error_suppressions = {} # The root directory used for deriving header guard CPP variable. # This is set by --root flag. _root = None +_root_debug = False # The allowed line length of files. # This is set by --linelength flag. @@ -459,8 +562,28 @@ _line_length = 80 # This is set by --extensions flag. _valid_extensions = set(['cc', 'h', 'cpp', 'cu', 'cuh', 'hpp']) +# Treat all headers starting with 'h' equally: .h, .hpp, .hxx etc. +# This is set by --headers flag. +_hpp_headers = set(['h']) + +# {str, bool}: a map from error categories to booleans which indicate if the +# category should be suppressed for every line. +_global_error_suppressions = {} + +def ProcessHppHeadersOption(val): + global _hpp_headers + try: + _hpp_headers = set(val.split(',')) + # Automatically append to extensions list so it does not have to be set 2 times + _valid_extensions.update(_hpp_headers) + except ValueError: + PrintUsage('Header extensions must be comma seperated list.') + +def IsHeaderExtension(file_extension): + return file_extension in _hpp_headers + def ParseNolintSuppressions(filename, raw_line, linenum, error): - """Updates the global list of error-suppressions. + """Updates the global list of line error-suppressions. Parses any NOLINT comments on the current line, updating the global error_suppressions store. Reports an error if the NOLINT comment @@ -472,42 +595,67 @@ def ParseNolintSuppressions(filename, raw_line, linenum, error): linenum: int, the number of the current line. error: function, an error handler. """ - # FIXME(adonovan): "NOLINT(" is misparsed as NOLINT(*). - matched = _RE_SUPPRESSION.search(raw_line) + matched = Search(r'\bNOLINT(NEXTLINE)?\b(\([^)]+\))?', raw_line) if matched: - category = matched.group(1) + if matched.group(1): + suppressed_line = linenum + 1 + else: + suppressed_line = linenum + category = matched.group(2) if category in (None, '(*)'): # => "suppress all" - _error_suppressions.setdefault(None, set()).add(linenum) + _error_suppressions.setdefault(None, set()).add(suppressed_line) else: if category.startswith('(') and category.endswith(')'): category = category[1:-1] if category in _ERROR_CATEGORIES: - _error_suppressions.setdefault(category, set()).add(linenum) - else: + _error_suppressions.setdefault(category, set()).add(suppressed_line) + elif category not in _LEGACY_ERROR_CATEGORIES: error(filename, linenum, 'readability/nolint', 5, 'Unknown NOLINT error category: %s' % category) +def ProcessGlobalSuppresions(lines): + """Updates the list of global error suppressions. + + Parses any lint directives in the file that have global effect. + + Args: + lines: An array of strings, each representing a line of the file, with the + last element being empty if the file is terminated with a newline. + """ + for line in lines: + if _SEARCH_C_FILE.search(line): + for category in _DEFAULT_C_SUPPRESSED_CATEGORIES: + _global_error_suppressions[category] = True + if _SEARCH_KERNEL_FILE.search(line): + for category in _DEFAULT_KERNEL_SUPPRESSED_CATEGORIES: + _global_error_suppressions[category] = True + + def ResetNolintSuppressions(): - "Resets the set of NOLINT suppressions to empty." + """Resets the set of NOLINT suppressions to empty.""" _error_suppressions.clear() + _global_error_suppressions.clear() def IsErrorSuppressedByNolint(category, linenum): """Returns true if the specified error category is suppressed on this line. Consults the global error_suppressions map populated by - ParseNolintSuppressions/ResetNolintSuppressions. + ParseNolintSuppressions/ProcessGlobalSuppresions/ResetNolintSuppressions. Args: category: str, the category of the error. linenum: int, the current line number. Returns: - bool, True iff the error should be suppressed due to a NOLINT comment. + bool, True iff the error should be suppressed due to a NOLINT comment or + global suppression. """ - return (linenum in _error_suppressions.get(category, set()) or + return (_global_error_suppressions.get(category, False) or + linenum in _error_suppressions.get(category, set()) or linenum in _error_suppressions.get(None, set())) + def Match(pattern, s): """Matches the string with the pattern, caching the compiled regexp.""" # The regexp compilation caching is inlined in both Match and Search for @@ -543,11 +691,17 @@ def Search(pattern, s): return _regexp_compile_cache[pattern].search(s) -class _IncludeState(dict): +def _IsSourceExtension(s): + """File extension (excluding dot) matches a source file extension.""" + return s in ('c', 'cc', 'cpp', 'cxx') + + +class _IncludeState(object): """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. + include_list contains list of lists of (header, line number) pairs. + It's a lists of lists rather than just one flat list to make it + easier to update across preprocessor boundaries. Call CheckNextIncludeOrder() once for each header in the file, passing in the type constants defined above. Calls in an illegal order will @@ -578,15 +732,42 @@ class _IncludeState(dict): } def __init__(self): - dict.__init__(self) - self.ResetSection() + self.include_list = [[]] + self.ResetSection('') + + def FindHeader(self, header): + """Check if a header has already been included. + + Args: + header: header to check. + Returns: + Line number of previous occurrence, or -1 if the header has not + been seen before. + """ + for section_list in self.include_list: + for f in section_list: + if f[0] == header: + return f[1] + return -1 - def ResetSection(self): + def ResetSection(self, directive): + """Reset section checking for preprocessor directive. + + Args: + directive: preprocessor directive (e.g. "if", "else"). + """ # The name of the current section. self._section = self._INITIAL_SECTION # The path of last found header. self._last_header = '' + # Update list of includes. Note that we never pop from the + # include list. + if directive in ('if', 'ifdef', 'ifndef'): + self.include_list.append([]) + elif directive in ('else', 'elif'): + self.include_list[-1] = [] + def SetLastHeader(self, header_path): self._last_header = header_path @@ -622,7 +803,7 @@ class _IncludeState(dict): # If previous line was a blank line, assume that the headers are # intentionally sorted the way they are. if (self._last_header > header_path and - not Match(r'^\s*$', clean_lines.elided[linenum - 1])): + Match(r'^\s*#\s*include\b', clean_lines.elided[linenum - 1])): return False return True @@ -688,8 +869,11 @@ class _CppLintState(object): self.error_count = 0 # global count of reported errors # filters to apply when emitting error messages self.filters = _DEFAULT_FILTERS[:] + # backup of filter list. Used to restore the state after each file. + self._filters_backup = self.filters[:] self.counting = 'total' # In what way are we counting errors? self.errors_by_category = {} # string to int dict storing error counts + self.quiet = False # Suppress non-error messagess? # output format: # "emacs" - format that emacs can parse (default) @@ -700,6 +884,12 @@ class _CppLintState(object): """Sets the output format for errors.""" self.output_format = output_format + def SetQuiet(self, quiet): + """Sets the module's quiet settings, and returns the previous setting.""" + last_quiet = self.quiet + self.quiet = quiet + return last_quiet + def SetVerboseLevel(self, level): """Sets the module's verbosity, and returns the previous setting.""" last_verbose_level = self.verbose_level @@ -726,6 +916,10 @@ class _CppLintState(object): """ # Default filters always have less priority than the flag ones. self.filters = _DEFAULT_FILTERS[:] + self.AddFilters(filters) + + def AddFilters(self, filters): + """ Adds more filters to the existing list of error-message filters. """ for filt in filters.split(','): clean_filt = filt.strip() if clean_filt: @@ -735,6 +929,14 @@ class _CppLintState(object): raise ValueError('Every filter in --filters must start with + or -' ' (%s does not)' % filt) + def BackupFilters(self): + """ Saves the current filter list to backup storage.""" + self._filters_backup = self.filters[:] + + def RestoreFilters(self): + """ Restores filters previously backed up.""" + self.filters = self._filters_backup[:] + def ResetErrorCounts(self): """Sets the module's error statistic back to zero.""" self.error_count = 0 @@ -755,7 +957,7 @@ class _CppLintState(object): for category, count in self.errors_by_category.iteritems(): sys.stderr.write('Category \'%s\' errors found: %d\n' % (category, count)) - sys.stderr.write('Total errors found: %d\n' % self.error_count) + sys.stdout.write('Total errors found: %d\n' % self.error_count) _cpplint_state = _CppLintState() @@ -769,6 +971,14 @@ def _SetOutputFormat(output_format): """Sets the module's output format.""" _cpplint_state.SetOutputFormat(output_format) +def _Quiet(): + """Return's the module's quiet setting.""" + return _cpplint_state.quiet + +def _SetQuiet(quiet): + """Set the module's quiet status, and return previous setting.""" + return _cpplint_state.SetQuiet(quiet) + def _VerboseLevel(): """Returns the module's verbosity setting.""" @@ -802,6 +1012,25 @@ def _SetFilters(filters): """ _cpplint_state.SetFilters(filters) +def _AddFilters(filters): + """Adds more filter overrides. + + Unlike _SetFilters, this function does not reset the current list of filters + available. + + Args: + filters: A string of comma-separated filters (eg "whitespace/indent"). + Each filter should start with + or -; else we die. + """ + _cpplint_state.AddFilters(filters) + +def _BackupFilters(): + """ Saves the current filter list to backup storage.""" + _cpplint_state.BackupFilters() + +def _RestoreFilters(): + """ Restores filters previously backed up.""" + _cpplint_state.RestoreFilters() class _FunctionState(object): """Tracks current function name and the number of lines in its body.""" @@ -837,6 +1066,9 @@ class _FunctionState(object): filename: The name of the current file. linenum: The number of the line to check. """ + if not self.in_a_function: + return + if Match(r'T(EST|est)', self.current_function): base_trigger = self._TEST_TRIGGER else: @@ -864,7 +1096,7 @@ class _IncludeError(Exception): pass -class FileInfo: +class FileInfo(object): """Provides utility functions for filenames. FileInfo provides easy access to the components of a file's path @@ -907,12 +1139,13 @@ class FileInfo: # Not SVN <= 1.6? Try to find a git, hg, or svn 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")) and - not os.path.exists(os.path.join(root_dir, ".hg")) and - not os.path.exists(os.path.join(root_dir, ".svn"))): - root_dir = os.path.dirname(root_dir) + root_dir = current_dir = os.path.dirname(fullname) + while current_dir != os.path.dirname(current_dir): + if (os.path.exists(os.path.join(current_dir, ".git")) or + os.path.exists(os.path.join(current_dir, ".hg")) or + os.path.exists(os.path.join(current_dir, ".svn"))): + root_dir = current_dir + current_dir = os.path.dirname(current_dir) if (os.path.exists(os.path.join(root_dir, ".git")) or os.path.exists(os.path.join(root_dir, ".hg")) or @@ -951,7 +1184,7 @@ class FileInfo: def IsSource(self): """File has a source file extension.""" - return self.Extension()[1:] in ('c', 'cc', 'cpp', 'cxx') + return _IsSourceExtension(self.Extension()[1:]) def _ShouldPrintError(category, confidence, linenum): @@ -962,6 +1195,7 @@ def _ShouldPrintError(category, confidence, linenum): # the verbosity level isn't high enough, or the filters filter it out. if IsErrorSuppressedByNolint(category, linenum): return False + if confidence < _cpplint_state.verbose_level: return False @@ -1006,8 +1240,8 @@ def Error(filename, linenum, category, confidence, message): if _ShouldPrintError(category, confidence, linenum): _cpplint_state.IncrementErrorCount(category) if _cpplint_state.output_format == 'vs7': - sys.stderr.write('%s(%s): %s [%s] [%d]\n' % ( - filename, linenum, message, category, confidence)) + sys.stderr.write('%s(%s): error cpplint: [%s] %s [%d]\n' % ( + filename, linenum, category, message, confidence)) elif _cpplint_state.output_format == 'eclipse': sys.stderr.write('%s:%s: warning: %s [%s] [%d]\n' % ( filename, linenum, message, category, confidence)) @@ -1019,11 +1253,9 @@ def Error(filename, linenum, category, confidence, message): # Matches standard C++ escape sequences 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. +# Match a single C style comment on the same line. +_RE_PATTERN_C_COMMENTS = r'/\*(?:[^*]|\*(?!/))*\*/' +# Matches multi-line C style 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. @@ -1032,10 +1264,10 @@ _RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES = re.compile(r"'.'") # 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) + r'(\s*' + _RE_PATTERN_C_COMMENTS + r'\s*$|' + + _RE_PATTERN_C_COMMENTS + r'\s+|' + + r'\s+' + _RE_PATTERN_C_COMMENTS + r'(?=\W)|' + + _RE_PATTERN_C_COMMENTS + r')') def IsCppString(line): @@ -1090,13 +1322,26 @@ def CleanseRawStrings(raw_lines): delimiter = None else: # Haven't found the end yet, append a blank line. - line = '' + line = '""' - else: + # Look for beginning of a raw string, and replace them with + # empty strings. This is done in a loop to handle multiple raw + # strings on the same line. + while delimiter is None: # Look for beginning of a raw string. # See 2.14.15 [lex.string] for syntax. - matched = Match(r'^(.*)\b(?:R|u8R|uR|UR|LR)"([^\s\\()]*)\((.*)$', line) - if matched: + # + # Once we have matched a raw string, we check the prefix of the + # line to make sure that the line is not part of a single line + # comment. It's done this way because we remove raw strings + # before removing comments as opposed to removing comments + # before removing raw strings. This is because there are some + # cpplint checks that requires the comments to be preserved, but + # we don't want to check comments that are inside raw strings. + matched = Match(r'^(.*?)\b(?:R|u8R|uR|UR|LR)"([^\s\\()]*)\((.*)$', line) + if (matched and + not Match(r'^([^\'"]|\'(\\.|[^\'])*\'|"(\\.|[^"])*")*//', + matched.group(1))): delimiter = ')' + matched.group(2) + '"' end = matched.group(3).find(delimiter) @@ -1108,6 +1353,8 @@ def CleanseRawStrings(raw_lines): else: # Start of a multi-line raw string line = matched.group(1) + '""' + else: + break lines_without_raw_strings.append(line) @@ -1141,7 +1388,7 @@ def RemoveMultiLineCommentsFromRange(lines, begin, end): # 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' + lines[i] = '/**/' def RemoveMultiLineComments(filename, lines, error): @@ -1177,12 +1424,14 @@ def CleanseComments(line): class CleansedLines(object): - """Holds 3 copies of all lines with different preprocessing applied to them. + """Holds 4 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 + 1) elided member contains lines without strings and comments. + 2) lines member contains lines without comments. 3) raw_lines member contains all the lines without processing. - All these three members are of <type 'list'>, and of the same length. + 4) lines_without_raw_strings member is same as raw_lines, but with C++11 raw + strings removed. + All these members are of <type 'list'>, and of the same length. """ def __init__(self, lines): @@ -1213,38 +1462,138 @@ class CleansedLines(object): 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 + if _RE_PATTERN_INCLUDE.match(elided): + return 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) + + # Replace quoted strings and digit separators. Both single quotes + # and double quotes are processed in the same loop, otherwise + # nested quotes wouldn't work. + collapsed = '' + while True: + # Find the first quote character + match = Match(r'^([^\'"]*)([\'"])(.*)$', elided) + if not match: + collapsed += elided + break + head, quote, tail = match.groups() + + if quote == '"': + # Collapse double quoted strings + second_quote = tail.find('"') + if second_quote >= 0: + collapsed += head + '""' + elided = tail[second_quote + 1:] + else: + # Unmatched double quote, don't bother processing the rest + # of the line since this is probably a multiline string. + collapsed += elided + break + else: + # Found single quote, check nearby text to eliminate digit separators. + # + # There is no special handling for floating point here, because + # the integer/fractional/exponent parts would all be parsed + # correctly as long as there are digits on both sides of the + # separator. So we are fine as long as we don't see something + # like "0.'3" (gcc 4.9.0 will not allow this literal). + if Search(r'\b(?:0[bBxX]?|[1-9])[0-9a-fA-F]*$', head): + match_literal = Match(r'^((?:\'?[0-9a-zA-Z_])*)(.*)$', "'" + tail) + collapsed += head + match_literal.group(1).replace("'", '') + elided = match_literal.group(2) + else: + second_quote = tail.find('\'') + if second_quote >= 0: + collapsed += head + "''" + elided = tail[second_quote + 1:] + else: + # Unmatched single quote + collapsed += elided + break + + return collapsed -def FindEndOfExpressionInLine(line, startpos, depth, startchar, endchar): - """Find the position just after the matching endchar. +def FindEndOfExpressionInLine(line, startpos, stack): + """Find the position just after the end of current parenthesized expression. Args: line: a CleansedLines line. startpos: start searching at this position. - depth: nesting level at startpos. - startchar: expression opening character. - endchar: expression closing character. + stack: nesting stack at startpos. Returns: - On finding matching endchar: (index just after matching endchar, 0) - Otherwise: (-1, new depth at end of this line) + On finding matching end: (index just after matching end, None) + On finding an unclosed expression: (-1, None) + Otherwise: (-1, new stack at end of this line) """ for i in xrange(startpos, len(line)): - if line[i] == startchar: - depth += 1 - elif line[i] == endchar: - depth -= 1 - if depth == 0: - return (i + 1, 0) - return (-1, depth) + char = line[i] + if char in '([{': + # Found start of parenthesized expression, push to expression stack + stack.append(char) + elif char == '<': + # Found potential start of template argument list + if i > 0 and line[i - 1] == '<': + # Left shift operator + if stack and stack[-1] == '<': + stack.pop() + if not stack: + return (-1, None) + elif i > 0 and Search(r'\boperator\s*$', line[0:i]): + # operator<, don't add to stack + continue + else: + # Tentative start of template argument list + stack.append('<') + elif char in ')]}': + # Found end of parenthesized expression. + # + # If we are currently expecting a matching '>', the pending '<' + # must have been an operator. Remove them from expression stack. + while stack and stack[-1] == '<': + stack.pop() + if not stack: + return (-1, None) + if ((stack[-1] == '(' and char == ')') or + (stack[-1] == '[' and char == ']') or + (stack[-1] == '{' and char == '}')): + stack.pop() + if not stack: + return (i + 1, None) + else: + # Mismatched parentheses + return (-1, None) + elif char == '>': + # Found potential end of template argument list. + + # Ignore "->" and operator functions + if (i > 0 and + (line[i - 1] == '-' or Search(r'\boperator\s*$', line[0:i - 1]))): + continue + + # Pop the stack if there is a matching '<'. Otherwise, ignore + # this '>' since it must be an operator. + if stack: + if stack[-1] == '<': + stack.pop() + if not stack: + return (i + 1, None) + elif char == ';': + # Found something that look like end of statements. If we are currently + # expecting a '>', the matching '<' must have been an operator, since + # template argument list should not contain statements. + while stack and stack[-1] == '<': + stack.pop() + if not stack: + return (-1, None) + + # Did not find end of expression or unbalanced parentheses on this line + return (-1, stack) def CloseExpression(clean_lines, linenum, pos): @@ -1253,6 +1602,11 @@ def CloseExpression(clean_lines, linenum, pos): If lines[linenum][pos] points to a '(' or '{' or '[' or '<', finds the linenum/pos that correspond to the closing of the expression. + TODO(unknown): cpplint spends a fair bit of time matching parentheses. + Ideally we would want to index all opening and closing parentheses once + and have CloseExpression be just a simple lookup, but due to preprocessor + tricks, this is not so easy. + Args: clean_lines: A CleansedLines instance containing the file. linenum: The number of the line to check. @@ -1266,35 +1620,28 @@ def CloseExpression(clean_lines, linenum, pos): """ line = clean_lines.elided[linenum] - startchar = line[pos] - if startchar not in '({[<': + if (line[pos] not in '({[<') or Match(r'<[<=]', line[pos:]): return (line, clean_lines.NumLines(), -1) - if startchar == '(': endchar = ')' - if startchar == '[': endchar = ']' - if startchar == '{': endchar = '}' - if startchar == '<': endchar = '>' # Check first line - (end_pos, num_open) = FindEndOfExpressionInLine( - line, pos, 0, startchar, endchar) + (end_pos, stack) = FindEndOfExpressionInLine(line, pos, []) if end_pos > -1: return (line, linenum, end_pos) # Continue scanning forward - while linenum < clean_lines.NumLines() - 1: + while stack and linenum < clean_lines.NumLines() - 1: linenum += 1 line = clean_lines.elided[linenum] - (end_pos, num_open) = FindEndOfExpressionInLine( - line, 0, num_open, startchar, endchar) + (end_pos, stack) = FindEndOfExpressionInLine(line, 0, stack) if end_pos > -1: return (line, linenum, end_pos) - # Did not find endchar before end of file, give up + # Did not find end of expression before end of file, give up return (line, clean_lines.NumLines(), -1) -def FindStartOfExpressionInLine(line, endpos, depth, startchar, endchar): - """Find position at the matching startchar. +def FindStartOfExpressionInLine(line, endpos, stack): + """Find position at the matching start of current expression. This is almost the reverse of FindEndOfExpressionInLine, but note that the input position and returned position differs by 1. @@ -1302,22 +1649,72 @@ def FindStartOfExpressionInLine(line, endpos, depth, startchar, endchar): Args: line: a CleansedLines line. endpos: start searching at this position. - depth: nesting level at endpos. - startchar: expression opening character. - endchar: expression closing character. + stack: nesting stack at endpos. Returns: - On finding matching startchar: (index at matching startchar, 0) - Otherwise: (-1, new depth at beginning of this line) + On finding matching start: (index at matching start, None) + On finding an unclosed expression: (-1, None) + Otherwise: (-1, new stack at beginning of this line) """ - for i in xrange(endpos, -1, -1): - if line[i] == endchar: - depth += 1 - elif line[i] == startchar: - depth -= 1 - if depth == 0: - return (i, 0) - return (-1, depth) + i = endpos + while i >= 0: + char = line[i] + if char in ')]}': + # Found end of expression, push to expression stack + stack.append(char) + elif char == '>': + # Found potential end of template argument list. + # + # Ignore it if it's a "->" or ">=" or "operator>" + if (i > 0 and + (line[i - 1] == '-' or + Match(r'\s>=\s', line[i - 1:]) or + Search(r'\boperator\s*$', line[0:i]))): + i -= 1 + else: + stack.append('>') + elif char == '<': + # Found potential start of template argument list + if i > 0 and line[i - 1] == '<': + # Left shift operator + i -= 1 + else: + # If there is a matching '>', we can pop the expression stack. + # Otherwise, ignore this '<' since it must be an operator. + if stack and stack[-1] == '>': + stack.pop() + if not stack: + return (i, None) + elif char in '([{': + # Found start of expression. + # + # If there are any unmatched '>' on the stack, they must be + # operators. Remove those. + while stack and stack[-1] == '>': + stack.pop() + if not stack: + return (-1, None) + if ((char == '(' and stack[-1] == ')') or + (char == '[' and stack[-1] == ']') or + (char == '{' and stack[-1] == '}')): + stack.pop() + if not stack: + return (i, None) + else: + # Mismatched parentheses + return (-1, None) + elif char == ';': + # Found something that look like end of statements. If we are currently + # expecting a '<', the matching '>' must have been an operator, since + # template argument list should not contain statements. + while stack and stack[-1] == '>': + stack.pop() + if not stack: + return (-1, None) + + i -= 1 + + return (-1, stack) def ReverseCloseExpression(clean_lines, linenum, pos): @@ -1338,30 +1735,23 @@ def ReverseCloseExpression(clean_lines, linenum, pos): return is the 'cleansed' line at linenum. """ line = clean_lines.elided[linenum] - endchar = line[pos] - if endchar not in ')}]>': + if line[pos] not in ')}]>': return (line, 0, -1) - if endchar == ')': startchar = '(' - if endchar == ']': startchar = '[' - if endchar == '}': startchar = '{' - if endchar == '>': startchar = '<' # Check last line - (start_pos, num_open) = FindStartOfExpressionInLine( - line, pos, 0, startchar, endchar) + (start_pos, stack) = FindStartOfExpressionInLine(line, pos, []) if start_pos > -1: return (line, linenum, start_pos) # Continue scanning backward - while linenum > 0: + while stack and linenum > 0: linenum -= 1 line = clean_lines.elided[linenum] - (start_pos, num_open) = FindStartOfExpressionInLine( - line, len(line) - 1, num_open, startchar, endchar) + (start_pos, stack) = FindStartOfExpressionInLine(line, len(line) - 1, stack) if start_pos > -1: return (line, linenum, start_pos) - # Did not find startchar before beginning of file, give up + # Did not find start of expression before beginning of file, give up return (line, 0, -1) @@ -1378,6 +1768,46 @@ def CheckForCopyright(filename, lines, error): 'You should have a line: "Copyright [year] <Copyright Owner>"') +def GetIndentLevel(line): + """Return the number of leading spaces in line. + + Args: + line: A string to check. + + Returns: + An integer count of leading spaces, possibly zero. + """ + indent = Match(r'^( *)\S', line) + if indent: + return len(indent.group(1)) + else: + return 0 + +def PathSplitToList(path): + """Returns the path split into a list by the separator. + + Args: + path: An absolute or relative path (e.g. '/a/b/c/' or '../a') + + Returns: + A list of path components (e.g. ['a', 'b', 'c]). + """ + lst = [] + while True: + (head, tail) = os.path.split(path) + if head == path: # absolute paths end + lst.append(head) + break + if tail == path: # relative paths end + lst.append(tail) + break + + path = head + lst.append(tail) + + lst.reverse() + return lst + def GetHeaderGuardCPPVariable(filename): """Returns the CPP variable that should be used as a header guard. @@ -1394,15 +1824,67 @@ def GetHeaderGuardCPPVariable(filename): # flymake. filename = re.sub(r'_flymake\.h$', '.h', filename) filename = re.sub(r'/\.flymake/([^/]*)$', r'/\1', filename) + # Replace 'c++' with 'cpp'. + filename = filename.replace('C++', 'cpp').replace('c++', 'cpp') fileinfo = FileInfo(filename) file_path_from_root = fileinfo.RepositoryName() - if _root: - file_path_from_root = re.sub('^' + _root + os.sep, '', file_path_from_root) - return re.sub(r'[-./\s]', '_', file_path_from_root).upper() + '_' + def FixupPathFromRoot(): + if _root_debug: + sys.stderr.write("\n_root fixup, _root = '%s', repository name = '%s'\n" + %(_root, fileinfo.RepositoryName())) + + # Process the file path with the --root flag if it was set. + if not _root: + if _root_debug: + sys.stderr.write("_root unspecified\n") + return file_path_from_root + + def StripListPrefix(lst, prefix): + # f(['x', 'y'], ['w, z']) -> None (not a valid prefix) + if lst[:len(prefix)] != prefix: + return None + # f(['a, 'b', 'c', 'd'], ['a', 'b']) -> ['c', 'd'] + return lst[(len(prefix)):] + + # root behavior: + # --root=subdir , lstrips subdir from the header guard + maybe_path = StripListPrefix(PathSplitToList(file_path_from_root), + PathSplitToList(_root)) + + if _root_debug: + sys.stderr.write("_root lstrip (maybe_path=%s, file_path_from_root=%s," + + " _root=%s)\n" %(maybe_path, file_path_from_root, _root)) + + if maybe_path: + return os.path.join(*maybe_path) + + # --root=.. , will prepend the outer directory to the header guard + full_path = fileinfo.FullName() + root_abspath = os.path.abspath(_root) + + maybe_path = StripListPrefix(PathSplitToList(full_path), + PathSplitToList(root_abspath)) + + if _root_debug: + sys.stderr.write("_root prepend (maybe_path=%s, full_path=%s, " + + "root_abspath=%s)\n" %(maybe_path, full_path, root_abspath)) + + if maybe_path: + return os.path.join(*maybe_path) + + if _root_debug: + sys.stderr.write("_root ignore, returning %s\n" %(file_path_from_root)) -def CheckForHeaderGuard(filename, lines, error): + # --root=FAKE_DIR is ignored + return file_path_from_root + + file_path_from_root = FixupPathFromRoot() + return re.sub(r'[^a-zA-Z0-9]', '_', file_path_from_root).upper() + '_' + + +def CheckForHeaderGuard(filename, clean_lines, error): """Checks that the file contains a header guard. Logs an error if no #ifndef header guard is present. For other @@ -1410,18 +1892,29 @@ def CheckForHeaderGuard(filename, lines, error): Args: filename: The name of the C++ header file. - lines: An array of strings, each representing a line of the file. + clean_lines: A CleansedLines instance containing the file. error: The function to call with any errors found. """ + # Don't check for header guards if there are error suppression + # comments somewhere in this file. + # + # Because this is silencing a warning for a nonexistent line, we + # only support the very specific NOLINT(build/header_guard) syntax, + # and not the general NOLINT or NOLINT(*) syntax. + raw_lines = clean_lines.lines_without_raw_strings + for i in raw_lines: + if Search(r'//\s*NOLINT\(build/header_guard\)', i): + return + cppvar = GetHeaderGuardCPPVariable(filename) - ifndef = None + ifndef = '' ifndef_linenum = 0 - define = None - endif = None + define = '' + endif = '' endif_linenum = 0 - for linenum, line in enumerate(lines): + for linenum, line in enumerate(raw_lines): linesplit = line.split() if len(linesplit) >= 2: # find the first occurrence of #ifndef and #define, save arg @@ -1436,18 +1929,12 @@ def CheckForHeaderGuard(filename, lines, error): endif = line endif_linenum = linenum - if not ifndef: + if not ifndef or not define or ifndef != define: error(filename, 0, 'build/header_guard', 5, 'No #ifndef header guard found, suggested CPP variable is: %s' % cppvar) return - if not define: - error(filename, 0, 'build/header_guard', 5, - 'No #define header guard found, suggested CPP variable is: %s' % - cppvar) - return - # The guard should be PATH_FILE_H_, but we also allow PATH_FILE_H__ # for backward compatibility. if ifndef != cppvar: @@ -1455,26 +1942,69 @@ def CheckForHeaderGuard(filename, lines, error): if ifndef != cppvar + '_': error_level = 5 - ParseNolintSuppressions(filename, lines[ifndef_linenum], ifndef_linenum, + ParseNolintSuppressions(filename, raw_lines[ifndef_linenum], ifndef_linenum, error) error(filename, ifndef_linenum, 'build/header_guard', error_level, '#ifndef header guard has wrong style, please use: %s' % cppvar) - if define != ifndef: - error(filename, 0, 'build/header_guard', 5, - '#ifndef and #define don\'t match, suggested CPP variable is: %s' % - cppvar) + # Check for "//" comments on endif line. + ParseNolintSuppressions(filename, raw_lines[endif_linenum], endif_linenum, + error) + match = Match(r'#endif\s*//\s*' + cppvar + r'(_)?\b', endif) + if match: + if match.group(1) == '_': + # Issue low severity warning for deprecated double trailing underscore + error(filename, endif_linenum, 'build/header_guard', 0, + '#endif line should be "#endif // %s"' % cppvar) return - if endif != ('#endif // %s' % cppvar): - error_level = 0 - if endif != ('#endif // %s' % (cppvar + '_')): - error_level = 5 + # Didn't find the corresponding "//" comment. If this file does not + # contain any "//" comments at all, it could be that the compiler + # only wants "/**/" comments, look for those instead. + no_single_line_comments = True + for i in xrange(1, len(raw_lines) - 1): + line = raw_lines[i] + if Match(r'^(?:(?:\'(?:\.|[^\'])*\')|(?:"(?:\.|[^"])*")|[^\'"])*//', line): + no_single_line_comments = False + break - ParseNolintSuppressions(filename, lines[endif_linenum], endif_linenum, - error) - error(filename, endif_linenum, 'build/header_guard', error_level, - '#endif line should be "#endif // %s"' % cppvar) + if no_single_line_comments: + match = Match(r'#endif\s*/\*\s*' + cppvar + r'(_)?\s*\*/', endif) + if match: + if match.group(1) == '_': + # Low severity warning for double trailing underscore + error(filename, endif_linenum, 'build/header_guard', 0, + '#endif line should be "#endif /* %s */"' % cppvar) + return + + # Didn't find anything + error(filename, endif_linenum, 'build/header_guard', 5, + '#endif line should be "#endif // %s"' % cppvar) + + +def CheckHeaderFileIncluded(filename, include_state, error): + """Logs an error if a .cc file does not include its header.""" + + # Do not check test files + fileinfo = FileInfo(filename) + if Search(_TEST_FILE_SUFFIX, fileinfo.BaseName()): + return + + headerfile = filename[0:len(filename) - len(fileinfo.Extension())] + '.h' + if not os.path.exists(headerfile): + return + headername = FileInfo(headerfile).RepositoryName() + first_include = 0 + for section_list in include_state.include_list: + for f in section_list: + if headername in f[0] or f[0] in headername: + return + if not first_include: + first_include = f[1] + + error(filename, first_include, 'build/include', 5, + '%s should include its header file %s' % (fileinfo.RepositoryName(), + headername)) def CheckForBadCharacters(filename, lines, error): @@ -1558,19 +2088,33 @@ def CheckForMultilineCommentsAndStrings(filename, clean_lines, linenum, error): 'Use C++11 raw strings or 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('), - ('strtok(', 'strtok_r('), - ('ttyname(', 'ttyname_r('), +# (non-threadsafe name, thread-safe alternative, validation pattern) +# +# The validation pattern is used to eliminate false positives such as: +# _rand(); // false positive due to substring match. +# ->rand(); // some member function rand(). +# ACMRandom rand(seed); // some variable named rand. +# ISAACRandom rand(); // another variable named rand. +# +# Basically we require the return value of these functions to be used +# in some expression context on the same line by matching on some +# operator before the function name. This eliminates constructors and +# member function calls. +_UNSAFE_FUNC_PREFIX = r'(?:[-+*/=%^&|(<]\s*|>\s+)' +_THREADING_LIST = ( + ('asctime(', 'asctime_r(', _UNSAFE_FUNC_PREFIX + r'asctime\([^)]+\)'), + ('ctime(', 'ctime_r(', _UNSAFE_FUNC_PREFIX + r'ctime\([^)]+\)'), + ('getgrgid(', 'getgrgid_r(', _UNSAFE_FUNC_PREFIX + r'getgrgid\([^)]+\)'), + ('getgrnam(', 'getgrnam_r(', _UNSAFE_FUNC_PREFIX + r'getgrnam\([^)]+\)'), + ('getlogin(', 'getlogin_r(', _UNSAFE_FUNC_PREFIX + r'getlogin\(\)'), + ('getpwnam(', 'getpwnam_r(', _UNSAFE_FUNC_PREFIX + r'getpwnam\([^)]+\)'), + ('getpwuid(', 'getpwuid_r(', _UNSAFE_FUNC_PREFIX + r'getpwuid\([^)]+\)'), + ('gmtime(', 'gmtime_r(', _UNSAFE_FUNC_PREFIX + r'gmtime\([^)]+\)'), + ('localtime(', 'localtime_r(', _UNSAFE_FUNC_PREFIX + r'localtime\([^)]+\)'), + ('rand(', 'rand_r(', _UNSAFE_FUNC_PREFIX + r'rand\(\)'), + ('strtok(', 'strtok_r(', + _UNSAFE_FUNC_PREFIX + r'strtok\([^)]+\)'), + ('ttyname(', 'ttyname_r(', _UNSAFE_FUNC_PREFIX + r'ttyname\([^)]+\)'), ) @@ -1590,14 +2134,13 @@ def CheckPosixThreading(filename, clean_lines, linenum, error): error: The function to call with any errors found. """ line = clean_lines.elided[linenum] - for single_thread_function, multithread_safe_function in threading_list: - ix = line.find(single_thread_function) - # Comparisons made explicit for clarity -- pylint: disable=g-explicit-bool-comparison - if ix >= 0 and (ix == 0 or (not line[ix - 1].isalnum() and - line[ix - 1] not in ('_', '.', '>'))): + for single_thread_func, multithread_safe_func, pattern in _THREADING_LIST: + # Additional pattern matching check to confirm that this is the + # function we are looking for + if Search(pattern, line): error(filename, linenum, 'runtime/threadsafe_fn', 2, - 'Consider using ' + multithread_safe_function + - '...) instead of ' + single_thread_function + + 'Consider using ' + multithread_safe_func + + '...) instead of ' + single_thread_func + '...) for improved thread safety.') @@ -1619,7 +2162,6 @@ def CheckVlogArguments(filename, clean_lines, linenum, error): 'VLOG() should be used with numeric verbosity level. ' 'Use LOG() if you want symbolic severity levels.') - # Matches invalid increment: *count++, which moves pointer instead of # incrementing a value. _RE_PATTERN_INVALID_INCREMENT = re.compile( @@ -1648,13 +2190,29 @@ def CheckInvalidIncrement(filename, clean_lines, linenum, error): 'Changing pointer instead of value (or unused value of operator*).') +def IsMacroDefinition(clean_lines, linenum): + if Search(r'^#define', clean_lines[linenum]): + return True + + if linenum > 0 and Search(r'\\$', clean_lines[linenum - 1]): + return True + + return False + + +def IsForwardClassDeclaration(clean_lines, linenum): + return Match(r'^\s*(\btemplate\b)*.*class\s+\w+;\s*$', clean_lines[linenum]) + + class _BlockInfo(object): """Stores information about a generic block of code.""" - def __init__(self, seen_open_brace): + def __init__(self, linenum, seen_open_brace): + self.starting_linenum = linenum self.seen_open_brace = seen_open_brace self.open_parentheses = 0 self.inline_asm = _NO_ASM + self.check_namespace_indentation = False def CheckBegin(self, filename, clean_lines, linenum, error): """Run checks that applies to text up to the opening brace. @@ -1684,15 +2242,33 @@ class _BlockInfo(object): """ pass + def IsBlockInfo(self): + """Returns true if this block is a _BlockInfo. + + This is convenient for verifying that an object is an instance of + a _BlockInfo, but not an instance of any of the derived classes. + + Returns: + True for this class, False for derived classes. + """ + return self.__class__ == _BlockInfo + + +class _ExternCInfo(_BlockInfo): + """Stores information about an 'extern "C"' block.""" + + def __init__(self, linenum): + _BlockInfo.__init__(self, linenum, True) + class _ClassInfo(_BlockInfo): """Stores information about a class.""" def __init__(self, name, class_or_struct, clean_lines, linenum): - _BlockInfo.__init__(self, False) + _BlockInfo.__init__(self, linenum, False) self.name = name - self.starting_linenum = linenum self.is_derived = False + self.check_namespace_indentation = True if class_or_struct == 'struct': self.access = 'public' self.is_struct = True @@ -1702,11 +2278,7 @@ class _ClassInfo(_BlockInfo): # Remember initial indentation level for this class. Using raw_lines here # instead of elided to account for leading comments. - initial_indent = Match(r'^( *)\S', clean_lines.raw_lines[linenum]) - if initial_indent: - self.class_indent = len(initial_indent.group(1)) - else: - self.class_indent = 0 + self.class_indent = GetIndentLevel(clean_lines.raw_lines[linenum]) # Try to find the end of the class. This will be confused by things like: # class A { @@ -1728,6 +2300,23 @@ class _ClassInfo(_BlockInfo): self.is_derived = True def CheckEnd(self, filename, clean_lines, linenum, error): + # If there is a DISALLOW macro, it should appear near the end of + # the class. + seen_last_thing_in_class = False + for i in xrange(linenum - 1, self.starting_linenum, -1): + match = Search( + r'\b(DISALLOW_COPY_AND_ASSIGN|DISALLOW_IMPLICIT_CONSTRUCTORS)\(' + + self.name + r'\)', + clean_lines.elided[i]) + if match: + if seen_last_thing_in_class: + error(filename, i, 'readability/constructors', 3, + match.group(1) + ' should be the last thing in the class') + break + + if not Match(r'^\s*$', clean_lines.elided[i]): + seen_last_thing_in_class = True + # Check that closing brace is aligned with beginning of the class. # Only do this if the closing brace is indented by only whitespaces. # This means we will not check single-line class definitions. @@ -1745,9 +2334,9 @@ class _NamespaceInfo(_BlockInfo): """Stores information about a namespace.""" def __init__(self, name, linenum): - _BlockInfo.__init__(self, False) + _BlockInfo.__init__(self, linenum, False) self.name = name or '' - self.starting_linenum = linenum + self.check_namespace_indentation = True def CheckEnd(self, filename, clean_lines, linenum, error): """Check end of namespace comments.""" @@ -1765,7 +2354,7 @@ class _NamespaceInfo(_BlockInfo): # deciding what these nontrivial things are, so this check is # triggered by namespace size only, which works most of the time. if (linenum - self.starting_linenum < 10 - and not Match(r'};*\s*(//|/\*).*\bnamespace\b', line)): + and not Match(r'^\s*};*\s*(//|/\*).*\bnamespace\b', line)): return # Look for matching comment at end of namespace. @@ -1782,17 +2371,24 @@ class _NamespaceInfo(_BlockInfo): # expected namespace. if self.name: # Named namespace - if not Match((r'};*\s*(//|/\*).*\bnamespace\s+' + re.escape(self.name) + - r'[\*/\.\\\s]* {$'), + if not Match((r'^\s*};*\s*(//|/\*).*\bnamespace\s+' + + re.escape(self.name) + r'[\*/\.\\\s]* {$'), line): error(filename, linenum, 'readability/namespace', 5, 'Namespace should be terminated with "// namespace %s {"' % self.name) else: # Anonymous namespace - if not Match(r'};*\s*(//|/\*).*\bnamespace[\*/\.\\\s]* {$', line): - error(filename, linenum, 'readability/namespace', 5, - 'Namespace should be terminated with "// namespace {"') + if not Match(r'^\s*};*\s*(//|/\*).*\bnamespace[\*/\.\\\s]* {$', line): + # If "// namespace anonymous" or "// anonymous namespace (more text)", + # mention "// anonymous namespace" as an acceptable form + if Match(r'^\s*}.*\b(namespace anonymous|anonymous namespace)\b', line): + error(filename, linenum, 'readability/namespace', 5, + 'Anonymous namespace should be terminated with "// namespace"' + ' or "// anonymous namespace"') + else: + error(filename, linenum, 'readability/namespace', 5, + 'Anonymous namespace should be terminated with "// namespace {"') class _PreprocessorInfo(object): @@ -1809,7 +2405,7 @@ class _PreprocessorInfo(object): self.seen_else = False -class _NestingState(object): +class NestingState(object): """Holds states related to parsing braces.""" def __init__(self): @@ -1821,6 +2417,17 @@ class _NestingState(object): # - _BlockInfo: some other type of block. self.stack = [] + # Top of the previous stack before each Update(). + # + # Because the nesting_stack is updated at the end of each line, we + # had to do some convoluted checks to find out what is the current + # scope at the beginning of the line. This check is simplified by + # saving the previous top of nesting stack. + # + # We could save the full stack, but we only need the top. Copying + # the full nesting stack would slow down cpplint by ~10%. + self.previous_stack_top = [] + # Stack of _PreprocessorInfo objects. self.pp_stack = [] @@ -1841,6 +2448,82 @@ class _NestingState(object): """ return self.stack and isinstance(self.stack[-1], _NamespaceInfo) + def InExternC(self): + """Check if we are currently one level inside an 'extern "C"' block. + + Returns: + True if top of the stack is an extern block, False otherwise. + """ + return self.stack and isinstance(self.stack[-1], _ExternCInfo) + + def InClassDeclaration(self): + """Check if we are currently one level inside a class or struct declaration. + + Returns: + True if top of the stack is a class/struct, False otherwise. + """ + return self.stack and isinstance(self.stack[-1], _ClassInfo) + + def InAsmBlock(self): + """Check if we are currently one level inside an inline ASM block. + + Returns: + True if the top of the stack is a block containing inline ASM. + """ + return self.stack and self.stack[-1].inline_asm != _NO_ASM + + def InTemplateArgumentList(self, clean_lines, linenum, pos): + """Check if current position is inside template argument list. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + pos: position just after the suspected template argument. + Returns: + True if (linenum, pos) is inside template arguments. + """ + while linenum < clean_lines.NumLines(): + # Find the earliest character that might indicate a template argument + line = clean_lines.elided[linenum] + match = Match(r'^[^{};=\[\]\.<>]*(.)', line[pos:]) + if not match: + linenum += 1 + pos = 0 + continue + token = match.group(1) + pos += len(match.group(0)) + + # These things do not look like template argument list: + # class Suspect { + # class Suspect x; } + if token in ('{', '}', ';'): return False + + # These things look like template argument list: + # template <class Suspect> + # template <class Suspect = default_value> + # template <class Suspect[]> + # template <class Suspect...> + if token in ('>', '=', '[', ']', '.'): return True + + # Check if token is an unmatched '<'. + # If not, move on to the next character. + if token != '<': + pos += 1 + if pos >= len(line): + linenum += 1 + pos = 0 + continue + + # We can't be sure if we just find a single '<', and need to + # find the matching '>'. + (_, end_line, end_pos) = CloseExpression(clean_lines, linenum, pos - 1) + if end_pos < 0: + # Not sure if template argument list or syntax error in file + return False + linenum = end_line + pos = end_pos + return False + def UpdatePreprocessor(self, line): """Update preprocessor stack. @@ -1897,6 +2580,7 @@ class _NestingState(object): # TODO(unknown): unexpected #endif, issue warning? pass + # TODO(unknown): Update() is too long, but we will refactor later. def Update(self, filename, clean_lines, linenum, error): """Update nesting state with current line. @@ -1908,7 +2592,17 @@ class _NestingState(object): """ line = clean_lines.elided[linenum] - # Update pp_stack first + # Remember top of the previous nesting stack. + # + # The stack is always pushed/popped and not modified in place, so + # we can just do a shallow copy instead of copy.deepcopy. Using + # deepcopy would slow down cpplint by ~28%. + if self.stack: + self.previous_stack_top = self.stack[-1] + else: + self.previous_stack_top = None + + # Update pp_stack self.UpdatePreprocessor(line) # Count parentheses. This is to avoid adding struct arguments to @@ -1959,32 +2653,27 @@ class _NestingState(object): # such as in: # class LOCKABLE API Object { # }; - # - # Templates with class arguments may confuse the parser, for example: - # template <class T - # class Comparator = less<T>, - # class Vector = vector<T> > - # class HeapQueue { - # - # Because this parser has no nesting state about templates, by the - # time it saw "class Comparator", it may think that it's a new class. - # Nested templates have a similar problem: - # template < - # typename ExportedType, - # typename TupleType, - # template <typename, typename> class ImplTemplate> - # - # To avoid these cases, we ignore classes that are followed by '=' or '>' class_decl_match = Match( - r'\s*(template\s*<[\w\s<>,:]*>\s*)?' - r'(class|struct)\s+([A-Z_]+\s+)*(\w+(?:::\w+)*)' - r'(([^=>]|<[^<>]*>|<[^<>]*<[^<>]*>\s*>)*)$', line) + r'^(\s*(?:template\s*<[\w\s<>,:]*>\s*)?' + r'(class|struct)\s+(?:[A-Z_]+\s+)*(\w+(?:::\w+)*))' + r'(.*)$', line) if (class_decl_match and (not self.stack or self.stack[-1].open_parentheses == 0)): - self.stack.append(_ClassInfo( - class_decl_match.group(4), class_decl_match.group(2), - clean_lines, linenum)) - line = class_decl_match.group(5) + # We do not want to accept classes that are actually template arguments: + # template <class Ignore1, + # class Ignore2 = Default<Args>, + # template <Args> class Ignore3> + # void Function() {}; + # + # To avoid template argument cases, we scan forward and look for + # an unmatched '>'. If we see one, assume we are inside a + # template argument list. + end_declaration = len(class_decl_match.group(1)) + if not self.InTemplateArgumentList(clean_lines, linenum, end_declaration): + self.stack.append(_ClassInfo( + class_decl_match.group(3), class_decl_match.group(2), + clean_lines, linenum)) + line = class_decl_match.group(4) # If we have not yet seen the opening brace for the innermost block, # run checks here. @@ -2029,10 +2718,13 @@ class _NestingState(object): # stack otherwise. if not self.SeenOpenBrace(): self.stack[-1].seen_open_brace = True + elif Match(r'^extern\s*"[^"]*"\s*\{', line): + self.stack.append(_ExternCInfo(linenum)) else: - self.stack.append(_BlockInfo(True)) + self.stack.append(_BlockInfo(linenum, True)) if _MATCH_ASM.match(line): self.stack[-1].inline_asm = _BLOCK_ASM + elif token == ';' or token == ')': # If we haven't seen an opening brace yet, but we already saw # a semicolon, this is probably a forward declaration. Pop @@ -2108,7 +2800,7 @@ def CheckForNonStandardConstructs(filename, clean_lines, linenum, filename: The name of the current file. clean_lines: A CleansedLines instance containing the file. linenum: The number of the line to check. - nesting_state: A _NestingState instance which maintains information about + nesting_state: A NestingState instance which maintains information about the current stack of nested blocks being parsed. error: A callable to which errors are reported, which takes 4 arguments: filename, line number, error level, and message @@ -2141,7 +2833,8 @@ def CheckForNonStandardConstructs(filename, clean_lines, linenum, r'\s+(register|static|extern|typedef)\b', line): error(filename, linenum, 'build/storage_class', 5, - 'Storage class (static, extern, typedef, etc) should be first.') + 'Storage-class specifier (static, extern, typedef, etc) should be ' + 'at the beginning of the declaration.') if Match(r'\s*#\s*endif\s*[^/\s]+', line): error(filename, linenum, 'build/endif_comment', 5, @@ -2181,26 +2874,79 @@ def CheckForNonStandardConstructs(filename, clean_lines, linenum, # Look for single-argument constructors that aren't marked explicit. # Technically a valid construct, but against style. - args = Match(r'\s+(?:inline\s+)?%s\s*\(([^,()]+)\)' - % re.escape(base_classname), - line) - if (args and - args.group(1) != 'void' and - not Match(r'(const\s+)?%s(\s+const)?\s*(?:<\w+>\s*)?&' - % re.escape(base_classname), args.group(1).strip())): - error(filename, linenum, 'runtime/explicit', 5, - 'Single-argument constructors should be marked explicit.') - - -def CheckSpacingForFunctionCall(filename, line, linenum, error): + explicit_constructor_match = Match( + r'\s+(?:(?:inline|constexpr)\s+)*(explicit\s+)?' + r'(?:(?:inline|constexpr)\s+)*%s\s*' + r'\(((?:[^()]|\([^()]*\))*)\)' + % re.escape(base_classname), + line) + + if explicit_constructor_match: + is_marked_explicit = explicit_constructor_match.group(1) + + if not explicit_constructor_match.group(2): + constructor_args = [] + else: + constructor_args = explicit_constructor_match.group(2).split(',') + + # collapse arguments so that commas in template parameter lists and function + # argument parameter lists don't split arguments in two + i = 0 + while i < len(constructor_args): + constructor_arg = constructor_args[i] + while (constructor_arg.count('<') > constructor_arg.count('>') or + constructor_arg.count('(') > constructor_arg.count(')')): + constructor_arg += ',' + constructor_args[i + 1] + del constructor_args[i + 1] + constructor_args[i] = constructor_arg + i += 1 + + defaulted_args = [arg for arg in constructor_args if '=' in arg] + noarg_constructor = (not constructor_args or # empty arg list + # 'void' arg specifier + (len(constructor_args) == 1 and + constructor_args[0].strip() == 'void')) + onearg_constructor = ((len(constructor_args) == 1 and # exactly one arg + not noarg_constructor) or + # all but at most one arg defaulted + (len(constructor_args) >= 1 and + not noarg_constructor and + len(defaulted_args) >= len(constructor_args) - 1)) + initializer_list_constructor = bool( + onearg_constructor and + Search(r'\bstd\s*::\s*initializer_list\b', constructor_args[0])) + copy_constructor = bool( + onearg_constructor and + Match(r'(const\s+)?%s(\s*<[^>]*>)?(\s+const)?\s*(?:<\w+>\s*)?&' + % re.escape(base_classname), constructor_args[0].strip())) + + if (not is_marked_explicit and + onearg_constructor and + not initializer_list_constructor and + not copy_constructor): + if defaulted_args: + error(filename, linenum, 'runtime/explicit', 5, + 'Constructors callable with one argument ' + 'should be marked explicit.') + else: + error(filename, linenum, 'runtime/explicit', 5, + 'Single-parameter constructors should be marked explicit.') + elif is_marked_explicit and not onearg_constructor: + if noarg_constructor: + error(filename, linenum, 'runtime/explicit', 5, + 'Zero-parameter constructors should not be marked explicit.') + + +def CheckSpacingForFunctionCall(filename, clean_lines, linenum, error): """Checks for the correctness of various spacing around function calls. Args: filename: The name of the current file. - line: The text of the line to check. + clean_lines: A CleansedLines instance containing the file. linenum: The number of the line to check. error: The function to call with any errors found. """ + line = clean_lines.elided[linenum] # Since function calls often occur inside if/for/while/switch # expressions - which have their own, more liberal conventions - we @@ -2243,10 +2989,18 @@ def CheckSpacingForFunctionCall(filename, line, linenum, error): error(filename, linenum, 'whitespace/parens', 2, 'Extra space after (') if (Search(r'\w\s+\(', fncall) and - not Search(r'#\s*define|typedef', fncall) and - not Search(r'\w\s+\((\w+::)*\*\w+\)\(', fncall)): - error(filename, linenum, 'whitespace/parens', 4, - 'Extra space before ( in function call') + not Search(r'_{0,2}asm_{0,2}\s+_{0,2}volatile_{0,2}\s+\(', fncall) and + not Search(r'#\s*define|typedef|using\s+\w+\s*=', fncall) and + not Search(r'\w\s+\((\w+::)*\*\w+\)\(', fncall) and + not Search(r'\bcase\s+\(', fncall)): + # TODO(unknown): Space after an operator function seem to be a common + # error, silence those for now by restricting them to highest verbosity. + if Search(r'\boperator_*\b', line): + error(filename, linenum, 'whitespace/parens', 0, + 'Extra space before ( in function call') + else: + error(filename, linenum, 'whitespace/parens', 4, + 'Extra space before ( in function call') # If the ) is followed only by a newline or a { + newline, assume it's # part of a control statement (if/while/etc), and don't complain if Search(r'[^)]\s+\)\s*[^{\s]', fncall): @@ -2275,12 +3029,26 @@ def IsBlankLine(line): return not line or line.isspace() +def CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line, + error): + is_namespace_indent_item = ( + len(nesting_state.stack) > 1 and + nesting_state.stack[-1].check_namespace_indentation and + isinstance(nesting_state.previous_stack_top, _NamespaceInfo) and + nesting_state.previous_stack_top == nesting_state.stack[-2]) + + if ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item, + clean_lines.elided, line): + CheckItemIndentationInNamespace(filename, clean_lines.elided, + line, error) + + def CheckForFunctionLengths(filename, clean_lines, linenum, 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 + https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions Uses a simplistic algorithm assuming other style guidelines (especially spacing) are followed. @@ -2300,8 +3068,6 @@ def CheckForFunctionLengths(filename, clean_lines, linenum, """ lines = clean_lines.lines line = lines[linenum] - raw = clean_lines.raw_lines - raw_line = raw[linenum] joined_line = '' starting_func = False @@ -2348,190 +3114,55 @@ def CheckForFunctionLengths(filename, clean_lines, linenum, _RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?') -def CheckComment(comment, filename, linenum, error): - """Checks for common mistakes in TODO comments. +def CheckComment(line, filename, linenum, next_line_start, error): + """Checks for common mistakes in comments. Args: - comment: The text of the comment from the line in question. + line: The line in question. filename: The name of the current file. linenum: The number of the line to check. + next_line_start: The first non-whitespace column of the next line. error: The function to call with any errors found. """ - match = _RE_PATTERN_TODO.match(comment) - if match: - # One whitespace is correct; zero whitespace is handled elsewhere. - leading_whitespace = match.group(1) - if len(leading_whitespace) > 1: - error(filename, linenum, 'whitespace/todo', 2, - 'Too many spaces before TODO') - - username = match.group(2) - if not username: - error(filename, linenum, 'readability/todo', 2, - 'Missing username in TODO; it should look like ' - '"// TODO(my_username): Stuff."') - - middle_whitespace = match.group(3) - # Comparisons made explicit for correctness -- pylint: disable=g-explicit-bool-comparison - if middle_whitespace != ' ' and middle_whitespace != '': - error(filename, linenum, 'whitespace/todo', 2, - 'TODO(my_username) should be followed by a space') - -def CheckAccess(filename, clean_lines, linenum, nesting_state, error): - """Checks for improper use of DISALLOW* macros. + commentpos = line.find('//') + if commentpos != -1: + # Check if the // may be in quotes. If so, ignore it + if re.sub(r'\\.', '', line[0:commentpos]).count('"') % 2 == 0: + # Allow one space for new scopes, two spaces otherwise: + if (commentpos >= 1 and + line[commentpos-1] not in string.whitespace): + error(filename, linenum, 'whitespace/comments', 2, + 'At least a single space is required between code and comments') - Args: - filename: The name of the current file. - clean_lines: A CleansedLines instance containing the file. - linenum: The number of the line to check. - nesting_state: A _NestingState instance which maintains information about - the current stack of nested blocks being parsed. - error: The function to call with any errors found. - """ - line = clean_lines.elided[linenum] # get rid of comments and strings - - matched = Match((r'\s*(DISALLOW_COPY_AND_ASSIGN|' - r'DISALLOW_EVIL_CONSTRUCTORS|' - r'DISALLOW_IMPLICIT_CONSTRUCTORS)'), line) - if not matched: - return - if nesting_state.stack and isinstance(nesting_state.stack[-1], _ClassInfo): - if nesting_state.stack[-1].access != 'private': - error(filename, linenum, 'readability/constructors', 3, - '%s must be in the private: section' % matched.group(1)) - - else: - # Found DISALLOW* macro outside a class declaration, or perhaps it - # was used inside a function when it should have been part of the - # class declaration. We could issue a warning here, but it - # probably resulted in a compiler error already. - pass - - -def FindNextMatchingAngleBracket(clean_lines, linenum, init_suffix): - """Find the corresponding > to close a template. - - Args: - clean_lines: A CleansedLines instance containing the file. - linenum: Current line number. - init_suffix: Remainder of the current line after the initial <. - - Returns: - True if a matching bracket exists. - """ - line = init_suffix - nesting_stack = ['<'] - while True: - # Find the next operator that can tell us whether < is used as an - # opening bracket or as a less-than operator. We only want to - # warn on the latter case. - # - # We could also check all other operators and terminate the search - # early, e.g. if we got something like this "a<b+c", the "<" is - # most likely a less-than operator, but then we will get false - # positives for default arguments and other template expressions. - match = Search(r'^[^<>(),;\[\]]*([<>(),;\[\]])(.*)$', line) - if match: - # Found an operator, update nesting stack - operator = match.group(1) - line = match.group(2) - - if nesting_stack[-1] == '<': - # Expecting closing angle bracket - if operator in ('<', '(', '['): - nesting_stack.append(operator) - elif operator == '>': - nesting_stack.pop() - if not nesting_stack: - # Found matching angle bracket - return True - elif operator == ',': - # Got a comma after a bracket, this is most likely a template - # argument. We have not seen a closing angle bracket yet, but - # it's probably a few lines later if we look for it, so just - # return early here. - return True - else: - # Got some other operator. - return False - - else: - # Expecting closing parenthesis or closing bracket - if operator in ('<', '(', '['): - nesting_stack.append(operator) - elif operator in (')', ']'): - # We don't bother checking for matching () or []. If we got - # something like (] or [), it would have been a syntax error. - nesting_stack.pop() - - else: - # Scan the next line - linenum += 1 - if linenum >= len(clean_lines.elided): - break - line = clean_lines.elided[linenum] - - # Exhausted all remaining lines and still no matching angle bracket. - # Most likely the input was incomplete, otherwise we should have - # seen a semicolon and returned early. - return True - - -def FindPreviousMatchingAngleBracket(clean_lines, linenum, init_prefix): - """Find the corresponding < that started a template. - - Args: - clean_lines: A CleansedLines instance containing the file. - linenum: Current line number. - init_prefix: Part of the current line before the initial >. - - Returns: - True if a matching bracket exists. - """ - line = init_prefix - nesting_stack = ['>'] - while True: - # Find the previous operator - match = Search(r'^(.*)([<>(),;\[\]])[^<>(),;\[\]]*$', line) - if match: - # Found an operator, update nesting stack - operator = match.group(2) - line = match.group(1) - - if nesting_stack[-1] == '>': - # Expecting opening angle bracket - if operator in ('>', ')', ']'): - nesting_stack.append(operator) - elif operator == '<': - nesting_stack.pop() - if not nesting_stack: - # Found matching angle bracket - return True - elif operator == ',': - # Got a comma before a bracket, this is most likely a - # template argument. The opening angle bracket is probably - # there if we look for it, so just return early here. - return True - else: - # Got some other operator. - return False - - else: - # Expecting opening parenthesis or opening bracket - if operator in ('>', ')', ']'): - nesting_stack.append(operator) - elif operator in ('(', '['): - nesting_stack.pop() - - else: - # Scan the previous line - linenum -= 1 - if linenum < 0: - break - line = clean_lines.elided[linenum] - - # Exhausted all earlier lines and still no matching angle bracket. - return False + # Checks for common mistakes in TODO comments. + comment = line[commentpos:] + match = _RE_PATTERN_TODO.match(comment) + if match: + # One whitespace is correct; zero whitespace is handled elsewhere. + leading_whitespace = match.group(1) + if len(leading_whitespace) > 1: + error(filename, linenum, 'whitespace/todo', 2, + 'Too many spaces before TODO') + + username = match.group(2) + if not username: + error(filename, linenum, 'readability/todo', 2, + 'Missing username in TODO; it should look like ' + '"// TODO(my_username): Stuff."') + + middle_whitespace = match.group(3) + # Comparisons made explicit for correctness -- pylint: disable=g-explicit-bool-comparison + if middle_whitespace != ' ' and middle_whitespace != '': + error(filename, linenum, 'whitespace/todo', 2, + 'TODO(my_username) should be followed by a space') + + # If the comment contains an alphanumeric character, there + # should be a space somewhere between it and the // unless + # it's a /// or //! Doxygen comment. + if (Match(r'//[^ ]*\w', comment) and + not Match(r'(///|//\!)(\s+|$)', comment)): + error(filename, linenum, 'whitespace/comments', 4, + 'Should have a space between // and comment') def CheckSpacing(filename, clean_lines, linenum, nesting_state, error): @@ -2547,7 +3178,7 @@ def CheckSpacing(filename, clean_lines, linenum, nesting_state, error): filename: The name of the current file. clean_lines: A CleansedLines instance containing the file. linenum: The number of the line to check. - nesting_state: A _NestingState instance which maintains information about + nesting_state: A NestingState instance which maintains information about the current stack of nested blocks being parsed. error: The function to call with any errors found. """ @@ -2570,7 +3201,12 @@ def CheckSpacing(filename, clean_lines, linenum, nesting_state, error): # } # # A warning about missing end of namespace comments will be issued instead. - if IsBlankLine(line) and not nesting_state.InNamespaceBody(): + # + # Also skip blank line checks for 'extern "C"' blocks, which are formatted + # like namespaces. + if (IsBlankLine(line) and + not nesting_state.InNamespaceBody() and + not nesting_state.InExternC()): elided = clean_lines.elided prev_line = elided[linenum - 1] prevbrace = prev_line.rfind('{') @@ -2633,51 +3269,64 @@ def CheckSpacing(filename, clean_lines, linenum, nesting_state, error): error(filename, linenum, 'whitespace/blank_line', 3, 'Do not leave a blank line after "%s:"' % matched.group(1)) - # Next, we complain if there's a comment too near the text - commentpos = line.find('//') - if commentpos != -1: - # Check if the // may be in quotes. If so, ignore it - # Comparisons made explicit for clarity -- pylint: disable=g-explicit-bool-comparison - if (line.count('"', 0, commentpos) - - line.count('\\"', 0, commentpos)) % 2 == 0: # not in quotes - # Allow one space for new scopes, two spaces otherwise: - if (commentpos >= 1 and - line[commentpos-1] not in string.whitespace): - error(filename, linenum, 'whitespace/comments', 2, - 'At least a single space is required between code and comments') - # There should always be a space between the // and the comment - commentend = commentpos + 2 - if commentend < len(line) and not line[commentend] == ' ': - # but some lines are exceptions -- e.g. if they're big - # comment delimiters like: - # //--------------- <TRUNCATED>
