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
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
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.
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
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
# - Check that base classes have virtual destructors
57
# put " // namespace" after } that closes a namespace, with
58
# namespace's name after 'namespace' if it is named.
59
# - Do not indent namespace contents
60
# - Avoid inlining non-trivial constructors in header files
61
# include base/basictypes.h if DISALLOW_EVIL_CONSTRUCTORS is used
62
# - Check for old-school (void) cast for call-sites of functions
63
# ignored return value
64
# - Check gUnit usage of anonymous namespace
65
# - Check for class declaration order (typedefs, consts, enums,
66
# ctor(s?), dtor, friend declarations, methods, member vars)
69
"""Does google-lint on c++ files.
71
The goal of this script is to identify places in the code that *may*
72
be in non-compliance with google style. It does not attempt to fix
73
up these problems -- the point is to educate. It does also not
74
attempt to find all problems, or to ensure that everything it does
75
find is legitimately a problem.
77
In particular, we can get very confused by /* and // inside strings!
78
We do a small hack, which is to ignore //'s with "'s after them on the
79
same line, but it is far from perfect (in either direction).
94
Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...]
95
[--counting=total|toplevel|detailed]
98
The style guidelines this tries to follow are those in
99
http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml
101
Every problem is given a confidence score from 1-5, with 5 meaning we are
102
certain of the problem, and 1 meaning it could be a legitimate construct.
103
This will miss some errors, and is not a substitute for a code review.
105
To suppress false-positive errors of a certain category, add a
106
'NOLINT(category)' comment to the line. NOLINT or NOLINT(*)
107
suppresses errors of all categories on that line.
109
The files passed in will be linted; at least one file must be provided.
110
Linted extensions are .C, .cpp, and .h. Other file types will be ignored.
115
By default, the output is formatted to ease emacs parsing. Visual Studio
116
compatible output (vs7) may also be used. Other formats are unsupported.
119
Specify a number 0-5 to restrict errors to certain verbosity levels.
122
Specify a comma-separated list of category-filters to apply: only
123
error messages whose category names pass the filters will be printed.
124
(Category names are printed with the message and look like
125
"[whitespace/indent]".) Filters are evaluated left to right.
126
"-FOO" and "FOO" means "do not print categories that start with FOO".
127
"+FOO" means "do print categories that start with FOO".
129
Examples: --filter=-whitespace,+whitespace/braces
130
--filter=whitespace,runtime/printf,+runtime/printf_format
131
--filter=-,+build/include_what_you_use
133
To see a list of all the categories used in cpplint, pass no arg:
136
counting=total|toplevel|detailed
137
The total number of errors found is always printed. If
138
'toplevel' is provided, then the count of errors in each of
139
the top-level categories like 'build' and 'whitespace' will
140
also be printed. If 'detailed' is provided, then a count
141
is provided for each category like 'build/class'.
144
# We categorize each error message we print. Here are the categories.
145
# We want an explicit list so we can list them all in cpplint --filter=.
146
# If you add a new error message with a new category, add it to the list
147
# here! cpplint_unittest.py should tell you if you forget to do this.
148
# \ used for clearer layout -- pylint: disable-msg=C6013
149
_ERROR_CATEGORIES = [
152
'build/endif_comment',
153
'build/explicit_make_pair',
154
'build/forward_decl',
155
'build/header_guard',
157
'build/include_alpha',
158
'build/include_order',
159
'build/include_what_you_use',
161
'build/printf_format',
162
'build/storage_class',
164
'readability/braces',
165
'readability/casting',
167
'readability/constructors',
168
'readability/fn_size',
169
'readability/function',
170
'readability/multiline_comment',
171
'readability/multiline_string',
172
'readability/nolint',
173
'readability/streams',
181
'runtime/invalid_increment',
182
'runtime/member_string_references',
186
'runtime/printf_format',
187
'runtime/references',
191
'runtime/threadsafe_fn',
193
'whitespace/blank_line',
196
'whitespace/comments',
197
'whitespace/end_of_line',
198
'whitespace/ending_newline',
201
'whitespace/line_length',
202
'whitespace/newline',
203
'whitespace/operators',
205
'whitespace/semicolon',
210
# The default state of the category filter. This is overrided by the --filter=
211
# flag. By default all errors are on, so only add here categories that should be
212
# off by default (i.e., categories that must be enabled by the --filter= flags).
213
# All entries here should start with a '-' or '+', as in the --filter= flag.
214
_DEFAULT_FILTERS = ['-build/include_alpha',
215
'-whitespace/braces',
216
'-whitespace/blank_line',
218
'-readability/streams',
219
'-build/header_guard',
222
'-readability/braces',
223
'-readability/multiline_comment',
227
# We used to check for high-bit characters, but after much discussion we
228
# decided those were OK, as long as they were in UTF-8 and didn't represent
229
# hard-coded international strings, which belong in a separate i18n file.
231
# Headers that we consider STL headers.
232
_STL_HEADERS = frozenset([
233
'algobase.h', 'algorithm', 'alloc.h', 'bitset', 'deque', 'exception',
234
'function.h', 'functional', 'hash_map', 'hash_map.h', 'hash_set',
235
'hash_set.h', 'iterator', 'list', 'list.h', 'map', 'memory', 'new',
236
'pair.h', 'pthread_alloc', 'queue', 'set', 'set.h', 'sstream', 'stack',
237
'stl_alloc.h', 'stl_relops.h', 'type_traits.h',
238
'utility', 'vector', 'vector.h',
242
# Non-STL C++ system headers.
243
_CPP_HEADERS = frozenset([
244
'algo.h', 'builtinbuf.h', 'bvector.h', 'cassert', 'cctype',
245
'cerrno', 'cfloat', 'ciso646', 'climits', 'clocale', 'cmath',
246
'complex', 'complex.h', 'csetjmp', 'csignal', 'cstdarg', 'cstddef',
247
'cstdio', 'cstdlib', 'cstring', 'ctime', 'cwchar', 'cwctype',
248
'defalloc.h', 'deque.h', 'editbuf.h', 'exception', 'fstream',
249
'fstream.h', 'hashtable.h', 'heap.h', 'indstream.h', 'iomanip',
250
'iomanip.h', 'ios', 'iosfwd', 'iostream', 'iostream.h', 'istream',
251
'istream.h', 'iterator.h', 'limits', 'map.h', 'multimap.h', 'multiset.h',
252
'numeric', 'ostream', 'ostream.h', 'parsestream.h', 'pfstream.h',
253
'PlotFile.h', 'procbuf.h', 'pthread_alloc.h', 'rope', 'rope.h',
254
'ropeimpl.h', 'SFile.h', 'slist', 'slist.h', 'stack.h', 'stdexcept',
255
'stdiostream.h', 'streambuf.h', 'stream.h', 'strfile.h', 'string',
256
'strstream', 'strstream.h', 'tempbuf.h', 'tree.h', 'typeinfo', 'valarray',
260
# Assertion macros. These are defined in base/logging.h and
261
# testing/base/gunit.h. Note that the _M versions need to come first
262
# for substring matching to work.
265
'EXPECT_TRUE_M', 'EXPECT_TRUE',
266
'ASSERT_TRUE_M', 'ASSERT_TRUE',
267
'EXPECT_FALSE_M', 'EXPECT_FALSE',
268
'ASSERT_FALSE_M', 'ASSERT_FALSE',
271
# Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE
272
_CHECK_REPLACEMENT = dict([(m, {}) for m in _CHECK_MACROS])
274
for op, replacement in [('==', 'EQ'), ('!=', 'NE'),
275
('>=', 'GE'), ('>', 'GT'),
276
('<=', 'LE'), ('<', 'LT')]:
277
_CHECK_REPLACEMENT['DCHECK'][op] = 'DCHECK_%s' % replacement
278
_CHECK_REPLACEMENT['CHECK'][op] = 'CHECK_%s' % replacement
279
_CHECK_REPLACEMENT['EXPECT_TRUE'][op] = 'EXPECT_%s' % replacement
280
_CHECK_REPLACEMENT['ASSERT_TRUE'][op] = 'ASSERT_%s' % replacement
281
_CHECK_REPLACEMENT['EXPECT_TRUE_M'][op] = 'EXPECT_%s_M' % replacement
282
_CHECK_REPLACEMENT['ASSERT_TRUE_M'][op] = 'ASSERT_%s_M' % replacement
284
for op, inv_replacement in [('==', 'NE'), ('!=', 'EQ'),
285
('>=', 'LT'), ('>', 'LE'),
286
('<=', 'GT'), ('<', 'GE')]:
287
_CHECK_REPLACEMENT['EXPECT_FALSE'][op] = 'EXPECT_%s' % inv_replacement
288
_CHECK_REPLACEMENT['ASSERT_FALSE'][op] = 'ASSERT_%s' % inv_replacement
289
_CHECK_REPLACEMENT['EXPECT_FALSE_M'][op] = 'EXPECT_%s_M' % inv_replacement
290
_CHECK_REPLACEMENT['ASSERT_FALSE_M'][op] = 'ASSERT_%s_M' % inv_replacement
293
# These constants define types of headers for use with
294
# _IncludeState.CheckNextIncludeOrder().
297
_LIKELY_MY_HEADER = 3
298
_POSSIBLE_MY_HEADER = 4
302
_regexp_compile_cache = {}
304
# Finds occurrences of NOLINT or NOLINT(...).
305
_RE_SUPPRESSION = re.compile(r'\bNOLINT\b(\([^)]*\))?')
307
# {str, set(int)}: a map from error categories to sets of linenumbers
308
# on which those errors are expected and should be suppressed.
309
_error_suppressions = {}
311
def ParseNolintSuppressions(filename, raw_line, linenum, error):
312
"""Updates the global list of error-suppressions.
314
Parses any NOLINT comments on the current line, updating the global
315
error_suppressions store. Reports an error if the NOLINT comment
319
filename: str, the name of the input file.
320
raw_line: str, the line of input text, with comments.
321
linenum: int, the number of the current line.
322
error: function, an error handler.
324
# FIXME(adonovan): "NOLINT(" is misparsed as NOLINT(*).
325
matched = _RE_SUPPRESSION.search(raw_line)
327
category = matched.group(1)
328
if category in (None, '(*)'): # => "suppress all"
329
_error_suppressions.setdefault(None, set()).add(linenum)
331
if category.startswith('(') and category.endswith(')'):
332
category = category[1:-1]
333
if category in _ERROR_CATEGORIES:
334
_error_suppressions.setdefault(category, set()).add(linenum)
336
error(filename, linenum, 'readability/nolint', 5,
337
'Unknown NOLINT error category: %s' % category)
340
def ResetNolintSuppressions():
341
"Resets the set of NOLINT suppressions to empty."
342
_error_suppressions.clear()
345
def IsErrorSuppressedByNolint(category, linenum):
346
"""Returns true if the specified error category is suppressed on this line.
348
Consults the global error_suppressions map populated by
349
ParseNolintSuppressions/ResetNolintSuppressions.
352
category: str, the category of the error.
353
linenum: int, the current line number.
355
bool, True iff the error should be suppressed due to a NOLINT comment.
357
return (linenum in _error_suppressions.get(category, set()) or
358
linenum in _error_suppressions.get(None, set()))
360
def Match(pattern, s):
361
"""Matches the string with the pattern, caching the compiled regexp."""
362
# The regexp compilation caching is inlined in both Match and Search for
363
# performance reasons; factoring it out into a separate function turns out
364
# to be noticeably expensive.
365
if not pattern in _regexp_compile_cache:
366
_regexp_compile_cache[pattern] = sre_compile.compile(pattern)
367
return _regexp_compile_cache[pattern].match(s)
370
def Search(pattern, s):
371
"""Searches the string for the pattern, caching the compiled regexp."""
372
if not pattern in _regexp_compile_cache:
373
_regexp_compile_cache[pattern] = sre_compile.compile(pattern)
374
return _regexp_compile_cache[pattern].search(s)
377
class _IncludeState(dict):
378
"""Tracks line numbers for includes, and the order in which includes appear.
380
As a dict, an _IncludeState object serves as a mapping between include
381
filename and line number on which that file was included.
383
Call CheckNextIncludeOrder() once for each header in the file, passing
384
in the type constants defined above. Calls in an illegal order will
385
raise an _IncludeError with an appropriate error message.
388
# self._section will move monotonically through this set. If it ever
389
# needs to move backwards, CheckNextIncludeOrder will raise an error.
397
_C_SYS_HEADER: 'C system header',
398
_CPP_SYS_HEADER: 'C++ system header',
399
_LIKELY_MY_HEADER: 'header this file implements',
400
_POSSIBLE_MY_HEADER: 'header this file may implement',
401
_OTHER_HEADER: 'other header',
404
_INITIAL_SECTION: "... nothing. (This can't be an error.)",
405
_MY_H_SECTION: 'a header this file implements',
406
_C_SECTION: 'C system header',
407
_CPP_SECTION: 'C++ system header',
408
_OTHER_H_SECTION: 'other header',
413
# The name of the current section.
414
self._section = self._INITIAL_SECTION
415
# The path of last found header.
416
self._last_header = ''
418
def CanonicalizeAlphabeticalOrder(self, header_path):
419
"""Returns a path canonicalized for alphabetical comparison.
421
- replaces "-" with "_" so they both cmp the same.
422
- removes '-inl' since we don't require them to be after the main header.
423
- lowercase everything, just in case.
426
header_path: Path to be canonicalized.
431
return header_path.replace('-inl.h', '.h').replace('-', '_').lower()
433
def IsInAlphabeticalOrder(self, header_path):
434
"""Check if a header is in alphabetical order with the previous header.
437
header_path: Header to be checked.
440
Returns true if the header is in alphabetical order.
442
canonical_header = self.CanonicalizeAlphabeticalOrder(header_path)
443
if self._last_header > canonical_header:
445
self._last_header = canonical_header
448
def CheckNextIncludeOrder(self, header_type):
449
"""Returns a non-empty error message if the next header is out of order.
451
This function also updates the internal state to be ready to check
455
header_type: One of the _XXX_HEADER constants defined above.
458
The empty string if the header is in the right order, or an
459
error message describing what's wrong.
462
error_message = ('Found %s after %s' %
463
(self._TYPE_NAMES[header_type],
464
self._SECTION_NAMES[self._section]))
466
last_section = self._section
468
if header_type == _C_SYS_HEADER:
469
if self._section <= self._C_SECTION:
470
self._section = self._C_SECTION
472
self._last_header = ''
474
elif header_type == _CPP_SYS_HEADER:
475
if self._section <= self._CPP_SECTION:
476
self._section = self._CPP_SECTION
478
self._last_header = ''
480
elif header_type == _LIKELY_MY_HEADER:
481
if self._section <= self._MY_H_SECTION:
482
self._section = self._MY_H_SECTION
484
self._section = self._OTHER_H_SECTION
485
elif header_type == _POSSIBLE_MY_HEADER:
486
if self._section <= self._MY_H_SECTION:
487
self._section = self._MY_H_SECTION
489
# This will always be the fallback because we're not sure
490
# enough that the header is associated with this file.
491
self._section = self._OTHER_H_SECTION
493
assert header_type == _OTHER_HEADER
494
self._section = self._OTHER_H_SECTION
496
if last_section != self._section:
497
self._last_header = ''
502
class _CppLintState(object):
503
"""Maintains module-wide state.."""
506
self.verbose_level = 1 # global setting.
507
self.error_count = 0 # global count of reported errors
508
# filters to apply when emitting error messages
509
self.filters = _DEFAULT_FILTERS[:]
510
self.counting = 'total' # In what way are we counting errors?
511
self.errors_by_category = {} # string to int dict storing error counts
514
# "emacs" - format that emacs can parse (default)
515
# "vs7" - format that Microsoft Visual Studio 7 can parse
516
self.output_format = 'emacs'
518
def SetOutputFormat(self, output_format):
519
"""Sets the output format for errors."""
520
self.output_format = output_format
522
def SetVerboseLevel(self, level):
523
"""Sets the module's verbosity, and returns the previous setting."""
524
last_verbose_level = self.verbose_level
525
self.verbose_level = level
526
return last_verbose_level
528
def SetCountingStyle(self, counting_style):
529
"""Sets the module's counting options."""
530
self.counting = counting_style
532
def SetFilters(self, filters):
533
"""Sets the error-message filters.
535
These filters are applied when deciding whether to emit a given
539
filters: A string of comma-separated filters (eg "+whitespace/indent").
540
Each filter should start with + or -; else we die.
543
ValueError: The comma-separated filters did not all start with '+' or '-'.
544
E.g. "-,+whitespace,-whitespace/indent,whitespace/badfilter"
546
# Default filters always have less priority than the flag ones.
547
self.filters = _DEFAULT_FILTERS[:]
548
for filt in filters.split(','):
549
clean_filt = filt.strip()
551
self.filters.append(clean_filt)
552
for filt in self.filters:
553
if not (filt.startswith('+') or filt.startswith('-')):
554
raise ValueError('Every filter in --filters must start with + or -'
555
' (%s does not)' % filt)
557
def ResetErrorCounts(self):
558
"""Sets the module's error statistic back to zero."""
560
self.errors_by_category = {}
562
def IncrementErrorCount(self, category):
563
"""Bumps the module's error statistic."""
564
self.error_count += 1
565
if self.counting in ('toplevel', 'detailed'):
566
if self.counting != 'detailed':
567
category = category.split('/')[0]
568
if category not in self.errors_by_category:
569
self.errors_by_category[category] = 0
570
self.errors_by_category[category] += 1
572
def PrintErrorCounts(self):
573
"""Print a summary of errors by category, and the total."""
574
for category, count in self.errors_by_category.iteritems():
575
sys.stderr.write('Category \'%s\' errors found: %d\n' %
577
sys.stderr.write('Total errors found: %d\n' % self.error_count)
579
_cpplint_state = _CppLintState()
583
"""Gets the module's output format."""
584
return _cpplint_state.output_format
587
def _SetOutputFormat(output_format):
588
"""Sets the module's output format."""
589
_cpplint_state.SetOutputFormat(output_format)
593
"""Returns the module's verbosity setting."""
594
return _cpplint_state.verbose_level
597
def _SetVerboseLevel(level):
598
"""Sets the module's verbosity, and returns the previous setting."""
599
return _cpplint_state.SetVerboseLevel(level)
602
def _SetCountingStyle(level):
603
"""Sets the module's counting options."""
604
_cpplint_state.SetCountingStyle(level)
608
"""Returns the module's list of output filters, as a list."""
609
return _cpplint_state.filters
612
def _SetFilters(filters):
613
"""Sets the module's error-message filters.
615
These filters are applied when deciding whether to emit a given
619
filters: A string of comma-separated filters (eg "whitespace/indent").
620
Each filter should start with + or -; else we die.
622
_cpplint_state.SetFilters(filters)
625
class _FunctionState(object):
626
"""Tracks current function name and the number of lines in its body."""
628
_NORMAL_TRIGGER = 250 # for --v=0, 500 for --v=1, etc.
629
_TEST_TRIGGER = 400 # about 50% more than _NORMAL_TRIGGER.
632
self.in_a_function = False
633
self.lines_in_function = 0
634
self.current_function = ''
636
def Begin(self, function_name):
637
"""Start analyzing function body.
640
function_name: The name of the function being tracked.
642
self.in_a_function = True
643
self.lines_in_function = 0
644
self.current_function = function_name
647
"""Count line in current function body."""
648
if self.in_a_function:
649
self.lines_in_function += 1
651
def Check(self, error, filename, linenum):
652
"""Report if too many lines in function body.
655
error: The function to call with any errors found.
656
filename: The name of the current file.
657
linenum: The number of the line to check.
659
if Match(r'T(EST|est)', self.current_function):
660
base_trigger = self._TEST_TRIGGER
662
base_trigger = self._NORMAL_TRIGGER
663
trigger = base_trigger * 2**_VerboseLevel()
665
if self.lines_in_function > trigger:
666
error_level = int(math.log(self.lines_in_function / base_trigger, 2))
667
# 50 => 0, 100 => 1, 200 => 2, 400 => 3, 800 => 4, 1600 => 5, ...
670
error(filename, linenum, 'readability/fn_size', error_level,
671
'Small and focused functions are preferred:'
672
' %s has %d non-comment lines'
673
' (error triggered by exceeding %d lines).' % (
674
self.current_function, self.lines_in_function, trigger))
677
"""Stop analyzing function body."""
678
self.in_a_function = False
681
class _IncludeError(Exception):
682
"""Indicates a problem with the include order in a file."""
687
"""Provides utility functions for filenames.
689
FileInfo provides easy access to the components of a file's path
690
relative to the project root.
693
def __init__(self, filename):
694
self._filename = filename
697
"""Make Windows paths like Unix."""
698
return os.path.abspath(self._filename).replace('\\', '/')
700
def RepositoryName(self):
701
"""FullName after removing the local path to the repository.
703
If we have a real absolute path name here we can try to do something smart:
704
detecting the root of the checkout and truncating /path/to/checkout from
705
the name so that we get header guards that don't include things like
706
"C:\Documents and Settings\..." or "/home/username/..." in them and thus
707
people on different computers who have checked the source out to different
708
locations won't see bogus errors.
710
fullname = self.FullName()
712
if os.path.exists(fullname):
713
project_dir = os.path.dirname(fullname)
715
if os.path.exists(os.path.join(project_dir, ".svn")):
716
# If there's a .svn file in the current directory, we recursively look
717
# up the directory tree for the top of the SVN checkout
718
root_dir = project_dir
719
one_up_dir = os.path.dirname(root_dir)
720
while os.path.exists(os.path.join(one_up_dir, ".svn")):
721
root_dir = os.path.dirname(root_dir)
722
one_up_dir = os.path.dirname(one_up_dir)
724
prefix = os.path.commonprefix([root_dir, project_dir])
725
return fullname[len(prefix) + 1:]
727
# Not SVN <= 1.6? Try to find a git, hg, or svn top level directory by
728
# searching up from the current path.
729
root_dir = os.path.dirname(fullname)
730
while (root_dir != os.path.dirname(root_dir) and
731
not os.path.exists(os.path.join(root_dir, ".git")) and
732
not os.path.exists(os.path.join(root_dir, ".hg")) and
733
not os.path.exists(os.path.join(root_dir, ".svn"))):
734
root_dir = os.path.dirname(root_dir)
736
if (os.path.exists(os.path.join(root_dir, ".git")) or
737
os.path.exists(os.path.join(root_dir, ".hg")) or
738
os.path.exists(os.path.join(root_dir, ".svn"))):
739
prefix = os.path.commonprefix([root_dir, project_dir])
740
return fullname[len(prefix) + 1:]
742
# Don't know what to do; header guard warnings may be wrong...
746
"""Splits the file into the directory, basename, and extension.
748
For 'chrome/browser/browser.cc', Split() would
749
return ('chrome/browser', 'browser', '.cc')
752
A tuple of (directory, basename, extension).
755
googlename = self.RepositoryName()
756
project, rest = os.path.split(googlename)
757
return (project,) + os.path.splitext(rest)
760
"""File base name - text after the final slash, before the final period."""
761
return self.Split()[1]
764
"""File extension - text following the final period."""
765
return self.Split()[2]
767
def NoExtension(self):
768
"""File has no source file extension."""
769
return '/'.join(self.Split()[0:2])
772
"""File has a source file extension."""
773
return self.Extension()[1:] in ('c', 'cc', 'cpp', 'cxx', 'C')
776
def _ShouldPrintError(category, confidence, linenum):
777
"""If confidence >= verbose, category passes filter and is not suppressed."""
779
# There are three ways we might decide not to print an error message:
780
# a "NOLINT(category)" comment appears in the source,
781
# the verbosity level isn't high enough, or the filters filter it out.
782
if IsErrorSuppressedByNolint(category, linenum):
784
if confidence < _cpplint_state.verbose_level:
788
for one_filter in _Filters():
789
if one_filter.startswith('-'):
790
if category.startswith(one_filter[1:]):
792
elif one_filter.startswith('+'):
793
if category.startswith(one_filter[1:]):
796
assert False # should have been checked for in SetFilter.
803
def Error(filename, linenum, category, confidence, message):
804
"""Logs the fact we've found a lint error.
806
We log where the error was found, and also our confidence in the error,
807
that is, how certain we are this is a legitimate style regression, and
808
not a misidentification or a use that's sometimes justified.
810
False positives can be suppressed by the use of
811
"cpplint(category)" comments on the offending line. These are
812
parsed into _error_suppressions.
815
filename: The name of the file containing the error.
816
linenum: The number of the line containing the error.
817
category: A string used to describe the "category" this bug
818
falls under: "whitespace", say, or "runtime". Categories
819
may have a hierarchy separated by slashes: "whitespace/indent".
820
confidence: A number from 1-5 representing a confidence score for
821
the error, with 5 meaning that we are certain of the problem,
822
and 1 meaning that it could be a legitimate construct.
823
message: The error message.
825
if _ShouldPrintError(category, confidence, linenum):
826
_cpplint_state.IncrementErrorCount(category)
827
if _cpplint_state.output_format == 'vs7':
828
sys.stderr.write('%s(%s): %s [%s] [%d]\n' % (
829
filename, linenum, message, category, confidence))
831
sys.stderr.write('%s:%s: %s [%s] [%d]\n' % (
832
filename, linenum, message, category, confidence))
835
# Matches standard C++ escape esequences per 2.13.2.3 of the C++ standard.
836
_RE_PATTERN_CLEANSE_LINE_ESCAPES = re.compile(
837
r'\\([abfnrtv?"\\\']|\d+|x[0-9a-fA-F]+)')
838
# Matches strings. Escape codes should already be removed by ESCAPES.
839
_RE_PATTERN_CLEANSE_LINE_DOUBLE_QUOTES = re.compile(r'"[^"]*"')
840
# Matches characters. Escape codes should already be removed by ESCAPES.
841
_RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES = re.compile(r"'.'")
842
# Matches multi-line C++ comments.
843
# This RE is a little bit more complicated than one might expect, because we
844
# have to take care of space removals tools so we can handle comments inside
846
# The current rule is: We only clear spaces from both sides when we're at the
847
# end of the line. Otherwise, we try to remove spaces from the right side,
848
# if this doesn't work we try on left side but only if there's a non-character
850
_RE_PATTERN_CLEANSE_LINE_C_COMMENTS = re.compile(
851
r"""(\s*/\*.*\*/\s*$|
854
/\*.*\*/)""", re.VERBOSE)
857
def IsCppString(line):
858
"""Does line terminate so, that the next symbol is in string constant.
860
This function does not consider single-line nor multi-line comments.
863
line: is a partial line of code starting from the 0..n.
866
True, if next character appended to 'line' is inside a
870
line = line.replace(r'\\', 'XX') # after this, \\" does not match to \"
871
return ((line.count('"') - line.count(r'\"') - line.count("'\"'")) & 1) == 1
874
def FindNextMultiLineCommentStart(lines, lineix):
875
"""Find the beginning marker for a multiline comment."""
876
while lineix < len(lines):
877
if lines[lineix].strip().startswith('/*'):
878
# Only return this marker if the comment goes beyond this line
879
if lines[lineix].strip().find('*/', 2) < 0:
885
def FindNextMultiLineCommentEnd(lines, lineix):
886
"""We are inside a comment, find the end marker."""
887
while lineix < len(lines):
888
if lines[lineix].strip().endswith('*/'):
894
def RemoveMultiLineCommentsFromRange(lines, begin, end):
895
"""Clears a range of lines for multi-line comments."""
896
# Having // dummy comments makes the lines non-empty, so we will not get
897
# unnecessary blank line warnings later in the code.
898
for i in range(begin, end):
899
lines[i] = '// dummy'
902
def RemoveMultiLineComments(filename, lines, error):
903
"""Removes multiline (c-style) comments from lines."""
905
while lineix < len(lines):
906
lineix_begin = FindNextMultiLineCommentStart(lines, lineix)
907
if lineix_begin >= len(lines):
909
lineix_end = FindNextMultiLineCommentEnd(lines, lineix_begin)
910
if lineix_end >= len(lines):
911
error(filename, lineix_begin + 1, 'readability/multiline_comment', 5,
912
'Could not find end of multi-line comment')
914
RemoveMultiLineCommentsFromRange(lines, lineix_begin, lineix_end + 1)
915
lineix = lineix_end + 1
918
def CleanseComments(line):
919
"""Removes //-comments and single-line C-style /* */ comments.
922
line: A line of C++ source.
925
The line with single-line comments removed.
927
commentpos = line.find('//')
928
if commentpos != -1 and not IsCppString(line[:commentpos]):
929
line = line[:commentpos].rstrip()
930
# get rid of /* ... */
931
return _RE_PATTERN_CLEANSE_LINE_C_COMMENTS.sub('', line)
934
class CleansedLines(object):
935
"""Holds 3 copies of all lines with different preprocessing applied to them.
937
1) elided member contains lines without strings and comments,
938
2) lines member contains lines without comments, and
939
3) raw member contains all the lines without processing.
940
All these three members are of <type 'list'>, and of the same length.
943
def __init__(self, lines):
946
self.raw_lines = lines
947
self.num_lines = len(lines)
948
for linenum in range(len(lines)):
949
self.lines.append(CleanseComments(lines[linenum]))
950
elided = self._CollapseStrings(lines[linenum])
951
self.elided.append(CleanseComments(elided))
954
"""Returns the number of lines represented."""
955
return self.num_lines
958
def _CollapseStrings(elided):
959
"""Collapses strings and chars on a line to simple "" or '' blocks.
961
We nix strings first so we're not fooled by text like '"http://"'
964
elided: The line being processed.
967
The line with collapsed strings.
969
if not _RE_PATTERN_INCLUDE.match(elided):
970
# Remove escaped characters first to make quote/single quote collapsing
971
# basic. Things that look like escaped characters shouldn't occur
972
# outside of strings and chars.
973
elided = _RE_PATTERN_CLEANSE_LINE_ESCAPES.sub('', elided)
974
elided = _RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES.sub("''", elided)
975
elided = _RE_PATTERN_CLEANSE_LINE_DOUBLE_QUOTES.sub('""', elided)
979
def CloseExpression(clean_lines, linenum, pos):
980
"""If input points to ( or { or [, finds the position that closes it.
982
If lines[linenum][pos] points to a '(' or '{' or '[', finds the
983
linenum/pos that correspond to the closing of the expression.
986
clean_lines: A CleansedLines instance containing the file.
987
linenum: The number of the line to check.
988
pos: A position on the line.
991
A tuple (line, linenum, pos) pointer *past* the closing brace, or
992
(line, len(lines), -1) if we never find a close. Note we ignore
993
strings and comments when matching; and the line we return is the
994
'cleansed' line at linenum.
997
line = clean_lines.elided[linenum]
998
startchar = line[pos]
999
if startchar not in '({[':
1000
return (line, clean_lines.NumLines(), -1)
1001
if startchar == '(': endchar = ')'
1002
if startchar == '[': endchar = ']'
1003
if startchar == '{': endchar = '}'
1005
num_open = line.count(startchar) - line.count(endchar)
1006
while linenum < clean_lines.NumLines() and num_open > 0:
1008
line = clean_lines.elided[linenum]
1009
num_open += line.count(startchar) - line.count(endchar)
1010
# OK, now find the endchar that actually got us back to even
1012
while num_open >= 0:
1013
endpos = line.rfind(')', 0, endpos)
1014
num_open -= 1 # chopped off another )
1015
return (line, linenum, endpos + 1)
1018
def CheckForCopyright(filename, lines, error):
1019
"""Logs an error if no Copyright message appears at the top of the file."""
1021
# We'll say it should occur by line 10. Don't forget there's a
1022
# dummy line at the front.
1023
for line in xrange(1, min(len(lines), 11)):
1024
if re.search(r'Copyright', lines[line], re.I): break
1025
else: # means no copyright line was found
1026
error(filename, 0, 'legal/copyright', 5,
1027
'No copyright message found. '
1028
'You should have a line: "Copyright [year] <Copyright Owner>"')
1031
def GetHeaderGuardCPPVariable(filename):
1032
"""Returns the CPP variable that should be used as a header guard.
1035
filename: The name of a C++ header file.
1038
The CPP variable that should be used as a header guard in the
1043
# Restores original filename in case that cpplint is invoked from Emacs's
1045
filename = re.sub(r'_flymake\.h$', '.h', filename)
1047
fileinfo = FileInfo(filename)
1048
return re.sub(r'[-./\s]', '_', fileinfo.RepositoryName()).upper() + '_'
1051
def CheckForHeaderGuard(filename, lines, error):
1052
"""Checks that the file contains a header guard.
1054
Logs an error if no #ifndef header guard is present. For other
1055
headers, checks that the full pathname is used.
1058
filename: The name of the C++ header file.
1059
lines: An array of strings, each representing a line of the file.
1060
error: The function to call with any errors found.
1063
cppvar = GetHeaderGuardCPPVariable(filename)
1070
for linenum, line in enumerate(lines):
1071
linesplit = line.split()
1072
if len(linesplit) >= 2:
1073
# find the first occurrence of #ifndef and #define, save arg
1074
if not ifndef and linesplit[0] == '#ifndef':
1075
# set ifndef to the header guard presented on the #ifndef line.
1076
ifndef = linesplit[1]
1077
ifndef_linenum = linenum
1078
if not define and linesplit[0] == '#define':
1079
define = linesplit[1]
1080
# find the last occurrence of #endif, save entire line
1081
if line.startswith('#endif'):
1083
endif_linenum = linenum
1085
if not ifndef or not define or ifndef != define:
1086
error(filename, 0, 'build/header_guard', 5,
1087
'No #ifndef header guard found, suggested CPP variable is: %s' %
1091
# The guard should be PATH_FILE_H_, but we also allow PATH_FILE_H__
1092
# for backward compatibility.
1093
if ifndef != cppvar:
1095
if ifndef != cppvar + '_':
1098
ParseNolintSuppressions(filename, lines[ifndef_linenum], ifndef_linenum,
1100
error(filename, ifndef_linenum, 'build/header_guard', error_level,
1101
'#ifndef header guard has wrong style, please use: %s' % cppvar)
1103
if endif != ('#endif // %s' % cppvar):
1105
if endif != ('#endif // %s' % (cppvar + '_')):
1108
ParseNolintSuppressions(filename, lines[endif_linenum], endif_linenum,
1110
error(filename, endif_linenum, 'build/header_guard', error_level,
1111
'#endif line should be "#endif // %s"' % cppvar)
1114
def CheckForUnicodeReplacementCharacters(filename, lines, error):
1115
"""Logs an error for each line containing Unicode replacement characters.
1117
These indicate that either the file contained invalid UTF-8 (likely)
1118
or Unicode replacement characters (which it shouldn't). Note that
1119
it's possible for this to throw off line numbering if the invalid
1120
UTF-8 occurred adjacent to a newline.
1123
filename: The name of the current file.
1124
lines: An array of strings, each representing a line of the file.
1125
error: The function to call with any errors found.
1127
for linenum, line in enumerate(lines):
1128
if u'\ufffd' in line:
1129
error(filename, linenum, 'readability/utf8', 5,
1130
'Line contains invalid UTF-8 (or Unicode replacement character).')
1133
def CheckForNewlineAtEOF(filename, lines, error):
1134
"""Logs an error if there is no newline char at the end of the file.
1137
filename: The name of the current file.
1138
lines: An array of strings, each representing a line of the file.
1139
error: The function to call with any errors found.
1142
# The array lines() was created by adding two newlines to the
1143
# original file (go figure), then splitting on \n.
1144
# To verify that the file ends in \n, we just have to make sure the
1145
# last-but-two element of lines() exists and is empty.
1146
if len(lines) < 3 or lines[-2]:
1147
error(filename, len(lines) - 2, 'whitespace/ending_newline', 5,
1148
'Could not find a newline character at the end of the file.')
1151
def CheckForMultilineCommentsAndStrings(filename, clean_lines, linenum, error):
1152
"""Logs an error if we see /* ... */ or "..." that extend past one line.
1154
/* ... */ comments are legit inside macros, for one line.
1155
Otherwise, we prefer // comments, so it's ok to warn about the
1156
other. Likewise, it's ok for strings to extend across multiple
1157
lines, as long as a line continuation character (backslash)
1158
terminates each line. Although not currently prohibited by the C++
1159
style guide, it's ugly and unnecessary. We don't do well with either
1160
in this lint program, so we warn about both.
1163
filename: The name of the current file.
1164
clean_lines: A CleansedLines instance containing the file.
1165
linenum: The number of the line to check.
1166
error: The function to call with any errors found.
1168
line = clean_lines.elided[linenum]
1170
# Remove all \\ (escaped backslashes) from the line. They are OK, and the
1171
# second (escaped) slash may trigger later \" detection erroneously.
1172
line = line.replace('\\\\', '')
1174
if line.count('/*') > line.count('*/'):
1175
error(filename, linenum, 'readability/multiline_comment', 5,
1176
'Complex multi-line /*...*/-style comment found. '
1177
'Lint may give bogus warnings. '
1178
'Consider replacing these with //-style comments, '
1179
'with #if 0...#endif, '
1180
'or with more clearly structured multi-line comments.')
1182
if (line.count('"') - line.count('\\"')) % 2:
1183
error(filename, linenum, 'readability/multiline_string', 5,
1184
'Multi-line string ("...") found. This lint script doesn\'t '
1185
'do well with such strings, and may give bogus warnings. They\'re '
1186
'ugly and unnecessary, and you should use concatenation instead".')
1190
('asctime(', 'asctime_r('),
1191
('ctime(', 'ctime_r('),
1192
('getgrgid(', 'getgrgid_r('),
1193
('getgrnam(', 'getgrnam_r('),
1194
('getlogin(', 'getlogin_r('),
1195
('getpwnam(', 'getpwnam_r('),
1196
('getpwuid(', 'getpwuid_r('),
1197
('gmtime(', 'gmtime_r('),
1198
('localtime(', 'localtime_r('),
1199
('rand(', 'rand_r('),
1200
('readdir(', 'readdir_r('),
1201
('strtok(', 'strtok_r('),
1202
('ttyname(', 'ttyname_r('),
1206
def CheckPosixThreading(filename, clean_lines, linenum, error):
1207
"""Checks for calls to thread-unsafe functions.
1209
Much code has been originally written without consideration of
1210
multi-threading. Also, engineers are relying on their old experience;
1211
they have learned posix before threading extensions were added. These
1212
tests guide the engineers to use thread-safe functions (when using
1216
filename: The name of the current file.
1217
clean_lines: A CleansedLines instance containing the file.
1218
linenum: The number of the line to check.
1219
error: The function to call with any errors found.
1221
line = clean_lines.elided[linenum]
1222
for single_thread_function, multithread_safe_function in threading_list:
1223
ix = line.find(single_thread_function)
1224
# Comparisons made explicit for clarity -- pylint: disable-msg=C6403
1225
if ix >= 0 and (ix == 0 or (not line[ix - 1].isalnum() and
1226
line[ix - 1] not in ('_', '.', '>'))):
1227
error(filename, linenum, 'runtime/threadsafe_fn', 2,
1228
'Consider using ' + multithread_safe_function +
1229
'...) instead of ' + single_thread_function +
1230
'...) for improved thread safety.')
1233
# Matches invalid increment: *count++, which moves pointer instead of
1234
# incrementing a value.
1235
_RE_PATTERN_INVALID_INCREMENT = re.compile(
1236
r'^\s*\*\w+(\+\+|--);')
1239
def CheckInvalidIncrement(filename, clean_lines, linenum, error):
1240
"""Checks for invalid increment *count++.
1242
For example following function:
1243
void increment_counter(int* count) {
1246
is invalid, because it effectively does count++, moving pointer, and should
1247
be replaced with ++*count, (*count)++ or *count += 1.
1250
filename: The name of the current file.
1251
clean_lines: A CleansedLines instance containing the file.
1252
linenum: The number of the line to check.
1253
error: The function to call with any errors found.
1255
line = clean_lines.elided[linenum]
1256
if _RE_PATTERN_INVALID_INCREMENT.match(line):
1257
error(filename, linenum, 'runtime/invalid_increment', 5,
1258
'Changing pointer instead of value (or unused value of operator*).')
1261
class _ClassInfo(object):
1262
"""Stores information about a class."""
1264
def __init__(self, name, clean_lines, linenum):
1266
self.linenum = linenum
1267
self.seen_open_brace = False
1268
self.is_derived = False
1269
self.virtual_method_linenumber = None
1270
self.has_virtual_destructor = False
1271
self.brace_depth = 0
1273
# Try to find the end of the class. This will be confused by things like:
1277
# But it's still good enough for CheckSectionSpacing.
1280
for i in range(linenum, clean_lines.NumLines()):
1281
line = clean_lines.lines[i]
1282
depth += line.count('{') - line.count('}')
1288
class _ClassState(object):
1289
"""Holds the current state of the parse relating to class declarations.
1291
It maintains a stack of _ClassInfos representing the parser's guess
1292
as to the current nesting of class declarations. The innermost class
1293
is at the top (back) of the stack. Typically, the stack will either
1294
be empty or have exactly one entry.
1298
self.classinfo_stack = []
1300
def CheckFinished(self, filename, error):
1301
"""Checks that all classes have been completely parsed.
1303
Call this when all lines in a file have been processed.
1305
filename: The name of the current file.
1306
error: The function to call with any errors found.
1308
if self.classinfo_stack:
1309
# Note: This test can result in false positives if #ifdef constructs
1310
# get in the way of brace matching. See the testBuildClass test in
1311
# cpplint_unittest.py for an example of this.
1312
error(filename, self.classinfo_stack[0].linenum, 'build/class', 5,
1313
'Failed to find complete declaration of class %s' %
1314
self.classinfo_stack[0].name)
1317
def CheckForNonStandardConstructs(filename, clean_lines, linenum,
1318
class_state, error):
1319
"""Logs an error if we see certain non-ANSI constructs ignored by gcc-2.
1321
Complain about several constructs which gcc-2 accepts, but which are
1322
not standard C++. Warning about these in lint is one way to ease the
1323
transition to new compilers.
1324
- put storage class first (e.g. "static const" instead of "const static").
1325
- "%lld" instead of %qd" in printf-type functions.
1326
- "%1$d" is non-standard in printf-type functions.
1327
- "\%" is an undefined character escape sequence.
1328
- text after #endif is not allowed.
1329
- invalid inner-style forward declaration.
1330
- >? and <? operators, and their >?= and <?= cousins.
1331
- classes with virtual methods need virtual destructors (compiler warning
1332
available, but not turned on yet.)
1334
Additionally, check for constructor/destructor style violations and reference
1335
members, as it is very convenient to do so while checking for
1339
filename: The name of the current file.
1340
clean_lines: A CleansedLines instance containing the file.
1341
linenum: The number of the line to check.
1342
class_state: A _ClassState instance which maintains information about
1343
the current stack of nested class declarations being parsed.
1344
error: A callable to which errors are reported, which takes 4 arguments:
1345
filename, line number, error level, and message
1348
# Remove comments from the line, but leave in strings for now.
1349
line = clean_lines.lines[linenum]
1351
if Search(r'printf\s*\(.*".*%[-+ ]?\d*q', line):
1352
error(filename, linenum, 'runtime/printf_format', 3,
1353
'%q in format strings is deprecated. Use %ll instead.')
1355
if Search(r'printf\s*\(.*".*%\d+\$', line):
1356
error(filename, linenum, 'runtime/printf_format', 2,
1357
'%N$ formats are unconventional. Try rewriting to avoid them.')
1359
# Remove escaped backslashes before looking for undefined escapes.
1360
line = line.replace('\\\\', '')
1362
if Search(r'("|\').*\\(%|\[|\(|{)', line):
1363
error(filename, linenum, 'build/printf_format', 3,
1364
'%, [, (, and { are undefined character escapes. Unescape them.')
1366
# For the rest, work with both comments and strings removed.
1367
line = clean_lines.elided[linenum]
1369
if Search(r'\b(const|volatile|void|char|short|int|long'
1370
r'|float|double|signed|unsigned'
1371
r'|schar|u?int8|u?int16|u?int32|u?int64)'
1372
r'\s+(auto|register|static|extern|typedef)\b',
1374
error(filename, linenum, 'build/storage_class', 5,
1375
'Storage class (static, extern, typedef, etc) should be first.')
1377
if Match(r'\s*#\s*endif\s*[^/\s]+', line):
1378
error(filename, linenum, 'build/endif_comment', 5,
1379
'Uncommented text after #endif is non-standard. Use a comment.')
1381
if Match(r'\s*class\s+(\w+\s*::\s*)+\w+\s*;', line):
1382
error(filename, linenum, 'build/forward_decl', 5,
1383
'Inner-style forward declarations are invalid. Remove this line.')
1385
if Search(r'(\w+|[+-]?\d+(\.\d*)?)\s*(<|>)\?=?\s*(\w+|[+-]?\d+)(\.\d*)?',
1387
error(filename, linenum, 'build/deprecated', 3,
1388
'>? and <? (max and min) operators are non-standard and deprecated.')
1390
if Search(r'^\s*const\s*string\s*&\s*\w+\s*;', line):
1391
# TODO(unknown): Could it be expanded safely to arbitrary references,
1392
# without triggering too many false positives? The first
1393
# attempt triggered 5 warnings for mostly benign code in the regtest, hence
1395
# Here's the original regexp, for the reference:
1396
# type_name = r'\w+((\s*::\s*\w+)|(\s*<\s*\w+?\s*>))?'
1397
# r'\s*const\s*' + type_name + '\s*&\s*\w+\s*;'
1398
error(filename, linenum, 'runtime/member_string_references', 2,
1399
'const string& members are dangerous. It is much better to use '
1400
'alternatives, such as pointers or simple constants.')
1402
# Track class entry and exit, and attempt to find cases within the
1403
# class declaration that don't meet the C++ style
1404
# guidelines. Tracking is very dependent on the code matching Google
1405
# style guidelines, but it seems to perform well enough in testing
1406
# to be a worthwhile addition to the checks.
1407
classinfo_stack = class_state.classinfo_stack
1408
# Look for a class declaration. The regexp accounts for decorated classes
1410
# class LOCKABLE API Object {
1412
class_decl_match = Match(
1413
r'\s*(template\s*<[\w\s<>,:]*>\s*)?'
1414
'(class|struct)\s+([A-Z_]+\s+)*(\w+(::\w+)*)', line)
1415
if class_decl_match:
1416
classinfo_stack.append(_ClassInfo(
1417
class_decl_match.group(4), clean_lines, linenum))
1419
# Everything else in this function uses the top of the stack if it's
1421
if not classinfo_stack:
1424
classinfo = classinfo_stack[-1]
1426
# If the opening brace hasn't been seen look for it and also
1427
# parent class declarations.
1428
if not classinfo.seen_open_brace:
1429
# If the line has a ';' in it, assume it's a forward declaration or
1430
# a single-line class declaration, which we won't process.
1431
if line.find(';') != -1:
1432
classinfo_stack.pop()
1434
classinfo.seen_open_brace = (line.find('{') != -1)
1435
# Look for a bare ':'
1436
if Search('(^|[^:]):($|[^:])', line):
1437
classinfo.is_derived = True
1438
if not classinfo.seen_open_brace:
1439
return # Everything else in this function is for after open brace
1441
# The class may have been declared with namespace or classname qualifiers.
1442
# The constructor and destructor will not have those qualifiers.
1443
base_classname = classinfo.name.split('::')[-1]
1445
# Look for single-argument constructors that aren't marked explicit.
1446
# Technically a valid construct, but against style.
1447
args = Match(r'\s+(?:inline\s+)?%s\s*\(([^,()]+)\)'
1448
% re.escape(base_classname),
1451
args.group(1) != 'void' and
1452
not Match(r'(const\s+)?%s\s*(?:<\w+>\s*)?&' % re.escape(base_classname),
1453
args.group(1).strip())):
1454
error(filename, linenum, 'runtime/explicit', 4,
1455
'Single-argument constructors should be marked explicit.')
1457
# Look for methods declared virtual.
1458
if Search(r'\bvirtual\b', line):
1459
classinfo.virtual_method_linenumber = linenum
1460
# Only look for a destructor declaration on the same line. It would
1461
# be extremely unlikely for the destructor declaration to occupy
1462
# more than one line.
1463
if Search(r'~%s\s*\(' % base_classname, line):
1464
classinfo.has_virtual_destructor = True
1466
# Look for class end.
1467
brace_depth = classinfo.brace_depth
1468
brace_depth = brace_depth + line.count('{') - line.count('}')
1469
if brace_depth <= 0:
1470
classinfo = classinfo_stack.pop()
1471
# Try to detect missing virtual destructor declarations.
1472
# For now, only warn if a non-derived class with virtual methods lacks
1473
# a virtual destructor. This is to make it less likely that people will
1474
# declare derived virtual destructors without declaring the base
1475
# destructor virtual.
1476
if ((classinfo.virtual_method_linenumber is not None) and
1477
(not classinfo.has_virtual_destructor) and
1478
(not classinfo.is_derived)): # Only warn for base classes
1479
error(filename, classinfo.linenum, 'runtime/virtual', 4,
1480
'The class %s probably needs a virtual destructor due to '
1481
'having virtual method(s), one declared at line %d.'
1482
% (classinfo.name, classinfo.virtual_method_linenumber))
1484
classinfo.brace_depth = brace_depth
1487
def CheckSpacingForFunctionCall(filename, line, linenum, error):
1488
"""Checks for the correctness of various spacing around function calls.
1491
filename: The name of the current file.
1492
line: The text of the line to check.
1493
linenum: The number of the line to check.
1494
error: The function to call with any errors found.
1497
# Since function calls often occur inside if/for/while/switch
1498
# expressions - which have their own, more liberal conventions - we
1499
# first see if we should be looking inside such an expression for a
1500
# function call, to which we can apply more strict standards.
1501
fncall = line # if there's no control flow construct, look at whole line
1502
for pattern in (r'\bif\s*\((.*)\)\s*{',
1503
r'\bfor\s*\((.*)\)\s*{',
1504
r'\bwhile\s*\((.*)\)\s*[{;]',
1505
r'\bswitch\s*\((.*)\)\s*{'):
1506
match = Search(pattern, line)
1508
fncall = match.group(1) # look inside the parens for function calls
1511
# Except in if/for/while/switch, there should never be space
1512
# immediately inside parens (eg "f( 3, 4 )"). We make an exception
1513
# for nested parens ( (a+b) + c ). Likewise, there should never be
1514
# a space before a ( when it's a function argument. I assume it's a
1515
# function argument when the char before the whitespace is legal in
1516
# a function name (alnum + _) and we're not starting a macro. Also ignore
1517
# pointers and references to arrays and functions coz they're too tricky:
1518
# we use a very simple way to recognize these:
1519
# " (something)(maybe-something)" or
1520
# " (something)(maybe-something," or
1521
# " (something)[something]"
1522
# Note that we assume the contents of [] to be short enough that
1523
# they'll never need to wrap.
1524
if ( # Ignore control structures.
1525
not Search(r'\b(if|for|while|switch|return|delete)\b', fncall) and
1526
# Ignore pointers/references to functions.
1527
not Search(r' \([^)]+\)\([^)]*(\)|,$)', fncall) and
1528
# Ignore pointers/references to arrays.
1529
not Search(r' \([^)]+\)\[[^\]]+\]', fncall)):
1530
if Search(r'\w\s*\(\s(?!\s*\\$)', fncall): # a ( used for a fn call
1531
error(filename, linenum, 'whitespace/parens', 4,
1532
'Extra space after ( in function call')
1533
elif Search(r'\(\s+(?!(\s*\\)|\()', fncall):
1534
error(filename, linenum, 'whitespace/parens', 2,
1535
'Extra space after (')
1536
if (Search(r'\w\s+\(', fncall) and
1537
not Search(r'#\s*define|typedef', fncall)):
1538
error(filename, linenum, 'whitespace/parens', 4,
1539
'Extra space before ( in function call')
1540
# If the ) is followed only by a newline or a { + newline, assume it's
1541
# part of a control statement (if/while/etc), and don't complain
1542
if Search(r'[^)]\s+\)\s*[^{\s]', fncall):
1543
# If the closing parenthesis is preceded by only whitespaces,
1544
# try to give a more descriptive error message.
1545
if Search(r'^\s+\)', fncall):
1546
error(filename, linenum, 'whitespace/parens', 2,
1547
'Closing ) should be moved to the previous line')
1549
error(filename, linenum, 'whitespace/parens', 2,
1550
'Extra space before )')
1553
def IsBlankLine(line):
1554
"""Returns true if the given line is blank.
1556
We consider a line to be blank if the line is empty or consists of
1560
line: A line of a string.
1563
True, if the given line is blank.
1565
return not line or line.isspace()
1568
def CheckForFunctionLengths(filename, clean_lines, linenum,
1569
function_state, error):
1570
"""Reports for long function bodies.
1572
For an overview why this is done, see:
1573
http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions
1575
Uses a simplistic algorithm assuming other style guidelines
1576
(especially spacing) are followed.
1577
Only checks unindented functions, so class members are unchecked.
1578
Trivial bodies are unchecked, so constructors with huge initializer lists
1580
Blank/comment lines are not counted so as to avoid encouraging the removal
1581
of vertical space and comments just to get through a lint check.
1582
NOLINT *on the last line of a function* disables this check.
1585
filename: The name of the current file.
1586
clean_lines: A CleansedLines instance containing the file.
1587
linenum: The number of the line to check.
1588
function_state: Current function name and lines in body so far.
1589
error: The function to call with any errors found.
1591
lines = clean_lines.lines
1592
line = lines[linenum]
1593
raw = clean_lines.raw_lines
1594
raw_line = raw[linenum]
1597
starting_func = False
1598
regexp = r'(\w(\w|::|\*|\&|\s)*)\(' # decls * & space::name( ...
1599
match_result = Match(regexp, line)
1601
# If the name is all caps and underscores, figure it's a macro and
1602
# ignore it, unless it's TEST or TEST_F.
1603
function_name = match_result.group(1).split()[-1]
1604
if function_name == 'TEST' or function_name == 'TEST_F' or (
1605
not Match(r'[A-Z_]+$', function_name)):
1606
starting_func = True
1610
for start_linenum in xrange(linenum, clean_lines.NumLines()):
1611
start_line = lines[start_linenum]
1612
joined_line += ' ' + start_line.lstrip()
1613
if Search(r'(;|})', start_line): # Declarations and trivial functions
1616
elif Search(r'{', start_line):
1618
function = Search(r'((\w|:)*)\(', line).group(1)
1619
if Match(r'TEST', function): # Handle TEST... macros
1620
parameter_regexp = Search(r'(\(.*\))', joined_line)
1621
if parameter_regexp: # Ignore bad syntax
1622
function += parameter_regexp.group(1)
1625
function_state.Begin(function)
1628
# No body for the function (or evidence of a non-function) was found.
1629
error(filename, linenum, 'readability/fn_size', 5,
1630
'Lint failed to find start of function body.')
1631
elif Match(r'^\}\s*$', line): # function end
1632
function_state.Check(error, filename, linenum)
1633
function_state.End()
1634
elif not Match(r'^\s*$', line):
1635
function_state.Count() # Count non-blank/non-comment lines.
1638
_RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?')
1641
def CheckComment(comment, filename, linenum, error):
1642
"""Checks for common mistakes in TODO comments.
1645
comment: The text of the comment from the line in question.
1646
filename: The name of the current file.
1647
linenum: The number of the line to check.
1648
error: The function to call with any errors found.
1650
match = _RE_PATTERN_TODO.match(comment)
1652
# One whitespace is correct; zero whitespace is handled elsewhere.
1653
leading_whitespace = match.group(1)
1654
if len(leading_whitespace) > 1:
1655
error(filename, linenum, 'whitespace/todo', 2,
1656
'Too many spaces before TODO')
1658
username = match.group(2)
1660
error(filename, linenum, 'readability/todo', 2,
1661
'Missing username in TODO; it should look like '
1662
'"// TODO(my_username): Stuff."')
1664
middle_whitespace = match.group(3)
1665
# Comparisons made explicit for correctness -- pylint: disable-msg=C6403
1666
if middle_whitespace != ' ' and middle_whitespace != '':
1667
error(filename, linenum, 'whitespace/todo', 2,
1668
'TODO(my_username) should be followed by a space')
1671
def CheckSpacing(filename, clean_lines, linenum, error):
1672
"""Checks for the correctness of various spacing issues in the code.
1674
Things we check for: spaces around operators, spaces after
1675
if/for/while/switch, no spaces around parens in function calls, two
1676
spaces between code and comment, don't start a block with a blank
1677
line, don't end a function with a blank line, don't add a blank line
1678
after public/protected/private, don't have too many blank lines in a row.
1681
filename: The name of the current file.
1682
clean_lines: A CleansedLines instance containing the file.
1683
linenum: The number of the line to check.
1684
error: The function to call with any errors found.
1687
raw = clean_lines.raw_lines
1690
# Before nixing comments, check if the line is blank for no good
1691
# reason. This includes the first line after a block is opened, and
1692
# blank lines at the end of a function (ie, right before a line like '}'
1693
if IsBlankLine(line):
1694
elided = clean_lines.elided
1695
prev_line = elided[linenum - 1]
1696
prevbrace = prev_line.rfind('{')
1697
# TODO(unknown): Don't complain if line before blank line, and line after,
1698
# both start with alnums and are indented the same amount.
1699
# This ignores whitespace at the start of a namespace block
1700
# because those are not usually indented.
1701
if (prevbrace != -1 and prev_line[prevbrace:].find('}') == -1
1702
and prev_line[:prevbrace].find('namespace') == -1):
1703
# OK, we have a blank line at the start of a code block. Before we
1704
# complain, we check if it is an exception to the rule: The previous
1705
# non-empty line has the parameters of a function header that are indented
1706
# 4 spaces (because they did not fit in a 80 column line when placed on
1707
# the same line as the function name). We also check for the case where
1708
# the previous line is indented 6 spaces, which may happen when the
1709
# initializers of a constructor do not fit into a 80 column line.
1711
if Match(r' {6}\w', prev_line): # Initializer list?
1712
# We are looking for the opening column of initializer list, which
1713
# should be indented 4 spaces to cause 6 space indentation afterwards.
1714
search_position = linenum-2
1715
while (search_position >= 0
1716
and Match(r' {6}\w', elided[search_position])):
1717
search_position -= 1
1718
exception = (search_position >= 0
1719
and elided[search_position][:5] == ' :')
1721
# Search for the function arguments or an initializer list. We use a
1722
# simple heuristic here: If the line is indented 4 spaces; and we have a
1723
# closing paren, without the opening paren, followed by an opening brace
1724
# or colon (for initializer lists) we assume that it is the last line of
1725
# a function header. If we have a colon indented 4 spaces, it is an
1727
exception = (Match(r' {4}\w[^\(]*\)\s*(const\s*)?(\{\s*$|:)',
1729
or Match(r' {4}:', prev_line))
1732
error(filename, linenum, 'whitespace/blank_line', 5,
1733
'Blank line at the start of a code block. Is this needed?')
1734
# This doesn't ignore whitespace at the end of a namespace block
1735
# because that is too hard without pairing open/close braces;
1736
# however, a special exception is made for namespace closing
1737
# brackets which have a comment containing "namespace".
1739
# Also, ignore blank lines at the end of a block in a long if-else
1742
# // Something followed by a blank line
1744
# } else if (condition2) {
1747
if linenum + 1 < clean_lines.NumLines():
1748
next_line = raw[linenum + 1]
1750
and Match(r'\s*}', next_line)
1751
and next_line.find('namespace') == -1
1752
and next_line.find('} else ') == -1):
1753
error(filename, linenum, 'whitespace/blank_line', 5,
1754
'Blank line at the end of a code block. Is this needed?')
1756
matched = Match(r'\s*(public|protected|private):', prev_line)
1758
error(filename, linenum, 'whitespace/blank_line', 3,
1759
'Do not leave a blank line after "%s:"' % matched.group(1))
1761
# Next, we complain if there's a comment too near the text
1762
commentpos = line.find('//')
1763
if commentpos != -1:
1764
# Check if the // may be in quotes. If so, ignore it
1765
# Comparisons made explicit for clarity -- pylint: disable-msg=C6403
1766
if (line.count('"', 0, commentpos) -
1767
line.count('\\"', 0, commentpos)) % 2 == 0: # not in quotes
1768
# Allow one space for new scopes, two spaces otherwise:
1769
if (not Match(r'^\s*{ //', line) and
1770
((commentpos >= 1 and
1771
line[commentpos-1] not in string.whitespace) or
1772
(commentpos >= 2 and
1773
line[commentpos-2] not in string.whitespace))):
1774
error(filename, linenum, 'whitespace/comments', 2,
1775
'At least two spaces is best between code and comments')
1776
# There should always be a space between the // and the comment
1777
commentend = commentpos + 2
1778
if commentend < len(line) and not line[commentend] == ' ':
1779
# but some lines are exceptions -- e.g. if they're big
1780
# comment delimiters like:
1781
# //----------------------------------------------------------
1782
# or are an empty C++ style Doxygen comment, like:
1784
# or they begin with multiple slashes followed by a space:
1785
# //////// Header comment
1786
match = (Search(r'[=/-]{4,}\s*$', line[commentend:]) or
1787
Search(r'^/$', line[commentend:]) or
1788
Search(r'^/+ ', line[commentend:]))
1790
error(filename, linenum, 'whitespace/comments', 4,
1791
'Should have a space between // and comment')
1792
CheckComment(line[commentpos:], filename, linenum, error)
1794
line = clean_lines.elided[linenum] # get rid of comments and strings
1796
# Don't try to do spacing checks for operator methods
1797
line = re.sub(r'operator(==|!=|<|<<|<=|>=|>>|>)\(', 'operator\(', line)
1799
# We allow no-spaces around = within an if: "if ( (a=Foo()) == 0 )".
1800
# Otherwise not. Note we only check for non-spaces on *both* sides;
1801
# sometimes people put non-spaces on one side when aligning ='s among
1802
# many lines (not that this is behavior that I approve of...)
1803
if Search(r'[\w.]=[\w.]', line) and not Search(r'\b(if|while) ', line):
1804
error(filename, linenum, 'whitespace/operators', 4,
1805
'Missing spaces around =')
1807
# It's ok not to have spaces around binary operators like + - * /, but if
1808
# there's too little whitespace, we get concerned. It's hard to tell,
1809
# though, so we punt on this one for now. TODO.
1811
# You should always have whitespace around binary operators.
1812
# Alas, we can't test < or > because they're legitimately used sans spaces
1813
# (a->b, vector<int> a). The only time we can tell is a < with no >, and
1814
# only if it's not template params list spilling into the next line.
1815
match = Search(r'[^<>=!\s](==|!=|<=|>=)[^<>=!\s]', line)
1817
# Note that while it seems that the '<[^<]*' term in the following
1818
# regexp could be simplified to '<.*', which would indeed match
1819
# the same class of strings, the [^<] means that searching for the
1820
# regexp takes linear rather than quadratic time.
1821
if not Search(r'<[^<]*,\s*$', line): # template params spill
1822
match = Search(r'[^<>=!\s](<)[^<>=!\s]([^>]|->)*$', line)
1824
error(filename, linenum, 'whitespace/operators', 3,
1825
'Missing spaces around %s' % match.group(1))
1826
# We allow no-spaces around << and >> when used like this: 10<<20, but
1827
# not otherwise (particularly, not when used as streams)
1828
match = Search(r'[^0-9\s](<<|>>)[^0-9\s]', line)
1830
error(filename, linenum, 'whitespace/operators', 3,
1831
'Missing spaces around %s' % match.group(1))
1833
# There shouldn't be space around unary operators
1834
match = Search(r'(!\s|~\s|[\s]--[\s;]|[\s]\+\+[\s;])', line)
1836
error(filename, linenum, 'whitespace/operators', 4,
1837
'Extra space for operator %s' % match.group(1))
1839
# A pet peeve of mine: no spaces after an if, while, switch, or for
1840
match = Search(r' (if\(|for\(|while\(|switch\()', line)
1842
error(filename, linenum, 'whitespace/parens', 5,
1843
'Missing space before ( in %s' % match.group(1))
1845
# For if/for/while/switch, the left and right parens should be
1846
# consistent about how many spaces are inside the parens, and
1847
# there should either be zero or one spaces inside the parens.
1848
# We don't want: "if ( foo)" or "if ( foo )".
1849
# Exception: "for ( ; foo; bar)" and "for (foo; bar; )" are allowed.
1850
match = Search(r'\b(if|for|while|switch)\s*'
1851
r'\(([ ]*)(.).*[^ ]+([ ]*)\)\s*{\s*$',
1854
if len(match.group(2)) != len(match.group(4)):
1855
if not (match.group(3) == ';' and
1856
len(match.group(2)) == 1 + len(match.group(4)) or
1857
not match.group(2) and Search(r'\bfor\s*\(.*; \)', line)):
1858
error(filename, linenum, 'whitespace/parens', 5,
1859
'Mismatching spaces inside () in %s' % match.group(1))
1860
if not len(match.group(2)) in [0, 1]:
1861
error(filename, linenum, 'whitespace/parens', 5,
1862
'Should have zero or one spaces inside ( and ) in %s' %
1865
# You should always have a space after a comma (either as fn arg or operator)
1866
if Search(r',[^\s]', line):
1867
error(filename, linenum, 'whitespace/comma', 3,
1868
'Missing space after ,')
1870
# You should always have a space after a semicolon
1871
# except for few corner cases
1872
# TODO(unknown): clarify if 'if (1) { return 1;}' is requires one more
1874
if Search(r';[^\s};\\)/]', line):
1875
error(filename, linenum, 'whitespace/semicolon', 3,
1876
'Missing space after ;')
1878
# Next we will look for issues with function calls.
1879
CheckSpacingForFunctionCall(filename, line, linenum, error)
1881
# Except after an opening paren, or after another opening brace (in case of
1882
# an initializer list, for instance), you should have spaces before your
1883
# braces. And since you should never have braces at the beginning of a line,
1884
# this is an easy test.
1885
if Search(r'[^ ({]{', line):
1886
error(filename, linenum, 'whitespace/braces', 5,
1887
'Missing space before {')
1889
# Make sure '} else {' has spaces.
1890
if Search(r'}else', line):
1891
error(filename, linenum, 'whitespace/braces', 5,
1892
'Missing space before else')
1894
# You shouldn't have spaces before your brackets, except maybe after
1895
# 'delete []' or 'new char * []'.
1896
if Search(r'\w\s+\[', line) and not Search(r'delete\s+\[', line):
1897
error(filename, linenum, 'whitespace/braces', 5,
1898
'Extra space before [')
1900
# You shouldn't have a space before a semicolon at the end of the line.
1901
# There's a special case for "for" since the style guide allows space before
1902
# the semicolon there.
1903
if Search(r':\s*;\s*$', line):
1904
error(filename, linenum, 'whitespace/semicolon', 5,
1905
'Semicolon defining empty statement. Use { } instead.')
1906
elif Search(r'^\s*;\s*$', line):
1907
error(filename, linenum, 'whitespace/semicolon', 5,
1908
'Line contains only semicolon. If this should be an empty statement, '
1910
elif (Search(r'\s+;\s*$', line) and
1911
not Search(r'\bfor\b', line)):
1912
error(filename, linenum, 'whitespace/semicolon', 5,
1913
'Extra space before last semicolon. If this should be an empty '
1914
'statement, use { } instead.')
1917
def CheckSectionSpacing(filename, clean_lines, class_info, linenum, error):
1918
"""Checks for additional blank line issues related to sections.
1920
Currently the only thing checked here is blank line before protected/private.
1923
filename: The name of the current file.
1924
clean_lines: A CleansedLines instance containing the file.
1925
class_info: A _ClassInfo objects.
1926
linenum: The number of the line to check.
1927
error: The function to call with any errors found.
1929
# Skip checks if the class is small, where small means 25 lines or less.
1930
# 25 lines seems like a good cutoff since that's the usual height of
1931
# terminals, and any class that can't fit in one screen can't really
1932
# be considered "small".
1934
# Also skip checks if we are on the first line. This accounts for
1935
# classes that look like
1936
# class Foo { public: ... };
1938
# If we didn't find the end of the class, last_line would be zero,
1939
# and the check will be skipped by the first condition.
1940
if (class_info.last_line - class_info.linenum <= 24 or
1941
linenum <= class_info.linenum):
1944
matched = Match(r'\s*(public|protected|private):', clean_lines.lines[linenum])
1946
# Issue warning if the line before public/protected/private was
1947
# not a blank line, but don't do this if the previous line contains
1948
# "class" or "struct". This can happen two ways:
1949
# - We are at the beginning of the class.
1950
# - We are forward-declaring an inner class that is semantically
1951
# private, but needed to be public for implementation reasons.
1952
prev_line = clean_lines.lines[linenum - 1]
1953
if (not IsBlankLine(prev_line) and
1954
not Search(r'\b(class|struct)\b', prev_line)):
1955
# Try a bit harder to find the beginning of the class. This is to
1956
# account for multi-line base-specifier lists, e.g.:
1959
end_class_head = class_info.linenum
1960
for i in range(class_info.linenum, linenum):
1961
if Search(r'\{\s*$', clean_lines.lines[i]):
1964
if end_class_head < linenum - 1:
1965
error(filename, linenum, 'whitespace/blank_line', 3,
1966
'"%s:" should be preceded by a blank line' % matched.group(1))
1969
def GetPreviousNonBlankLine(clean_lines, linenum):
1970
"""Return the most recent non-blank line and its line number.
1973
clean_lines: A CleansedLines instance containing the file contents.
1974
linenum: The number of the line to check.
1977
A tuple with two elements. The first element is the contents of the last
1978
non-blank line before the current line, or the empty string if this is the
1979
first non-blank line. The second is the line number of that line, or -1
1980
if this is the first non-blank line.
1983
prevlinenum = linenum - 1
1984
while prevlinenum >= 0:
1985
prevline = clean_lines.elided[prevlinenum]
1986
if not IsBlankLine(prevline): # if not a blank line...
1987
return (prevline, prevlinenum)
1992
def CheckBraces(filename, clean_lines, linenum, error):
1993
"""Looks for misplaced braces (e.g. at the end of line).
1996
filename: The name of the current file.
1997
clean_lines: A CleansedLines instance containing the file.
1998
linenum: The number of the line to check.
1999
error: The function to call with any errors found.
2002
line = clean_lines.elided[linenum] # get rid of comments and strings
2004
if Match(r'\s*{\s*$', line):
2005
# We allow an open brace to start a line in the case where someone
2006
# is using braces in a block to explicitly create a new scope,
2007
# which is commonly used to control the lifetime of
2008
# stack-allocated variables. We don't detect this perfectly: we
2009
# just don't complain if the last non-whitespace character on the
2010
# previous non-blank line is ';', ':', '{', or '}'.
2011
prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0]
2012
if not Search(r'[;:}{]\s*$', prevline):
2013
error(filename, linenum, 'whitespace/braces', 4,
2014
'{ should almost always be at the end of the previous line')
2016
# An else clause should be on the same line as the preceding closing brace.
2017
if Match(r'\s*else\s*', line):
2018
prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0]
2019
if Match(r'\s*}\s*$', prevline):
2020
error(filename, linenum, 'whitespace/newline', 4,
2021
'An else should appear on the same line as the preceding }')
2023
# If braces come on one side of an else, they should be on both.
2024
# However, we have to worry about "else if" that spans multiple lines!
2025
if Search(r'}\s*else[^{]*$', line) or Match(r'[^}]*else\s*{', line):
2026
if Search(r'}\s*else if([^{]*)$', line): # could be multi-line if
2027
# find the ( after the if
2028
pos = line.find('else if')
2029
pos = line.find('(', pos)
2031
(endline, _, endpos) = CloseExpression(clean_lines, linenum, pos)
2032
if endline[endpos:].find('{') == -1: # must be brace after if
2033
error(filename, linenum, 'readability/braces', 5,
2034
'If an else has a brace on one side, it should have it on both')
2035
else: # common case: else not followed by a multi-line if
2036
error(filename, linenum, 'readability/braces', 5,
2037
'If an else has a brace on one side, it should have it on both')
2039
# Likewise, an else should never have the else clause on the same line
2040
if Search(r'\belse [^\s{]', line) and not Search(r'\belse if\b', line):
2041
error(filename, linenum, 'whitespace/newline', 4,
2042
'Else clause should never be on same line as else (use 2 lines)')
2044
# In the same way, a do/while should never be on one line
2045
if Match(r'\s*do [^\s{]', line):
2046
error(filename, linenum, 'whitespace/newline', 4,
2047
'do/while clauses should not be on a single line')
2049
# Braces shouldn't be followed by a ; unless they're defining a struct
2050
# or initializing an array.
2051
# We can't tell in general, but we can for some common cases.
2052
prevlinenum = linenum
2054
(prevline, prevlinenum) = GetPreviousNonBlankLine(clean_lines, prevlinenum)
2055
if Match(r'\s+{.*}\s*;', line) and not prevline.count(';'):
2056
line = prevline + line
2059
if (Search(r'{.*}\s*;', line) and
2060
line.count('{') == line.count('}') and
2061
not Search(r'struct|class|enum|\s*=\s*{', line)):
2062
error(filename, linenum, 'readability/braces', 4,
2063
"You don't need a ; after a }")
2066
def ReplaceableCheck(operator, macro, line):
2067
"""Determine whether a basic CHECK can be replaced with a more specific one.
2069
For example suggest using CHECK_EQ instead of CHECK(a == b) and
2070
similarly for CHECK_GE, CHECK_GT, CHECK_LE, CHECK_LT, CHECK_NE.
2073
operator: The C++ operator used in the CHECK.
2074
macro: The CHECK or EXPECT macro being called.
2075
line: The current source line.
2078
True if the CHECK can be replaced with a more specific one.
2081
# This matches decimal and hex integers, strings, and chars (in that order).
2082
match_constant = r'([-+]?(\d+|0[xX][0-9a-fA-F]+)[lLuU]{0,3}|".*"|\'.*\')'
2084
# Expression to match two sides of the operator with something that
2085
# looks like a literal, since CHECK(x == iterator) won't compile.
2086
# This means we can't catch all the cases where a more specific
2087
# CHECK is possible, but it's less annoying than dealing with
2088
# extraneous warnings.
2089
match_this = (r'\s*' + macro + r'\((\s*' +
2090
match_constant + r'\s*' + operator + r'[^<>].*|'
2091
r'.*[^<>]' + operator + r'\s*' + match_constant +
2094
# Don't complain about CHECK(x == NULL) or similar because
2095
# CHECK_EQ(x, NULL) won't compile (requires a cast).
2096
# Also, don't complain about more complex boolean expressions
2097
# involving && or || such as CHECK(a == b || c == d).
2098
return Match(match_this, line) and not Search(r'NULL|&&|\|\|', line)
2101
def CheckCheck(filename, clean_lines, linenum, error):
2102
"""Checks the use of CHECK and EXPECT macros.
2105
filename: The name of the current file.
2106
clean_lines: A CleansedLines instance containing the file.
2107
linenum: The number of the line to check.
2108
error: The function to call with any errors found.
2111
# Decide the set of replacement macros that should be suggested
2112
raw_lines = clean_lines.raw_lines
2114
for macro in _CHECK_MACROS:
2115
if raw_lines[linenum].find(macro) >= 0:
2116
current_macro = macro
2118
if not current_macro:
2119
# Don't waste time here if line doesn't contain 'CHECK' or 'EXPECT'
2122
line = clean_lines.elided[linenum] # get rid of comments and strings
2124
# Encourage replacing plain CHECKs with CHECK_EQ/CHECK_NE/etc.
2125
for operator in ['==', '!=', '>=', '>', '<=', '<']:
2126
if ReplaceableCheck(operator, current_macro, line):
2127
error(filename, linenum, 'readability/check', 2,
2128
'Consider using %s instead of %s(a %s b)' % (
2129
_CHECK_REPLACEMENT[current_macro][operator],
2130
current_macro, operator))
2134
def GetLineWidth(line):
2135
"""Determines the width of the line in column positions.
2138
line: A string, which may be a Unicode string.
2141
The width of the line in column positions, accounting for Unicode
2142
combining characters and wide characters.
2144
if isinstance(line, unicode):
2146
for uc in unicodedata.normalize('NFC', line):
2147
if unicodedata.east_asian_width(uc) in ('W', 'F'):
2149
elif not unicodedata.combining(uc):
2156
def CheckStyle(filename, clean_lines, linenum, file_extension, class_state,
2158
"""Checks rules from the 'C++ style rules' section of cppguide.html.
2160
Most of these rules are hard to test (naming, comment style), but we
2161
do what we can. In particular we check for 2-space indents, line lengths,
2162
tab usage, spaces inside code, etc.
2165
filename: The name of the current file.
2166
clean_lines: A CleansedLines instance containing the file.
2167
linenum: The number of the line to check.
2168
file_extension: The extension (without the dot) of the filename.
2169
error: The function to call with any errors found.
2172
raw_lines = clean_lines.raw_lines
2173
line = raw_lines[linenum]
2175
if line.find('\t') != -1:
2176
error(filename, linenum, 'whitespace/tab', 5,
2177
'Tab found; better to use spaces')
2179
# One or three blank spaces at the beginning of the line is weird; it's
2180
# hard to reconcile that with 2-space indents.
2181
# NOTE: here are the conditions rob pike used for his tests. Mine aren't
2182
# as sophisticated, but it may be worth becoming so: RLENGTH==initial_spaces
2183
# if(RLENGTH > 20) complain = 0;
2184
# if(match($0, " +(error|private|public|protected):")) complain = 0;
2185
# if(match(prev, "&& *$")) complain = 0;
2186
# if(match(prev, "\\|\\| *$")) complain = 0;
2187
# if(match(prev, "[\",=><] *$")) complain = 0;
2188
# if(match($0, " <<")) complain = 0;
2189
# if(match(prev, " +for \\(")) complain = 0;
2190
# if(prevodd && match(prevprev, " +for \\(")) complain = 0;
2192
cleansed_line = clean_lines.elided[linenum]
2193
while initial_spaces < len(line) and line[initial_spaces] == ' ':
2195
if line and line[-1].isspace():
2196
error(filename, linenum, 'whitespace/end_of_line', 4,
2197
'Line ends in whitespace. Consider deleting these extra spaces.')
2198
# There are certain situations we allow one space, notably for labels
2199
elif ((initial_spaces == 1 or initial_spaces == 3) and
2200
not Match(r'\s*\w+\s*:\s*$', cleansed_line)):
2201
error(filename, linenum, 'whitespace/indent', 3,
2202
'Weird number of spaces at line-start. '
2203
'Are you using a 2-space indent?')
2204
# Labels should always be indented at least one space.
2205
elif not initial_spaces and line[:2] != '//' and Search(r'[^:]:\s*$',
2207
error(filename, linenum, 'whitespace/labels', 4,
2208
'Labels should always be indented at least one space. '
2209
'If this is a member-initializer list in a constructor or '
2210
'the base class list in a class definition, the colon should '
2211
'be on the following line.')
2214
# Check if the line is a header guard.
2215
is_header_guard = False
2216
if file_extension == 'h':
2217
cppvar = GetHeaderGuardCPPVariable(filename)
2218
if (line.startswith('#ifndef %s' % cppvar) or
2219
line.startswith('#define %s' % cppvar) or
2220
line.startswith('#endif // %s' % cppvar)):
2221
is_header_guard = True
2222
# #include lines and header guards can be long, since there's no clean way to
2225
# URLs can be long too. It's possible to split these, but it makes them
2226
# harder to cut&paste.
2228
# The "$Id:...$" comment may also get very long without it being the
2230
if (not line.startswith('#include') and not is_header_guard and
2231
not Match(r'^\s*//.*http(s?)://\S*$', line) and
2232
not Match(r'^// \$Id:.*#[0-9]+ \$$', line)):
2233
line_width = GetLineWidth(line)
2234
if line_width > 100:
2235
error(filename, linenum, 'whitespace/line_length', 4,
2236
'Lines should very rarely be longer than 100 characters')
2237
elif line_width > 80:
2238
error(filename, linenum, 'whitespace/line_length', 2,
2239
'Lines should be <= 80 characters long')
2241
if (cleansed_line.count(';') > 1 and
2242
# for loops are allowed two ;'s (and may run over two lines).
2243
cleansed_line.find('for') == -1 and
2244
(GetPreviousNonBlankLine(clean_lines, linenum)[0].find('for') == -1 or
2245
GetPreviousNonBlankLine(clean_lines, linenum)[0].find(';') != -1) and
2246
# It's ok to have many commands in a switch case that fits in 1 line
2247
not ((cleansed_line.find('case ') != -1 or
2248
cleansed_line.find('default:') != -1) and
2249
cleansed_line.find('break;') != -1)):
2250
error(filename, linenum, 'whitespace/newline', 4,
2251
'More than one command on the same line')
2253
# Some more style checks
2254
CheckBraces(filename, clean_lines, linenum, error)
2255
CheckSpacing(filename, clean_lines, linenum, error)
2256
CheckCheck(filename, clean_lines, linenum, error)
2257
if class_state and class_state.classinfo_stack:
2258
CheckSectionSpacing(filename, clean_lines,
2259
class_state.classinfo_stack[-1], linenum, error)
2262
_RE_PATTERN_INCLUDE_NEW_STYLE = re.compile(r'#include +"[^/]+\.h"')
2263
_RE_PATTERN_INCLUDE = re.compile(r'^\s*#\s*include\s*([<"])([^>"]*)[>"].*$')
2264
# Matches the first component of a filename delimited by -s and _s. That is:
2265
# _RE_FIRST_COMPONENT.match('foo').group(0) == 'foo'
2266
# _RE_FIRST_COMPONENT.match('foo.cc').group(0) == 'foo'
2267
# _RE_FIRST_COMPONENT.match('foo-bar_baz.cc').group(0) == 'foo'
2268
# _RE_FIRST_COMPONENT.match('foo_bar-baz.cc').group(0) == 'foo'
2269
_RE_FIRST_COMPONENT = re.compile(r'^[^-_.]+')
2272
def _DropCommonSuffixes(filename):
2273
"""Drops common suffixes like _test.cc or -inl.h from filename.
2276
>>> _DropCommonSuffixes('foo/foo-inl.h')
2278
>>> _DropCommonSuffixes('foo/bar/foo.cc')
2280
>>> _DropCommonSuffixes('foo/foo_internal.h')
2282
>>> _DropCommonSuffixes('foo/foo_unusualinternal.h')
2283
'foo/foo_unusualinternal'
2286
filename: The input filename.
2289
The filename with the common suffix removed.
2291
for suffix in ('test.cc', 'regtest.cc', 'unittest.cc',
2292
'inl.h', 'impl.h', 'internal.h'):
2293
if (filename.endswith(suffix) and len(filename) > len(suffix) and
2294
filename[-len(suffix) - 1] in ('-', '_')):
2295
return filename[:-len(suffix) - 1]
2296
return os.path.splitext(filename)[0]
2299
def _IsTestFilename(filename):
2300
"""Determines if the given filename has a suffix that identifies it as a test.
2303
filename: The input filename.
2306
True if 'filename' looks like a test, False otherwise.
2308
if (filename.endswith('_test.cc') or
2309
filename.endswith('_unittest.cc') or
2310
filename.endswith('_regtest.cc')):
2316
def _ClassifyInclude(fileinfo, include, is_system):
2317
"""Figures out what kind of header 'include' is.
2320
fileinfo: The current file cpplint is running over. A FileInfo instance.
2321
include: The path to a #included file.
2322
is_system: True if the #include used <> rather than "".
2325
One of the _XXX_HEADER constants.
2328
>>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'stdio.h', True)
2330
>>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'string', True)
2332
>>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/foo.h', False)
2334
>>> _ClassifyInclude(FileInfo('foo/foo_unknown_extension.cc'),
2335
... 'bar/foo_other_ext.h', False)
2337
>>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/bar.h', False)
2340
# This is a list of all standard c++ header files, except
2341
# those already checked for above.
2342
is_stl_h = include in _STL_HEADERS
2343
is_cpp_h = is_stl_h or include in _CPP_HEADERS
2347
return _CPP_SYS_HEADER
2349
return _C_SYS_HEADER
2351
# If the target file and the include we're checking share a
2352
# basename when we drop common extensions, and the include
2353
# lives in . , then it's likely to be owned by the target file.
2354
target_dir, target_base = (
2355
os.path.split(_DropCommonSuffixes(fileinfo.RepositoryName())))
2356
include_dir, include_base = os.path.split(_DropCommonSuffixes(include))
2357
if target_base == include_base and (
2358
include_dir == target_dir or
2359
include_dir == os.path.normpath(target_dir + '/../public')):
2360
return _LIKELY_MY_HEADER
2362
# If the target and include share some initial basename
2363
# component, it's possible the target is implementing the
2364
# include, so it's allowed to be first, but we'll never
2365
# complain if it's not there.
2366
target_first_component = _RE_FIRST_COMPONENT.match(target_base)
2367
include_first_component = _RE_FIRST_COMPONENT.match(include_base)
2368
if (target_first_component and include_first_component and
2369
target_first_component.group(0) ==
2370
include_first_component.group(0)):
2371
return _POSSIBLE_MY_HEADER
2373
return _OTHER_HEADER
2377
def CheckIncludeLine(filename, clean_lines, linenum, include_state, error):
2378
"""Check rules that are applicable to #include lines.
2380
Strings on #include lines are NOT removed from elided line, to make
2381
certain tasks easier. However, to prevent false positives, checks
2382
applicable to #include lines in CheckLanguage must be put here.
2385
filename: The name of the current file.
2386
clean_lines: A CleansedLines instance containing the file.
2387
linenum: The number of the line to check.
2388
include_state: An _IncludeState instance in which the headers are inserted.
2389
error: The function to call with any errors found.
2391
fileinfo = FileInfo(filename)
2393
line = clean_lines.lines[linenum]
2395
# "include" should use the new style "foo/bar.h" instead of just "bar.h"
2396
if _RE_PATTERN_INCLUDE_NEW_STYLE.search(line):
2397
error(filename, linenum, 'build/include', 4,
2398
'Include the directory when naming .h files')
2400
# we shouldn't include a file more than once. actually, there are a
2401
# handful of instances where doing so is okay, but in general it's
2403
match = _RE_PATTERN_INCLUDE.search(line)
2405
include = match.group(2)
2406
is_system = (match.group(1) == '<')
2407
if include in include_state:
2408
error(filename, linenum, 'build/include', 4,
2409
'"%s" already included at %s:%s' %
2410
(include, filename, include_state[include]))
2412
include_state[include] = linenum
2414
# We want to ensure that headers appear in the right order:
2415
# 1) for foo.cc, foo.h (preferred location)
2417
# 3) cpp system files
2418
# 4) for foo.cc, foo.h (deprecated location)
2419
# 5) other google headers
2421
# We classify each include statement as one of those 5 types
2422
# using a number of techniques. The include_state object keeps
2423
# track of the highest type seen, and complains if we see a
2424
# lower type after that.
2425
error_message = include_state.CheckNextIncludeOrder(
2426
_ClassifyInclude(fileinfo, include, is_system))
2428
error(filename, linenum, 'build/include_order', 4,
2429
'%s. Should be: %s.h, c system, c++ system, other.' %
2430
(error_message, fileinfo.BaseName()))
2431
if not include_state.IsInAlphabeticalOrder(include):
2432
error(filename, linenum, 'build/include_alpha', 4,
2433
'Include "%s" not in alphabetical order' % include)
2435
# Look for any of the stream classes that are part of standard C++.
2436
match = _RE_PATTERN_INCLUDE.match(line)
2438
include = match.group(2)
2439
if Match(r'(f|ind|io|i|o|parse|pf|stdio|str|)?stream$', include):
2440
# Many unit tests use cout, so we exempt them.
2441
if not _IsTestFilename(filename):
2442
error(filename, linenum, 'readability/streams', 3,
2443
'Streams are highly discouraged.')
2446
def _GetTextInside(text, start_pattern):
2447
"""Retrieves all the text between matching open and close parentheses.
2449
Given a string of lines and a regular expression string, retrieve all the text
2450
following the expression and between opening punctuation symbols like
2451
(, [, or {, and the matching close-punctuation symbol. This properly nested
2452
occurrences of the punctuations, so for the text like
2453
printf(a(), b(c()));
2454
a call to _GetTextInside(text, r'printf\(') will return 'a(), b(c())'.
2455
start_pattern must match string having an open punctuation symbol at the end.
2458
text: The lines to extract text. Its comments and strings must be elided.
2459
It can be single line and can span multiple lines.
2460
start_pattern: The regexp string indicating where to start extracting
2464
None if either the opening string or ending punctuation could not be found.
2466
# TODO(sugawarayu): Audit cpplint.py to see what places could be profitably
2467
# rewritten to use _GetTextInside (and use inferior regexp matching today).
2469
# Give opening punctuations to get the matching close-punctuations.
2470
matching_punctuation = {'(': ')', '{': '}', '[': ']'}
2471
closing_punctuation = set(matching_punctuation.itervalues())
2473
# Find the position to start extracting text.
2474
match = re.search(start_pattern, text, re.M)
2475
if not match: # start_pattern not found in text.
2477
start_position = match.end(0)
2479
assert start_position > 0, (
2480
'start_pattern must ends with an opening punctuation.')
2481
assert text[start_position - 1] in matching_punctuation, (
2482
'start_pattern must ends with an opening punctuation.')
2483
# Stack of closing punctuations we expect to have in text after position.
2484
punctuation_stack = [matching_punctuation[text[start_position - 1]]]
2485
position = start_position
2486
while punctuation_stack and position < len(text):
2487
if text[position] == punctuation_stack[-1]:
2488
punctuation_stack.pop()
2489
elif text[position] in closing_punctuation:
2490
# A closing punctuation without matching opening punctuations.
2492
elif text[position] in matching_punctuation:
2493
punctuation_stack.append(matching_punctuation[text[position]])
2495
if punctuation_stack:
2496
# Opening punctuations left without matching close-punctuations.
2498
# punctuations match.
2499
return text[start_position:position - 1]
2502
def CheckLanguage(filename, clean_lines, linenum, file_extension, include_state,
2504
"""Checks rules from the 'C++ language rules' section of cppguide.html.
2506
Some of these rules are hard to test (function overloading, using
2507
uint32 inappropriately), but we do the best we can.
2510
filename: The name of the current file.
2511
clean_lines: A CleansedLines instance containing the file.
2512
linenum: The number of the line to check.
2513
file_extension: The extension (without the dot) of the filename.
2514
include_state: An _IncludeState instance in which the headers are inserted.
2515
error: The function to call with any errors found.
2517
# If the line is empty or consists of entirely a comment, no need to
2519
line = clean_lines.elided[linenum]
2523
match = _RE_PATTERN_INCLUDE.search(line)
2525
CheckIncludeLine(filename, clean_lines, linenum, include_state, error)
2528
# Create an extended_line, which is the concatenation of the current and
2529
# next lines, for more effective checking of code that may span more than one
2531
if linenum + 1 < clean_lines.NumLines():
2532
extended_line = line + clean_lines.elided[linenum + 1]
2534
extended_line = line
2536
# Make Windows paths like Unix.
2537
fullname = os.path.abspath(filename).replace('\\', '/')
2539
# TODO(unknown): figure out if they're using default arguments in fn proto.
2541
# Check for non-const references in functions. This is tricky because &
2542
# is also used to take the address of something. We allow <> for templates,
2543
# (ignoring whatever is between the braces) and : for classes.
2544
# These are complicated re's. They try to capture the following:
2545
# paren (for fn-prototype start), typename, &, varname. For the const
2546
# version, we're willing for const to be before typename or after
2547
# Don't check the implementation on same line.
2548
fnline = line.split('{', 1)[0]
2549
if (len(re.findall(r'\([^()]*\b(?:[\w:]|<[^()]*>)+(\s?&|&\s?)\w+', fnline)) >
2550
len(re.findall(r'\([^()]*\bconst\s+(?:typename\s+)?(?:struct\s+)?'
2551
r'(?:[\w:]|<[^()]*>)+(\s?&|&\s?)\w+', fnline)) +
2552
len(re.findall(r'\([^()]*\b(?:[\w:]|<[^()]*>)+\s+const(\s?&|&\s?)[\w]+',
2555
# We allow non-const references in a few standard places, like functions
2556
# called "swap()" or iostream operators like "<<" or ">>".
2558
r'(swap|Swap|operator[<>][<>])\s*\(\s*(?:[\w:]|<.*>)+\s*&',
2560
error(filename, linenum, 'runtime/references', 2,
2561
'Is this a non-const reference? '
2562
'If so, make const or use a pointer.')
2564
# Check to see if they're using an conversion function cast.
2565
# I just try to capture the most common basic types, though there are more.
2566
# Parameterless conversion functions, such as bool(), are allowed as they are
2567
# probably a member operator declaration or default constructor.
2569
r'(\bnew\s+)?\b' # Grab 'new' operator, if it's there
2570
r'(int|float|double|bool|char|int32|uint32|int64|uint64)\([^)]', line)
2572
# gMock methods are defined using some variant of MOCK_METHODx(name, type)
2573
# where type may be float(), int(string), etc. Without context they are
2574
# virtually indistinguishable from int(x) casts. Likewise, gMock's
2575
# MockCallback takes a template parameter of the form return_type(arg_type),
2576
# which looks much like the cast we're trying to detect.
2577
if (match.group(1) is None and # If new operator, then this isn't a cast
2578
not (Match(r'^\s*MOCK_(CONST_)?METHOD\d+(_T)?\(', line) or
2579
Match(r'^\s*MockCallback<.*>', line))):
2580
error(filename, linenum, 'readability/casting', 4,
2581
'Using deprecated casting style. '
2582
'Use static_cast<%s>(...) instead' %
2585
CheckCStyleCast(filename, linenum, line, clean_lines.raw_lines[linenum],
2587
r'\((int|float|double|bool|char|u?int(16|32|64))\)', error)
2589
# This doesn't catch all cases. Consider (const char * const)"hello".
2591
# (char *) "foo" should always be a const_cast (reinterpret_cast won't
2593
if CheckCStyleCast(filename, linenum, line, clean_lines.raw_lines[linenum],
2594
'const_cast', r'\((char\s?\*+\s?)\)\s*"', error):
2597
# Check pointer casts for other than string constants
2598
CheckCStyleCast(filename, linenum, line, clean_lines.raw_lines[linenum],
2599
'reinterpret_cast', r'\((\w+\s?\*+\s?)\)', error)
2601
# In addition, we look for people taking the address of a cast. This
2602
# is dangerous -- casts can assign to temporaries, so the pointer doesn't
2603
# point where you think.
2605
r'(&\([^)]+\)[\w(])|(&(static|dynamic|reinterpret)_cast\b)', line):
2606
error(filename, linenum, 'runtime/casting', 4,
2607
('Are you taking an address of a cast? '
2608
'This is dangerous: could be a temp var. '
2609
'Take the address before doing the cast, rather than after'))
2611
# Check for people declaring static/global STL strings at the top level.
2612
# This is dangerous because the C++ language does not guarantee that
2613
# globals with constructors are initialized before the first access.
2615
r'((?:|static +)(?:|const +))string +([a-zA-Z0-9_:]+)\b(.*)',
2617
# Make sure it's not a function.
2618
# Function template specialization looks like: "string foo<Type>(...".
2619
# Class template definitions look like: "string Foo<Type>::Method(...".
2620
if match and not Match(r'\s*(<.*>)?(::[a-zA-Z0-9_]+)?\s*\(([^"]|$)',
2622
error(filename, linenum, 'runtime/string', 4,
2623
'For a static/global string constant, use a C style string instead: '
2625
(match.group(1), match.group(2)))
2627
# Check that we're not using RTTI outside of testing code.
2628
if Search(r'\bdynamic_cast<', line) and not _IsTestFilename(filename):
2629
error(filename, linenum, 'runtime/rtti', 5,
2630
'Do not use dynamic_cast<>. If you need to cast within a class '
2631
"hierarchy, use static_cast<> to upcast. Google doesn't support "
2634
if Search(r'\b([A-Za-z0-9_]*_)\(\1\)', line):
2635
error(filename, linenum, 'runtime/init', 4,
2636
'You seem to be initializing a member variable with itself.')
2638
if file_extension == 'h':
2639
# TODO(unknown): check that 1-arg constructors are explicit.
2640
# How to tell it's a constructor?
2641
# (handled in CheckForNonStandardConstructs for now)
2642
# TODO(unknown): check that classes have DISALLOW_EVIL_CONSTRUCTORS
2646
# Check if people are using the verboten C basic types. The only exception
2647
# we regularly allow is "unsigned short port" for port.
2648
if Search(r'\bshort port\b', line):
2649
if not Search(r'\bunsigned short port\b', line):
2650
error(filename, linenum, 'runtime/int', 4,
2651
'Use "unsigned short" for ports, not "short"')
2653
match = Search(r'\b(short|long(?! +double)|long long)\b', line)
2655
error(filename, linenum, 'runtime/int', 4,
2656
'Use int16/int64/etc, rather than the C type %s' % match.group(1))
2658
# When snprintf is used, the second argument shouldn't be a literal.
2659
match = Search(r'snprintf\s*\(([^,]*),\s*([0-9]*)\s*,', line)
2660
if match and match.group(2) != '0':
2661
# If 2nd arg is zero, snprintf is used to calculate size.
2662
error(filename, linenum, 'runtime/printf', 3,
2663
'If you can, use sizeof(%s) instead of %s as the 2nd arg '
2664
'to snprintf.' % (match.group(1), match.group(2)))
2666
# Check if some verboten C functions are being used.
2667
if Search(r'\bsprintf\b', line):
2668
error(filename, linenum, 'runtime/printf', 5,
2669
'Never use sprintf. Use snprintf instead.')
2670
match = Search(r'\b(strcpy|strcat)\b', line)
2672
error(filename, linenum, 'runtime/printf', 4,
2673
'Almost always, snprintf is better than %s' % match.group(1))
2675
if Search(r'\bsscanf\b', line):
2676
error(filename, linenum, 'runtime/printf', 1,
2677
'sscanf can be ok, but is slow and can overflow buffers.')
2679
# Check if some verboten operator overloading is going on
2680
# TODO(unknown): catch out-of-line unary operator&:
2682
# int operator&(const X& x) { return 42; } // unary operator&
2683
# The trick is it's hard to tell apart from binary operator&:
2684
# class Y { int operator&(const Y& x) { return 23; } }; // binary operator&
2685
if Search(r'\boperator\s*&\s*\(\s*\)', line):
2686
error(filename, linenum, 'runtime/operator', 4,
2687
'Unary operator& is dangerous. Do not use it.')
2689
# Check for suspicious usage of "if" like
2691
if Search(r'\}\s*if\s*\(', line):
2692
error(filename, linenum, 'readability/braces', 4,
2693
'Did you mean "else if"? If not, start a new line for "if".')
2695
# Check for potential format string bugs like printf(foo).
2696
# We constrain the pattern not to pick things like DocidForPrintf(foo).
2697
# Not perfect but it can catch printf(foo.c_str()) and printf(foo->c_str())
2698
# TODO(sugawarayu): Catch the following case. Need to change the calling
2699
# convention of the whole function to process multiple line to handle it.
2701
# boy_this_is_a_really_long_variable_that_cannot_fit_on_the_prev_line);
2702
printf_args = _GetTextInside(line, r'(?i)\b(string)?printf\s*\(')
2704
match = Match(r'([\w.\->()]+)$', printf_args)
2706
function_name = re.search(r'\b((?:string)?printf)\s*\(',
2707
line, re.I).group(1)
2708
error(filename, linenum, 'runtime/printf', 4,
2709
'Potential format string bug. Do %s("%%s", %s) instead.'
2710
% (function_name, match.group(1)))
2712
# Check for potential memset bugs like memset(buf, sizeof(buf), 0).
2713
match = Search(r'memset\s*\(([^,]*),\s*([^,]*),\s*0\s*\)', line)
2714
if match and not Match(r"^''|-?[0-9]+|0x[0-9A-Fa-f]$", match.group(2)):
2715
error(filename, linenum, 'runtime/memset', 4,
2716
'Did you mean "memset(%s, 0, %s)"?'
2717
% (match.group(1), match.group(2)))
2719
if Search(r'\busing namespace\b', line):
2720
error(filename, linenum, 'build/namespaces', 5,
2721
'Do not use namespace using-directives. '
2722
'Use using-declarations instead.')
2724
# Detect variable-length arrays.
2725
match = Match(r'\s*(.+::)?(\w+) [a-z]\w*\[(.+)];', line)
2726
if (match and match.group(2) != 'return' and match.group(2) != 'delete' and
2727
match.group(3).find(']') == -1):
2728
# Split the size using space and arithmetic operators as delimiters.
2729
# If any of the resulting tokens are not compile time constants then
2731
tokens = re.split(r'\s|\+|\-|\*|\/|<<|>>]', match.group(3))
2739
if Search(r'sizeof\(.+\)', tok): continue
2740
if Search(r'arraysize\(\w+\)', tok): continue
2742
tok = tok.lstrip('(')
2743
tok = tok.rstrip(')')
2744
if not tok: continue
2745
if Match(r'\d+', tok): continue
2746
if Match(r'0[xX][0-9a-fA-F]+', tok): continue
2747
if Match(r'k[A-Z0-9]\w*', tok): continue
2748
if Match(r'(.+::)?k[A-Z0-9]\w*', tok): continue
2749
if Match(r'(.+::)?[A-Z][A-Z0-9_]*', tok): continue
2750
# A catch all for tricky sizeof cases, including 'sizeof expression',
2751
# 'sizeof(*type)', 'sizeof(const type)', 'sizeof(struct StructName)'
2752
# requires skipping the next token because we split on ' ' and '*'.
2753
if tok.startswith('sizeof'):
2759
error(filename, linenum, 'runtime/arrays', 1,
2760
'Do not use variable-length arrays. Use an appropriately named '
2761
"('k' followed by CamelCase) compile-time constant for the size.")
2763
# If DISALLOW_EVIL_CONSTRUCTORS, DISALLOW_COPY_AND_ASSIGN, or
2764
# DISALLOW_IMPLICIT_CONSTRUCTORS is present, then it should be the last thing
2765
# in the class declaration.
2768
r'(DISALLOW_(EVIL_CONSTRUCTORS|COPY_AND_ASSIGN|IMPLICIT_CONSTRUCTORS))'
2771
if match and linenum + 1 < clean_lines.NumLines():
2772
next_line = clean_lines.elided[linenum + 1]
2773
# We allow some, but not all, declarations of variables to be present
2774
# in the statement that defines the class. The [\w\*,\s]* fragment of
2775
# the regular expression below allows users to declare instances of
2776
# the class or pointers to instances, but not less common types such
2777
# as function pointers or arrays. It's a tradeoff between allowing
2778
# reasonable code and avoiding trying to parse more C++ using regexps.
2779
if not Search(r'^\s*}[\w\*,\s]*;', next_line):
2780
error(filename, linenum, 'readability/constructors', 3,
2781
match.group(1) + ' should be the last thing in the class')
2783
# Check for use of unnamed namespaces in header files. Registration
2784
# macros are typically OK, so we allow use of "namespace {" on lines
2785
# that end with backslashes.
2786
if (file_extension == 'h'
2787
and Search(r'\bnamespace\s*{', line)
2788
and line[-1] != '\\'):
2789
error(filename, linenum, 'build/namespaces', 4,
2790
'Do not use unnamed namespaces in header files. See '
2791
'http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Namespaces'
2792
' for more information.')
2795
def CheckCStyleCast(filename, linenum, line, raw_line, cast_type, pattern,
2797
"""Checks for a C-style cast by looking for the pattern.
2799
This also handles sizeof(type) warnings, due to similarity of content.
2802
filename: The name of the current file.
2803
linenum: The number of the line to check.
2804
line: The line of code to check.
2805
raw_line: The raw line of code to check, with comments.
2806
cast_type: The string for the C++ cast to recommend. This is either
2807
reinterpret_cast, static_cast, or const_cast, depending.
2808
pattern: The regular expression used to find C-style casts.
2809
error: The function to call with any errors found.
2812
True if an error was emitted.
2815
match = Search(pattern, line)
2820
sizeof_match = Match(r'.*sizeof\s*$', line[0:match.start(1) - 1])
2822
error(filename, linenum, 'runtime/sizeof', 1,
2823
'Using sizeof(type). Use sizeof(varname) instead if possible')
2826
remainder = line[match.end(0):]
2828
# The close paren is for function pointers as arguments to a function.
2829
# eg, void foo(void (*bar)(int));
2830
# The semicolon check is a more basic function check; also possibly a
2831
# function pointer typedef.
2832
# eg, void foo(int); or void foo(int) const;
2833
# The equals check is for function pointer assignment.
2834
# eg, void *(*foo)(int) = ...
2835
# The > is for MockCallback<...> ...
2837
# Right now, this will only catch cases where there's a single argument, and
2838
# it's unnamed. It should probably be expanded to check for multiple
2839
# arguments with some unnamed.
2840
function_match = Match(r'\s*(\)|=|(const)?\s*(;|\{|throw\(\)|>))', remainder)
2842
if (not function_match.group(3) or
2843
function_match.group(3) == ';' or
2844
('MockCallback<' not in raw_line and
2845
'/*' not in raw_line)):
2846
error(filename, linenum, 'readability/function', 3,
2847
'All parameters should be named in a function')
2850
# At this point, all that should be left is actual casts.
2851
error(filename, linenum, 'readability/casting', 4,
2852
'Using C-style cast. Use %s<%s>(...) instead' %
2853
(cast_type, match.group(1)))
2858
_HEADERS_CONTAINING_TEMPLATES = (
2859
('<deque>', ('deque',)),
2860
('<functional>', ('unary_function', 'binary_function',
2861
'plus', 'minus', 'multiplies', 'divides', 'modulus',
2863
'equal_to', 'not_equal_to', 'greater', 'less',
2864
'greater_equal', 'less_equal',
2865
'logical_and', 'logical_or', 'logical_not',
2866
'unary_negate', 'not1', 'binary_negate', 'not2',
2867
'bind1st', 'bind2nd',
2868
'pointer_to_unary_function',
2869
'pointer_to_binary_function',
2871
'mem_fun_t', 'mem_fun', 'mem_fun1_t', 'mem_fun1_ref_t',
2873
'const_mem_fun_t', 'const_mem_fun1_t',
2874
'const_mem_fun_ref_t', 'const_mem_fun1_ref_t',
2877
('<limits>', ('numeric_limits',)),
2878
('<list>', ('list',)),
2879
('<map>', ('map', 'multimap',)),
2880
('<memory>', ('allocator',)),
2881
('<queue>', ('queue', 'priority_queue',)),
2882
('<set>', ('set', 'multiset',)),
2883
('<stack>', ('stack',)),
2884
('<string>', ('char_traits', 'basic_string',)),
2885
('<utility>', ('pair',)),
2886
('<vector>', ('vector',)),
2889
# Note: std::hash is their hash, ::hash is our hash
2890
('<hash_map>', ('hash_map', 'hash_multimap',)),
2891
('<hash_set>', ('hash_set', 'hash_multiset',)),
2892
('<slist>', ('slist',)),
2895
_RE_PATTERN_STRING = re.compile(r'\bstring\b')
2897
_re_pattern_algorithm_header = []
2898
for _template in ('copy', 'max', 'min', 'min_element', 'sort', 'swap',
2900
# Match max<type>(..., ...), max(..., ...), but not foo->max, foo.max or
2902
_re_pattern_algorithm_header.append(
2903
(re.compile(r'[^>.]\b' + _template + r'(<.*?>)?\([^\)]'),
2907
_re_pattern_templates = []
2908
for _header, _templates in _HEADERS_CONTAINING_TEMPLATES:
2909
for _template in _templates:
2910
_re_pattern_templates.append(
2911
(re.compile(r'(\<|\b)' + _template + r'\s*\<'),
2916
def FilesBelongToSameModule(filename_cc, filename_h):
2917
"""Check if these two filenames belong to the same module.
2919
The concept of a 'module' here is a as follows:
2920
foo.h, foo-inl.h, foo.cc, foo_test.cc and foo_unittest.cc belong to the
2921
same 'module' if they are in the same directory.
2922
some/path/public/xyzzy and some/path/internal/xyzzy are also considered
2923
to belong to the same module here.
2925
If the filename_cc contains a longer path than the filename_h, for example,
2926
'/absolute/path/to/base/sysinfo.cc', and this file would include
2927
'base/sysinfo.h', this function also produces the prefix needed to open the
2928
header. This is used by the caller of this function to more robustly open the
2929
header file. We don't have access to the real include paths in this context,
2930
so we need this guesswork here.
2932
Known bugs: tools/base/bar.cc and base/bar.h belong to the same module
2933
according to this implementation. Because of this, this function gives
2934
some false positives. This should be sufficiently rare in practice.
2937
filename_cc: is the path for the .cc file
2938
filename_h: is the path for the header path
2941
Tuple with a bool and a string:
2942
bool: True if filename_cc and filename_h belong to the same module.
2943
string: the additional prefix needed to open the header file.
2946
if not filename_cc.endswith('.C'):
2948
filename_cc = filename_cc[:-len('.C')]
2949
if filename_cc.endswith('_unittest'):
2950
filename_cc = filename_cc[:-len('_unittest')]
2951
elif filename_cc.endswith('_test'):
2952
filename_cc = filename_cc[:-len('_test')]
2953
filename_cc = filename_cc.replace('/public/', '/')
2954
filename_cc = filename_cc.replace('/internal/', '/')
2956
if not filename_h.endswith('.h'):
2958
filename_h = filename_h[:-len('.h')]
2959
if filename_h.endswith('-inl'):
2960
filename_h = filename_h[:-len('-inl')]
2961
filename_h = filename_h.replace('/public/', '/')
2962
filename_h = filename_h.replace('/internal/', '/')
2964
files_belong_to_same_module = filename_cc.endswith(filename_h)
2966
if files_belong_to_same_module:
2967
common_path = filename_cc[:-len(filename_h)]
2968
return files_belong_to_same_module, common_path
2971
def UpdateIncludeState(filename, include_state, io=codecs):
2972
"""Fill up the include_state with new includes found from the file.
2975
filename: the name of the header to read.
2976
include_state: an _IncludeState instance in which the headers are inserted.
2977
io: The io factory to use to read the file. Provided for testability.
2980
True if a header was succesfully added. False otherwise.
2984
headerfile = io.open(filename, 'r', 'utf8', 'replace')
2988
for line in headerfile:
2990
clean_line = CleanseComments(line)
2991
match = _RE_PATTERN_INCLUDE.search(clean_line)
2993
include = match.group(2)
2994
# The value formatting is cute, but not really used right now.
2995
# What matters here is that the key is in include_state.
2996
include_state.setdefault(include, '%s:%d' % (filename, linenum))
3000
def CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error,
3002
"""Reports for missing stl includes.
3004
This function will output warnings to make sure you are including the headers
3005
necessary for the stl containers and functions that you use. We only give one
3006
reason to include a header. For example, if you use both equal_to<> and
3007
less<> in a .h file, only one (the latter in the file) of these will be
3008
reported as a reason to include the <functional>.
3011
filename: The name of the current file.
3012
clean_lines: A CleansedLines instance containing the file.
3013
include_state: An _IncludeState instance.
3014
error: The function to call with any errors found.
3015
io: The IO factory to use to read the header file. Provided for unittest
3018
required = {} # A map of header name to linenumber and the template entity.
3019
# Example of required: { '<functional>': (1219, 'less<>') }
3021
for linenum in xrange(clean_lines.NumLines()):
3022
line = clean_lines.elided[linenum]
3023
if not line or line[0] == '#':
3026
# String is special -- it is a non-templatized type in STL.
3027
matched = _RE_PATTERN_STRING.search(line)
3029
# Don't warn about strings in non-STL namespaces:
3030
# (We check only the first match per line; good enough.)
3031
prefix = line[:matched.start()]
3032
if prefix.endswith('std::') or not prefix.endswith('::'):
3033
required['<string>'] = (linenum, 'string')
3035
for pattern, template, header in _re_pattern_algorithm_header:
3036
if pattern.search(line):
3037
required[header] = (linenum, template)
3039
# The following function is just a speed up, no semantics are changed.
3040
if not '<' in line: # Reduces the cpu time usage by skipping lines.
3043
for pattern, template, header in _re_pattern_templates:
3044
if pattern.search(line):
3045
required[header] = (linenum, template)
3047
# The policy is that if you #include something in foo.h you don't need to
3048
# include it again in foo.cc. Here, we will look at possible includes.
3049
# Let's copy the include_state so it is only messed up within this function.
3050
include_state = include_state.copy()
3052
# Did we find the header for this file (if any) and succesfully load it?
3053
header_found = False
3055
# Use the absolute path so that matching works properly.
3056
abs_filename = os.path.abspath(filename)
3058
# For Emacs's flymake.
3059
# If cpplint is invoked from Emacs's flymake, a temporary file is generated
3060
# by flymake and that file name might end with '_flymake.cc'. In that case,
3061
# restore original file name here so that the corresponding header file can be
3063
# e.g. If the file name is 'foo_flymake.cc', we should search for 'foo.h'
3064
# instead of 'foo_flymake.h'
3065
abs_filename = re.sub(r'_flymake\.cc$', '.cc', abs_filename)
3067
# include_state is modified during iteration, so we iterate over a copy of
3069
header_keys = include_state.keys()
3070
for header in header_keys:
3071
(same_module, common_path) = FilesBelongToSameModule(abs_filename, header)
3072
fullpath = common_path + header
3073
if same_module and UpdateIncludeState(fullpath, include_state, io):
3076
# If we can't find the header file for a .cc, assume it's because we don't
3077
# know where to look. In that case we'll give up as we're not sure they
3078
# didn't include it in the .h file.
3079
# TODO(unknown): Do a better job of finding .h files so we are confident that
3080
# not having the .h file means there isn't one.
3081
if filename.endswith('.C') and not header_found:
3084
# All the lines have been processed, report the errors found.
3085
for required_header_unstripped in required:
3086
template = required[required_header_unstripped][1]
3087
if required_header_unstripped.strip('<>"') not in include_state:
3088
error(filename, required[required_header_unstripped][0],
3089
'build/include_what_you_use', 4,
3090
'Add #include ' + required_header_unstripped + ' for ' + template)
3093
_RE_PATTERN_EXPLICIT_MAKEPAIR = re.compile(r'\bmake_pair\s*<')
3096
def CheckMakePairUsesDeduction(filename, clean_lines, linenum, error):
3097
"""Check that make_pair's template arguments are deduced.
3099
G++ 4.6 in C++0x mode fails badly if make_pair's template arguments are
3100
specified explicitly, and such use isn't intended in any case.
3103
filename: The name of the current file.
3104
clean_lines: A CleansedLines instance containing the file.
3105
linenum: The number of the line to check.
3106
error: The function to call with any errors found.
3108
raw = clean_lines.raw_lines
3110
match = _RE_PATTERN_EXPLICIT_MAKEPAIR.search(line)
3112
error(filename, linenum, 'build/explicit_make_pair',
3113
4, # 4 = high confidence
3114
'Omit template arguments from make_pair OR use pair directly OR'
3115
' if appropriate, construct a pair directly')
3118
def ProcessLine(filename, file_extension,
3119
clean_lines, line, include_state, function_state,
3120
class_state, error, extra_check_functions=[]):
3121
"""Processes a single line in the file.
3124
filename: Filename of the file that is being processed.
3125
file_extension: The extension (dot not included) of the file.
3126
clean_lines: An array of strings, each representing a line of the file,
3127
with comments stripped.
3128
line: Number of line being processed.
3129
include_state: An _IncludeState instance in which the headers are inserted.
3130
function_state: A _FunctionState instance which counts function lines, etc.
3131
class_state: A _ClassState instance which maintains information about
3132
the current stack of nested class declarations being parsed.
3133
error: A callable to which errors are reported, which takes 4 arguments:
3134
filename, line number, error level, and message
3135
extra_check_functions: An array of additional check functions that will be
3136
run on each source line. Each function takes 4
3137
arguments: filename, clean_lines, line, error
3139
raw_lines = clean_lines.raw_lines
3140
ParseNolintSuppressions(filename, raw_lines[line], line, error)
3141
CheckForFunctionLengths(filename, clean_lines, line, function_state, error)
3142
CheckForMultilineCommentsAndStrings(filename, clean_lines, line, error)
3143
CheckStyle(filename, clean_lines, line, file_extension, class_state, error)
3144
CheckLanguage(filename, clean_lines, line, file_extension, include_state,
3146
CheckForNonStandardConstructs(filename, clean_lines, line,
3148
CheckPosixThreading(filename, clean_lines, line, error)
3149
CheckInvalidIncrement(filename, clean_lines, line, error)
3150
CheckMakePairUsesDeduction(filename, clean_lines, line, error)
3151
for check_fn in extra_check_functions:
3152
check_fn(filename, clean_lines, line, error)
3154
def ProcessFileData(filename, file_extension, lines, error,
3155
extra_check_functions=[]):
3156
"""Performs lint checks and reports any errors to the given error function.
3159
filename: Filename of the file that is being processed.
3160
file_extension: The extension (dot not included) of the file.
3161
lines: An array of strings, each representing a line of the file, with the
3162
last element being empty if the file is terminated with a newline.
3163
error: A callable to which errors are reported, which takes 4 arguments:
3164
filename, line number, error level, and message
3165
extra_check_functions: An array of additional check functions that will be
3166
run on each source line. Each function takes 4
3167
arguments: filename, clean_lines, line, error
3169
lines = (['// marker so line numbers and indices both start at 1'] + lines +
3170
['// marker so line numbers end in a known way'])
3172
include_state = _IncludeState()
3173
function_state = _FunctionState()
3174
class_state = _ClassState()
3176
ResetNolintSuppressions()
3178
CheckForCopyright(filename, lines, error)
3180
if file_extension == 'h':
3181
CheckForHeaderGuard(filename, lines, error)
3183
RemoveMultiLineComments(filename, lines, error)
3184
clean_lines = CleansedLines(lines)
3185
for line in xrange(clean_lines.NumLines()):
3186
ProcessLine(filename, file_extension, clean_lines, line,
3187
include_state, function_state, class_state, error,
3188
extra_check_functions)
3189
class_state.CheckFinished(filename, error)
3191
CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error)
3193
# We check here rather than inside ProcessLine so that we see raw
3194
# lines rather than "cleaned" lines.
3195
CheckForUnicodeReplacementCharacters(filename, lines, error)
3197
CheckForNewlineAtEOF(filename, lines, error)
3199
def ProcessFile(filename, vlevel, extra_check_functions=[]):
3200
"""Does google-lint on a single file.
3203
filename: The name of the file to parse.
3205
vlevel: The level of errors to report. Every error of confidence
3206
>= verbose_level will be reported. 0 is a good default.
3208
extra_check_functions: An array of additional check functions that will be
3209
run on each source line. Each function takes 4
3210
arguments: filename, clean_lines, line, error
3213
_SetVerboseLevel(vlevel)
3216
# Support the UNIX convention of using "-" for stdin. Note that
3217
# we are not opening the file with universal newline support
3218
# (which codecs doesn't support anyway), so the resulting lines do
3219
# contain trailing '\r' characters if we are reading a file that
3221
# If after the split a trailing '\r' is present, it is removed
3222
# below. If it is not expected to be present (i.e. os.linesep !=
3223
# '\r\n' as in Windows), a warning is issued below if this file
3227
lines = codecs.StreamReaderWriter(sys.stdin,
3228
codecs.getreader('utf8'),
3229
codecs.getwriter('utf8'),
3230
'replace').read().split('\n')
3232
lines = codecs.open(filename, 'r', 'utf8', 'replace').read().split('\n')
3234
carriage_return_found = False
3235
# Remove trailing '\r'.
3236
for linenum in range(len(lines)):
3237
if lines[linenum].endswith('\r'):
3238
lines[linenum] = lines[linenum].rstrip('\r')
3239
carriage_return_found = True
3243
"Skipping input '%s': Can't open for reading\n" % filename)
3246
# Note, if no dot is found, this will give the entire filename as the ext.
3247
file_extension = filename[filename.rfind('.') + 1:]
3249
# When reading from stdin, the extension is unknown, so no cpplint tests
3250
# should rely on the extension.
3251
if (filename != '-' and file_extension != 'C' and file_extension != 'h'
3252
and file_extension != 'cpp'):
3253
sys.stderr.write('Ignoring %s; not a .C or .h file\n' % filename)
3255
ProcessFileData(filename, file_extension, lines, Error,
3256
extra_check_functions)
3257
if carriage_return_found and os.linesep != '\r\n':
3258
# Use 0 for linenum since outputting only one error for potentially
3260
Error(filename, 0, 'whitespace/newline', 1,
3261
'One or more unexpected \\r (^M) found;'
3262
'better to use only a \\n')
3264
sys.stderr.write('Done processing %s\n' % filename)
3267
def PrintUsage(message):
3268
"""Prints a brief usage string and exits, optionally with an error message.
3271
message: The optional error message.
3273
sys.stderr.write(_USAGE)
3275
sys.exit('\nFATAL ERROR: ' + message)
3280
def PrintCategories():
3281
"""Prints a list of all the error-categories used by error messages.
3283
These are the categories used to filter messages via --filter.
3285
sys.stderr.write(''.join(' %s\n' % cat for cat in _ERROR_CATEGORIES))
3289
def ParseArguments(args):
3290
"""Parses the command line arguments.
3292
This may set the output format and verbosity level as side-effects.
3295
args: The command line arguments:
3298
The list of filenames to lint.
3301
(opts, filenames) = getopt.getopt(args, '', ['help', 'output=', 'verbose=',
3304
except getopt.GetoptError:
3305
PrintUsage('Invalid arguments.')
3307
verbosity = _VerboseLevel()
3308
output_format = _OutputFormat()
3312
for (opt, val) in opts:
3315
elif opt == '--output':
3316
if not val in ('emacs', 'vs7'):
3317
PrintUsage('The only allowed output formats are emacs and vs7.')
3319
elif opt == '--verbose':
3320
verbosity = int(val)
3321
elif opt == '--filter':
3325
elif opt == '--counting':
3326
if val not in ('total', 'toplevel', 'detailed'):
3327
PrintUsage('Valid counting options are total, toplevel, and detailed')
3328
counting_style = val
3331
PrintUsage('No files were specified.')
3333
_SetOutputFormat(output_format)
3334
_SetVerboseLevel(verbosity)
3335
_SetFilters(filters)
3336
_SetCountingStyle(counting_style)
3342
filenames = ParseArguments(sys.argv[1:])
3344
# Change stderr to write with replacement characters so we don't die
3345
# if we try to print something containing non-ASCII characters.
3346
sys.stderr = codecs.StreamReaderWriter(sys.stderr,
3347
codecs.getreader('utf8'),
3348
codecs.getwriter('utf8'),
3351
_cpplint_state.ResetErrorCounts()
3352
for filename in filenames:
3353
ProcessFile(filename, _cpplint_state.verbose_level)
3354
_cpplint_state.PrintErrorCounts()
3356
sys.exit(_cpplint_state.error_count > 0)
3359
if __name__ == '__main__':