Merge "vp9_find_mv_refs_idx: remove unused split_count"
[aom.git] / tools / cpplint.py
blob159dbbb07e5b2e3f46a2dff2edf1ac8db6439a97
1 #!/usr/bin/python
3 # Copyright (c) 2009 Google Inc. All rights reserved.
5 # Redistribution and use in source and binary forms, with or without
6 # modification, are permitted provided that the following conditions are
7 # met:
9 # * Redistributions of source code must retain the above copyright
10 # notice, this list of conditions and the following disclaimer.
11 # * Redistributions in binary form must reproduce the above
12 # copyright notice, this list of conditions and the following disclaimer
13 # in the documentation and/or other materials provided with the
14 # distribution.
15 # * Neither the name of Google Inc. nor the names of its
16 # contributors may be used to endorse or promote products derived from
17 # this software without specific prior written permission.
19 # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 # "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 # LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22 # A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23 # OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 # SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25 # LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 # DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 # THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31 # Here are some issues that I've had people identify in my code during reviews,
32 # that I think are possible to flag automatically in a lint tool. If these were
33 # caught by lint, it would save time both for myself and that of my reviewers.
34 # Most likely, some of these are beyond the scope of the current lint framework,
35 # but I think it is valuable to retain these wish-list items even if they cannot
36 # be immediately implemented.
38 # Suggestions
39 # -----------
40 # - Check for no 'explicit' for multi-arg ctor
41 # - Check for boolean assign RHS in parens
42 # - Check for ctor initializer-list colon position and spacing
43 # - Check that if there's a ctor, there should be a dtor
44 # - Check accessors that return non-pointer member variables are
45 # declared const
46 # - Check accessors that return non-const pointer member vars are
47 # *not* declared const
48 # - Check for using public includes for testing
49 # - Check for spaces between brackets in one-line inline method
50 # - Check for no assert()
51 # - Check for spaces surrounding operators
52 # - Check for 0 in pointer context (should be NULL)
53 # - Check for 0 in char context (should be '\0')
54 # - Check for camel-case method name conventions for methods
55 # that are not simple inline getters and setters
56 # - Do not indent namespace contents
57 # - Avoid inlining non-trivial constructors in header files
58 # - Check for old-school (void) cast for call-sites of functions
59 # ignored return value
60 # - Check gUnit usage of anonymous namespace
61 # - Check for class declaration order (typedefs, consts, enums,
62 # ctor(s?), dtor, friend declarations, methods, member vars)
65 """Does google-lint on c++ files.
67 The goal of this script is to identify places in the code that *may*
68 be in non-compliance with google style. It does not attempt to fix
69 up these problems -- the point is to educate. It does also not
70 attempt to find all problems, or to ensure that everything it does
71 find is legitimately a problem.
73 In particular, we can get very confused by /* and // inside strings!
74 We do a small hack, which is to ignore //'s with "'s after them on the
75 same line, but it is far from perfect (in either direction).
76 """
78 import codecs
79 import copy
80 import getopt
81 import math # for log
82 import os
83 import re
84 import sre_compile
85 import string
86 import sys
87 import unicodedata
90 _USAGE = """
91 Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...]
92 [--counting=total|toplevel|detailed]
93 <file> [file] ...
95 The style guidelines this tries to follow are those in
96 http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml
98 Every problem is given a confidence score from 1-5, with 5 meaning we are
99 certain of the problem, and 1 meaning it could be a legitimate construct.
100 This will miss some errors, and is not a substitute for a code review.
102 To suppress false-positive errors of a certain category, add a
103 'NOLINT(category)' comment to the line. NOLINT or NOLINT(*)
104 suppresses errors of all categories on that line.
106 The files passed in will be linted; at least one file must be provided.
107 Linted extensions are .cc, .cpp, and .h. Other file types will be ignored.
109 Flags:
111 output=vs7
112 By default, the output is formatted to ease emacs parsing. Visual Studio
113 compatible output (vs7) may also be used. Other formats are unsupported.
115 verbose=#
116 Specify a number 0-5 to restrict errors to certain verbosity levels.
118 filter=-x,+y,...
119 Specify a comma-separated list of category-filters to apply: only
120 error messages whose category names pass the filters will be printed.
121 (Category names are printed with the message and look like
122 "[whitespace/indent]".) Filters are evaluated left to right.
123 "-FOO" and "FOO" means "do not print categories that start with FOO".
124 "+FOO" means "do print categories that start with FOO".
126 Examples: --filter=-whitespace,+whitespace/braces
127 --filter=whitespace,runtime/printf,+runtime/printf_format
128 --filter=-,+build/include_what_you_use
130 To see a list of all the categories used in cpplint, pass no arg:
131 --filter=
133 counting=total|toplevel|detailed
134 The total number of errors found is always printed. If
135 'toplevel' is provided, then the count of errors in each of
136 the top-level categories like 'build' and 'whitespace' will
137 also be printed. If 'detailed' is provided, then a count
138 is provided for each category like 'build/class'.
140 root=subdir
141 The root directory used for deriving header guard CPP variable.
142 By default, the header guard CPP variable is calculated as the relative
143 path to the directory that contains .git, .hg, or .svn. When this flag
144 is specified, the relative path is calculated from the specified
145 directory. If the specified directory does not exist, this flag is
146 ignored.
148 Examples:
149 Assuing that src/.git exists, the header guard CPP variables for
150 src/chrome/browser/ui/browser.h are:
152 No flag => CHROME_BROWSER_UI_BROWSER_H_
153 --root=chrome => BROWSER_UI_BROWSER_H_
154 --root=chrome/browser => UI_BROWSER_H_
157 # We categorize each error message we print. Here are the categories.
158 # We want an explicit list so we can list them all in cpplint --filter=.
159 # If you add a new error message with a new category, add it to the list
160 # here! cpplint_unittest.py should tell you if you forget to do this.
161 # \ used for clearer layout -- pylint: disable-msg=C6013
162 _ERROR_CATEGORIES = [
163 'build/class',
164 'build/deprecated',
165 'build/endif_comment',
166 'build/explicit_make_pair',
167 'build/forward_decl',
168 'build/header_guard',
169 'build/include',
170 'build/include_alpha',
171 'build/include_order',
172 'build/include_what_you_use',
173 'build/namespaces',
174 'build/printf_format',
175 'build/storage_class',
176 'legal/copyright',
177 'readability/alt_tokens',
178 'readability/braces',
179 'readability/casting',
180 'readability/check',
181 'readability/constructors',
182 'readability/fn_size',
183 'readability/function',
184 'readability/multiline_comment',
185 'readability/multiline_string',
186 'readability/namespace',
187 'readability/nolint',
188 'readability/streams',
189 'readability/todo',
190 'readability/utf8',
191 'runtime/arrays',
192 'runtime/casting',
193 'runtime/explicit',
194 'runtime/int',
195 'runtime/init',
196 'runtime/invalid_increment',
197 'runtime/member_string_references',
198 'runtime/memset',
199 'runtime/operator',
200 'runtime/printf',
201 'runtime/printf_format',
202 'runtime/references',
203 'runtime/rtti',
204 'runtime/sizeof',
205 'runtime/string',
206 'runtime/threadsafe_fn',
207 'whitespace/blank_line',
208 'whitespace/braces',
209 'whitespace/comma',
210 'whitespace/comments',
211 'whitespace/empty_loop_body',
212 'whitespace/end_of_line',
213 'whitespace/ending_newline',
214 'whitespace/forcolon',
215 'whitespace/indent',
216 'whitespace/labels',
217 'whitespace/line_length',
218 'whitespace/newline',
219 'whitespace/operators',
220 'whitespace/parens',
221 'whitespace/semicolon',
222 'whitespace/tab',
223 'whitespace/todo'
226 # The default state of the category filter. This is overrided by the --filter=
227 # flag. By default all errors are on, so only add here categories that should be
228 # off by default (i.e., categories that must be enabled by the --filter= flags).
229 # All entries here should start with a '-' or '+', as in the --filter= flag.
230 _DEFAULT_FILTERS = ['-build/include_alpha']
232 # We used to check for high-bit characters, but after much discussion we
233 # decided those were OK, as long as they were in UTF-8 and didn't represent
234 # hard-coded international strings, which belong in a separate i18n file.
236 # Headers that we consider STL headers.
237 _STL_HEADERS = frozenset([
238 'algobase.h', 'algorithm', 'alloc.h', 'bitset', 'deque', 'exception',
239 'function.h', 'functional', 'hash_map', 'hash_map.h', 'hash_set',
240 'hash_set.h', 'iterator', 'list', 'list.h', 'map', 'memory', 'new',
241 'pair.h', 'pthread_alloc', 'queue', 'set', 'set.h', 'sstream', 'stack',
242 'stl_alloc.h', 'stl_relops.h', 'type_traits.h',
243 'utility', 'vector', 'vector.h',
247 # Non-STL C++ system headers.
248 _CPP_HEADERS = frozenset([
249 'algo.h', 'builtinbuf.h', 'bvector.h', 'cassert', 'cctype',
250 'cerrno', 'cfloat', 'ciso646', 'climits', 'clocale', 'cmath',
251 'complex', 'complex.h', 'csetjmp', 'csignal', 'cstdarg', 'cstddef',
252 'cstdio', 'cstdlib', 'cstring', 'ctime', 'cwchar', 'cwctype',
253 'defalloc.h', 'deque.h', 'editbuf.h', 'exception', 'fstream',
254 'fstream.h', 'hashtable.h', 'heap.h', 'indstream.h', 'iomanip',
255 'iomanip.h', 'ios', 'iosfwd', 'iostream', 'iostream.h', 'istream',
256 'istream.h', 'iterator.h', 'limits', 'map.h', 'multimap.h', 'multiset.h',
257 'numeric', 'ostream', 'ostream.h', 'parsestream.h', 'pfstream.h',
258 'PlotFile.h', 'procbuf.h', 'pthread_alloc.h', 'rope', 'rope.h',
259 'ropeimpl.h', 'SFile.h', 'slist', 'slist.h', 'stack.h', 'stdexcept',
260 'stdiostream.h', 'streambuf.h', 'stream.h', 'strfile.h', 'string',
261 'strstream', 'strstream.h', 'tempbuf.h', 'tree.h', 'typeinfo', 'valarray',
265 # Assertion macros. These are defined in base/logging.h and
266 # testing/base/gunit.h. Note that the _M versions need to come first
267 # for substring matching to work.
268 _CHECK_MACROS = [
269 'DCHECK', 'CHECK',
270 'EXPECT_TRUE_M', 'EXPECT_TRUE',
271 'ASSERT_TRUE_M', 'ASSERT_TRUE',
272 'EXPECT_FALSE_M', 'EXPECT_FALSE',
273 'ASSERT_FALSE_M', 'ASSERT_FALSE',
276 # Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE
277 _CHECK_REPLACEMENT = dict([(m, {}) for m in _CHECK_MACROS])
279 for op, replacement in [('==', 'EQ'), ('!=', 'NE'),
280 ('>=', 'GE'), ('>', 'GT'),
281 ('<=', 'LE'), ('<', 'LT')]:
282 _CHECK_REPLACEMENT['DCHECK'][op] = 'DCHECK_%s' % replacement
283 _CHECK_REPLACEMENT['CHECK'][op] = 'CHECK_%s' % replacement
284 _CHECK_REPLACEMENT['EXPECT_TRUE'][op] = 'EXPECT_%s' % replacement
285 _CHECK_REPLACEMENT['ASSERT_TRUE'][op] = 'ASSERT_%s' % replacement
286 _CHECK_REPLACEMENT['EXPECT_TRUE_M'][op] = 'EXPECT_%s_M' % replacement
287 _CHECK_REPLACEMENT['ASSERT_TRUE_M'][op] = 'ASSERT_%s_M' % replacement
289 for op, inv_replacement in [('==', 'NE'), ('!=', 'EQ'),
290 ('>=', 'LT'), ('>', 'LE'),
291 ('<=', 'GT'), ('<', 'GE')]:
292 _CHECK_REPLACEMENT['EXPECT_FALSE'][op] = 'EXPECT_%s' % inv_replacement
293 _CHECK_REPLACEMENT['ASSERT_FALSE'][op] = 'ASSERT_%s' % inv_replacement
294 _CHECK_REPLACEMENT['EXPECT_FALSE_M'][op] = 'EXPECT_%s_M' % inv_replacement
295 _CHECK_REPLACEMENT['ASSERT_FALSE_M'][op] = 'ASSERT_%s_M' % inv_replacement
297 # Alternative tokens and their replacements. For full list, see section 2.5
298 # Alternative tokens [lex.digraph] in the C++ standard.
300 # Digraphs (such as '%:') are not included here since it's a mess to
301 # match those on a word boundary.
302 _ALT_TOKEN_REPLACEMENT = {
303 'and': '&&',
304 'bitor': '|',
305 'or': '||',
306 'xor': '^',
307 'compl': '~',
308 'bitand': '&',
309 'and_eq': '&=',
310 'or_eq': '|=',
311 'xor_eq': '^=',
312 'not': '!',
313 'not_eq': '!='
316 # Compile regular expression that matches all the above keywords. The "[ =()]"
317 # bit is meant to avoid matching these keywords outside of boolean expressions.
319 # False positives include C-style multi-line comments (http://go/nsiut )
320 # and multi-line strings (http://go/beujw ), but those have always been
321 # troublesome for cpplint.
322 _ALT_TOKEN_REPLACEMENT_PATTERN = re.compile(
323 r'[ =()](' + ('|'.join(_ALT_TOKEN_REPLACEMENT.keys())) + r')(?=[ (]|$)')
326 # These constants define types of headers for use with
327 # _IncludeState.CheckNextIncludeOrder().
328 _C_SYS_HEADER = 1
329 _CPP_SYS_HEADER = 2
330 _LIKELY_MY_HEADER = 3
331 _POSSIBLE_MY_HEADER = 4
332 _OTHER_HEADER = 5
334 # These constants define the current inline assembly state
335 _NO_ASM = 0 # Outside of inline assembly block
336 _INSIDE_ASM = 1 # Inside inline assembly block
337 _END_ASM = 2 # Last line of inline assembly block
338 _BLOCK_ASM = 3 # The whole block is an inline assembly block
340 # Match start of assembly blocks
341 _MATCH_ASM = re.compile(r'^\s*(?:asm|_asm|__asm|__asm__)'
342 r'(?:\s+(volatile|__volatile__))?'
343 r'\s*[{(]')
346 _regexp_compile_cache = {}
348 # Finds occurrences of NOLINT or NOLINT(...).
349 _RE_SUPPRESSION = re.compile(r'\bNOLINT\b(\([^)]*\))?')
351 # {str, set(int)}: a map from error categories to sets of linenumbers
352 # on which those errors are expected and should be suppressed.
353 _error_suppressions = {}
355 # The root directory used for deriving header guard CPP variable.
356 # This is set by --root flag.
357 _root = None
359 def ParseNolintSuppressions(filename, raw_line, linenum, error):
360 """Updates the global list of error-suppressions.
362 Parses any NOLINT comments on the current line, updating the global
363 error_suppressions store. Reports an error if the NOLINT comment
364 was malformed.
366 Args:
367 filename: str, the name of the input file.
368 raw_line: str, the line of input text, with comments.
369 linenum: int, the number of the current line.
370 error: function, an error handler.
372 # FIXME(adonovan): "NOLINT(" is misparsed as NOLINT(*).
373 matched = _RE_SUPPRESSION.search(raw_line)
374 if matched:
375 category = matched.group(1)
376 if category in (None, '(*)'): # => "suppress all"
377 _error_suppressions.setdefault(None, set()).add(linenum)
378 else:
379 if category.startswith('(') and category.endswith(')'):
380 category = category[1:-1]
381 if category in _ERROR_CATEGORIES:
382 _error_suppressions.setdefault(category, set()).add(linenum)
383 else:
384 error(filename, linenum, 'readability/nolint', 5,
385 'Unknown NOLINT error category: %s' % category)
388 def ResetNolintSuppressions():
389 "Resets the set of NOLINT suppressions to empty."
390 _error_suppressions.clear()
393 def IsErrorSuppressedByNolint(category, linenum):
394 """Returns true if the specified error category is suppressed on this line.
396 Consults the global error_suppressions map populated by
397 ParseNolintSuppressions/ResetNolintSuppressions.
399 Args:
400 category: str, the category of the error.
401 linenum: int, the current line number.
402 Returns:
403 bool, True iff the error should be suppressed due to a NOLINT comment.
405 return (linenum in _error_suppressions.get(category, set()) or
406 linenum in _error_suppressions.get(None, set()))
408 def Match(pattern, s):
409 """Matches the string with the pattern, caching the compiled regexp."""
410 # The regexp compilation caching is inlined in both Match and Search for
411 # performance reasons; factoring it out into a separate function turns out
412 # to be noticeably expensive.
413 if not pattern in _regexp_compile_cache:
414 _regexp_compile_cache[pattern] = sre_compile.compile(pattern)
415 return _regexp_compile_cache[pattern].match(s)
418 def Search(pattern, s):
419 """Searches the string for the pattern, caching the compiled regexp."""
420 if not pattern in _regexp_compile_cache:
421 _regexp_compile_cache[pattern] = sre_compile.compile(pattern)
422 return _regexp_compile_cache[pattern].search(s)
425 class _IncludeState(dict):
426 """Tracks line numbers for includes, and the order in which includes appear.
428 As a dict, an _IncludeState object serves as a mapping between include
429 filename and line number on which that file was included.
431 Call CheckNextIncludeOrder() once for each header in the file, passing
432 in the type constants defined above. Calls in an illegal order will
433 raise an _IncludeError with an appropriate error message.
436 # self._section will move monotonically through this set. If it ever
437 # needs to move backwards, CheckNextIncludeOrder will raise an error.
438 _INITIAL_SECTION = 0
439 _MY_H_SECTION = 1
440 _C_SECTION = 2
441 _CPP_SECTION = 3
442 _OTHER_H_SECTION = 4
444 _TYPE_NAMES = {
445 _C_SYS_HEADER: 'C system header',
446 _CPP_SYS_HEADER: 'C++ system header',
447 _LIKELY_MY_HEADER: 'header this file implements',
448 _POSSIBLE_MY_HEADER: 'header this file may implement',
449 _OTHER_HEADER: 'other header',
451 _SECTION_NAMES = {
452 _INITIAL_SECTION: "... nothing. (This can't be an error.)",
453 _MY_H_SECTION: 'a header this file implements',
454 _C_SECTION: 'C system header',
455 _CPP_SECTION: 'C++ system header',
456 _OTHER_H_SECTION: 'other header',
459 def __init__(self):
460 dict.__init__(self)
461 # The name of the current section.
462 self._section = self._INITIAL_SECTION
463 # The path of last found header.
464 self._last_header = ''
466 def CanonicalizeAlphabeticalOrder(self, header_path):
467 """Returns a path canonicalized for alphabetical comparison.
469 - replaces "-" with "_" so they both cmp the same.
470 - removes '-inl' since we don't require them to be after the main header.
471 - lowercase everything, just in case.
473 Args:
474 header_path: Path to be canonicalized.
476 Returns:
477 Canonicalized path.
479 return header_path.replace('-inl.h', '.h').replace('-', '_').lower()
481 def IsInAlphabeticalOrder(self, header_path):
482 """Check if a header is in alphabetical order with the previous header.
484 Args:
485 header_path: Header to be checked.
487 Returns:
488 Returns true if the header is in alphabetical order.
490 canonical_header = self.CanonicalizeAlphabeticalOrder(header_path)
491 if self._last_header > canonical_header:
492 return False
493 self._last_header = canonical_header
494 return True
496 def CheckNextIncludeOrder(self, header_type):
497 """Returns a non-empty error message if the next header is out of order.
499 This function also updates the internal state to be ready to check
500 the next include.
502 Args:
503 header_type: One of the _XXX_HEADER constants defined above.
505 Returns:
506 The empty string if the header is in the right order, or an
507 error message describing what's wrong.
510 error_message = ('Found %s after %s' %
511 (self._TYPE_NAMES[header_type],
512 self._SECTION_NAMES[self._section]))
514 last_section = self._section
516 if header_type == _C_SYS_HEADER:
517 if self._section <= self._C_SECTION:
518 self._section = self._C_SECTION
519 else:
520 self._last_header = ''
521 return error_message
522 elif header_type == _CPP_SYS_HEADER:
523 if self._section <= self._CPP_SECTION:
524 self._section = self._CPP_SECTION
525 else:
526 self._last_header = ''
527 return error_message
528 elif header_type == _LIKELY_MY_HEADER:
529 if self._section <= self._MY_H_SECTION:
530 self._section = self._MY_H_SECTION
531 else:
532 self._section = self._OTHER_H_SECTION
533 elif header_type == _POSSIBLE_MY_HEADER:
534 if self._section <= self._MY_H_SECTION:
535 self._section = self._MY_H_SECTION
536 else:
537 # This will always be the fallback because we're not sure
538 # enough that the header is associated with this file.
539 self._section = self._OTHER_H_SECTION
540 else:
541 assert header_type == _OTHER_HEADER
542 self._section = self._OTHER_H_SECTION
544 if last_section != self._section:
545 self._last_header = ''
547 return ''
550 class _CppLintState(object):
551 """Maintains module-wide state.."""
553 def __init__(self):
554 self.verbose_level = 1 # global setting.
555 self.error_count = 0 # global count of reported errors
556 # filters to apply when emitting error messages
557 self.filters = _DEFAULT_FILTERS[:]
558 self.counting = 'total' # In what way are we counting errors?
559 self.errors_by_category = {} # string to int dict storing error counts
561 # output format:
562 # "emacs" - format that emacs can parse (default)
563 # "vs7" - format that Microsoft Visual Studio 7 can parse
564 self.output_format = 'emacs'
566 def SetOutputFormat(self, output_format):
567 """Sets the output format for errors."""
568 self.output_format = output_format
570 def SetVerboseLevel(self, level):
571 """Sets the module's verbosity, and returns the previous setting."""
572 last_verbose_level = self.verbose_level
573 self.verbose_level = level
574 return last_verbose_level
576 def SetCountingStyle(self, counting_style):
577 """Sets the module's counting options."""
578 self.counting = counting_style
580 def SetFilters(self, filters):
581 """Sets the error-message filters.
583 These filters are applied when deciding whether to emit a given
584 error message.
586 Args:
587 filters: A string of comma-separated filters (eg "+whitespace/indent").
588 Each filter should start with + or -; else we die.
590 Raises:
591 ValueError: The comma-separated filters did not all start with '+' or '-'.
592 E.g. "-,+whitespace,-whitespace/indent,whitespace/badfilter"
594 # Default filters always have less priority than the flag ones.
595 self.filters = _DEFAULT_FILTERS[:]
596 for filt in filters.split(','):
597 clean_filt = filt.strip()
598 if clean_filt:
599 self.filters.append(clean_filt)
600 for filt in self.filters:
601 if not (filt.startswith('+') or filt.startswith('-')):
602 raise ValueError('Every filter in --filters must start with + or -'
603 ' (%s does not)' % filt)
605 def ResetErrorCounts(self):
606 """Sets the module's error statistic back to zero."""
607 self.error_count = 0
608 self.errors_by_category = {}
610 def IncrementErrorCount(self, category):
611 """Bumps the module's error statistic."""
612 self.error_count += 1
613 if self.counting in ('toplevel', 'detailed'):
614 if self.counting != 'detailed':
615 category = category.split('/')[0]
616 if category not in self.errors_by_category:
617 self.errors_by_category[category] = 0
618 self.errors_by_category[category] += 1
620 def PrintErrorCounts(self):
621 """Print a summary of errors by category, and the total."""
622 for category, count in self.errors_by_category.iteritems():
623 sys.stderr.write('Category \'%s\' errors found: %d\n' %
624 (category, count))
625 sys.stderr.write('Total errors found: %d\n' % self.error_count)
627 _cpplint_state = _CppLintState()
630 def _OutputFormat():
631 """Gets the module's output format."""
632 return _cpplint_state.output_format
635 def _SetOutputFormat(output_format):
636 """Sets the module's output format."""
637 _cpplint_state.SetOutputFormat(output_format)
640 def _VerboseLevel():
641 """Returns the module's verbosity setting."""
642 return _cpplint_state.verbose_level
645 def _SetVerboseLevel(level):
646 """Sets the module's verbosity, and returns the previous setting."""
647 return _cpplint_state.SetVerboseLevel(level)
650 def _SetCountingStyle(level):
651 """Sets the module's counting options."""
652 _cpplint_state.SetCountingStyle(level)
655 def _Filters():
656 """Returns the module's list of output filters, as a list."""
657 return _cpplint_state.filters
660 def _SetFilters(filters):
661 """Sets the module's error-message filters.
663 These filters are applied when deciding whether to emit a given
664 error message.
666 Args:
667 filters: A string of comma-separated filters (eg "whitespace/indent").
668 Each filter should start with + or -; else we die.
670 _cpplint_state.SetFilters(filters)
673 class _FunctionState(object):
674 """Tracks current function name and the number of lines in its body."""
676 _NORMAL_TRIGGER = 250 # for --v=0, 500 for --v=1, etc.
677 _TEST_TRIGGER = 400 # about 50% more than _NORMAL_TRIGGER.
679 def __init__(self):
680 self.in_a_function = False
681 self.lines_in_function = 0
682 self.current_function = ''
684 def Begin(self, function_name):
685 """Start analyzing function body.
687 Args:
688 function_name: The name of the function being tracked.
690 self.in_a_function = True
691 self.lines_in_function = 0
692 self.current_function = function_name
694 def Count(self):
695 """Count line in current function body."""
696 if self.in_a_function:
697 self.lines_in_function += 1
699 def Check(self, error, filename, linenum):
700 """Report if too many lines in function body.
702 Args:
703 error: The function to call with any errors found.
704 filename: The name of the current file.
705 linenum: The number of the line to check.
707 if Match(r'T(EST|est)', self.current_function):
708 base_trigger = self._TEST_TRIGGER
709 else:
710 base_trigger = self._NORMAL_TRIGGER
711 trigger = base_trigger * 2**_VerboseLevel()
713 if self.lines_in_function > trigger:
714 error_level = int(math.log(self.lines_in_function / base_trigger, 2))
715 # 50 => 0, 100 => 1, 200 => 2, 400 => 3, 800 => 4, 1600 => 5, ...
716 if error_level > 5:
717 error_level = 5
718 error(filename, linenum, 'readability/fn_size', error_level,
719 'Small and focused functions are preferred:'
720 ' %s has %d non-comment lines'
721 ' (error triggered by exceeding %d lines).' % (
722 self.current_function, self.lines_in_function, trigger))
724 def End(self):
725 """Stop analyzing function body."""
726 self.in_a_function = False
729 class _IncludeError(Exception):
730 """Indicates a problem with the include order in a file."""
731 pass
734 class FileInfo:
735 """Provides utility functions for filenames.
737 FileInfo provides easy access to the components of a file's path
738 relative to the project root.
741 def __init__(self, filename):
742 self._filename = filename
744 def FullName(self):
745 """Make Windows paths like Unix."""
746 return os.path.abspath(self._filename).replace('\\', '/')
748 def RepositoryName(self):
749 """FullName after removing the local path to the repository.
751 If we have a real absolute path name here we can try to do something smart:
752 detecting the root of the checkout and truncating /path/to/checkout from
753 the name so that we get header guards that don't include things like
754 "C:\Documents and Settings\..." or "/home/username/..." in them and thus
755 people on different computers who have checked the source out to different
756 locations won't see bogus errors.
758 fullname = self.FullName()
760 if os.path.exists(fullname):
761 project_dir = os.path.dirname(fullname)
763 if os.path.exists(os.path.join(project_dir, ".svn")):
764 # If there's a .svn file in the current directory, we recursively look
765 # up the directory tree for the top of the SVN checkout
766 root_dir = project_dir
767 one_up_dir = os.path.dirname(root_dir)
768 while os.path.exists(os.path.join(one_up_dir, ".svn")):
769 root_dir = os.path.dirname(root_dir)
770 one_up_dir = os.path.dirname(one_up_dir)
772 prefix = os.path.commonprefix([root_dir, project_dir])
773 return fullname[len(prefix) + 1:]
775 # Not SVN <= 1.6? Try to find a git, hg, or svn top level directory by
776 # searching up from the current path.
777 root_dir = os.path.dirname(fullname)
778 while (root_dir != os.path.dirname(root_dir) and
779 not os.path.exists(os.path.join(root_dir, ".git")) and
780 not os.path.exists(os.path.join(root_dir, ".hg")) and
781 not os.path.exists(os.path.join(root_dir, ".svn"))):
782 root_dir = os.path.dirname(root_dir)
784 if (os.path.exists(os.path.join(root_dir, ".git")) or
785 os.path.exists(os.path.join(root_dir, ".hg")) or
786 os.path.exists(os.path.join(root_dir, ".svn"))):
787 prefix = os.path.commonprefix([root_dir, project_dir])
788 return fullname[len(prefix) + 1:]
790 # Don't know what to do; header guard warnings may be wrong...
791 return fullname
793 def Split(self):
794 """Splits the file into the directory, basename, and extension.
796 For 'chrome/browser/browser.cc', Split() would
797 return ('chrome/browser', 'browser', '.cc')
799 Returns:
800 A tuple of (directory, basename, extension).
803 googlename = self.RepositoryName()
804 project, rest = os.path.split(googlename)
805 return (project,) + os.path.splitext(rest)
807 def BaseName(self):
808 """File base name - text after the final slash, before the final period."""
809 return self.Split()[1]
811 def Extension(self):
812 """File extension - text following the final period."""
813 return self.Split()[2]
815 def NoExtension(self):
816 """File has no source file extension."""
817 return '/'.join(self.Split()[0:2])
819 def IsSource(self):
820 """File has a source file extension."""
821 return self.Extension()[1:] in ('c', 'cc', 'cpp', 'cxx')
824 def _ShouldPrintError(category, confidence, linenum):
825 """If confidence >= verbose, category passes filter and is not suppressed."""
827 # There are three ways we might decide not to print an error message:
828 # a "NOLINT(category)" comment appears in the source,
829 # the verbosity level isn't high enough, or the filters filter it out.
830 if IsErrorSuppressedByNolint(category, linenum):
831 return False
832 if confidence < _cpplint_state.verbose_level:
833 return False
835 is_filtered = False
836 for one_filter in _Filters():
837 if one_filter.startswith('-'):
838 if category.startswith(one_filter[1:]):
839 is_filtered = True
840 elif one_filter.startswith('+'):
841 if category.startswith(one_filter[1:]):
842 is_filtered = False
843 else:
844 assert False # should have been checked for in SetFilter.
845 if is_filtered:
846 return False
848 return True
851 def Error(filename, linenum, category, confidence, message):
852 """Logs the fact we've found a lint error.
854 We log where the error was found, and also our confidence in the error,
855 that is, how certain we are this is a legitimate style regression, and
856 not a misidentification or a use that's sometimes justified.
858 False positives can be suppressed by the use of
859 "cpplint(category)" comments on the offending line. These are
860 parsed into _error_suppressions.
862 Args:
863 filename: The name of the file containing the error.
864 linenum: The number of the line containing the error.
865 category: A string used to describe the "category" this bug
866 falls under: "whitespace", say, or "runtime". Categories
867 may have a hierarchy separated by slashes: "whitespace/indent".
868 confidence: A number from 1-5 representing a confidence score for
869 the error, with 5 meaning that we are certain of the problem,
870 and 1 meaning that it could be a legitimate construct.
871 message: The error message.
873 if _ShouldPrintError(category, confidence, linenum):
874 _cpplint_state.IncrementErrorCount(category)
875 if _cpplint_state.output_format == 'vs7':
876 sys.stderr.write('%s(%s): %s [%s] [%d]\n' % (
877 filename, linenum, message, category, confidence))
878 else:
879 sys.stderr.write('%s:%s: %s [%s] [%d]\n' % (
880 filename, linenum, message, category, confidence))
883 # Matches standard C++ escape esequences per 2.13.2.3 of the C++ standard.
884 _RE_PATTERN_CLEANSE_LINE_ESCAPES = re.compile(
885 r'\\([abfnrtv?"\\\']|\d+|x[0-9a-fA-F]+)')
886 # Matches strings. Escape codes should already be removed by ESCAPES.
887 _RE_PATTERN_CLEANSE_LINE_DOUBLE_QUOTES = re.compile(r'"[^"]*"')
888 # Matches characters. Escape codes should already be removed by ESCAPES.
889 _RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES = re.compile(r"'.'")
890 # Matches multi-line C++ comments.
891 # This RE is a little bit more complicated than one might expect, because we
892 # have to take care of space removals tools so we can handle comments inside
893 # statements better.
894 # The current rule is: We only clear spaces from both sides when we're at the
895 # end of the line. Otherwise, we try to remove spaces from the right side,
896 # if this doesn't work we try on left side but only if there's a non-character
897 # on the right.
898 _RE_PATTERN_CLEANSE_LINE_C_COMMENTS = re.compile(
899 r"""(\s*/\*.*\*/\s*$|
900 /\*.*\*/\s+|
901 \s+/\*.*\*/(?=\W)|
902 /\*.*\*/)""", re.VERBOSE)
905 def IsCppString(line):
906 """Does line terminate so, that the next symbol is in string constant.
908 This function does not consider single-line nor multi-line comments.
910 Args:
911 line: is a partial line of code starting from the 0..n.
913 Returns:
914 True, if next character appended to 'line' is inside a
915 string constant.
918 line = line.replace(r'\\', 'XX') # after this, \\" does not match to \"
919 return ((line.count('"') - line.count(r'\"') - line.count("'\"'")) & 1) == 1
922 def FindNextMultiLineCommentStart(lines, lineix):
923 """Find the beginning marker for a multiline comment."""
924 while lineix < len(lines):
925 if lines[lineix].strip().startswith('/*'):
926 # Only return this marker if the comment goes beyond this line
927 if lines[lineix].strip().find('*/', 2) < 0:
928 return lineix
929 lineix += 1
930 return len(lines)
933 def FindNextMultiLineCommentEnd(lines, lineix):
934 """We are inside a comment, find the end marker."""
935 while lineix < len(lines):
936 if lines[lineix].strip().endswith('*/'):
937 return lineix
938 lineix += 1
939 return len(lines)
942 def RemoveMultiLineCommentsFromRange(lines, begin, end):
943 """Clears a range of lines for multi-line comments."""
944 # Having // dummy comments makes the lines non-empty, so we will not get
945 # unnecessary blank line warnings later in the code.
946 for i in range(begin, end):
947 lines[i] = '// dummy'
950 def RemoveMultiLineComments(filename, lines, error):
951 """Removes multiline (c-style) comments from lines."""
952 lineix = 0
953 while lineix < len(lines):
954 lineix_begin = FindNextMultiLineCommentStart(lines, lineix)
955 if lineix_begin >= len(lines):
956 return
957 lineix_end = FindNextMultiLineCommentEnd(lines, lineix_begin)
958 if lineix_end >= len(lines):
959 error(filename, lineix_begin + 1, 'readability/multiline_comment', 5,
960 'Could not find end of multi-line comment')
961 return
962 RemoveMultiLineCommentsFromRange(lines, lineix_begin, lineix_end + 1)
963 lineix = lineix_end + 1
966 def CleanseComments(line):
967 """Removes //-comments and single-line C-style /* */ comments.
969 Args:
970 line: A line of C++ source.
972 Returns:
973 The line with single-line comments removed.
975 commentpos = line.find('//')
976 if commentpos != -1 and not IsCppString(line[:commentpos]):
977 line = line[:commentpos].rstrip()
978 # get rid of /* ... */
979 return _RE_PATTERN_CLEANSE_LINE_C_COMMENTS.sub('', line)
982 class CleansedLines(object):
983 """Holds 3 copies of all lines with different preprocessing applied to them.
985 1) elided member contains lines without strings and comments,
986 2) lines member contains lines without comments, and
987 3) raw_lines member contains all the lines without processing.
988 All these three members are of <type 'list'>, and of the same length.
991 def __init__(self, lines):
992 self.elided = []
993 self.lines = []
994 self.raw_lines = lines
995 self.num_lines = len(lines)
996 for linenum in range(len(lines)):
997 self.lines.append(CleanseComments(lines[linenum]))
998 elided = self._CollapseStrings(lines[linenum])
999 self.elided.append(CleanseComments(elided))
1001 def NumLines(self):
1002 """Returns the number of lines represented."""
1003 return self.num_lines
1005 @staticmethod
1006 def _CollapseStrings(elided):
1007 """Collapses strings and chars on a line to simple "" or '' blocks.
1009 We nix strings first so we're not fooled by text like '"http://"'
1011 Args:
1012 elided: The line being processed.
1014 Returns:
1015 The line with collapsed strings.
1017 if not _RE_PATTERN_INCLUDE.match(elided):
1018 # Remove escaped characters first to make quote/single quote collapsing
1019 # basic. Things that look like escaped characters shouldn't occur
1020 # outside of strings and chars.
1021 elided = _RE_PATTERN_CLEANSE_LINE_ESCAPES.sub('', elided)
1022 elided = _RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES.sub("''", elided)
1023 elided = _RE_PATTERN_CLEANSE_LINE_DOUBLE_QUOTES.sub('""', elided)
1024 return elided
1027 def FindEndOfExpressionInLine(line, startpos, depth, startchar, endchar):
1028 """Find the position just after the matching endchar.
1030 Args:
1031 line: a CleansedLines line.
1032 startpos: start searching at this position.
1033 depth: nesting level at startpos.
1034 startchar: expression opening character.
1035 endchar: expression closing character.
1037 Returns:
1038 Index just after endchar.
1040 for i in xrange(startpos, len(line)):
1041 if line[i] == startchar:
1042 depth += 1
1043 elif line[i] == endchar:
1044 depth -= 1
1045 if depth == 0:
1046 return i + 1
1047 return -1
1050 def CloseExpression(clean_lines, linenum, pos):
1051 """If input points to ( or { or [, finds the position that closes it.
1053 If lines[linenum][pos] points to a '(' or '{' or '[', finds the
1054 linenum/pos that correspond to the closing of the expression.
1056 Args:
1057 clean_lines: A CleansedLines instance containing the file.
1058 linenum: The number of the line to check.
1059 pos: A position on the line.
1061 Returns:
1062 A tuple (line, linenum, pos) pointer *past* the closing brace, or
1063 (line, len(lines), -1) if we never find a close. Note we ignore
1064 strings and comments when matching; and the line we return is the
1065 'cleansed' line at linenum.
1068 line = clean_lines.elided[linenum]
1069 startchar = line[pos]
1070 if startchar not in '({[':
1071 return (line, clean_lines.NumLines(), -1)
1072 if startchar == '(': endchar = ')'
1073 if startchar == '[': endchar = ']'
1074 if startchar == '{': endchar = '}'
1076 # Check first line
1077 end_pos = FindEndOfExpressionInLine(line, pos, 0, startchar, endchar)
1078 if end_pos > -1:
1079 return (line, linenum, end_pos)
1080 tail = line[pos:]
1081 num_open = tail.count(startchar) - tail.count(endchar)
1082 while linenum < clean_lines.NumLines() - 1:
1083 linenum += 1
1084 line = clean_lines.elided[linenum]
1085 delta = line.count(startchar) - line.count(endchar)
1086 if num_open + delta <= 0:
1087 return (line, linenum,
1088 FindEndOfExpressionInLine(line, 0, num_open, startchar, endchar))
1089 num_open += delta
1091 # Did not find endchar before end of file, give up
1092 return (line, clean_lines.NumLines(), -1)
1094 def CheckForCopyright(filename, lines, error):
1095 """Logs an error if no Copyright message appears at the top of the file."""
1097 # We'll say it should occur by line 10. Don't forget there's a
1098 # dummy line at the front.
1099 for line in xrange(1, min(len(lines), 11)):
1100 if re.search(r'Copyright', lines[line], re.I): break
1101 else: # means no copyright line was found
1102 error(filename, 0, 'legal/copyright', 5,
1103 'No copyright message found. '
1104 'You should have a line: "Copyright [year] <Copyright Owner>"')
1107 def GetHeaderGuardCPPVariable(filename):
1108 """Returns the CPP variable that should be used as a header guard.
1110 Args:
1111 filename: The name of a C++ header file.
1113 Returns:
1114 The CPP variable that should be used as a header guard in the
1115 named file.
1119 # Restores original filename in case that cpplint is invoked from Emacs's
1120 # flymake.
1121 filename = re.sub(r'_flymake\.h$', '.h', filename)
1122 filename = re.sub(r'/\.flymake/([^/]*)$', r'/\1', filename)
1124 fileinfo = FileInfo(filename)
1125 file_path_from_root = fileinfo.RepositoryName()
1126 if _root:
1127 file_path_from_root = re.sub('^' + _root + os.sep, '', file_path_from_root)
1128 return re.sub(r'[-./\s]', '_', file_path_from_root).upper() + '_'
1131 def CheckForHeaderGuard(filename, lines, error):
1132 """Checks that the file contains a header guard.
1134 Logs an error if no #ifndef header guard is present. For other
1135 headers, checks that the full pathname is used.
1137 Args:
1138 filename: The name of the C++ header file.
1139 lines: An array of strings, each representing a line of the file.
1140 error: The function to call with any errors found.
1143 cppvar = GetHeaderGuardCPPVariable(filename)
1145 ifndef = None
1146 ifndef_linenum = 0
1147 define = None
1148 endif = None
1149 endif_linenum = 0
1150 for linenum, line in enumerate(lines):
1151 linesplit = line.split()
1152 if len(linesplit) >= 2:
1153 # find the first occurrence of #ifndef and #define, save arg
1154 if not ifndef and linesplit[0] == '#ifndef':
1155 # set ifndef to the header guard presented on the #ifndef line.
1156 ifndef = linesplit[1]
1157 ifndef_linenum = linenum
1158 if not define and linesplit[0] == '#define':
1159 define = linesplit[1]
1160 # find the last occurrence of #endif, save entire line
1161 if line.startswith('#endif'):
1162 endif = line
1163 endif_linenum = linenum
1165 if not ifndef:
1166 error(filename, 0, 'build/header_guard', 5,
1167 'No #ifndef header guard found, suggested CPP variable is: %s' %
1168 cppvar)
1169 return
1171 if not define:
1172 error(filename, 0, 'build/header_guard', 5,
1173 'No #define header guard found, suggested CPP variable is: %s' %
1174 cppvar)
1175 return
1177 # The guard should be PATH_FILE_H_, but we also allow PATH_FILE_H__
1178 # for backward compatibility.
1179 if ifndef != cppvar:
1180 error_level = 0
1181 if ifndef != cppvar + '_':
1182 error_level = 5
1184 ParseNolintSuppressions(filename, lines[ifndef_linenum], ifndef_linenum,
1185 error)
1186 error(filename, ifndef_linenum, 'build/header_guard', error_level,
1187 '#ifndef header guard has wrong style, please use: %s' % cppvar)
1189 if define != ifndef:
1190 error(filename, 0, 'build/header_guard', 5,
1191 '#ifndef and #define don\'t match, suggested CPP variable is: %s' %
1192 cppvar)
1193 return
1195 if endif != ('#endif // %s' % cppvar):
1196 error_level = 0
1197 if endif != ('#endif // %s' % (cppvar + '_')):
1198 error_level = 5
1200 ParseNolintSuppressions(filename, lines[endif_linenum], endif_linenum,
1201 error)
1202 error(filename, endif_linenum, 'build/header_guard', error_level,
1203 '#endif line should be "#endif // %s"' % cppvar)
1206 def CheckForUnicodeReplacementCharacters(filename, lines, error):
1207 """Logs an error for each line containing Unicode replacement characters.
1209 These indicate that either the file contained invalid UTF-8 (likely)
1210 or Unicode replacement characters (which it shouldn't). Note that
1211 it's possible for this to throw off line numbering if the invalid
1212 UTF-8 occurred adjacent to a newline.
1214 Args:
1215 filename: The name of the current file.
1216 lines: An array of strings, each representing a line of the file.
1217 error: The function to call with any errors found.
1219 for linenum, line in enumerate(lines):
1220 if u'\ufffd' in line:
1221 error(filename, linenum, 'readability/utf8', 5,
1222 'Line contains invalid UTF-8 (or Unicode replacement character).')
1225 def CheckForNewlineAtEOF(filename, lines, error):
1226 """Logs an error if there is no newline char at the end of the file.
1228 Args:
1229 filename: The name of the current file.
1230 lines: An array of strings, each representing a line of the file.
1231 error: The function to call with any errors found.
1234 # The array lines() was created by adding two newlines to the
1235 # original file (go figure), then splitting on \n.
1236 # To verify that the file ends in \n, we just have to make sure the
1237 # last-but-two element of lines() exists and is empty.
1238 if len(lines) < 3 or lines[-2]:
1239 error(filename, len(lines) - 2, 'whitespace/ending_newline', 5,
1240 'Could not find a newline character at the end of the file.')
1243 def CheckForMultilineCommentsAndStrings(filename, clean_lines, linenum, error):
1244 """Logs an error if we see /* ... */ or "..." that extend past one line.
1246 /* ... */ comments are legit inside macros, for one line.
1247 Otherwise, we prefer // comments, so it's ok to warn about the
1248 other. Likewise, it's ok for strings to extend across multiple
1249 lines, as long as a line continuation character (backslash)
1250 terminates each line. Although not currently prohibited by the C++
1251 style guide, it's ugly and unnecessary. We don't do well with either
1252 in this lint program, so we warn about both.
1254 Args:
1255 filename: The name of the current file.
1256 clean_lines: A CleansedLines instance containing the file.
1257 linenum: The number of the line to check.
1258 error: The function to call with any errors found.
1260 line = clean_lines.elided[linenum]
1262 # Remove all \\ (escaped backslashes) from the line. They are OK, and the
1263 # second (escaped) slash may trigger later \" detection erroneously.
1264 line = line.replace('\\\\', '')
1266 if line.count('/*') > line.count('*/'):
1267 error(filename, linenum, 'readability/multiline_comment', 5,
1268 'Complex multi-line /*...*/-style comment found. '
1269 'Lint may give bogus warnings. '
1270 'Consider replacing these with //-style comments, '
1271 'with #if 0...#endif, '
1272 'or with more clearly structured multi-line comments.')
1274 if (line.count('"') - line.count('\\"')) % 2:
1275 error(filename, linenum, 'readability/multiline_string', 5,
1276 'Multi-line string ("...") found. This lint script doesn\'t '
1277 'do well with such strings, and may give bogus warnings. They\'re '
1278 'ugly and unnecessary, and you should use concatenation instead".')
1281 threading_list = (
1282 ('asctime(', 'asctime_r('),
1283 ('ctime(', 'ctime_r('),
1284 ('getgrgid(', 'getgrgid_r('),
1285 ('getgrnam(', 'getgrnam_r('),
1286 ('getlogin(', 'getlogin_r('),
1287 ('getpwnam(', 'getpwnam_r('),
1288 ('getpwuid(', 'getpwuid_r('),
1289 ('gmtime(', 'gmtime_r('),
1290 ('localtime(', 'localtime_r('),
1291 ('rand(', 'rand_r('),
1292 ('readdir(', 'readdir_r('),
1293 ('strtok(', 'strtok_r('),
1294 ('ttyname(', 'ttyname_r('),
1298 def CheckPosixThreading(filename, clean_lines, linenum, error):
1299 """Checks for calls to thread-unsafe functions.
1301 Much code has been originally written without consideration of
1302 multi-threading. Also, engineers are relying on their old experience;
1303 they have learned posix before threading extensions were added. These
1304 tests guide the engineers to use thread-safe functions (when using
1305 posix directly).
1307 Args:
1308 filename: The name of the current file.
1309 clean_lines: A CleansedLines instance containing the file.
1310 linenum: The number of the line to check.
1311 error: The function to call with any errors found.
1313 line = clean_lines.elided[linenum]
1314 for single_thread_function, multithread_safe_function in threading_list:
1315 ix = line.find(single_thread_function)
1316 # Comparisons made explicit for clarity -- pylint: disable-msg=C6403
1317 if ix >= 0 and (ix == 0 or (not line[ix - 1].isalnum() and
1318 line[ix - 1] not in ('_', '.', '>'))):
1319 error(filename, linenum, 'runtime/threadsafe_fn', 2,
1320 'Consider using ' + multithread_safe_function +
1321 '...) instead of ' + single_thread_function +
1322 '...) for improved thread safety.')
1325 # Matches invalid increment: *count++, which moves pointer instead of
1326 # incrementing a value.
1327 _RE_PATTERN_INVALID_INCREMENT = re.compile(
1328 r'^\s*\*\w+(\+\+|--);')
1331 def CheckInvalidIncrement(filename, clean_lines, linenum, error):
1332 """Checks for invalid increment *count++.
1334 For example following function:
1335 void increment_counter(int* count) {
1336 *count++;
1338 is invalid, because it effectively does count++, moving pointer, and should
1339 be replaced with ++*count, (*count)++ or *count += 1.
1341 Args:
1342 filename: The name of the current file.
1343 clean_lines: A CleansedLines instance containing the file.
1344 linenum: The number of the line to check.
1345 error: The function to call with any errors found.
1347 line = clean_lines.elided[linenum]
1348 if _RE_PATTERN_INVALID_INCREMENT.match(line):
1349 error(filename, linenum, 'runtime/invalid_increment', 5,
1350 'Changing pointer instead of value (or unused value of operator*).')
1353 class _BlockInfo(object):
1354 """Stores information about a generic block of code."""
1356 def __init__(self, seen_open_brace):
1357 self.seen_open_brace = seen_open_brace
1358 self.open_parentheses = 0
1359 self.inline_asm = _NO_ASM
1361 def CheckBegin(self, filename, clean_lines, linenum, error):
1362 """Run checks that applies to text up to the opening brace.
1364 This is mostly for checking the text after the class identifier
1365 and the "{", usually where the base class is specified. For other
1366 blocks, there isn't much to check, so we always pass.
1368 Args:
1369 filename: The name of the current file.
1370 clean_lines: A CleansedLines instance containing the file.
1371 linenum: The number of the line to check.
1372 error: The function to call with any errors found.
1374 pass
1376 def CheckEnd(self, filename, clean_lines, linenum, error):
1377 """Run checks that applies to text after the closing brace.
1379 This is mostly used for checking end of namespace comments.
1381 Args:
1382 filename: The name of the current file.
1383 clean_lines: A CleansedLines instance containing the file.
1384 linenum: The number of the line to check.
1385 error: The function to call with any errors found.
1387 pass
1390 class _ClassInfo(_BlockInfo):
1391 """Stores information about a class."""
1393 def __init__(self, name, class_or_struct, clean_lines, linenum):
1394 _BlockInfo.__init__(self, False)
1395 self.name = name
1396 self.starting_linenum = linenum
1397 self.is_derived = False
1398 if class_or_struct == 'struct':
1399 self.access = 'public'
1400 else:
1401 self.access = 'private'
1403 # Try to find the end of the class. This will be confused by things like:
1404 # class A {
1405 # } *x = { ...
1407 # But it's still good enough for CheckSectionSpacing.
1408 self.last_line = 0
1409 depth = 0
1410 for i in range(linenum, clean_lines.NumLines()):
1411 line = clean_lines.elided[i]
1412 depth += line.count('{') - line.count('}')
1413 if not depth:
1414 self.last_line = i
1415 break
1417 def CheckBegin(self, filename, clean_lines, linenum, error):
1418 # Look for a bare ':'
1419 if Search('(^|[^:]):($|[^:])', clean_lines.elided[linenum]):
1420 self.is_derived = True
1423 class _NamespaceInfo(_BlockInfo):
1424 """Stores information about a namespace."""
1426 def __init__(self, name, linenum):
1427 _BlockInfo.__init__(self, False)
1428 self.name = name or ''
1429 self.starting_linenum = linenum
1431 def CheckEnd(self, filename, clean_lines, linenum, error):
1432 """Check end of namespace comments."""
1433 line = clean_lines.raw_lines[linenum]
1435 # Check how many lines is enclosed in this namespace. Don't issue
1436 # warning for missing namespace comments if there aren't enough
1437 # lines. However, do apply checks if there is already an end of
1438 # namespace comment and it's incorrect.
1440 # TODO(unknown): We always want to check end of namespace comments
1441 # if a namespace is large, but sometimes we also want to apply the
1442 # check if a short namespace contained nontrivial things (something
1443 # other than forward declarations). There is currently no logic on
1444 # deciding what these nontrivial things are, so this check is
1445 # triggered by namespace size only, which works most of the time.
1446 if (linenum - self.starting_linenum < 10
1447 and not Match(r'};*\s*(//|/\*).*\bnamespace\b', line)):
1448 return
1450 # Look for matching comment at end of namespace.
1452 # Note that we accept C style "/* */" comments for terminating
1453 # namespaces, so that code that terminate namespaces inside
1454 # preprocessor macros can be cpplint clean. Example: http://go/nxpiz
1456 # We also accept stuff like "// end of namespace <name>." with the
1457 # period at the end.
1459 # Besides these, we don't accept anything else, otherwise we might
1460 # get false negatives when existing comment is a substring of the
1461 # expected namespace. Example: http://go/ldkdc, http://cl/23548205
1462 if self.name:
1463 # Named namespace
1464 if not Match((r'};*\s*(//|/\*).*\bnamespace\s+' + re.escape(self.name) +
1465 r'[\*/\.\\\s]*$'),
1466 line):
1467 error(filename, linenum, 'readability/namespace', 5,
1468 'Namespace should be terminated with "// namespace %s"' %
1469 self.name)
1470 else:
1471 # Anonymous namespace
1472 if not Match(r'};*\s*(//|/\*).*\bnamespace[\*/\.\\\s]*$', line):
1473 error(filename, linenum, 'readability/namespace', 5,
1474 'Namespace should be terminated with "// namespace"')
1477 class _PreprocessorInfo(object):
1478 """Stores checkpoints of nesting stacks when #if/#else is seen."""
1480 def __init__(self, stack_before_if):
1481 # The entire nesting stack before #if
1482 self.stack_before_if = stack_before_if
1484 # The entire nesting stack up to #else
1485 self.stack_before_else = []
1487 # Whether we have already seen #else or #elif
1488 self.seen_else = False
1491 class _NestingState(object):
1492 """Holds states related to parsing braces."""
1494 def __init__(self):
1495 # Stack for tracking all braces. An object is pushed whenever we
1496 # see a "{", and popped when we see a "}". Only 3 types of
1497 # objects are possible:
1498 # - _ClassInfo: a class or struct.
1499 # - _NamespaceInfo: a namespace.
1500 # - _BlockInfo: some other type of block.
1501 self.stack = []
1503 # Stack of _PreprocessorInfo objects.
1504 self.pp_stack = []
1506 def SeenOpenBrace(self):
1507 """Check if we have seen the opening brace for the innermost block.
1509 Returns:
1510 True if we have seen the opening brace, False if the innermost
1511 block is still expecting an opening brace.
1513 return (not self.stack) or self.stack[-1].seen_open_brace
1515 def InNamespaceBody(self):
1516 """Check if we are currently one level inside a namespace body.
1518 Returns:
1519 True if top of the stack is a namespace block, False otherwise.
1521 return self.stack and isinstance(self.stack[-1], _NamespaceInfo)
1523 def UpdatePreprocessor(self, line):
1524 """Update preprocessor stack.
1526 We need to handle preprocessors due to classes like this:
1527 #ifdef SWIG
1528 struct ResultDetailsPageElementExtensionPoint {
1529 #else
1530 struct ResultDetailsPageElementExtensionPoint : public Extension {
1531 #endif
1532 (see http://go/qwddn for original example)
1534 We make the following assumptions (good enough for most files):
1535 - Preprocessor condition evaluates to true from #if up to first
1536 #else/#elif/#endif.
1538 - Preprocessor condition evaluates to false from #else/#elif up
1539 to #endif. We still perform lint checks on these lines, but
1540 these do not affect nesting stack.
1542 Args:
1543 line: current line to check.
1545 if Match(r'^\s*#\s*(if|ifdef|ifndef)\b', line):
1546 # Beginning of #if block, save the nesting stack here. The saved
1547 # stack will allow us to restore the parsing state in the #else case.
1548 self.pp_stack.append(_PreprocessorInfo(copy.deepcopy(self.stack)))
1549 elif Match(r'^\s*#\s*(else|elif)\b', line):
1550 # Beginning of #else block
1551 if self.pp_stack:
1552 if not self.pp_stack[-1].seen_else:
1553 # This is the first #else or #elif block. Remember the
1554 # whole nesting stack up to this point. This is what we
1555 # keep after the #endif.
1556 self.pp_stack[-1].seen_else = True
1557 self.pp_stack[-1].stack_before_else = copy.deepcopy(self.stack)
1559 # Restore the stack to how it was before the #if
1560 self.stack = copy.deepcopy(self.pp_stack[-1].stack_before_if)
1561 else:
1562 # TODO(unknown): unexpected #else, issue warning?
1563 pass
1564 elif Match(r'^\s*#\s*endif\b', line):
1565 # End of #if or #else blocks.
1566 if self.pp_stack:
1567 # If we saw an #else, we will need to restore the nesting
1568 # stack to its former state before the #else, otherwise we
1569 # will just continue from where we left off.
1570 if self.pp_stack[-1].seen_else:
1571 # Here we can just use a shallow copy since we are the last
1572 # reference to it.
1573 self.stack = self.pp_stack[-1].stack_before_else
1574 # Drop the corresponding #if
1575 self.pp_stack.pop()
1576 else:
1577 # TODO(unknown): unexpected #endif, issue warning?
1578 pass
1580 def Update(self, filename, clean_lines, linenum, error):
1581 """Update nesting state with current line.
1583 Args:
1584 filename: The name of the current file.
1585 clean_lines: A CleansedLines instance containing the file.
1586 linenum: The number of the line to check.
1587 error: The function to call with any errors found.
1589 line = clean_lines.elided[linenum]
1591 # Update pp_stack first
1592 self.UpdatePreprocessor(line)
1594 # Count parentheses. This is to avoid adding struct arguments to
1595 # the nesting stack.
1596 if self.stack:
1597 inner_block = self.stack[-1]
1598 depth_change = line.count('(') - line.count(')')
1599 inner_block.open_parentheses += depth_change
1601 # Also check if we are starting or ending an inline assembly block.
1602 if inner_block.inline_asm in (_NO_ASM, _END_ASM):
1603 if (depth_change != 0 and
1604 inner_block.open_parentheses == 1 and
1605 _MATCH_ASM.match(line)):
1606 # Enter assembly block
1607 inner_block.inline_asm = _INSIDE_ASM
1608 else:
1609 # Not entering assembly block. If previous line was _END_ASM,
1610 # we will now shift to _NO_ASM state.
1611 inner_block.inline_asm = _NO_ASM
1612 elif (inner_block.inline_asm == _INSIDE_ASM and
1613 inner_block.open_parentheses == 0):
1614 # Exit assembly block
1615 inner_block.inline_asm = _END_ASM
1617 # Consume namespace declaration at the beginning of the line. Do
1618 # this in a loop so that we catch same line declarations like this:
1619 # namespace proto2 { namespace bridge { class MessageSet; } }
1620 while True:
1621 # Match start of namespace. The "\b\s*" below catches namespace
1622 # declarations even if it weren't followed by a whitespace, this
1623 # is so that we don't confuse our namespace checker. The
1624 # missing spaces will be flagged by CheckSpacing.
1625 namespace_decl_match = Match(r'^\s*namespace\b\s*([:\w]+)?(.*)$', line)
1626 if not namespace_decl_match:
1627 break
1629 new_namespace = _NamespaceInfo(namespace_decl_match.group(1), linenum)
1630 self.stack.append(new_namespace)
1632 line = namespace_decl_match.group(2)
1633 if line.find('{') != -1:
1634 new_namespace.seen_open_brace = True
1635 line = line[line.find('{') + 1:]
1637 # Look for a class declaration in whatever is left of the line
1638 # after parsing namespaces. The regexp accounts for decorated classes
1639 # such as in:
1640 # class LOCKABLE API Object {
1641 # };
1643 # Templates with class arguments may confuse the parser, for example:
1644 # template <class T
1645 # class Comparator = less<T>,
1646 # class Vector = vector<T> >
1647 # class HeapQueue {
1649 # Because this parser has no nesting state about templates, by the
1650 # time it saw "class Comparator", it may think that it's a new class.
1651 # Nested templates have a similar problem:
1652 # template <
1653 # typename ExportedType,
1654 # typename TupleType,
1655 # template <typename, typename> class ImplTemplate>
1657 # To avoid these cases, we ignore classes that are followed by '=' or '>'
1658 class_decl_match = Match(
1659 r'\s*(template\s*<[\w\s<>,:]*>\s*)?'
1660 '(class|struct)\s+([A-Z_]+\s+)*(\w+(?:::\w+)*)'
1661 '(([^=>]|<[^<>]*>)*)$', line)
1662 if (class_decl_match and
1663 (not self.stack or self.stack[-1].open_parentheses == 0)):
1664 self.stack.append(_ClassInfo(
1665 class_decl_match.group(4), class_decl_match.group(2),
1666 clean_lines, linenum))
1667 line = class_decl_match.group(5)
1669 # If we have not yet seen the opening brace for the innermost block,
1670 # run checks here.
1671 if not self.SeenOpenBrace():
1672 self.stack[-1].CheckBegin(filename, clean_lines, linenum, error)
1674 # Update access control if we are inside a class/struct
1675 if self.stack and isinstance(self.stack[-1], _ClassInfo):
1676 access_match = Match(r'\s*(public|private|protected)\s*:', line)
1677 if access_match:
1678 self.stack[-1].access = access_match.group(1)
1680 # Consume braces or semicolons from what's left of the line
1681 while True:
1682 # Match first brace, semicolon, or closed parenthesis.
1683 matched = Match(r'^[^{;)}]*([{;)}])(.*)$', line)
1684 if not matched:
1685 break
1687 token = matched.group(1)
1688 if token == '{':
1689 # If namespace or class hasn't seen a opening brace yet, mark
1690 # namespace/class head as complete. Push a new block onto the
1691 # stack otherwise.
1692 if not self.SeenOpenBrace():
1693 self.stack[-1].seen_open_brace = True
1694 else:
1695 self.stack.append(_BlockInfo(True))
1696 if _MATCH_ASM.match(line):
1697 self.stack[-1].inline_asm = _BLOCK_ASM
1698 elif token == ';' or token == ')':
1699 # If we haven't seen an opening brace yet, but we already saw
1700 # a semicolon, this is probably a forward declaration. Pop
1701 # the stack for these.
1703 # Similarly, if we haven't seen an opening brace yet, but we
1704 # already saw a closing parenthesis, then these are probably
1705 # function arguments with extra "class" or "struct" keywords.
1706 # Also pop these stack for these.
1707 if not self.SeenOpenBrace():
1708 self.stack.pop()
1709 else: # token == '}'
1710 # Perform end of block checks and pop the stack.
1711 if self.stack:
1712 self.stack[-1].CheckEnd(filename, clean_lines, linenum, error)
1713 self.stack.pop()
1714 line = matched.group(2)
1716 def InnermostClass(self):
1717 """Get class info on the top of the stack.
1719 Returns:
1720 A _ClassInfo object if we are inside a class, or None otherwise.
1722 for i in range(len(self.stack), 0, -1):
1723 classinfo = self.stack[i - 1]
1724 if isinstance(classinfo, _ClassInfo):
1725 return classinfo
1726 return None
1728 def CheckClassFinished(self, filename, error):
1729 """Checks that all classes have been completely parsed.
1731 Call this when all lines in a file have been processed.
1732 Args:
1733 filename: The name of the current file.
1734 error: The function to call with any errors found.
1736 # Note: This test can result in false positives if #ifdef constructs
1737 # get in the way of brace matching. See the testBuildClass test in
1738 # cpplint_unittest.py for an example of this.
1739 for obj in self.stack:
1740 if isinstance(obj, _ClassInfo):
1741 error(filename, obj.starting_linenum, 'build/class', 5,
1742 'Failed to find complete declaration of class %s' %
1743 obj.name)
1746 def CheckForNonStandardConstructs(filename, clean_lines, linenum,
1747 nesting_state, error):
1748 """Logs an error if we see certain non-ANSI constructs ignored by gcc-2.
1750 Complain about several constructs which gcc-2 accepts, but which are
1751 not standard C++. Warning about these in lint is one way to ease the
1752 transition to new compilers.
1753 - put storage class first (e.g. "static const" instead of "const static").
1754 - "%lld" instead of %qd" in printf-type functions.
1755 - "%1$d" is non-standard in printf-type functions.
1756 - "\%" is an undefined character escape sequence.
1757 - text after #endif is not allowed.
1758 - invalid inner-style forward declaration.
1759 - >? and <? operators, and their >?= and <?= cousins.
1761 Additionally, check for constructor/destructor style violations and reference
1762 members, as it is very convenient to do so while checking for
1763 gcc-2 compliance.
1765 Args:
1766 filename: The name of the current file.
1767 clean_lines: A CleansedLines instance containing the file.
1768 linenum: The number of the line to check.
1769 nesting_state: A _NestingState instance which maintains information about
1770 the current stack of nested blocks being parsed.
1771 error: A callable to which errors are reported, which takes 4 arguments:
1772 filename, line number, error level, and message
1775 # Remove comments from the line, but leave in strings for now.
1776 line = clean_lines.lines[linenum]
1778 if Search(r'printf\s*\(.*".*%[-+ ]?\d*q', line):
1779 error(filename, linenum, 'runtime/printf_format', 3,
1780 '%q in format strings is deprecated. Use %ll instead.')
1782 if Search(r'printf\s*\(.*".*%\d+\$', line):
1783 error(filename, linenum, 'runtime/printf_format', 2,
1784 '%N$ formats are unconventional. Try rewriting to avoid them.')
1786 # Remove escaped backslashes before looking for undefined escapes.
1787 line = line.replace('\\\\', '')
1789 if Search(r'("|\').*\\(%|\[|\(|{)', line):
1790 error(filename, linenum, 'build/printf_format', 3,
1791 '%, [, (, and { are undefined character escapes. Unescape them.')
1793 # For the rest, work with both comments and strings removed.
1794 line = clean_lines.elided[linenum]
1796 if Search(r'\b(const|volatile|void|char|short|int|long'
1797 r'|float|double|signed|unsigned'
1798 r'|schar|u?int8|u?int16|u?int32|u?int64)'
1799 r'\s+(register|static|extern|typedef)\b',
1800 line):
1801 error(filename, linenum, 'build/storage_class', 5,
1802 'Storage class (static, extern, typedef, etc) should be first.')
1804 if Match(r'\s*#\s*endif\s*[^/\s]+', line):
1805 error(filename, linenum, 'build/endif_comment', 5,
1806 'Uncommented text after #endif is non-standard. Use a comment.')
1808 if Match(r'\s*class\s+(\w+\s*::\s*)+\w+\s*;', line):
1809 error(filename, linenum, 'build/forward_decl', 5,
1810 'Inner-style forward declarations are invalid. Remove this line.')
1812 if Search(r'(\w+|[+-]?\d+(\.\d*)?)\s*(<|>)\?=?\s*(\w+|[+-]?\d+)(\.\d*)?',
1813 line):
1814 error(filename, linenum, 'build/deprecated', 3,
1815 '>? and <? (max and min) operators are non-standard and deprecated.')
1817 if Search(r'^\s*const\s*string\s*&\s*\w+\s*;', line):
1818 # TODO(unknown): Could it be expanded safely to arbitrary references,
1819 # without triggering too many false positives? The first
1820 # attempt triggered 5 warnings for mostly benign code in the regtest, hence
1821 # the restriction.
1822 # Here's the original regexp, for the reference:
1823 # type_name = r'\w+((\s*::\s*\w+)|(\s*<\s*\w+?\s*>))?'
1824 # r'\s*const\s*' + type_name + '\s*&\s*\w+\s*;'
1825 error(filename, linenum, 'runtime/member_string_references', 2,
1826 'const string& members are dangerous. It is much better to use '
1827 'alternatives, such as pointers or simple constants.')
1829 # Everything else in this function operates on class declarations.
1830 # Return early if the top of the nesting stack is not a class, or if
1831 # the class head is not completed yet.
1832 classinfo = nesting_state.InnermostClass()
1833 if not classinfo or not classinfo.seen_open_brace:
1834 return
1836 # The class may have been declared with namespace or classname qualifiers.
1837 # The constructor and destructor will not have those qualifiers.
1838 base_classname = classinfo.name.split('::')[-1]
1840 # Look for single-argument constructors that aren't marked explicit.
1841 # Technically a valid construct, but against style.
1842 args = Match(r'\s+(?:inline\s+)?%s\s*\(([^,()]+)\)'
1843 % re.escape(base_classname),
1844 line)
1845 if (args and
1846 args.group(1) != 'void' and
1847 not Match(r'(const\s+)?%s\s*(?:<\w+>\s*)?&' % re.escape(base_classname),
1848 args.group(1).strip())):
1849 error(filename, linenum, 'runtime/explicit', 5,
1850 'Single-argument constructors should be marked explicit.')
1853 def CheckSpacingForFunctionCall(filename, line, linenum, error):
1854 """Checks for the correctness of various spacing around function calls.
1856 Args:
1857 filename: The name of the current file.
1858 line: The text of the line to check.
1859 linenum: The number of the line to check.
1860 error: The function to call with any errors found.
1863 # Since function calls often occur inside if/for/while/switch
1864 # expressions - which have their own, more liberal conventions - we
1865 # first see if we should be looking inside such an expression for a
1866 # function call, to which we can apply more strict standards.
1867 fncall = line # if there's no control flow construct, look at whole line
1868 for pattern in (r'\bif\s*\((.*)\)\s*{',
1869 r'\bfor\s*\((.*)\)\s*{',
1870 r'\bwhile\s*\((.*)\)\s*[{;]',
1871 r'\bswitch\s*\((.*)\)\s*{'):
1872 match = Search(pattern, line)
1873 if match:
1874 fncall = match.group(1) # look inside the parens for function calls
1875 break
1877 # Except in if/for/while/switch, there should never be space
1878 # immediately inside parens (eg "f( 3, 4 )"). We make an exception
1879 # for nested parens ( (a+b) + c ). Likewise, there should never be
1880 # a space before a ( when it's a function argument. I assume it's a
1881 # function argument when the char before the whitespace is legal in
1882 # a function name (alnum + _) and we're not starting a macro. Also ignore
1883 # pointers and references to arrays and functions coz they're too tricky:
1884 # we use a very simple way to recognize these:
1885 # " (something)(maybe-something)" or
1886 # " (something)(maybe-something," or
1887 # " (something)[something]"
1888 # Note that we assume the contents of [] to be short enough that
1889 # they'll never need to wrap.
1890 if ( # Ignore control structures.
1891 not Search(r'\b(if|for|while|switch|return|delete)\b', fncall) and
1892 # Ignore pointers/references to functions.
1893 not Search(r' \([^)]+\)\([^)]*(\)|,$)', fncall) and
1894 # Ignore pointers/references to arrays.
1895 not Search(r' \([^)]+\)\[[^\]]+\]', fncall)):
1896 if Search(r'\w\s*\(\s(?!\s*\\$)', fncall): # a ( used for a fn call
1897 error(filename, linenum, 'whitespace/parens', 4,
1898 'Extra space after ( in function call')
1899 elif Search(r'\(\s+(?!(\s*\\)|\()', fncall):
1900 error(filename, linenum, 'whitespace/parens', 2,
1901 'Extra space after (')
1902 if (Search(r'\w\s+\(', fncall) and
1903 not Search(r'#\s*define|typedef', fncall) and
1904 not Search(r'\w\s+\((\w+::)?\*\w+\)\(', fncall)):
1905 error(filename, linenum, 'whitespace/parens', 4,
1906 'Extra space before ( in function call')
1907 # If the ) is followed only by a newline or a { + newline, assume it's
1908 # part of a control statement (if/while/etc), and don't complain
1909 if Search(r'[^)]\s+\)\s*[^{\s]', fncall):
1910 # If the closing parenthesis is preceded by only whitespaces,
1911 # try to give a more descriptive error message.
1912 if Search(r'^\s+\)', fncall):
1913 error(filename, linenum, 'whitespace/parens', 2,
1914 'Closing ) should be moved to the previous line')
1915 else:
1916 error(filename, linenum, 'whitespace/parens', 2,
1917 'Extra space before )')
1920 def IsBlankLine(line):
1921 """Returns true if the given line is blank.
1923 We consider a line to be blank if the line is empty or consists of
1924 only white spaces.
1926 Args:
1927 line: A line of a string.
1929 Returns:
1930 True, if the given line is blank.
1932 return not line or line.isspace()
1935 def CheckForFunctionLengths(filename, clean_lines, linenum,
1936 function_state, error):
1937 """Reports for long function bodies.
1939 For an overview why this is done, see:
1940 http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions
1942 Uses a simplistic algorithm assuming other style guidelines
1943 (especially spacing) are followed.
1944 Only checks unindented functions, so class members are unchecked.
1945 Trivial bodies are unchecked, so constructors with huge initializer lists
1946 may be missed.
1947 Blank/comment lines are not counted so as to avoid encouraging the removal
1948 of vertical space and comments just to get through a lint check.
1949 NOLINT *on the last line of a function* disables this check.
1951 Args:
1952 filename: The name of the current file.
1953 clean_lines: A CleansedLines instance containing the file.
1954 linenum: The number of the line to check.
1955 function_state: Current function name and lines in body so far.
1956 error: The function to call with any errors found.
1958 lines = clean_lines.lines
1959 line = lines[linenum]
1960 raw = clean_lines.raw_lines
1961 raw_line = raw[linenum]
1962 joined_line = ''
1964 starting_func = False
1965 regexp = r'(\w(\w|::|\*|\&|\s)*)\(' # decls * & space::name( ...
1966 match_result = Match(regexp, line)
1967 if match_result:
1968 # If the name is all caps and underscores, figure it's a macro and
1969 # ignore it, unless it's TEST or TEST_F.
1970 function_name = match_result.group(1).split()[-1]
1971 if function_name == 'TEST' or function_name == 'TEST_F' or (
1972 not Match(r'[A-Z_]+$', function_name)):
1973 starting_func = True
1975 if starting_func:
1976 body_found = False
1977 for start_linenum in xrange(linenum, clean_lines.NumLines()):
1978 start_line = lines[start_linenum]
1979 joined_line += ' ' + start_line.lstrip()
1980 if Search(r'(;|})', start_line): # Declarations and trivial functions
1981 body_found = True
1982 break # ... ignore
1983 elif Search(r'{', start_line):
1984 body_found = True
1985 function = Search(r'((\w|:)*)\(', line).group(1)
1986 if Match(r'TEST', function): # Handle TEST... macros
1987 parameter_regexp = Search(r'(\(.*\))', joined_line)
1988 if parameter_regexp: # Ignore bad syntax
1989 function += parameter_regexp.group(1)
1990 else:
1991 function += '()'
1992 function_state.Begin(function)
1993 break
1994 if not body_found:
1995 # No body for the function (or evidence of a non-function) was found.
1996 error(filename, linenum, 'readability/fn_size', 5,
1997 'Lint failed to find start of function body.')
1998 elif Match(r'^\}\s*$', line): # function end
1999 function_state.Check(error, filename, linenum)
2000 function_state.End()
2001 elif not Match(r'^\s*$', line):
2002 function_state.Count() # Count non-blank/non-comment lines.
2005 _RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?')
2008 def CheckComment(comment, filename, linenum, error):
2009 """Checks for common mistakes in TODO comments.
2011 Args:
2012 comment: The text of the comment from the line in question.
2013 filename: The name of the current file.
2014 linenum: The number of the line to check.
2015 error: The function to call with any errors found.
2017 match = _RE_PATTERN_TODO.match(comment)
2018 if match:
2019 # One whitespace is correct; zero whitespace is handled elsewhere.
2020 leading_whitespace = match.group(1)
2021 if len(leading_whitespace) > 1:
2022 error(filename, linenum, 'whitespace/todo', 2,
2023 'Too many spaces before TODO')
2025 username = match.group(2)
2026 if not username:
2027 error(filename, linenum, 'readability/todo', 2,
2028 'Missing username in TODO; it should look like '
2029 '"// TODO(my_username): Stuff."')
2031 middle_whitespace = match.group(3)
2032 # Comparisons made explicit for correctness -- pylint: disable-msg=C6403
2033 if middle_whitespace != ' ' and middle_whitespace != '':
2034 error(filename, linenum, 'whitespace/todo', 2,
2035 'TODO(my_username) should be followed by a space')
2037 def CheckAccess(filename, clean_lines, linenum, nesting_state, error):
2038 """Checks for improper use of DISALLOW* macros.
2040 Args:
2041 filename: The name of the current file.
2042 clean_lines: A CleansedLines instance containing the file.
2043 linenum: The number of the line to check.
2044 nesting_state: A _NestingState instance which maintains information about
2045 the current stack of nested blocks being parsed.
2046 error: The function to call with any errors found.
2048 line = clean_lines.elided[linenum] # get rid of comments and strings
2050 matched = Match((r'\s*(DISALLOW_COPY_AND_ASSIGN|'
2051 r'DISALLOW_EVIL_CONSTRUCTORS|'
2052 r'DISALLOW_IMPLICIT_CONSTRUCTORS)'), line)
2053 if not matched:
2054 return
2055 if nesting_state.stack and isinstance(nesting_state.stack[-1], _ClassInfo):
2056 if nesting_state.stack[-1].access != 'private':
2057 error(filename, linenum, 'readability/constructors', 3,
2058 '%s must be in the private: section' % matched.group(1))
2060 else:
2061 # Found DISALLOW* macro outside a class declaration, or perhaps it
2062 # was used inside a function when it should have been part of the
2063 # class declaration. We could issue a warning here, but it
2064 # probably resulted in a compiler error already.
2065 pass
2068 def FindNextMatchingAngleBracket(clean_lines, linenum, init_suffix):
2069 """Find the corresponding > to close a template.
2071 Args:
2072 clean_lines: A CleansedLines instance containing the file.
2073 linenum: Current line number.
2074 init_suffix: Remainder of the current line after the initial <.
2076 Returns:
2077 True if a matching bracket exists.
2079 line = init_suffix
2080 nesting_stack = ['<']
2081 while True:
2082 # Find the next operator that can tell us whether < is used as an
2083 # opening bracket or as a less-than operator. We only want to
2084 # warn on the latter case.
2086 # We could also check all other operators and terminate the search
2087 # early, e.g. if we got something like this "a<b+c", the "<" is
2088 # most likely a less-than operator, but then we will get false
2089 # positives for default arguments (e.g. http://go/prccd) and
2090 # other template expressions (e.g. http://go/oxcjq).
2091 match = Search(r'^[^<>(),;\[\]]*([<>(),;\[\]])(.*)$', line)
2092 if match:
2093 # Found an operator, update nesting stack
2094 operator = match.group(1)
2095 line = match.group(2)
2097 if nesting_stack[-1] == '<':
2098 # Expecting closing angle bracket
2099 if operator in ('<', '(', '['):
2100 nesting_stack.append(operator)
2101 elif operator == '>':
2102 nesting_stack.pop()
2103 if not nesting_stack:
2104 # Found matching angle bracket
2105 return True
2106 elif operator == ',':
2107 # Got a comma after a bracket, this is most likely a template
2108 # argument. We have not seen a closing angle bracket yet, but
2109 # it's probably a few lines later if we look for it, so just
2110 # return early here.
2111 return True
2112 else:
2113 # Got some other operator.
2114 return False
2116 else:
2117 # Expecting closing parenthesis or closing bracket
2118 if operator in ('<', '(', '['):
2119 nesting_stack.append(operator)
2120 elif operator in (')', ']'):
2121 # We don't bother checking for matching () or []. If we got
2122 # something like (] or [), it would have been a syntax error.
2123 nesting_stack.pop()
2125 else:
2126 # Scan the next line
2127 linenum += 1
2128 if linenum >= len(clean_lines.elided):
2129 break
2130 line = clean_lines.elided[linenum]
2132 # Exhausted all remaining lines and still no matching angle bracket.
2133 # Most likely the input was incomplete, otherwise we should have
2134 # seen a semicolon and returned early.
2135 return True
2138 def FindPreviousMatchingAngleBracket(clean_lines, linenum, init_prefix):
2139 """Find the corresponding < that started a template.
2141 Args:
2142 clean_lines: A CleansedLines instance containing the file.
2143 linenum: Current line number.
2144 init_prefix: Part of the current line before the initial >.
2146 Returns:
2147 True if a matching bracket exists.
2149 line = init_prefix
2150 nesting_stack = ['>']
2151 while True:
2152 # Find the previous operator
2153 match = Search(r'^(.*)([<>(),;\[\]])[^<>(),;\[\]]*$', line)
2154 if match:
2155 # Found an operator, update nesting stack
2156 operator = match.group(2)
2157 line = match.group(1)
2159 if nesting_stack[-1] == '>':
2160 # Expecting opening angle bracket
2161 if operator in ('>', ')', ']'):
2162 nesting_stack.append(operator)
2163 elif operator == '<':
2164 nesting_stack.pop()
2165 if not nesting_stack:
2166 # Found matching angle bracket
2167 return True
2168 elif operator == ',':
2169 # Got a comma before a bracket, this is most likely a
2170 # template argument. The opening angle bracket is probably
2171 # there if we look for it, so just return early here.
2172 return True
2173 else:
2174 # Got some other operator.
2175 return False
2177 else:
2178 # Expecting opening parenthesis or opening bracket
2179 if operator in ('>', ')', ']'):
2180 nesting_stack.append(operator)
2181 elif operator in ('(', '['):
2182 nesting_stack.pop()
2184 else:
2185 # Scan the previous line
2186 linenum -= 1
2187 if linenum < 0:
2188 break
2189 line = clean_lines.elided[linenum]
2191 # Exhausted all earlier lines and still no matching angle bracket.
2192 return False
2195 def CheckSpacing(filename, clean_lines, linenum, nesting_state, error):
2196 """Checks for the correctness of various spacing issues in the code.
2198 Things we check for: spaces around operators, spaces after
2199 if/for/while/switch, no spaces around parens in function calls, two
2200 spaces between code and comment, don't start a block with a blank
2201 line, don't end a function with a blank line, don't add a blank line
2202 after public/protected/private, don't have too many blank lines in a row.
2204 Args:
2205 filename: The name of the current file.
2206 clean_lines: A CleansedLines instance containing the file.
2207 linenum: The number of the line to check.
2208 nesting_state: A _NestingState instance which maintains information about
2209 the current stack of nested blocks being parsed.
2210 error: The function to call with any errors found.
2213 raw = clean_lines.raw_lines
2214 line = raw[linenum]
2216 # Before nixing comments, check if the line is blank for no good
2217 # reason. This includes the first line after a block is opened, and
2218 # blank lines at the end of a function (ie, right before a line like '}'
2220 # Skip all the blank line checks if we are immediately inside a
2221 # namespace body. In other words, don't issue blank line warnings
2222 # for this block:
2223 # namespace {
2227 # A warning about missing end of namespace comments will be issued instead.
2228 if IsBlankLine(line) and not nesting_state.InNamespaceBody():
2229 elided = clean_lines.elided
2230 prev_line = elided[linenum - 1]
2231 prevbrace = prev_line.rfind('{')
2232 # TODO(unknown): Don't complain if line before blank line, and line after,
2233 # both start with alnums and are indented the same amount.
2234 # This ignores whitespace at the start of a namespace block
2235 # because those are not usually indented.
2236 if prevbrace != -1 and prev_line[prevbrace:].find('}') == -1:
2237 # OK, we have a blank line at the start of a code block. Before we
2238 # complain, we check if it is an exception to the rule: The previous
2239 # non-empty line has the parameters of a function header that are indented
2240 # 4 spaces (because they did not fit in a 80 column line when placed on
2241 # the same line as the function name). We also check for the case where
2242 # the previous line is indented 6 spaces, which may happen when the
2243 # initializers of a constructor do not fit into a 80 column line.
2244 exception = False
2245 if Match(r' {6}\w', prev_line): # Initializer list?
2246 # We are looking for the opening column of initializer list, which
2247 # should be indented 4 spaces to cause 6 space indentation afterwards.
2248 search_position = linenum-2
2249 while (search_position >= 0
2250 and Match(r' {6}\w', elided[search_position])):
2251 search_position -= 1
2252 exception = (search_position >= 0
2253 and elided[search_position][:5] == ' :')
2254 else:
2255 # Search for the function arguments or an initializer list. We use a
2256 # simple heuristic here: If the line is indented 4 spaces; and we have a
2257 # closing paren, without the opening paren, followed by an opening brace
2258 # or colon (for initializer lists) we assume that it is the last line of
2259 # a function header. If we have a colon indented 4 spaces, it is an
2260 # initializer list.
2261 exception = (Match(r' {4}\w[^\(]*\)\s*(const\s*)?(\{\s*$|:)',
2262 prev_line)
2263 or Match(r' {4}:', prev_line))
2265 if not exception:
2266 error(filename, linenum, 'whitespace/blank_line', 2,
2267 'Blank line at the start of a code block. Is this needed?')
2268 # Ignore blank lines at the end of a block in a long if-else
2269 # chain, like this:
2270 # if (condition1) {
2271 # // Something followed by a blank line
2273 # } else if (condition2) {
2274 # // Something else
2276 if linenum + 1 < clean_lines.NumLines():
2277 next_line = raw[linenum + 1]
2278 if (next_line
2279 and Match(r'\s*}', next_line)
2280 and next_line.find('} else ') == -1):
2281 error(filename, linenum, 'whitespace/blank_line', 3,
2282 'Blank line at the end of a code block. Is this needed?')
2284 matched = Match(r'\s*(public|protected|private):', prev_line)
2285 if matched:
2286 error(filename, linenum, 'whitespace/blank_line', 3,
2287 'Do not leave a blank line after "%s:"' % matched.group(1))
2289 # Next, we complain if there's a comment too near the text
2290 commentpos = line.find('//')
2291 if commentpos != -1:
2292 # Check if the // may be in quotes. If so, ignore it
2293 # Comparisons made explicit for clarity -- pylint: disable-msg=C6403
2294 if (line.count('"', 0, commentpos) -
2295 line.count('\\"', 0, commentpos)) % 2 == 0: # not in quotes
2296 # Allow one space for new scopes, two spaces otherwise:
2297 if (not Match(r'^\s*{ //', line) and
2298 ((commentpos >= 1 and
2299 line[commentpos-1] not in string.whitespace) or
2300 (commentpos >= 2 and
2301 line[commentpos-2] not in string.whitespace))):
2302 error(filename, linenum, 'whitespace/comments', 2,
2303 'At least two spaces is best between code and comments')
2304 # There should always be a space between the // and the comment
2305 commentend = commentpos + 2
2306 if commentend < len(line) and not line[commentend] == ' ':
2307 # but some lines are exceptions -- e.g. if they're big
2308 # comment delimiters like:
2309 # //----------------------------------------------------------
2310 # or are an empty C++ style Doxygen comment, like:
2311 # ///
2312 # or they begin with multiple slashes followed by a space:
2313 # //////// Header comment
2314 match = (Search(r'[=/-]{4,}\s*$', line[commentend:]) or
2315 Search(r'^/$', line[commentend:]) or
2316 Search(r'^/+ ', line[commentend:]))
2317 if not match:
2318 error(filename, linenum, 'whitespace/comments', 4,
2319 'Should have a space between // and comment')
2320 CheckComment(line[commentpos:], filename, linenum, error)
2322 line = clean_lines.elided[linenum] # get rid of comments and strings
2324 # Don't try to do spacing checks for operator methods
2325 line = re.sub(r'operator(==|!=|<|<<|<=|>=|>>|>)\(', 'operator\(', line)
2327 # We allow no-spaces around = within an if: "if ( (a=Foo()) == 0 )".
2328 # Otherwise not. Note we only check for non-spaces on *both* sides;
2329 # sometimes people put non-spaces on one side when aligning ='s among
2330 # many lines (not that this is behavior that I approve of...)
2331 if Search(r'[\w.]=[\w.]', line) and not Search(r'\b(if|while) ', line):
2332 error(filename, linenum, 'whitespace/operators', 4,
2333 'Missing spaces around =')
2335 # It's ok not to have spaces around binary operators like + - * /, but if
2336 # there's too little whitespace, we get concerned. It's hard to tell,
2337 # though, so we punt on this one for now. TODO.
2339 # You should always have whitespace around binary operators.
2341 # Check <= and >= first to avoid false positives with < and >, then
2342 # check non-include lines for spacing around < and >.
2343 match = Search(r'[^<>=!\s](==|!=|<=|>=)[^<>=!\s]', line)
2344 if match:
2345 error(filename, linenum, 'whitespace/operators', 3,
2346 'Missing spaces around %s' % match.group(1))
2347 # We allow no-spaces around << when used like this: 10<<20, but
2348 # not otherwise (particularly, not when used as streams)
2349 match = Search(r'(\S)(?:L|UL|ULL|l|ul|ull)?<<(\S)', line)
2350 if match and not (match.group(1).isdigit() and match.group(2).isdigit()):
2351 error(filename, linenum, 'whitespace/operators', 3,
2352 'Missing spaces around <<')
2353 elif not Match(r'#.*include', line):
2354 # Avoid false positives on ->
2355 reduced_line = line.replace('->', '')
2357 # Look for < that is not surrounded by spaces. This is only
2358 # triggered if both sides are missing spaces, even though
2359 # technically should should flag if at least one side is missing a
2360 # space. This is done to avoid some false positives with shifts.
2361 match = Search(r'[^\s<]<([^\s=<].*)', reduced_line)
2362 if (match and
2363 not FindNextMatchingAngleBracket(clean_lines, linenum, match.group(1))):
2364 error(filename, linenum, 'whitespace/operators', 3,
2365 'Missing spaces around <')
2367 # Look for > that is not surrounded by spaces. Similar to the
2368 # above, we only trigger if both sides are missing spaces to avoid
2369 # false positives with shifts.
2370 match = Search(r'^(.*[^\s>])>[^\s=>]', reduced_line)
2371 if (match and
2372 not FindPreviousMatchingAngleBracket(clean_lines, linenum,
2373 match.group(1))):
2374 error(filename, linenum, 'whitespace/operators', 3,
2375 'Missing spaces around >')
2377 # We allow no-spaces around >> for almost anything. This is because
2378 # C++11 allows ">>" to close nested templates, which accounts for
2379 # most cases when ">>" is not followed by a space.
2381 # We still warn on ">>" followed by alpha character, because that is
2382 # likely due to ">>" being used for right shifts, e.g.:
2383 # value >> alpha
2385 # When ">>" is used to close templates, the alphanumeric letter that
2386 # follows would be part of an identifier, and there should still be
2387 # a space separating the template type and the identifier.
2388 # type<type<type>> alpha
2389 match = Search(r'>>[a-zA-Z_]', line)
2390 if match:
2391 error(filename, linenum, 'whitespace/operators', 3,
2392 'Missing spaces around >>')
2394 # There shouldn't be space around unary operators
2395 match = Search(r'(!\s|~\s|[\s]--[\s;]|[\s]\+\+[\s;])', line)
2396 if match:
2397 error(filename, linenum, 'whitespace/operators', 4,
2398 'Extra space for operator %s' % match.group(1))
2400 # A pet peeve of mine: no spaces after an if, while, switch, or for
2401 match = Search(r' (if\(|for\(|while\(|switch\()', line)
2402 if match:
2403 error(filename, linenum, 'whitespace/parens', 5,
2404 'Missing space before ( in %s' % match.group(1))
2406 # For if/for/while/switch, the left and right parens should be
2407 # consistent about how many spaces are inside the parens, and
2408 # there should either be zero or one spaces inside the parens.
2409 # We don't want: "if ( foo)" or "if ( foo )".
2410 # Exception: "for ( ; foo; bar)" and "for (foo; bar; )" are allowed.
2411 match = Search(r'\b(if|for|while|switch)\s*'
2412 r'\(([ ]*)(.).*[^ ]+([ ]*)\)\s*{\s*$',
2413 line)
2414 if match:
2415 if len(match.group(2)) != len(match.group(4)):
2416 if not (match.group(3) == ';' and
2417 len(match.group(2)) == 1 + len(match.group(4)) or
2418 not match.group(2) and Search(r'\bfor\s*\(.*; \)', line)):
2419 error(filename, linenum, 'whitespace/parens', 5,
2420 'Mismatching spaces inside () in %s' % match.group(1))
2421 if not len(match.group(2)) in [0, 1]:
2422 error(filename, linenum, 'whitespace/parens', 5,
2423 'Should have zero or one spaces inside ( and ) in %s' %
2424 match.group(1))
2426 # You should always have a space after a comma (either as fn arg or operator)
2427 if Search(r',[^\s]', line):
2428 error(filename, linenum, 'whitespace/comma', 3,
2429 'Missing space after ,')
2431 # You should always have a space after a semicolon
2432 # except for few corner cases
2433 # TODO(unknown): clarify if 'if (1) { return 1;}' is requires one more
2434 # space after ;
2435 if Search(r';[^\s};\\)/]', line):
2436 error(filename, linenum, 'whitespace/semicolon', 3,
2437 'Missing space after ;')
2439 # Next we will look for issues with function calls.
2440 CheckSpacingForFunctionCall(filename, line, linenum, error)
2442 # Except after an opening paren, or after another opening brace (in case of
2443 # an initializer list, for instance), you should have spaces before your
2444 # braces. And since you should never have braces at the beginning of a line,
2445 # this is an easy test.
2446 if Search(r'[^ ({]{', line):
2447 error(filename, linenum, 'whitespace/braces', 5,
2448 'Missing space before {')
2450 # Make sure '} else {' has spaces.
2451 if Search(r'}else', line):
2452 error(filename, linenum, 'whitespace/braces', 5,
2453 'Missing space before else')
2455 # You shouldn't have spaces before your brackets, except maybe after
2456 # 'delete []' or 'new char * []'.
2457 if Search(r'\w\s+\[', line) and not Search(r'delete\s+\[', line):
2458 error(filename, linenum, 'whitespace/braces', 5,
2459 'Extra space before [')
2461 # You shouldn't have a space before a semicolon at the end of the line.
2462 # There's a special case for "for" since the style guide allows space before
2463 # the semicolon there.
2464 if Search(r':\s*;\s*$', line):
2465 error(filename, linenum, 'whitespace/semicolon', 5,
2466 'Semicolon defining empty statement. Use {} instead.')
2467 elif Search(r'^\s*;\s*$', line):
2468 error(filename, linenum, 'whitespace/semicolon', 5,
2469 'Line contains only semicolon. If this should be an empty statement, '
2470 'use {} instead.')
2471 elif (Search(r'\s+;\s*$', line) and
2472 not Search(r'\bfor\b', line)):
2473 error(filename, linenum, 'whitespace/semicolon', 5,
2474 'Extra space before last semicolon. If this should be an empty '
2475 'statement, use {} instead.')
2477 # In range-based for, we wanted spaces before and after the colon, but
2478 # not around "::" tokens that might appear.
2479 if (Search('for *\(.*[^:]:[^: ]', line) or
2480 Search('for *\(.*[^: ]:[^:]', line)):
2481 error(filename, linenum, 'whitespace/forcolon', 2,
2482 'Missing space around colon in range-based for loop')
2485 def CheckSectionSpacing(filename, clean_lines, class_info, linenum, error):
2486 """Checks for additional blank line issues related to sections.
2488 Currently the only thing checked here is blank line before protected/private.
2490 Args:
2491 filename: The name of the current file.
2492 clean_lines: A CleansedLines instance containing the file.
2493 class_info: A _ClassInfo objects.
2494 linenum: The number of the line to check.
2495 error: The function to call with any errors found.
2497 # Skip checks if the class is small, where small means 25 lines or less.
2498 # 25 lines seems like a good cutoff since that's the usual height of
2499 # terminals, and any class that can't fit in one screen can't really
2500 # be considered "small".
2502 # Also skip checks if we are on the first line. This accounts for
2503 # classes that look like
2504 # class Foo { public: ... };
2506 # If we didn't find the end of the class, last_line would be zero,
2507 # and the check will be skipped by the first condition.
2508 if (class_info.last_line - class_info.starting_linenum <= 24 or
2509 linenum <= class_info.starting_linenum):
2510 return
2512 matched = Match(r'\s*(public|protected|private):', clean_lines.lines[linenum])
2513 if matched:
2514 # Issue warning if the line before public/protected/private was
2515 # not a blank line, but don't do this if the previous line contains
2516 # "class" or "struct". This can happen two ways:
2517 # - We are at the beginning of the class.
2518 # - We are forward-declaring an inner class that is semantically
2519 # private, but needed to be public for implementation reasons.
2520 # Also ignores cases where the previous line ends with a backslash as can be
2521 # common when defining classes in C macros.
2522 prev_line = clean_lines.lines[linenum - 1]
2523 if (not IsBlankLine(prev_line) and
2524 not Search(r'\b(class|struct)\b', prev_line) and
2525 not Search(r'\\$', prev_line)):
2526 # Try a bit harder to find the beginning of the class. This is to
2527 # account for multi-line base-specifier lists, e.g.:
2528 # class Derived
2529 # : public Base {
2530 end_class_head = class_info.starting_linenum
2531 for i in range(class_info.starting_linenum, linenum):
2532 if Search(r'\{\s*$', clean_lines.lines[i]):
2533 end_class_head = i
2534 break
2535 if end_class_head < linenum - 1:
2536 error(filename, linenum, 'whitespace/blank_line', 3,
2537 '"%s:" should be preceded by a blank line' % matched.group(1))
2540 def GetPreviousNonBlankLine(clean_lines, linenum):
2541 """Return the most recent non-blank line and its line number.
2543 Args:
2544 clean_lines: A CleansedLines instance containing the file contents.
2545 linenum: The number of the line to check.
2547 Returns:
2548 A tuple with two elements. The first element is the contents of the last
2549 non-blank line before the current line, or the empty string if this is the
2550 first non-blank line. The second is the line number of that line, or -1
2551 if this is the first non-blank line.
2554 prevlinenum = linenum - 1
2555 while prevlinenum >= 0:
2556 prevline = clean_lines.elided[prevlinenum]
2557 if not IsBlankLine(prevline): # if not a blank line...
2558 return (prevline, prevlinenum)
2559 prevlinenum -= 1
2560 return ('', -1)
2563 def CheckBraces(filename, clean_lines, linenum, error):
2564 """Looks for misplaced braces (e.g. at the end of line).
2566 Args:
2567 filename: The name of the current file.
2568 clean_lines: A CleansedLines instance containing the file.
2569 linenum: The number of the line to check.
2570 error: The function to call with any errors found.
2573 line = clean_lines.elided[linenum] # get rid of comments and strings
2575 if Match(r'\s*{\s*$', line):
2576 # We allow an open brace to start a line in the case where someone
2577 # is using braces in a block to explicitly create a new scope,
2578 # which is commonly used to control the lifetime of
2579 # stack-allocated variables. We don't detect this perfectly: we
2580 # just don't complain if the last non-whitespace character on the
2581 # previous non-blank line is ';', ':', '{', or '}', or if the previous
2582 # line starts a preprocessor block.
2583 prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0]
2584 if (not Search(r'[;:}{]\s*$', prevline) and
2585 not Match(r'\s*#', prevline)):
2586 error(filename, linenum, 'whitespace/braces', 4,
2587 '{ should almost always be at the end of the previous line')
2589 # An else clause should be on the same line as the preceding closing brace.
2590 if Match(r'\s*else\s*', line):
2591 prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0]
2592 if Match(r'\s*}\s*$', prevline):
2593 error(filename, linenum, 'whitespace/newline', 4,
2594 'An else should appear on the same line as the preceding }')
2596 # If braces come on one side of an else, they should be on both.
2597 # However, we have to worry about "else if" that spans multiple lines!
2598 if Search(r'}\s*else[^{]*$', line) or Match(r'[^}]*else\s*{', line):
2599 if Search(r'}\s*else if([^{]*)$', line): # could be multi-line if
2600 # find the ( after the if
2601 pos = line.find('else if')
2602 pos = line.find('(', pos)
2603 if pos > 0:
2604 (endline, _, endpos) = CloseExpression(clean_lines, linenum, pos)
2605 if endline[endpos:].find('{') == -1: # must be brace after if
2606 error(filename, linenum, 'readability/braces', 5,
2607 'If an else has a brace on one side, it should have it on both')
2608 else: # common case: else not followed by a multi-line if
2609 error(filename, linenum, 'readability/braces', 5,
2610 'If an else has a brace on one side, it should have it on both')
2612 # Likewise, an else should never have the else clause on the same line
2613 if Search(r'\belse [^\s{]', line) and not Search(r'\belse if\b', line):
2614 error(filename, linenum, 'whitespace/newline', 4,
2615 'Else clause should never be on same line as else (use 2 lines)')
2617 # In the same way, a do/while should never be on one line
2618 if Match(r'\s*do [^\s{]', line):
2619 error(filename, linenum, 'whitespace/newline', 4,
2620 'do/while clauses should not be on a single line')
2622 # Braces shouldn't be followed by a ; unless they're defining a struct
2623 # or initializing an array.
2624 # We can't tell in general, but we can for some common cases.
2625 prevlinenum = linenum
2626 while True:
2627 (prevline, prevlinenum) = GetPreviousNonBlankLine(clean_lines, prevlinenum)
2628 if Match(r'\s+{.*}\s*;', line) and not prevline.count(';'):
2629 line = prevline + line
2630 else:
2631 break
2632 if (Search(r'{.*}\s*;', line) and
2633 line.count('{') == line.count('}') and
2634 not Search(r'struct|class|enum|\s*=\s*{', line)):
2635 error(filename, linenum, 'readability/braces', 4,
2636 "You don't need a ; after a }")
2639 def CheckEmptyLoopBody(filename, clean_lines, linenum, error):
2640 """Loop for empty loop body with only a single semicolon.
2642 Args:
2643 filename: The name of the current file.
2644 clean_lines: A CleansedLines instance containing the file.
2645 linenum: The number of the line to check.
2646 error: The function to call with any errors found.
2649 # Search for loop keywords at the beginning of the line. Because only
2650 # whitespaces are allowed before the keywords, this will also ignore most
2651 # do-while-loops, since those lines should start with closing brace.
2652 line = clean_lines.elided[linenum]
2653 if Match(r'\s*(for|while)\s*\(', line):
2654 # Find the end of the conditional expression
2655 (end_line, end_linenum, end_pos) = CloseExpression(
2656 clean_lines, linenum, line.find('('))
2658 # Output warning if what follows the condition expression is a semicolon.
2659 # No warning for all other cases, including whitespace or newline, since we
2660 # have a separate check for semicolons preceded by whitespace.
2661 if end_pos >= 0 and Match(r';', end_line[end_pos:]):
2662 error(filename, end_linenum, 'whitespace/empty_loop_body', 5,
2663 'Empty loop bodies should use {} or continue')
2666 def ReplaceableCheck(operator, macro, line):
2667 """Determine whether a basic CHECK can be replaced with a more specific one.
2669 For example suggest using CHECK_EQ instead of CHECK(a == b) and
2670 similarly for CHECK_GE, CHECK_GT, CHECK_LE, CHECK_LT, CHECK_NE.
2672 Args:
2673 operator: The C++ operator used in the CHECK.
2674 macro: The CHECK or EXPECT macro being called.
2675 line: The current source line.
2677 Returns:
2678 True if the CHECK can be replaced with a more specific one.
2681 # This matches decimal and hex integers, strings, and chars (in that order).
2682 match_constant = r'([-+]?(\d+|0[xX][0-9a-fA-F]+)[lLuU]{0,3}|".*"|\'.*\')'
2684 # Expression to match two sides of the operator with something that
2685 # looks like a literal, since CHECK(x == iterator) won't compile.
2686 # This means we can't catch all the cases where a more specific
2687 # CHECK is possible, but it's less annoying than dealing with
2688 # extraneous warnings.
2689 match_this = (r'\s*' + macro + r'\((\s*' +
2690 match_constant + r'\s*' + operator + r'[^<>].*|'
2691 r'.*[^<>]' + operator + r'\s*' + match_constant +
2692 r'\s*\))')
2694 # Don't complain about CHECK(x == NULL) or similar because
2695 # CHECK_EQ(x, NULL) won't compile (requires a cast).
2696 # Also, don't complain about more complex boolean expressions
2697 # involving && or || such as CHECK(a == b || c == d).
2698 return Match(match_this, line) and not Search(r'NULL|&&|\|\|', line)
2701 def CheckCheck(filename, clean_lines, linenum, error):
2702 """Checks the use of CHECK and EXPECT macros.
2704 Args:
2705 filename: The name of the current file.
2706 clean_lines: A CleansedLines instance containing the file.
2707 linenum: The number of the line to check.
2708 error: The function to call with any errors found.
2711 # Decide the set of replacement macros that should be suggested
2712 raw_lines = clean_lines.raw_lines
2713 current_macro = ''
2714 for macro in _CHECK_MACROS:
2715 if raw_lines[linenum].find(macro) >= 0:
2716 current_macro = macro
2717 break
2718 if not current_macro:
2719 # Don't waste time here if line doesn't contain 'CHECK' or 'EXPECT'
2720 return
2722 line = clean_lines.elided[linenum] # get rid of comments and strings
2724 # Encourage replacing plain CHECKs with CHECK_EQ/CHECK_NE/etc.
2725 for operator in ['==', '!=', '>=', '>', '<=', '<']:
2726 if ReplaceableCheck(operator, current_macro, line):
2727 error(filename, linenum, 'readability/check', 2,
2728 'Consider using %s instead of %s(a %s b)' % (
2729 _CHECK_REPLACEMENT[current_macro][operator],
2730 current_macro, operator))
2731 break
2734 def CheckAltTokens(filename, clean_lines, linenum, error):
2735 """Check alternative keywords being used in boolean expressions.
2737 Args:
2738 filename: The name of the current file.
2739 clean_lines: A CleansedLines instance containing the file.
2740 linenum: The number of the line to check.
2741 error: The function to call with any errors found.
2743 line = clean_lines.elided[linenum]
2745 # Avoid preprocessor lines
2746 if Match(r'^\s*#', line):
2747 return
2749 # Last ditch effort to avoid multi-line comments. This will not help
2750 # if the comment started before the current line or ended after the
2751 # current line, but it catches most of the false positives. At least,
2752 # it provides a way to workaround this warning for people who use
2753 # multi-line comments in preprocessor macros.
2755 # TODO(unknown): remove this once cpplint has better support for
2756 # multi-line comments.
2757 if line.find('/*') >= 0 or line.find('*/') >= 0:
2758 return
2760 for match in _ALT_TOKEN_REPLACEMENT_PATTERN.finditer(line):
2761 error(filename, linenum, 'readability/alt_tokens', 2,
2762 'Use operator %s instead of %s' % (
2763 _ALT_TOKEN_REPLACEMENT[match.group(1)], match.group(1)))
2766 def GetLineWidth(line):
2767 """Determines the width of the line in column positions.
2769 Args:
2770 line: A string, which may be a Unicode string.
2772 Returns:
2773 The width of the line in column positions, accounting for Unicode
2774 combining characters and wide characters.
2776 if isinstance(line, unicode):
2777 width = 0
2778 for uc in unicodedata.normalize('NFC', line):
2779 if unicodedata.east_asian_width(uc) in ('W', 'F'):
2780 width += 2
2781 elif not unicodedata.combining(uc):
2782 width += 1
2783 return width
2784 else:
2785 return len(line)
2788 def CheckStyle(filename, clean_lines, linenum, file_extension, nesting_state,
2789 error):
2790 """Checks rules from the 'C++ style rules' section of cppguide.html.
2792 Most of these rules are hard to test (naming, comment style), but we
2793 do what we can. In particular we check for 2-space indents, line lengths,
2794 tab usage, spaces inside code, etc.
2796 Args:
2797 filename: The name of the current file.
2798 clean_lines: A CleansedLines instance containing the file.
2799 linenum: The number of the line to check.
2800 file_extension: The extension (without the dot) of the filename.
2801 nesting_state: A _NestingState instance which maintains information about
2802 the current stack of nested blocks being parsed.
2803 error: The function to call with any errors found.
2806 raw_lines = clean_lines.raw_lines
2807 line = raw_lines[linenum]
2809 if line.find('\t') != -1:
2810 error(filename, linenum, 'whitespace/tab', 1,
2811 'Tab found; better to use spaces')
2813 # One or three blank spaces at the beginning of the line is weird; it's
2814 # hard to reconcile that with 2-space indents.
2815 # NOTE: here are the conditions rob pike used for his tests. Mine aren't
2816 # as sophisticated, but it may be worth becoming so: RLENGTH==initial_spaces
2817 # if(RLENGTH > 20) complain = 0;
2818 # if(match($0, " +(error|private|public|protected):")) complain = 0;
2819 # if(match(prev, "&& *$")) complain = 0;
2820 # if(match(prev, "\\|\\| *$")) complain = 0;
2821 # if(match(prev, "[\",=><] *$")) complain = 0;
2822 # if(match($0, " <<")) complain = 0;
2823 # if(match(prev, " +for \\(")) complain = 0;
2824 # if(prevodd && match(prevprev, " +for \\(")) complain = 0;
2825 initial_spaces = 0
2826 cleansed_line = clean_lines.elided[linenum]
2827 while initial_spaces < len(line) and line[initial_spaces] == ' ':
2828 initial_spaces += 1
2829 if line and line[-1].isspace():
2830 error(filename, linenum, 'whitespace/end_of_line', 4,
2831 'Line ends in whitespace. Consider deleting these extra spaces.')
2832 # There are certain situations we allow one space, notably for labels
2833 elif ((initial_spaces == 1 or initial_spaces == 3) and
2834 not Match(r'\s*\w+\s*:\s*$', cleansed_line)):
2835 error(filename, linenum, 'whitespace/indent', 3,
2836 'Weird number of spaces at line-start. '
2837 'Are you using a 2-space indent?')
2838 # Labels should always be indented at least one space.
2839 elif not initial_spaces and line[:2] != '//' and Search(r'[^:]:\s*$',
2840 line):
2841 error(filename, linenum, 'whitespace/labels', 4,
2842 'Labels should always be indented at least one space. '
2843 'If this is a member-initializer list in a constructor or '
2844 'the base class list in a class definition, the colon should '
2845 'be on the following line.')
2848 # Check if the line is a header guard.
2849 is_header_guard = False
2850 if file_extension == 'h':
2851 cppvar = GetHeaderGuardCPPVariable(filename)
2852 if (line.startswith('#ifndef %s' % cppvar) or
2853 line.startswith('#define %s' % cppvar) or
2854 line.startswith('#endif // %s' % cppvar)):
2855 is_header_guard = True
2856 # #include lines and header guards can be long, since there's no clean way to
2857 # split them.
2859 # URLs can be long too. It's possible to split these, but it makes them
2860 # harder to cut&paste.
2862 # The "$Id:...$" comment may also get very long without it being the
2863 # developers fault.
2864 if (not line.startswith('#include') and not is_header_guard and
2865 not Match(r'^\s*//.*http(s?)://\S*$', line) and
2866 not Match(r'^// \$Id:.*#[0-9]+ \$$', line)):
2867 line_width = GetLineWidth(line)
2868 if line_width > 100:
2869 error(filename, linenum, 'whitespace/line_length', 4,
2870 'Lines should very rarely be longer than 100 characters')
2871 elif line_width > 80:
2872 error(filename, linenum, 'whitespace/line_length', 2,
2873 'Lines should be <= 80 characters long')
2875 if (cleansed_line.count(';') > 1 and
2876 # for loops are allowed two ;'s (and may run over two lines).
2877 cleansed_line.find('for') == -1 and
2878 (GetPreviousNonBlankLine(clean_lines, linenum)[0].find('for') == -1 or
2879 GetPreviousNonBlankLine(clean_lines, linenum)[0].find(';') != -1) and
2880 # It's ok to have many commands in a switch case that fits in 1 line
2881 not ((cleansed_line.find('case ') != -1 or
2882 cleansed_line.find('default:') != -1) and
2883 cleansed_line.find('break;') != -1)):
2884 error(filename, linenum, 'whitespace/newline', 0,
2885 'More than one command on the same line')
2887 # Some more style checks
2888 CheckBraces(filename, clean_lines, linenum, error)
2889 CheckEmptyLoopBody(filename, clean_lines, linenum, error)
2890 CheckAccess(filename, clean_lines, linenum, nesting_state, error)
2891 CheckSpacing(filename, clean_lines, linenum, nesting_state, error)
2892 CheckCheck(filename, clean_lines, linenum, error)
2893 CheckAltTokens(filename, clean_lines, linenum, error)
2894 classinfo = nesting_state.InnermostClass()
2895 if classinfo:
2896 CheckSectionSpacing(filename, clean_lines, classinfo, linenum, error)
2899 _RE_PATTERN_INCLUDE_NEW_STYLE = re.compile(r'#include +"[^/]+\.h"')
2900 _RE_PATTERN_INCLUDE = re.compile(r'^\s*#\s*include\s*([<"])([^>"]*)[>"].*$')
2901 # Matches the first component of a filename delimited by -s and _s. That is:
2902 # _RE_FIRST_COMPONENT.match('foo').group(0) == 'foo'
2903 # _RE_FIRST_COMPONENT.match('foo.cc').group(0) == 'foo'
2904 # _RE_FIRST_COMPONENT.match('foo-bar_baz.cc').group(0) == 'foo'
2905 # _RE_FIRST_COMPONENT.match('foo_bar-baz.cc').group(0) == 'foo'
2906 _RE_FIRST_COMPONENT = re.compile(r'^[^-_.]+')
2909 def _DropCommonSuffixes(filename):
2910 """Drops common suffixes like _test.cc or -inl.h from filename.
2912 For example:
2913 >>> _DropCommonSuffixes('foo/foo-inl.h')
2914 'foo/foo'
2915 >>> _DropCommonSuffixes('foo/bar/foo.cc')
2916 'foo/bar/foo'
2917 >>> _DropCommonSuffixes('foo/foo_internal.h')
2918 'foo/foo'
2919 >>> _DropCommonSuffixes('foo/foo_unusualinternal.h')
2920 'foo/foo_unusualinternal'
2922 Args:
2923 filename: The input filename.
2925 Returns:
2926 The filename with the common suffix removed.
2928 for suffix in ('test.cc', 'regtest.cc', 'unittest.cc',
2929 'inl.h', 'impl.h', 'internal.h'):
2930 if (filename.endswith(suffix) and len(filename) > len(suffix) and
2931 filename[-len(suffix) - 1] in ('-', '_')):
2932 return filename[:-len(suffix) - 1]
2933 return os.path.splitext(filename)[0]
2936 def _IsTestFilename(filename):
2937 """Determines if the given filename has a suffix that identifies it as a test.
2939 Args:
2940 filename: The input filename.
2942 Returns:
2943 True if 'filename' looks like a test, False otherwise.
2945 if (filename.endswith('_test.cc') or
2946 filename.endswith('_unittest.cc') or
2947 filename.endswith('_regtest.cc')):
2948 return True
2949 else:
2950 return False
2953 def _ClassifyInclude(fileinfo, include, is_system):
2954 """Figures out what kind of header 'include' is.
2956 Args:
2957 fileinfo: The current file cpplint is running over. A FileInfo instance.
2958 include: The path to a #included file.
2959 is_system: True if the #include used <> rather than "".
2961 Returns:
2962 One of the _XXX_HEADER constants.
2964 For example:
2965 >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'stdio.h', True)
2966 _C_SYS_HEADER
2967 >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'string', True)
2968 _CPP_SYS_HEADER
2969 >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/foo.h', False)
2970 _LIKELY_MY_HEADER
2971 >>> _ClassifyInclude(FileInfo('foo/foo_unknown_extension.cc'),
2972 ... 'bar/foo_other_ext.h', False)
2973 _POSSIBLE_MY_HEADER
2974 >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/bar.h', False)
2975 _OTHER_HEADER
2977 # This is a list of all standard c++ header files, except
2978 # those already checked for above.
2979 is_stl_h = include in _STL_HEADERS
2980 is_cpp_h = is_stl_h or include in _CPP_HEADERS
2982 if is_system:
2983 if is_cpp_h:
2984 return _CPP_SYS_HEADER
2985 else:
2986 return _C_SYS_HEADER
2988 # If the target file and the include we're checking share a
2989 # basename when we drop common extensions, and the include
2990 # lives in . , then it's likely to be owned by the target file.
2991 target_dir, target_base = (
2992 os.path.split(_DropCommonSuffixes(fileinfo.RepositoryName())))
2993 include_dir, include_base = os.path.split(_DropCommonSuffixes(include))
2994 if target_base == include_base and (
2995 include_dir == target_dir or
2996 include_dir == os.path.normpath(target_dir + '/../public')):
2997 return _LIKELY_MY_HEADER
2999 # If the target and include share some initial basename
3000 # component, it's possible the target is implementing the
3001 # include, so it's allowed to be first, but we'll never
3002 # complain if it's not there.
3003 target_first_component = _RE_FIRST_COMPONENT.match(target_base)
3004 include_first_component = _RE_FIRST_COMPONENT.match(include_base)
3005 if (target_first_component and include_first_component and
3006 target_first_component.group(0) ==
3007 include_first_component.group(0)):
3008 return _POSSIBLE_MY_HEADER
3010 return _OTHER_HEADER
3014 def CheckIncludeLine(filename, clean_lines, linenum, include_state, error):
3015 """Check rules that are applicable to #include lines.
3017 Strings on #include lines are NOT removed from elided line, to make
3018 certain tasks easier. However, to prevent false positives, checks
3019 applicable to #include lines in CheckLanguage must be put here.
3021 Args:
3022 filename: The name of the current file.
3023 clean_lines: A CleansedLines instance containing the file.
3024 linenum: The number of the line to check.
3025 include_state: An _IncludeState instance in which the headers are inserted.
3026 error: The function to call with any errors found.
3028 fileinfo = FileInfo(filename)
3030 line = clean_lines.lines[linenum]
3032 # "include" should use the new style "foo/bar.h" instead of just "bar.h"
3033 if _RE_PATTERN_INCLUDE_NEW_STYLE.search(line):
3034 error(filename, linenum, 'build/include', 4,
3035 'Include the directory when naming .h files')
3037 # we shouldn't include a file more than once. actually, there are a
3038 # handful of instances where doing so is okay, but in general it's
3039 # not.
3040 match = _RE_PATTERN_INCLUDE.search(line)
3041 if match:
3042 include = match.group(2)
3043 is_system = (match.group(1) == '<')
3044 if include in include_state:
3045 error(filename, linenum, 'build/include', 4,
3046 '"%s" already included at %s:%s' %
3047 (include, filename, include_state[include]))
3048 else:
3049 include_state[include] = linenum
3051 # We want to ensure that headers appear in the right order:
3052 # 1) for foo.cc, foo.h (preferred location)
3053 # 2) c system files
3054 # 3) cpp system files
3055 # 4) for foo.cc, foo.h (deprecated location)
3056 # 5) other google headers
3058 # We classify each include statement as one of those 5 types
3059 # using a number of techniques. The include_state object keeps
3060 # track of the highest type seen, and complains if we see a
3061 # lower type after that.
3062 error_message = include_state.CheckNextIncludeOrder(
3063 _ClassifyInclude(fileinfo, include, is_system))
3064 if error_message:
3065 error(filename, linenum, 'build/include_order', 4,
3066 '%s. Should be: %s.h, c system, c++ system, other.' %
3067 (error_message, fileinfo.BaseName()))
3068 if not include_state.IsInAlphabeticalOrder(include):
3069 error(filename, linenum, 'build/include_alpha', 4,
3070 'Include "%s" not in alphabetical order' % include)
3072 # Look for any of the stream classes that are part of standard C++.
3073 match = _RE_PATTERN_INCLUDE.match(line)
3074 if match:
3075 include = match.group(2)
3076 if Match(r'(f|ind|io|i|o|parse|pf|stdio|str|)?stream$', include):
3077 # Many unit tests use cout, so we exempt them.
3078 if not _IsTestFilename(filename):
3079 error(filename, linenum, 'readability/streams', 3,
3080 'Streams are highly discouraged.')
3083 def _GetTextInside(text, start_pattern):
3084 """Retrieves all the text between matching open and close parentheses.
3086 Given a string of lines and a regular expression string, retrieve all the text
3087 following the expression and between opening punctuation symbols like
3088 (, [, or {, and the matching close-punctuation symbol. This properly nested
3089 occurrences of the punctuations, so for the text like
3090 printf(a(), b(c()));
3091 a call to _GetTextInside(text, r'printf\(') will return 'a(), b(c())'.
3092 start_pattern must match string having an open punctuation symbol at the end.
3094 Args:
3095 text: The lines to extract text. Its comments and strings must be elided.
3096 It can be single line and can span multiple lines.
3097 start_pattern: The regexp string indicating where to start extracting
3098 the text.
3099 Returns:
3100 The extracted text.
3101 None if either the opening string or ending punctuation could not be found.
3103 # TODO(sugawarayu): Audit cpplint.py to see what places could be profitably
3104 # rewritten to use _GetTextInside (and use inferior regexp matching today).
3106 # Give opening punctuations to get the matching close-punctuations.
3107 matching_punctuation = {'(': ')', '{': '}', '[': ']'}
3108 closing_punctuation = set(matching_punctuation.itervalues())
3110 # Find the position to start extracting text.
3111 match = re.search(start_pattern, text, re.M)
3112 if not match: # start_pattern not found in text.
3113 return None
3114 start_position = match.end(0)
3116 assert start_position > 0, (
3117 'start_pattern must ends with an opening punctuation.')
3118 assert text[start_position - 1] in matching_punctuation, (
3119 'start_pattern must ends with an opening punctuation.')
3120 # Stack of closing punctuations we expect to have in text after position.
3121 punctuation_stack = [matching_punctuation[text[start_position - 1]]]
3122 position = start_position
3123 while punctuation_stack and position < len(text):
3124 if text[position] == punctuation_stack[-1]:
3125 punctuation_stack.pop()
3126 elif text[position] in closing_punctuation:
3127 # A closing punctuation without matching opening punctuations.
3128 return None
3129 elif text[position] in matching_punctuation:
3130 punctuation_stack.append(matching_punctuation[text[position]])
3131 position += 1
3132 if punctuation_stack:
3133 # Opening punctuations left without matching close-punctuations.
3134 return None
3135 # punctuations match.
3136 return text[start_position:position - 1]
3139 def CheckLanguage(filename, clean_lines, linenum, file_extension, include_state,
3140 error):
3141 """Checks rules from the 'C++ language rules' section of cppguide.html.
3143 Some of these rules are hard to test (function overloading, using
3144 uint32 inappropriately), but we do the best we can.
3146 Args:
3147 filename: The name of the current file.
3148 clean_lines: A CleansedLines instance containing the file.
3149 linenum: The number of the line to check.
3150 file_extension: The extension (without the dot) of the filename.
3151 include_state: An _IncludeState instance in which the headers are inserted.
3152 error: The function to call with any errors found.
3154 # If the line is empty or consists of entirely a comment, no need to
3155 # check it.
3156 line = clean_lines.elided[linenum]
3157 if not line:
3158 return
3160 match = _RE_PATTERN_INCLUDE.search(line)
3161 if match:
3162 CheckIncludeLine(filename, clean_lines, linenum, include_state, error)
3163 return
3165 # Create an extended_line, which is the concatenation of the current and
3166 # next lines, for more effective checking of code that may span more than one
3167 # line.
3168 if linenum + 1 < clean_lines.NumLines():
3169 extended_line = line + clean_lines.elided[linenum + 1]
3170 else:
3171 extended_line = line
3173 # Make Windows paths like Unix.
3174 fullname = os.path.abspath(filename).replace('\\', '/')
3176 # TODO(unknown): figure out if they're using default arguments in fn proto.
3178 # Check for non-const references in functions. This is tricky because &
3179 # is also used to take the address of something. We allow <> for templates,
3180 # (ignoring whatever is between the braces) and : for classes.
3181 # These are complicated re's. They try to capture the following:
3182 # paren (for fn-prototype start), typename, &, varname. For the const
3183 # version, we're willing for const to be before typename or after
3184 # Don't check the implementation on same line.
3185 fnline = line.split('{', 1)[0]
3186 if (len(re.findall(r'\([^()]*\b(?:[\w:]|<[^()]*>)+(\s?&|&\s?)\w+', fnline)) >
3187 len(re.findall(r'\([^()]*\bconst\s+(?:typename\s+)?(?:struct\s+)?'
3188 r'(?:[\w:]|<[^()]*>)+(\s?&|&\s?)\w+', fnline)) +
3189 len(re.findall(r'\([^()]*\b(?:[\w:]|<[^()]*>)+\s+const(\s?&|&\s?)[\w]+',
3190 fnline))):
3192 # We allow non-const references in a few standard places, like functions
3193 # called "swap()" or iostream operators like "<<" or ">>". We also filter
3194 # out for loops, which lint otherwise mistakenly thinks are functions.
3195 if not Search(
3196 r'(for|swap|Swap|operator[<>][<>])\s*\(\s*'
3197 r'(?:(?:typename\s*)?[\w:]|<.*>)+\s*&',
3198 fnline):
3199 error(filename, linenum, 'runtime/references', 2,
3200 'Is this a non-const reference? '
3201 'If so, make const or use a pointer.')
3203 # Check to see if they're using an conversion function cast.
3204 # I just try to capture the most common basic types, though there are more.
3205 # Parameterless conversion functions, such as bool(), are allowed as they are
3206 # probably a member operator declaration or default constructor.
3207 match = Search(
3208 r'(\bnew\s+)?\b' # Grab 'new' operator, if it's there
3209 r'(int|float|double|bool|char|int32|uint32|int64|uint64)\([^)]', line)
3210 if match:
3211 # gMock methods are defined using some variant of MOCK_METHODx(name, type)
3212 # where type may be float(), int(string), etc. Without context they are
3213 # virtually indistinguishable from int(x) casts. Likewise, gMock's
3214 # MockCallback takes a template parameter of the form return_type(arg_type),
3215 # which looks much like the cast we're trying to detect.
3216 if (match.group(1) is None and # If new operator, then this isn't a cast
3217 not (Match(r'^\s*MOCK_(CONST_)?METHOD\d+(_T)?\(', line) or
3218 Match(r'^\s*MockCallback<.*>', line))):
3219 # Try a bit harder to catch gmock lines: the only place where
3220 # something looks like an old-style cast is where we declare the
3221 # return type of the mocked method, and the only time when we
3222 # are missing context is if MOCK_METHOD was split across
3223 # multiple lines (for example http://go/hrfhr ), so we only need
3224 # to check the previous line for MOCK_METHOD.
3225 if (linenum == 0 or
3226 not Match(r'^\s*MOCK_(CONST_)?METHOD\d+(_T)?\(\S+,\s*$',
3227 clean_lines.elided[linenum - 1])):
3228 error(filename, linenum, 'readability/casting', 4,
3229 'Using deprecated casting style. '
3230 'Use static_cast<%s>(...) instead' %
3231 match.group(2))
3233 CheckCStyleCast(filename, linenum, line, clean_lines.raw_lines[linenum],
3234 'static_cast',
3235 r'\((int|float|double|bool|char|u?int(16|32|64))\)', error)
3237 # This doesn't catch all cases. Consider (const char * const)"hello".
3239 # (char *) "foo" should always be a const_cast (reinterpret_cast won't
3240 # compile).
3241 if CheckCStyleCast(filename, linenum, line, clean_lines.raw_lines[linenum],
3242 'const_cast', r'\((char\s?\*+\s?)\)\s*"', error):
3243 pass
3244 else:
3245 # Check pointer casts for other than string constants
3246 CheckCStyleCast(filename, linenum, line, clean_lines.raw_lines[linenum],
3247 'reinterpret_cast', r'\((\w+\s?\*+\s?)\)', error)
3249 # In addition, we look for people taking the address of a cast. This
3250 # is dangerous -- casts can assign to temporaries, so the pointer doesn't
3251 # point where you think.
3252 if Search(
3253 r'(&\([^)]+\)[\w(])|(&(static|dynamic|reinterpret)_cast\b)', line):
3254 error(filename, linenum, 'runtime/casting', 4,
3255 ('Are you taking an address of a cast? '
3256 'This is dangerous: could be a temp var. '
3257 'Take the address before doing the cast, rather than after'))
3259 # Check for people declaring static/global STL strings at the top level.
3260 # This is dangerous because the C++ language does not guarantee that
3261 # globals with constructors are initialized before the first access.
3262 match = Match(
3263 r'((?:|static +)(?:|const +))string +([a-zA-Z0-9_:]+)\b(.*)',
3264 line)
3265 # Make sure it's not a function.
3266 # Function template specialization looks like: "string foo<Type>(...".
3267 # Class template definitions look like: "string Foo<Type>::Method(...".
3268 if match and not Match(r'\s*(<.*>)?(::[a-zA-Z0-9_]+)?\s*\(([^"]|$)',
3269 match.group(3)):
3270 error(filename, linenum, 'runtime/string', 4,
3271 'For a static/global string constant, use a C style string instead: '
3272 '"%schar %s[]".' %
3273 (match.group(1), match.group(2)))
3275 # Check that we're not using RTTI outside of testing code.
3276 if Search(r'\bdynamic_cast<', line) and not _IsTestFilename(filename):
3277 error(filename, linenum, 'runtime/rtti', 5,
3278 'Do not use dynamic_cast<>. If you need to cast within a class '
3279 "hierarchy, use static_cast<> to upcast. Google doesn't support "
3280 'RTTI.')
3282 if Search(r'\b([A-Za-z0-9_]*_)\(\1\)', line):
3283 error(filename, linenum, 'runtime/init', 4,
3284 'You seem to be initializing a member variable with itself.')
3286 if file_extension == 'h':
3287 # TODO(unknown): check that 1-arg constructors are explicit.
3288 # How to tell it's a constructor?
3289 # (handled in CheckForNonStandardConstructs for now)
3290 # TODO(unknown): check that classes have DISALLOW_EVIL_CONSTRUCTORS
3291 # (level 1 error)
3292 pass
3294 # Check if people are using the verboten C basic types. The only exception
3295 # we regularly allow is "unsigned short port" for port.
3296 if Search(r'\bshort port\b', line):
3297 if not Search(r'\bunsigned short port\b', line):
3298 error(filename, linenum, 'runtime/int', 4,
3299 'Use "unsigned short" for ports, not "short"')
3300 else:
3301 match = Search(r'\b(short|long(?! +double)|long long)\b', line)
3302 if match:
3303 error(filename, linenum, 'runtime/int', 4,
3304 'Use int16/int64/etc, rather than the C type %s' % match.group(1))
3306 # When snprintf is used, the second argument shouldn't be a literal.
3307 match = Search(r'snprintf\s*\(([^,]*),\s*([0-9]*)\s*,', line)
3308 if match and match.group(2) != '0':
3309 # If 2nd arg is zero, snprintf is used to calculate size.
3310 error(filename, linenum, 'runtime/printf', 3,
3311 'If you can, use sizeof(%s) instead of %s as the 2nd arg '
3312 'to snprintf.' % (match.group(1), match.group(2)))
3314 # Check if some verboten C functions are being used.
3315 if Search(r'\bsprintf\b', line):
3316 error(filename, linenum, 'runtime/printf', 5,
3317 'Never use sprintf. Use snprintf instead.')
3318 match = Search(r'\b(strcpy|strcat)\b', line)
3319 if match:
3320 error(filename, linenum, 'runtime/printf', 4,
3321 'Almost always, snprintf is better than %s' % match.group(1))
3323 if Search(r'\bsscanf\b', line):
3324 error(filename, linenum, 'runtime/printf', 1,
3325 'sscanf can be ok, but is slow and can overflow buffers.')
3327 # Check if some verboten operator overloading is going on
3328 # TODO(unknown): catch out-of-line unary operator&:
3329 # class X {};
3330 # int operator&(const X& x) { return 42; } // unary operator&
3331 # The trick is it's hard to tell apart from binary operator&:
3332 # class Y { int operator&(const Y& x) { return 23; } }; // binary operator&
3333 if Search(r'\boperator\s*&\s*\(\s*\)', line):
3334 error(filename, linenum, 'runtime/operator', 4,
3335 'Unary operator& is dangerous. Do not use it.')
3337 # Check for suspicious usage of "if" like
3338 # } if (a == b) {
3339 if Search(r'\}\s*if\s*\(', line):
3340 error(filename, linenum, 'readability/braces', 4,
3341 'Did you mean "else if"? If not, start a new line for "if".')
3343 # Check for potential format string bugs like printf(foo).
3344 # We constrain the pattern not to pick things like DocidForPrintf(foo).
3345 # Not perfect but it can catch printf(foo.c_str()) and printf(foo->c_str())
3346 # TODO(sugawarayu): Catch the following case. Need to change the calling
3347 # convention of the whole function to process multiple line to handle it.
3348 # printf(
3349 # boy_this_is_a_really_long_variable_that_cannot_fit_on_the_prev_line);
3350 printf_args = _GetTextInside(line, r'(?i)\b(string)?printf\s*\(')
3351 if printf_args:
3352 match = Match(r'([\w.\->()]+)$', printf_args)
3353 if match and match.group(1) != '__VA_ARGS__':
3354 function_name = re.search(r'\b((?:string)?printf)\s*\(',
3355 line, re.I).group(1)
3356 error(filename, linenum, 'runtime/printf', 4,
3357 'Potential format string bug. Do %s("%%s", %s) instead.'
3358 % (function_name, match.group(1)))
3360 # Check for potential memset bugs like memset(buf, sizeof(buf), 0).
3361 match = Search(r'memset\s*\(([^,]*),\s*([^,]*),\s*0\s*\)', line)
3362 if match and not Match(r"^''|-?[0-9]+|0x[0-9A-Fa-f]$", match.group(2)):
3363 error(filename, linenum, 'runtime/memset', 4,
3364 'Did you mean "memset(%s, 0, %s)"?'
3365 % (match.group(1), match.group(2)))
3367 if Search(r'\busing namespace\b', line):
3368 error(filename, linenum, 'build/namespaces', 5,
3369 'Do not use namespace using-directives. '
3370 'Use using-declarations instead.')
3372 # Detect variable-length arrays.
3373 match = Match(r'\s*(.+::)?(\w+) [a-z]\w*\[(.+)];', line)
3374 if (match and match.group(2) != 'return' and match.group(2) != 'delete' and
3375 match.group(3).find(']') == -1):
3376 # Split the size using space and arithmetic operators as delimiters.
3377 # If any of the resulting tokens are not compile time constants then
3378 # report the error.
3379 tokens = re.split(r'\s|\+|\-|\*|\/|<<|>>]', match.group(3))
3380 is_const = True
3381 skip_next = False
3382 for tok in tokens:
3383 if skip_next:
3384 skip_next = False
3385 continue
3387 if Search(r'sizeof\(.+\)', tok): continue
3388 if Search(r'arraysize\(\w+\)', tok): continue
3390 tok = tok.lstrip('(')
3391 tok = tok.rstrip(')')
3392 if not tok: continue
3393 if Match(r'\d+', tok): continue
3394 if Match(r'0[xX][0-9a-fA-F]+', tok): continue
3395 if Match(r'k[A-Z0-9]\w*', tok): continue
3396 if Match(r'(.+::)?k[A-Z0-9]\w*', tok): continue
3397 if Match(r'(.+::)?[A-Z][A-Z0-9_]*', tok): continue
3398 # A catch all for tricky sizeof cases, including 'sizeof expression',
3399 # 'sizeof(*type)', 'sizeof(const type)', 'sizeof(struct StructName)'
3400 # requires skipping the next token because we split on ' ' and '*'.
3401 if tok.startswith('sizeof'):
3402 skip_next = True
3403 continue
3404 is_const = False
3405 break
3406 if not is_const:
3407 error(filename, linenum, 'runtime/arrays', 1,
3408 'Do not use variable-length arrays. Use an appropriately named '
3409 "('k' followed by CamelCase) compile-time constant for the size.")
3411 # If DISALLOW_EVIL_CONSTRUCTORS, DISALLOW_COPY_AND_ASSIGN, or
3412 # DISALLOW_IMPLICIT_CONSTRUCTORS is present, then it should be the last thing
3413 # in the class declaration.
3414 match = Match(
3415 (r'\s*'
3416 r'(DISALLOW_(EVIL_CONSTRUCTORS|COPY_AND_ASSIGN|IMPLICIT_CONSTRUCTORS))'
3417 r'\(.*\);$'),
3418 line)
3419 if match and linenum + 1 < clean_lines.NumLines():
3420 next_line = clean_lines.elided[linenum + 1]
3421 # We allow some, but not all, declarations of variables to be present
3422 # in the statement that defines the class. The [\w\*,\s]* fragment of
3423 # the regular expression below allows users to declare instances of
3424 # the class or pointers to instances, but not less common types such
3425 # as function pointers or arrays. It's a tradeoff between allowing
3426 # reasonable code and avoiding trying to parse more C++ using regexps.
3427 if not Search(r'^\s*}[\w\*,\s]*;', next_line):
3428 error(filename, linenum, 'readability/constructors', 3,
3429 match.group(1) + ' should be the last thing in the class')
3431 # Check for use of unnamed namespaces in header files. Registration
3432 # macros are typically OK, so we allow use of "namespace {" on lines
3433 # that end with backslashes.
3434 if (file_extension == 'h'
3435 and Search(r'\bnamespace\s*{', line)
3436 and line[-1] != '\\'):
3437 error(filename, linenum, 'build/namespaces', 4,
3438 'Do not use unnamed namespaces in header files. See '
3439 'http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Namespaces'
3440 ' for more information.')
3443 def CheckCStyleCast(filename, linenum, line, raw_line, cast_type, pattern,
3444 error):
3445 """Checks for a C-style cast by looking for the pattern.
3447 This also handles sizeof(type) warnings, due to similarity of content.
3449 Args:
3450 filename: The name of the current file.
3451 linenum: The number of the line to check.
3452 line: The line of code to check.
3453 raw_line: The raw line of code to check, with comments.
3454 cast_type: The string for the C++ cast to recommend. This is either
3455 reinterpret_cast, static_cast, or const_cast, depending.
3456 pattern: The regular expression used to find C-style casts.
3457 error: The function to call with any errors found.
3459 Returns:
3460 True if an error was emitted.
3461 False otherwise.
3463 match = Search(pattern, line)
3464 if not match:
3465 return False
3467 # e.g., sizeof(int)
3468 sizeof_match = Match(r'.*sizeof\s*$', line[0:match.start(1) - 1])
3469 if sizeof_match:
3470 error(filename, linenum, 'runtime/sizeof', 1,
3471 'Using sizeof(type). Use sizeof(varname) instead if possible')
3472 return True
3474 # operator++(int) and operator--(int)
3475 if (line[0:match.start(1) - 1].endswith(' operator++') or
3476 line[0:match.start(1) - 1].endswith(' operator--')):
3477 return False
3479 remainder = line[match.end(0):]
3481 # The close paren is for function pointers as arguments to a function.
3482 # eg, void foo(void (*bar)(int));
3483 # The semicolon check is a more basic function check; also possibly a
3484 # function pointer typedef.
3485 # eg, void foo(int); or void foo(int) const;
3486 # The equals check is for function pointer assignment.
3487 # eg, void *(*foo)(int) = ...
3488 # The > is for MockCallback<...> ...
3490 # Right now, this will only catch cases where there's a single argument, and
3491 # it's unnamed. It should probably be expanded to check for multiple
3492 # arguments with some unnamed.
3493 function_match = Match(r'\s*(\)|=|(const)?\s*(;|\{|throw\(\)|>))', remainder)
3494 if function_match:
3495 if (not function_match.group(3) or
3496 function_match.group(3) == ';' or
3497 ('MockCallback<' not in raw_line and
3498 '/*' not in raw_line)):
3499 error(filename, linenum, 'readability/function', 3,
3500 'All parameters should be named in a function')
3501 return True
3503 # At this point, all that should be left is actual casts.
3504 error(filename, linenum, 'readability/casting', 4,
3505 'Using C-style cast. Use %s<%s>(...) instead' %
3506 (cast_type, match.group(1)))
3508 return True
3511 _HEADERS_CONTAINING_TEMPLATES = (
3512 ('<deque>', ('deque',)),
3513 ('<functional>', ('unary_function', 'binary_function',
3514 'plus', 'minus', 'multiplies', 'divides', 'modulus',
3515 'negate',
3516 'equal_to', 'not_equal_to', 'greater', 'less',
3517 'greater_equal', 'less_equal',
3518 'logical_and', 'logical_or', 'logical_not',
3519 'unary_negate', 'not1', 'binary_negate', 'not2',
3520 'bind1st', 'bind2nd',
3521 'pointer_to_unary_function',
3522 'pointer_to_binary_function',
3523 'ptr_fun',
3524 'mem_fun_t', 'mem_fun', 'mem_fun1_t', 'mem_fun1_ref_t',
3525 'mem_fun_ref_t',
3526 'const_mem_fun_t', 'const_mem_fun1_t',
3527 'const_mem_fun_ref_t', 'const_mem_fun1_ref_t',
3528 'mem_fun_ref',
3530 ('<limits>', ('numeric_limits',)),
3531 ('<list>', ('list',)),
3532 ('<map>', ('map', 'multimap',)),
3533 ('<memory>', ('allocator',)),
3534 ('<queue>', ('queue', 'priority_queue',)),
3535 ('<set>', ('set', 'multiset',)),
3536 ('<stack>', ('stack',)),
3537 ('<string>', ('char_traits', 'basic_string',)),
3538 ('<utility>', ('pair',)),
3539 ('<vector>', ('vector',)),
3541 # gcc extensions.
3542 # Note: std::hash is their hash, ::hash is our hash
3543 ('<hash_map>', ('hash_map', 'hash_multimap',)),
3544 ('<hash_set>', ('hash_set', 'hash_multiset',)),
3545 ('<slist>', ('slist',)),
3548 _RE_PATTERN_STRING = re.compile(r'\bstring\b')
3550 _re_pattern_algorithm_header = []
3551 for _template in ('copy', 'max', 'min', 'min_element', 'sort', 'swap',
3552 'transform'):
3553 # Match max<type>(..., ...), max(..., ...), but not foo->max, foo.max or
3554 # type::max().
3555 _re_pattern_algorithm_header.append(
3556 (re.compile(r'[^>.]\b' + _template + r'(<.*?>)?\([^\)]'),
3557 _template,
3558 '<algorithm>'))
3560 _re_pattern_templates = []
3561 for _header, _templates in _HEADERS_CONTAINING_TEMPLATES:
3562 for _template in _templates:
3563 _re_pattern_templates.append(
3564 (re.compile(r'(\<|\b)' + _template + r'\s*\<'),
3565 _template + '<>',
3566 _header))
3569 def FilesBelongToSameModule(filename_cc, filename_h):
3570 """Check if these two filenames belong to the same module.
3572 The concept of a 'module' here is a as follows:
3573 foo.h, foo-inl.h, foo.cc, foo_test.cc and foo_unittest.cc belong to the
3574 same 'module' if they are in the same directory.
3575 some/path/public/xyzzy and some/path/internal/xyzzy are also considered
3576 to belong to the same module here.
3578 If the filename_cc contains a longer path than the filename_h, for example,
3579 '/absolute/path/to/base/sysinfo.cc', and this file would include
3580 'base/sysinfo.h', this function also produces the prefix needed to open the
3581 header. This is used by the caller of this function to more robustly open the
3582 header file. We don't have access to the real include paths in this context,
3583 so we need this guesswork here.
3585 Known bugs: tools/base/bar.cc and base/bar.h belong to the same module
3586 according to this implementation. Because of this, this function gives
3587 some false positives. This should be sufficiently rare in practice.
3589 Args:
3590 filename_cc: is the path for the .cc file
3591 filename_h: is the path for the header path
3593 Returns:
3594 Tuple with a bool and a string:
3595 bool: True if filename_cc and filename_h belong to the same module.
3596 string: the additional prefix needed to open the header file.
3599 if not filename_cc.endswith('.cc'):
3600 return (False, '')
3601 filename_cc = filename_cc[:-len('.cc')]
3602 if filename_cc.endswith('_unittest'):
3603 filename_cc = filename_cc[:-len('_unittest')]
3604 elif filename_cc.endswith('_test'):
3605 filename_cc = filename_cc[:-len('_test')]
3606 filename_cc = filename_cc.replace('/public/', '/')
3607 filename_cc = filename_cc.replace('/internal/', '/')
3609 if not filename_h.endswith('.h'):
3610 return (False, '')
3611 filename_h = filename_h[:-len('.h')]
3612 if filename_h.endswith('-inl'):
3613 filename_h = filename_h[:-len('-inl')]
3614 filename_h = filename_h.replace('/public/', '/')
3615 filename_h = filename_h.replace('/internal/', '/')
3617 files_belong_to_same_module = filename_cc.endswith(filename_h)
3618 common_path = ''
3619 if files_belong_to_same_module:
3620 common_path = filename_cc[:-len(filename_h)]
3621 return files_belong_to_same_module, common_path
3624 def UpdateIncludeState(filename, include_state, io=codecs):
3625 """Fill up the include_state with new includes found from the file.
3627 Args:
3628 filename: the name of the header to read.
3629 include_state: an _IncludeState instance in which the headers are inserted.
3630 io: The io factory to use to read the file. Provided for testability.
3632 Returns:
3633 True if a header was succesfully added. False otherwise.
3635 headerfile = None
3636 try:
3637 headerfile = io.open(filename, 'r', 'utf8', 'replace')
3638 except IOError:
3639 return False
3640 linenum = 0
3641 for line in headerfile:
3642 linenum += 1
3643 clean_line = CleanseComments(line)
3644 match = _RE_PATTERN_INCLUDE.search(clean_line)
3645 if match:
3646 include = match.group(2)
3647 # The value formatting is cute, but not really used right now.
3648 # What matters here is that the key is in include_state.
3649 include_state.setdefault(include, '%s:%d' % (filename, linenum))
3650 return True
3653 def CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error,
3654 io=codecs):
3655 """Reports for missing stl includes.
3657 This function will output warnings to make sure you are including the headers
3658 necessary for the stl containers and functions that you use. We only give one
3659 reason to include a header. For example, if you use both equal_to<> and
3660 less<> in a .h file, only one (the latter in the file) of these will be
3661 reported as a reason to include the <functional>.
3663 Args:
3664 filename: The name of the current file.
3665 clean_lines: A CleansedLines instance containing the file.
3666 include_state: An _IncludeState instance.
3667 error: The function to call with any errors found.
3668 io: The IO factory to use to read the header file. Provided for unittest
3669 injection.
3671 required = {} # A map of header name to linenumber and the template entity.
3672 # Example of required: { '<functional>': (1219, 'less<>') }
3674 for linenum in xrange(clean_lines.NumLines()):
3675 line = clean_lines.elided[linenum]
3676 if not line or line[0] == '#':
3677 continue
3679 # String is special -- it is a non-templatized type in STL.
3680 matched = _RE_PATTERN_STRING.search(line)
3681 if matched:
3682 # Don't warn about strings in non-STL namespaces:
3683 # (We check only the first match per line; good enough.)
3684 prefix = line[:matched.start()]
3685 if prefix.endswith('std::') or not prefix.endswith('::'):
3686 required['<string>'] = (linenum, 'string')
3688 for pattern, template, header in _re_pattern_algorithm_header:
3689 if pattern.search(line):
3690 required[header] = (linenum, template)
3692 # The following function is just a speed up, no semantics are changed.
3693 if not '<' in line: # Reduces the cpu time usage by skipping lines.
3694 continue
3696 for pattern, template, header in _re_pattern_templates:
3697 if pattern.search(line):
3698 required[header] = (linenum, template)
3700 # The policy is that if you #include something in foo.h you don't need to
3701 # include it again in foo.cc. Here, we will look at possible includes.
3702 # Let's copy the include_state so it is only messed up within this function.
3703 include_state = include_state.copy()
3705 # Did we find the header for this file (if any) and succesfully load it?
3706 header_found = False
3708 # Use the absolute path so that matching works properly.
3709 abs_filename = FileInfo(filename).FullName()
3711 # For Emacs's flymake.
3712 # If cpplint is invoked from Emacs's flymake, a temporary file is generated
3713 # by flymake and that file name might end with '_flymake.cc'. In that case,
3714 # restore original file name here so that the corresponding header file can be
3715 # found.
3716 # e.g. If the file name is 'foo_flymake.cc', we should search for 'foo.h'
3717 # instead of 'foo_flymake.h'
3718 abs_filename = re.sub(r'_flymake\.cc$', '.cc', abs_filename)
3720 # include_state is modified during iteration, so we iterate over a copy of
3721 # the keys.
3722 header_keys = include_state.keys()
3723 for header in header_keys:
3724 (same_module, common_path) = FilesBelongToSameModule(abs_filename, header)
3725 fullpath = common_path + header
3726 if same_module and UpdateIncludeState(fullpath, include_state, io):
3727 header_found = True
3729 # If we can't find the header file for a .cc, assume it's because we don't
3730 # know where to look. In that case we'll give up as we're not sure they
3731 # didn't include it in the .h file.
3732 # TODO(unknown): Do a better job of finding .h files so we are confident that
3733 # not having the .h file means there isn't one.
3734 if filename.endswith('.cc') and not header_found:
3735 return
3737 # All the lines have been processed, report the errors found.
3738 for required_header_unstripped in required:
3739 template = required[required_header_unstripped][1]
3740 if required_header_unstripped.strip('<>"') not in include_state:
3741 error(filename, required[required_header_unstripped][0],
3742 'build/include_what_you_use', 4,
3743 'Add #include ' + required_header_unstripped + ' for ' + template)
3746 _RE_PATTERN_EXPLICIT_MAKEPAIR = re.compile(r'\bmake_pair\s*<')
3749 def CheckMakePairUsesDeduction(filename, clean_lines, linenum, error):
3750 """Check that make_pair's template arguments are deduced.
3752 G++ 4.6 in C++0x mode fails badly if make_pair's template arguments are
3753 specified explicitly, and such use isn't intended in any case.
3755 Args:
3756 filename: The name of the current file.
3757 clean_lines: A CleansedLines instance containing the file.
3758 linenum: The number of the line to check.
3759 error: The function to call with any errors found.
3761 raw = clean_lines.raw_lines
3762 line = raw[linenum]
3763 match = _RE_PATTERN_EXPLICIT_MAKEPAIR.search(line)
3764 if match:
3765 error(filename, linenum, 'build/explicit_make_pair',
3766 4, # 4 = high confidence
3767 'For C++11-compatibility, omit template arguments from make_pair'
3768 ' OR use pair directly OR if appropriate, construct a pair directly')
3771 def ProcessLine(filename, file_extension, clean_lines, line,
3772 include_state, function_state, nesting_state, error,
3773 extra_check_functions=[]):
3774 """Processes a single line in the file.
3776 Args:
3777 filename: Filename of the file that is being processed.
3778 file_extension: The extension (dot not included) of the file.
3779 clean_lines: An array of strings, each representing a line of the file,
3780 with comments stripped.
3781 line: Number of line being processed.
3782 include_state: An _IncludeState instance in which the headers are inserted.
3783 function_state: A _FunctionState instance which counts function lines, etc.
3784 nesting_state: A _NestingState instance which maintains information about
3785 the current stack of nested blocks being parsed.
3786 error: A callable to which errors are reported, which takes 4 arguments:
3787 filename, line number, error level, and message
3788 extra_check_functions: An array of additional check functions that will be
3789 run on each source line. Each function takes 4
3790 arguments: filename, clean_lines, line, error
3792 raw_lines = clean_lines.raw_lines
3793 ParseNolintSuppressions(filename, raw_lines[line], line, error)
3794 nesting_state.Update(filename, clean_lines, line, error)
3795 if nesting_state.stack and nesting_state.stack[-1].inline_asm != _NO_ASM:
3796 return
3797 CheckForFunctionLengths(filename, clean_lines, line, function_state, error)
3798 CheckForMultilineCommentsAndStrings(filename, clean_lines, line, error)
3799 CheckStyle(filename, clean_lines, line, file_extension, nesting_state, error)
3800 CheckLanguage(filename, clean_lines, line, file_extension, include_state,
3801 error)
3802 CheckForNonStandardConstructs(filename, clean_lines, line,
3803 nesting_state, error)
3804 CheckPosixThreading(filename, clean_lines, line, error)
3805 CheckInvalidIncrement(filename, clean_lines, line, error)
3806 CheckMakePairUsesDeduction(filename, clean_lines, line, error)
3807 for check_fn in extra_check_functions:
3808 check_fn(filename, clean_lines, line, error)
3810 def ProcessFileData(filename, file_extension, lines, error,
3811 extra_check_functions=[]):
3812 """Performs lint checks and reports any errors to the given error function.
3814 Args:
3815 filename: Filename of the file that is being processed.
3816 file_extension: The extension (dot not included) of the file.
3817 lines: An array of strings, each representing a line of the file, with the
3818 last element being empty if the file is terminated with a newline.
3819 error: A callable to which errors are reported, which takes 4 arguments:
3820 filename, line number, error level, and message
3821 extra_check_functions: An array of additional check functions that will be
3822 run on each source line. Each function takes 4
3823 arguments: filename, clean_lines, line, error
3825 lines = (['// marker so line numbers and indices both start at 1'] + lines +
3826 ['// marker so line numbers end in a known way'])
3828 include_state = _IncludeState()
3829 function_state = _FunctionState()
3830 nesting_state = _NestingState()
3832 ResetNolintSuppressions()
3834 CheckForCopyright(filename, lines, error)
3836 if file_extension == 'h':
3837 CheckForHeaderGuard(filename, lines, error)
3839 RemoveMultiLineComments(filename, lines, error)
3840 clean_lines = CleansedLines(lines)
3841 for line in xrange(clean_lines.NumLines()):
3842 ProcessLine(filename, file_extension, clean_lines, line,
3843 include_state, function_state, nesting_state, error,
3844 extra_check_functions)
3845 nesting_state.CheckClassFinished(filename, error)
3847 CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error)
3849 # We check here rather than inside ProcessLine so that we see raw
3850 # lines rather than "cleaned" lines.
3851 CheckForUnicodeReplacementCharacters(filename, lines, error)
3853 CheckForNewlineAtEOF(filename, lines, error)
3855 def ProcessFile(filename, vlevel, extra_check_functions=[]):
3856 """Does google-lint on a single file.
3858 Args:
3859 filename: The name of the file to parse.
3861 vlevel: The level of errors to report. Every error of confidence
3862 >= verbose_level will be reported. 0 is a good default.
3864 extra_check_functions: An array of additional check functions that will be
3865 run on each source line. Each function takes 4
3866 arguments: filename, clean_lines, line, error
3869 _SetVerboseLevel(vlevel)
3871 try:
3872 # Support the UNIX convention of using "-" for stdin. Note that
3873 # we are not opening the file with universal newline support
3874 # (which codecs doesn't support anyway), so the resulting lines do
3875 # contain trailing '\r' characters if we are reading a file that
3876 # has CRLF endings.
3877 # If after the split a trailing '\r' is present, it is removed
3878 # below. If it is not expected to be present (i.e. os.linesep !=
3879 # '\r\n' as in Windows), a warning is issued below if this file
3880 # is processed.
3882 if filename == '-':
3883 lines = codecs.StreamReaderWriter(sys.stdin,
3884 codecs.getreader('utf8'),
3885 codecs.getwriter('utf8'),
3886 'replace').read().split('\n')
3887 else:
3888 lines = codecs.open(filename, 'r', 'utf8', 'replace').read().split('\n')
3890 carriage_return_found = False
3891 # Remove trailing '\r'.
3892 for linenum in range(len(lines)):
3893 if lines[linenum].endswith('\r'):
3894 lines[linenum] = lines[linenum].rstrip('\r')
3895 carriage_return_found = True
3897 except IOError:
3898 sys.stderr.write(
3899 "Skipping input '%s': Can't open for reading\n" % filename)
3900 return
3902 # Note, if no dot is found, this will give the entire filename as the ext.
3903 file_extension = filename[filename.rfind('.') + 1:]
3905 # When reading from stdin, the extension is unknown, so no cpplint tests
3906 # should rely on the extension.
3907 if (filename != '-' and file_extension != 'cc' and file_extension != 'h'
3908 and file_extension != 'cpp'):
3909 sys.stderr.write('Ignoring %s; not a .cc or .h file\n' % filename)
3910 else:
3911 ProcessFileData(filename, file_extension, lines, Error,
3912 extra_check_functions)
3913 if carriage_return_found and os.linesep != '\r\n':
3914 # Use 0 for linenum since outputting only one error for potentially
3915 # several lines.
3916 Error(filename, 0, 'whitespace/newline', 1,
3917 'One or more unexpected \\r (^M) found;'
3918 'better to use only a \\n')
3920 sys.stderr.write('Done processing %s\n' % filename)
3923 def PrintUsage(message):
3924 """Prints a brief usage string and exits, optionally with an error message.
3926 Args:
3927 message: The optional error message.
3929 sys.stderr.write(_USAGE)
3930 if message:
3931 sys.exit('\nFATAL ERROR: ' + message)
3932 else:
3933 sys.exit(1)
3936 def PrintCategories():
3937 """Prints a list of all the error-categories used by error messages.
3939 These are the categories used to filter messages via --filter.
3941 sys.stderr.write(''.join(' %s\n' % cat for cat in _ERROR_CATEGORIES))
3942 sys.exit(0)
3945 def ParseArguments(args):
3946 """Parses the command line arguments.
3948 This may set the output format and verbosity level as side-effects.
3950 Args:
3951 args: The command line arguments:
3953 Returns:
3954 The list of filenames to lint.
3956 try:
3957 (opts, filenames) = getopt.getopt(args, '', ['help', 'output=', 'verbose=',
3958 'counting=',
3959 'filter=',
3960 'root='])
3961 except getopt.GetoptError:
3962 PrintUsage('Invalid arguments.')
3964 verbosity = _VerboseLevel()
3965 output_format = _OutputFormat()
3966 filters = ''
3967 counting_style = ''
3969 for (opt, val) in opts:
3970 if opt == '--help':
3971 PrintUsage(None)
3972 elif opt == '--output':
3973 if not val in ('emacs', 'vs7'):
3974 PrintUsage('The only allowed output formats are emacs and vs7.')
3975 output_format = val
3976 elif opt == '--verbose':
3977 verbosity = int(val)
3978 elif opt == '--filter':
3979 filters = val
3980 if not filters:
3981 PrintCategories()
3982 elif opt == '--counting':
3983 if val not in ('total', 'toplevel', 'detailed'):
3984 PrintUsage('Valid counting options are total, toplevel, and detailed')
3985 counting_style = val
3986 elif opt == '--root':
3987 global _root
3988 _root = val
3990 if not filenames:
3991 PrintUsage('No files were specified.')
3993 _SetOutputFormat(output_format)
3994 _SetVerboseLevel(verbosity)
3995 _SetFilters(filters)
3996 _SetCountingStyle(counting_style)
3998 return filenames
4001 def main():
4002 filenames = ParseArguments(sys.argv[1:])
4004 # Change stderr to write with replacement characters so we don't die
4005 # if we try to print something containing non-ASCII characters.
4006 sys.stderr = codecs.StreamReaderWriter(sys.stderr,
4007 codecs.getreader('utf8'),
4008 codecs.getwriter('utf8'),
4009 'replace')
4011 _cpplint_state.ResetErrorCounts()
4012 for filename in filenames:
4013 ProcessFile(filename, _cpplint_state.verbose_level)
4014 _cpplint_state.PrintErrorCounts()
4016 sys.exit(_cpplint_state.error_count > 0)
4019 if __name__ == '__main__':
4020 main()