blob: 52cb7d0cabed3c3122c320d879d2d9d01fab986b [file] [log] [blame]
avakulenko@google.com554223d2014-12-04 22:00:20 +00001#!/usr/bin/env python
erg@google.com4e00b9a2009-01-12 23:05:11 +00002#
erg@google.com8f91ab22011-09-06 21:04:45 +00003# Copyright (c) 2009 Google Inc. All rights reserved.
erg@google.com4e00b9a2009-01-12 23:05:11 +00004#
erg@google.com969161c2009-06-26 22:06:46 +00005# Redistribution and use in source and binary forms, with or without
6# modification, are permitted provided that the following conditions are
7# met:
erg@google.com4e00b9a2009-01-12 23:05:11 +00008#
erg@google.com969161c2009-06-26 22:06:46 +00009# * Redistributions of source code must retain the above copyright
10# notice, this list of conditions and the following disclaimer.
11# * Redistributions in binary form must reproduce the above
12# copyright notice, this list of conditions and the following disclaimer
13# in the documentation and/or other materials provided with the
14# distribution.
15# * Neither the name of Google Inc. nor the names of its
16# contributors may be used to endorse or promote products derived from
17# this software without specific prior written permission.
erg@google.com4e00b9a2009-01-12 23:05:11 +000018#
erg@google.com969161c2009-06-26 22:06:46 +000019# 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.
erg@google.com4e00b9a2009-01-12 23:05:11 +000030
erg@google.com4e00b9a2009-01-12 23:05:11 +000031"""Does google-lint on c++ files.
32
33The goal of this script is to identify places in the code that *may*
34be in non-compliance with google style. It does not attempt to fix
35up these problems -- the point is to educate. It does also not
36attempt to find all problems, or to ensure that everything it does
37find is legitimately a problem.
38
39In particular, we can get very confused by /* and // inside strings!
40We do a small hack, which is to ignore //'s with "'s after them on the
41same line, but it is far from perfect (in either direction).
42"""
43
44import codecs
erg@google.comd350fe52013-01-14 17:51:48 +000045import copy
erg@google.com4e00b9a2009-01-12 23:05:11 +000046import getopt
47import math # for log
48import os
49import re
50import sre_compile
51import string
52import sys
53import unicodedata
54
55
56_USAGE = """
57Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...]
erg@google.comab53edf2013-11-05 22:23:37 +000058 [--counting=total|toplevel|detailed] [--root=subdir]
LukeCz7197a242016-09-24 13:27:35 -050059 [--linelength=digits] [--headers=x,y,...]
erg@google.com4e00b9a2009-01-12 23:05:11 +000060 <file> [file] ...
61
62 The style guidelines this tries to follow are those in
Ackermann Yuriy79692902016-04-01 21:41:34 +130063 https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml
erg@google.com4e00b9a2009-01-12 23:05:11 +000064
65 Every problem is given a confidence score from 1-5, with 5 meaning we are
66 certain of the problem, and 1 meaning it could be a legitimate construct.
67 This will miss some errors, and is not a substitute for a code review.
68
erg+personal@google.com05189642010-04-30 20:43:03 +000069 To suppress false-positive errors of a certain category, add a
70 'NOLINT(category)' comment to the line. NOLINT or NOLINT(*)
71 suppresses errors of all categories on that line.
erg@google.com4e00b9a2009-01-12 23:05:11 +000072
73 The files passed in will be linted; at least one file must be provided.
erg@google.com19680272013-12-16 22:48:54 +000074 Default linted extensions are .cc, .cpp, .cu, .cuh and .h. Change the
75 extensions with the --extensions flag.
erg@google.com4e00b9a2009-01-12 23:05:11 +000076
77 Flags:
78
79 output=vs7
80 By default, the output is formatted to ease emacs parsing. Visual Studio
81 compatible output (vs7) may also be used. Other formats are unsupported.
82
83 verbose=#
84 Specify a number 0-5 to restrict errors to certain verbosity levels.
85
86 filter=-x,+y,...
87 Specify a comma-separated list of category-filters to apply: only
88 error messages whose category names pass the filters will be printed.
89 (Category names are printed with the message and look like
90 "[whitespace/indent]".) Filters are evaluated left to right.
91 "-FOO" and "FOO" means "do not print categories that start with FOO".
92 "+FOO" means "do print categories that start with FOO".
93
94 Examples: --filter=-whitespace,+whitespace/braces
95 --filter=whitespace,runtime/printf,+runtime/printf_format
96 --filter=-,+build/include_what_you_use
97
98 To see a list of all the categories used in cpplint, pass no arg:
99 --filter=
erg@google.coma868d2d2009-10-09 21:18:45 +0000100
101 counting=total|toplevel|detailed
102 The total number of errors found is always printed. If
103 'toplevel' is provided, then the count of errors in each of
104 the top-level categories like 'build' and 'whitespace' will
105 also be printed. If 'detailed' is provided, then a count
106 is provided for each category like 'build/class'.
erg@google.com4d70a882013-04-16 21:06:32 +0000107
108 root=subdir
109 The root directory used for deriving header guard CPP variable.
110 By default, the header guard CPP variable is calculated as the relative
111 path to the directory that contains .git, .hg, or .svn. When this flag
112 is specified, the relative path is calculated from the specified
113 directory. If the specified directory does not exist, this flag is
114 ignored.
115
116 Examples:
avakulenko@google.com02af6282014-06-04 18:53:25 +0000117 Assuming that src/.git exists, the header guard CPP variables for
erg@google.com4d70a882013-04-16 21:06:32 +0000118 src/chrome/browser/ui/browser.h are:
119
120 No flag => CHROME_BROWSER_UI_BROWSER_H_
121 --root=chrome => BROWSER_UI_BROWSER_H_
122 --root=chrome/browser => UI_BROWSER_H_
erg@google.comab53edf2013-11-05 22:23:37 +0000123
124 linelength=digits
125 This is the allowed line length for the project. The default value is
126 80 characters.
127
128 Examples:
129 --linelength=120
erg@google.com19680272013-12-16 22:48:54 +0000130
131 extensions=extension,extension,...
132 The allowed file extensions that cpplint will check
133
134 Examples:
135 --extensions=hpp,cpp
erg@google.com7430eef2014-07-28 22:33:46 +0000136
LukeCz7197a242016-09-24 13:27:35 -0500137 headers=x,y,...
138 The header extensions that cpplint will treat as .h in checks. Values are
139 automatically added to --extensions list.
140
141 Examples:
142 --headers=hpp,hxx
143 --headers=hpp
144
erg@google.com7430eef2014-07-28 22:33:46 +0000145 cpplint.py supports per-directory configurations specified in CPPLINT.cfg
146 files. CPPLINT.cfg file can contain a number of key=value pairs.
147 Currently the following options are supported:
148
149 set noparent
150 filter=+filter1,-filter2,...
151 exclude_files=regex
avakulenko@google.com310681b2014-08-22 19:38:55 +0000152 linelength=80
Fabian Guera2322e4f2016-05-01 17:36:30 +0200153 root=subdir
LukeCz7197a242016-09-24 13:27:35 -0500154 headers=x,y,...
erg@google.com7430eef2014-07-28 22:33:46 +0000155
156 "set noparent" option prevents cpplint from traversing directory tree
157 upwards looking for more .cfg files in parent directories. This option
158 is usually placed in the top-level project directory.
159
160 The "filter" option is similar in function to --filter flag. It specifies
161 message filters in addition to the |_DEFAULT_FILTERS| and those specified
162 through --filter command-line flag.
163
164 "exclude_files" allows to specify a regular expression to be matched against
165 a file name. If the expression matches, the file is skipped and not run
166 through liner.
167
avakulenko@google.com310681b2014-08-22 19:38:55 +0000168 "linelength" allows to specify the allowed line length for the project.
169
Fabian Guera2322e4f2016-05-01 17:36:30 +0200170 The "root" option is similar in function to the --root flag (see example
171 above).
LukeCz7197a242016-09-24 13:27:35 -0500172
173 The "headers" option is similar in function to the --headers flag
174 (see example above).
Fabian Guera2322e4f2016-05-01 17:36:30 +0200175
erg@google.com7430eef2014-07-28 22:33:46 +0000176 CPPLINT.cfg has an effect on files in the same directory and all
177 sub-directories, unless overridden by a nested configuration file.
178
179 Example file:
180 filter=-build/include_order,+build/include_alpha
181 exclude_files=.*\.cc
182
183 The above example disables build/include_order warning and enables
184 build/include_alpha as well as excludes all .cc from being
185 processed by linter, in the current directory (where the .cfg
186 file is located) and all sub-directories.
erg@google.com4e00b9a2009-01-12 23:05:11 +0000187"""
188
189# We categorize each error message we print. Here are the categories.
190# We want an explicit list so we can list them all in cpplint --filter=.
191# If you add a new error message with a new category, add it to the list
192# here! cpplint_unittest.py should tell you if you forget to do this.
erg+personal@google.com05189642010-04-30 20:43:03 +0000193_ERROR_CATEGORIES = [
avakulenko@google.com554223d2014-12-04 22:00:20 +0000194 'build/class',
195 'build/c++11',
Alex Vakulenko01e47232016-05-06 13:54:15 -0700196 'build/c++14',
197 'build/c++tr1',
avakulenko@google.com554223d2014-12-04 22:00:20 +0000198 'build/deprecated',
199 'build/endif_comment',
200 'build/explicit_make_pair',
201 'build/forward_decl',
202 'build/header_guard',
203 'build/include',
204 'build/include_alpha',
205 'build/include_order',
206 'build/include_what_you_use',
207 'build/namespaces',
208 'build/printf_format',
209 'build/storage_class',
210 'legal/copyright',
211 'readability/alt_tokens',
212 'readability/braces',
213 'readability/casting',
214 'readability/check',
215 'readability/constructors',
216 'readability/fn_size',
avakulenko@google.com554223d2014-12-04 22:00:20 +0000217 'readability/inheritance',
218 'readability/multiline_comment',
219 'readability/multiline_string',
220 'readability/namespace',
221 'readability/nolint',
222 'readability/nul',
223 'readability/strings',
224 'readability/todo',
225 'readability/utf8',
226 'runtime/arrays',
227 'runtime/casting',
228 'runtime/explicit',
229 'runtime/int',
230 'runtime/init',
231 'runtime/invalid_increment',
232 'runtime/member_string_references',
233 'runtime/memset',
234 'runtime/indentation_namespace',
235 'runtime/operator',
236 'runtime/printf',
237 'runtime/printf_format',
238 'runtime/references',
239 'runtime/string',
240 'runtime/threadsafe_fn',
241 'runtime/vlog',
242 'whitespace/blank_line',
243 'whitespace/braces',
244 'whitespace/comma',
245 'whitespace/comments',
246 'whitespace/empty_conditional_body',
Alex Vakulenko01e47232016-05-06 13:54:15 -0700247 'whitespace/empty_if_body',
avakulenko@google.com554223d2014-12-04 22:00:20 +0000248 'whitespace/empty_loop_body',
249 'whitespace/end_of_line',
250 'whitespace/ending_newline',
251 'whitespace/forcolon',
252 'whitespace/indent',
253 'whitespace/line_length',
254 'whitespace/newline',
255 'whitespace/operators',
256 'whitespace/parens',
257 'whitespace/semicolon',
258 'whitespace/tab',
259 'whitespace/todo',
260 ]
261
262# These error categories are no longer enforced by cpplint, but for backwards-
263# compatibility they may still appear in NOLINT comments.
264_LEGACY_ERROR_CATEGORIES = [
265 'readability/streams',
Alex Vakulenko01e47232016-05-06 13:54:15 -0700266 'readability/function',
avakulenko@google.com554223d2014-12-04 22:00:20 +0000267 ]
erg@google.com4e00b9a2009-01-12 23:05:11 +0000268
avakulenko@google.com02af6282014-06-04 18:53:25 +0000269# The default state of the category filter. This is overridden by the --filter=
erg@google.come35f7652009-06-19 20:52:09 +0000270# flag. By default all errors are on, so only add here categories that should be
271# off by default (i.e., categories that must be enabled by the --filter= flags).
272# All entries here should start with a '-' or '+', as in the --filter= flag.
erg@google.com8a95ecc2011-09-08 00:45:54 +0000273_DEFAULT_FILTERS = ['-build/include_alpha']
erg@google.come35f7652009-06-19 20:52:09 +0000274
Alex Vakulenko01e47232016-05-06 13:54:15 -0700275# The default list of categories suppressed for C (not C++) files.
276_DEFAULT_C_SUPPRESSED_CATEGORIES = [
277 'readability/casting',
278 ]
279
280# The default list of categories suppressed for Linux Kernel files.
281_DEFAULT_KERNEL_SUPPRESSED_CATEGORIES = [
282 'whitespace/tab',
283 ]
284
erg@google.com4e00b9a2009-01-12 23:05:11 +0000285# We used to check for high-bit characters, but after much discussion we
286# decided those were OK, as long as they were in UTF-8 and didn't represent
erg@google.com8a95ecc2011-09-08 00:45:54 +0000287# hard-coded international strings, which belong in a separate i18n file.
erg@google.com4e00b9a2009-01-12 23:05:11 +0000288
erg@google.comfd5da632013-10-25 17:39:45 +0000289# C++ headers
erg@google.com4e00b9a2009-01-12 23:05:11 +0000290_CPP_HEADERS = frozenset([
erg@google.comfd5da632013-10-25 17:39:45 +0000291 # Legacy
292 'algobase.h',
293 'algo.h',
294 'alloc.h',
295 'builtinbuf.h',
296 'bvector.h',
297 'complex.h',
298 'defalloc.h',
299 'deque.h',
300 'editbuf.h',
301 'fstream.h',
302 'function.h',
303 'hash_map',
304 'hash_map.h',
305 'hash_set',
306 'hash_set.h',
307 'hashtable.h',
308 'heap.h',
309 'indstream.h',
310 'iomanip.h',
311 'iostream.h',
312 'istream.h',
313 'iterator.h',
314 'list.h',
315 'map.h',
316 'multimap.h',
317 'multiset.h',
318 'ostream.h',
319 'pair.h',
320 'parsestream.h',
321 'pfstream.h',
322 'procbuf.h',
323 'pthread_alloc',
324 'pthread_alloc.h',
325 'rope',
326 'rope.h',
327 'ropeimpl.h',
328 'set.h',
329 'slist',
330 'slist.h',
331 'stack.h',
332 'stdiostream.h',
333 'stl_alloc.h',
334 'stl_relops.h',
335 'streambuf.h',
336 'stream.h',
337 'strfile.h',
338 'strstream.h',
339 'tempbuf.h',
340 'tree.h',
341 'type_traits.h',
342 'vector.h',
343 # 17.6.1.2 C++ library headers
344 'algorithm',
345 'array',
346 'atomic',
347 'bitset',
348 'chrono',
349 'codecvt',
350 'complex',
351 'condition_variable',
352 'deque',
353 'exception',
354 'forward_list',
355 'fstream',
356 'functional',
357 'future',
358 'initializer_list',
359 'iomanip',
360 'ios',
361 'iosfwd',
362 'iostream',
363 'istream',
364 'iterator',
365 'limits',
366 'list',
367 'locale',
368 'map',
369 'memory',
370 'mutex',
371 'new',
372 'numeric',
373 'ostream',
374 'queue',
375 'random',
376 'ratio',
377 'regex',
Alex Vakulenko01e47232016-05-06 13:54:15 -0700378 'scoped_allocator',
erg@google.comfd5da632013-10-25 17:39:45 +0000379 'set',
380 'sstream',
381 'stack',
382 'stdexcept',
383 'streambuf',
384 'string',
385 'strstream',
386 'system_error',
387 'thread',
388 'tuple',
389 'typeindex',
390 'typeinfo',
391 'type_traits',
392 'unordered_map',
393 'unordered_set',
394 'utility',
erg@google.com5d00c562013-07-12 19:57:05 +0000395 'valarray',
erg@google.comfd5da632013-10-25 17:39:45 +0000396 'vector',
397 # 17.6.1.2 C++ headers for C library facilities
398 'cassert',
399 'ccomplex',
400 'cctype',
401 'cerrno',
402 'cfenv',
403 'cfloat',
404 'cinttypes',
405 'ciso646',
406 'climits',
407 'clocale',
408 'cmath',
409 'csetjmp',
410 'csignal',
411 'cstdalign',
412 'cstdarg',
413 'cstdbool',
414 'cstddef',
415 'cstdint',
416 'cstdio',
417 'cstdlib',
418 'cstring',
419 'ctgmath',
420 'ctime',
421 'cuchar',
422 'cwchar',
423 'cwctype',
erg@google.com4e00b9a2009-01-12 23:05:11 +0000424 ])
425
Alex Vakulenko01e47232016-05-06 13:54:15 -0700426# Type names
427_TYPES = re.compile(
428 r'^(?:'
429 # [dcl.type.simple]
430 r'(char(16_t|32_t)?)|wchar_t|'
431 r'bool|short|int|long|signed|unsigned|float|double|'
432 # [support.types]
433 r'(ptrdiff_t|size_t|max_align_t|nullptr_t)|'
434 # [cstdint.syn]
435 r'(u?int(_fast|_least)?(8|16|32|64)_t)|'
436 r'(u?int(max|ptr)_t)|'
437 r')$')
438
avakulenko@google.com02af6282014-06-04 18:53:25 +0000439
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000440# These headers are excluded from [build/include] and [build/include_order]
441# checks:
442# - Anything not following google file name conventions (containing an
443# uppercase character, such as Python.h or nsStringAPI.h, for example).
444# - Lua headers.
445_THIRD_PARTY_HEADERS_PATTERN = re.compile(
446 r'^(?:[^/]*[A-Z][^/]*\.h|lua\.h|lauxlib\.h|lualib\.h)$')
447
Alex Vakulenko01e47232016-05-06 13:54:15 -0700448# Pattern for matching FileInfo.BaseName() against test file name
449_TEST_FILE_SUFFIX = r'(_test|_unittest|_regtest)$'
450
451# Pattern that matches only complete whitespace, possibly across multiple lines.
452_EMPTY_CONDITIONAL_BODY_PATTERN = re.compile(r'^\s*$', re.DOTALL)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000453
erg@google.com4e00b9a2009-01-12 23:05:11 +0000454# Assertion macros. These are defined in base/logging.h and
Alex Vakulenko01e47232016-05-06 13:54:15 -0700455# testing/base/public/gunit.h.
erg@google.com4e00b9a2009-01-12 23:05:11 +0000456_CHECK_MACROS = [
erg@google.come35f7652009-06-19 20:52:09 +0000457 'DCHECK', 'CHECK',
Alex Vakulenko01e47232016-05-06 13:54:15 -0700458 'EXPECT_TRUE', 'ASSERT_TRUE',
459 'EXPECT_FALSE', 'ASSERT_FALSE',
erg@google.com4e00b9a2009-01-12 23:05:11 +0000460 ]
461
erg@google.come35f7652009-06-19 20:52:09 +0000462# Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE
erg@google.com4e00b9a2009-01-12 23:05:11 +0000463_CHECK_REPLACEMENT = dict([(m, {}) for m in _CHECK_MACROS])
464
465for op, replacement in [('==', 'EQ'), ('!=', 'NE'),
466 ('>=', 'GE'), ('>', 'GT'),
467 ('<=', 'LE'), ('<', 'LT')]:
erg@google.come35f7652009-06-19 20:52:09 +0000468 _CHECK_REPLACEMENT['DCHECK'][op] = 'DCHECK_%s' % replacement
erg@google.com4e00b9a2009-01-12 23:05:11 +0000469 _CHECK_REPLACEMENT['CHECK'][op] = 'CHECK_%s' % replacement
470 _CHECK_REPLACEMENT['EXPECT_TRUE'][op] = 'EXPECT_%s' % replacement
471 _CHECK_REPLACEMENT['ASSERT_TRUE'][op] = 'ASSERT_%s' % replacement
erg@google.com4e00b9a2009-01-12 23:05:11 +0000472
473for op, inv_replacement in [('==', 'NE'), ('!=', 'EQ'),
474 ('>=', 'LT'), ('>', 'LE'),
475 ('<=', 'GT'), ('<', 'GE')]:
476 _CHECK_REPLACEMENT['EXPECT_FALSE'][op] = 'EXPECT_%s' % inv_replacement
477 _CHECK_REPLACEMENT['ASSERT_FALSE'][op] = 'ASSERT_%s' % inv_replacement
erg@google.com4e00b9a2009-01-12 23:05:11 +0000478
erg@google.comd350fe52013-01-14 17:51:48 +0000479# Alternative tokens and their replacements. For full list, see section 2.5
480# Alternative tokens [lex.digraph] in the C++ standard.
481#
482# Digraphs (such as '%:') are not included here since it's a mess to
483# match those on a word boundary.
484_ALT_TOKEN_REPLACEMENT = {
485 'and': '&&',
486 'bitor': '|',
487 'or': '||',
488 'xor': '^',
489 'compl': '~',
490 'bitand': '&',
491 'and_eq': '&=',
492 'or_eq': '|=',
493 'xor_eq': '^=',
494 'not': '!',
495 'not_eq': '!='
496 }
497
498# Compile regular expression that matches all the above keywords. The "[ =()]"
499# bit is meant to avoid matching these keywords outside of boolean expressions.
500#
erg@google.comc6671232013-10-25 21:44:03 +0000501# False positives include C-style multi-line comments and multi-line strings
502# but those have always been troublesome for cpplint.
erg@google.comd350fe52013-01-14 17:51:48 +0000503_ALT_TOKEN_REPLACEMENT_PATTERN = re.compile(
504 r'[ =()](' + ('|'.join(_ALT_TOKEN_REPLACEMENT.keys())) + r')(?=[ (]|$)')
505
erg@google.com4e00b9a2009-01-12 23:05:11 +0000506
507# These constants define types of headers for use with
508# _IncludeState.CheckNextIncludeOrder().
509_C_SYS_HEADER = 1
510_CPP_SYS_HEADER = 2
511_LIKELY_MY_HEADER = 3
512_POSSIBLE_MY_HEADER = 4
513_OTHER_HEADER = 5
514
erg@google.comd350fe52013-01-14 17:51:48 +0000515# These constants define the current inline assembly state
516_NO_ASM = 0 # Outside of inline assembly block
517_INSIDE_ASM = 1 # Inside inline assembly block
518_END_ASM = 2 # Last line of inline assembly block
519_BLOCK_ASM = 3 # The whole block is an inline assembly block
520
521# Match start of assembly blocks
522_MATCH_ASM = re.compile(r'^\s*(?:asm|_asm|__asm|__asm__)'
523 r'(?:\s+(volatile|__volatile__))?'
524 r'\s*[{(]')
525
Alex Vakulenko01e47232016-05-06 13:54:15 -0700526# Match strings that indicate we're working on a C (not C++) file.
527_SEARCH_C_FILE = re.compile(r'\b(?:LINT_C_FILE|'
528 r'vim?:\s*.*(\s*|:)filetype=c(\s*|:|$))')
529
530# Match string that indicates we're working on a Linux Kernel file.
531_SEARCH_KERNEL_FILE = re.compile(r'\b(?:LINT_KERNEL_FILE)')
erg@google.com4e00b9a2009-01-12 23:05:11 +0000532
533_regexp_compile_cache = {}
534
erg+personal@google.com05189642010-04-30 20:43:03 +0000535# {str, set(int)}: a map from error categories to sets of linenumbers
536# on which those errors are expected and should be suppressed.
537_error_suppressions = {}
538
erg@google.com4d70a882013-04-16 21:06:32 +0000539# The root directory used for deriving header guard CPP variable.
540# This is set by --root flag.
541_root = None
542
erg@google.comab53edf2013-11-05 22:23:37 +0000543# The allowed line length of files.
544# This is set by --linelength flag.
545_line_length = 80
546
erg@google.com19680272013-12-16 22:48:54 +0000547# The allowed extensions for file names
548# This is set by --extensions flag.
549_valid_extensions = set(['cc', 'h', 'cpp', 'cu', 'cuh'])
550
LukeCz7197a242016-09-24 13:27:35 -0500551# Treat all headers starting with 'h' equally: .h, .hpp, .hxx etc.
552# This is set by --headers flag.
LukeCz8920b132016-09-26 19:40:47 -0500553_hpp_headers = set(['h'])
LukeCz7197a242016-09-24 13:27:35 -0500554
Alex Vakulenko01e47232016-05-06 13:54:15 -0700555# {str, bool}: a map from error categories to booleans which indicate if the
556# category should be suppressed for every line.
557_global_error_suppressions = {}
558
LukeCz7197a242016-09-24 13:27:35 -0500559def ProcessHppHeadersOption(val):
560 global _hpp_headers
561 try:
562 _hpp_headers = set(val.split(','))
563 # Automatically append to extensions list so it does not have to be set 2 times
564 _valid_extensions.update(_hpp_headers)
565 except ValueError:
566 PrintUsage('Header extensions must be comma seperated list.')
567
568def IsHeaderExtension(file_extension):
LukeCz8920b132016-09-26 19:40:47 -0500569 return file_extension in _hpp_headers
Alex Vakulenko01e47232016-05-06 13:54:15 -0700570
erg+personal@google.com05189642010-04-30 20:43:03 +0000571def ParseNolintSuppressions(filename, raw_line, linenum, error):
Alex Vakulenko01e47232016-05-06 13:54:15 -0700572 """Updates the global list of line error-suppressions.
erg+personal@google.com05189642010-04-30 20:43:03 +0000573
574 Parses any NOLINT comments on the current line, updating the global
575 error_suppressions store. Reports an error if the NOLINT comment
576 was malformed.
577
578 Args:
579 filename: str, the name of the input file.
580 raw_line: str, the line of input text, with comments.
581 linenum: int, the number of the current line.
582 error: function, an error handler.
583 """
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000584 matched = Search(r'\bNOLINT(NEXTLINE)?\b(\([^)]+\))?', raw_line)
erg@google.com8a95ecc2011-09-08 00:45:54 +0000585 if matched:
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000586 if matched.group(1):
587 suppressed_line = linenum + 1
588 else:
589 suppressed_line = linenum
590 category = matched.group(2)
erg+personal@google.com05189642010-04-30 20:43:03 +0000591 if category in (None, '(*)'): # => "suppress all"
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000592 _error_suppressions.setdefault(None, set()).add(suppressed_line)
erg+personal@google.com05189642010-04-30 20:43:03 +0000593 else:
594 if category.startswith('(') and category.endswith(')'):
595 category = category[1:-1]
596 if category in _ERROR_CATEGORIES:
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000597 _error_suppressions.setdefault(category, set()).add(suppressed_line)
avakulenko@google.com554223d2014-12-04 22:00:20 +0000598 elif category not in _LEGACY_ERROR_CATEGORIES:
erg+personal@google.com05189642010-04-30 20:43:03 +0000599 error(filename, linenum, 'readability/nolint', 5,
erg@google.com8a95ecc2011-09-08 00:45:54 +0000600 'Unknown NOLINT error category: %s' % category)
erg+personal@google.com05189642010-04-30 20:43:03 +0000601
602
Alex Vakulenko01e47232016-05-06 13:54:15 -0700603def ProcessGlobalSuppresions(lines):
604 """Updates the list of global error suppressions.
605
606 Parses any lint directives in the file that have global effect.
607
608 Args:
609 lines: An array of strings, each representing a line of the file, with the
610 last element being empty if the file is terminated with a newline.
611 """
612 for line in lines:
613 if _SEARCH_C_FILE.search(line):
614 for category in _DEFAULT_C_SUPPRESSED_CATEGORIES:
615 _global_error_suppressions[category] = True
616 if _SEARCH_KERNEL_FILE.search(line):
617 for category in _DEFAULT_KERNEL_SUPPRESSED_CATEGORIES:
618 _global_error_suppressions[category] = True
619
620
erg+personal@google.com05189642010-04-30 20:43:03 +0000621def ResetNolintSuppressions():
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000622 """Resets the set of NOLINT suppressions to empty."""
erg+personal@google.com05189642010-04-30 20:43:03 +0000623 _error_suppressions.clear()
Alex Vakulenko01e47232016-05-06 13:54:15 -0700624 _global_error_suppressions.clear()
erg+personal@google.com05189642010-04-30 20:43:03 +0000625
626
627def IsErrorSuppressedByNolint(category, linenum):
628 """Returns true if the specified error category is suppressed on this line.
629
630 Consults the global error_suppressions map populated by
Alex Vakulenko01e47232016-05-06 13:54:15 -0700631 ParseNolintSuppressions/ProcessGlobalSuppresions/ResetNolintSuppressions.
erg+personal@google.com05189642010-04-30 20:43:03 +0000632
633 Args:
634 category: str, the category of the error.
635 linenum: int, the current line number.
636 Returns:
Alex Vakulenko01e47232016-05-06 13:54:15 -0700637 bool, True iff the error should be suppressed due to a NOLINT comment or
638 global suppression.
erg+personal@google.com05189642010-04-30 20:43:03 +0000639 """
Alex Vakulenko01e47232016-05-06 13:54:15 -0700640 return (_global_error_suppressions.get(category, False) or
641 linenum in _error_suppressions.get(category, set()) or
erg+personal@google.com05189642010-04-30 20:43:03 +0000642 linenum in _error_suppressions.get(None, set()))
erg@google.com4e00b9a2009-01-12 23:05:11 +0000643
avakulenko@google.com02af6282014-06-04 18:53:25 +0000644
erg@google.com4e00b9a2009-01-12 23:05:11 +0000645def Match(pattern, s):
646 """Matches the string with the pattern, caching the compiled regexp."""
647 # The regexp compilation caching is inlined in both Match and Search for
648 # performance reasons; factoring it out into a separate function turns out
649 # to be noticeably expensive.
erg@google.comc6671232013-10-25 21:44:03 +0000650 if pattern not in _regexp_compile_cache:
erg@google.com4e00b9a2009-01-12 23:05:11 +0000651 _regexp_compile_cache[pattern] = sre_compile.compile(pattern)
652 return _regexp_compile_cache[pattern].match(s)
653
654
erg@google.comfd5da632013-10-25 17:39:45 +0000655def ReplaceAll(pattern, rep, s):
656 """Replaces instances of pattern in a string with a replacement.
657
658 The compiled regex is kept in a cache shared by Match and Search.
659
660 Args:
661 pattern: regex pattern
662 rep: replacement text
663 s: search string
664
665 Returns:
666 string with replacements made (or original string if no replacements)
667 """
668 if pattern not in _regexp_compile_cache:
669 _regexp_compile_cache[pattern] = sre_compile.compile(pattern)
670 return _regexp_compile_cache[pattern].sub(rep, s)
671
672
erg@google.com4e00b9a2009-01-12 23:05:11 +0000673def Search(pattern, s):
674 """Searches the string for the pattern, caching the compiled regexp."""
erg@google.comc6671232013-10-25 21:44:03 +0000675 if pattern not in _regexp_compile_cache:
erg@google.com4e00b9a2009-01-12 23:05:11 +0000676 _regexp_compile_cache[pattern] = sre_compile.compile(pattern)
677 return _regexp_compile_cache[pattern].search(s)
678
679
Alex Vakulenko01e47232016-05-06 13:54:15 -0700680def _IsSourceExtension(s):
681 """File extension (excluding dot) matches a source file extension."""
682 return s in ('c', 'cc', 'cpp', 'cxx')
683
684
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000685class _IncludeState(object):
erg@google.com4e00b9a2009-01-12 23:05:11 +0000686 """Tracks line numbers for includes, and the order in which includes appear.
687
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000688 include_list contains list of lists of (header, line number) pairs.
689 It's a lists of lists rather than just one flat list to make it
690 easier to update across preprocessor boundaries.
erg@google.com4e00b9a2009-01-12 23:05:11 +0000691
692 Call CheckNextIncludeOrder() once for each header in the file, passing
693 in the type constants defined above. Calls in an illegal order will
694 raise an _IncludeError with an appropriate error message.
695
696 """
697 # self._section will move monotonically through this set. If it ever
698 # needs to move backwards, CheckNextIncludeOrder will raise an error.
699 _INITIAL_SECTION = 0
700 _MY_H_SECTION = 1
701 _C_SECTION = 2
702 _CPP_SECTION = 3
703 _OTHER_H_SECTION = 4
704
705 _TYPE_NAMES = {
706 _C_SYS_HEADER: 'C system header',
707 _CPP_SYS_HEADER: 'C++ system header',
708 _LIKELY_MY_HEADER: 'header this file implements',
709 _POSSIBLE_MY_HEADER: 'header this file may implement',
710 _OTHER_HEADER: 'other header',
711 }
712 _SECTION_NAMES = {
713 _INITIAL_SECTION: "... nothing. (This can't be an error.)",
714 _MY_H_SECTION: 'a header this file implements',
715 _C_SECTION: 'C system header',
716 _CPP_SECTION: 'C++ system header',
717 _OTHER_H_SECTION: 'other header',
718 }
719
720 def __init__(self):
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000721 self.include_list = [[]]
722 self.ResetSection('')
erg@google.com2aa59982013-10-28 19:09:25 +0000723
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000724 def FindHeader(self, header):
725 """Check if a header has already been included.
726
727 Args:
728 header: header to check.
729 Returns:
730 Line number of previous occurrence, or -1 if the header has not
731 been seen before.
732 """
733 for section_list in self.include_list:
734 for f in section_list:
735 if f[0] == header:
736 return f[1]
737 return -1
738
739 def ResetSection(self, directive):
740 """Reset section checking for preprocessor directive.
741
742 Args:
743 directive: preprocessor directive (e.g. "if", "else").
744 """
erg@google.coma868d2d2009-10-09 21:18:45 +0000745 # The name of the current section.
erg@google.com4e00b9a2009-01-12 23:05:11 +0000746 self._section = self._INITIAL_SECTION
erg@google.coma868d2d2009-10-09 21:18:45 +0000747 # The path of last found header.
748 self._last_header = ''
749
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +0000750 # Update list of includes. Note that we never pop from the
751 # include list.
752 if directive in ('if', 'ifdef', 'ifndef'):
753 self.include_list.append([])
754 elif directive in ('else', 'elif'):
755 self.include_list[-1] = []
756
erg@google.comfd5da632013-10-25 17:39:45 +0000757 def SetLastHeader(self, header_path):
758 self._last_header = header_path
759
erg@google.coma868d2d2009-10-09 21:18:45 +0000760 def CanonicalizeAlphabeticalOrder(self, header_path):
erg@google.com8a95ecc2011-09-08 00:45:54 +0000761 """Returns a path canonicalized for alphabetical comparison.
erg@google.coma868d2d2009-10-09 21:18:45 +0000762
763 - replaces "-" with "_" so they both cmp the same.
764 - removes '-inl' since we don't require them to be after the main header.
765 - lowercase everything, just in case.
766
767 Args:
768 header_path: Path to be canonicalized.
769
770 Returns:
771 Canonicalized path.
772 """
773 return header_path.replace('-inl.h', '.h').replace('-', '_').lower()
774
erg@google.comfd5da632013-10-25 17:39:45 +0000775 def IsInAlphabeticalOrder(self, clean_lines, linenum, header_path):
erg@google.coma868d2d2009-10-09 21:18:45 +0000776 """Check if a header is in alphabetical order with the previous header.
777
778 Args:
erg@google.comfd5da632013-10-25 17:39:45 +0000779 clean_lines: A CleansedLines instance containing the file.
780 linenum: The number of the line to check.
781 header_path: Canonicalized header to be checked.
erg@google.coma868d2d2009-10-09 21:18:45 +0000782
783 Returns:
784 Returns true if the header is in alphabetical order.
785 """
erg@google.comfd5da632013-10-25 17:39:45 +0000786 # If previous section is different from current section, _last_header will
787 # be reset to empty string, so it's always less than current header.
788 #
789 # If previous line was a blank line, assume that the headers are
790 # intentionally sorted the way they are.
791 if (self._last_header > header_path and
avakulenko@google.com554223d2014-12-04 22:00:20 +0000792 Match(r'^\s*#\s*include\b', clean_lines.elided[linenum - 1])):
erg@google.coma868d2d2009-10-09 21:18:45 +0000793 return False
erg@google.coma868d2d2009-10-09 21:18:45 +0000794 return True
erg@google.com4e00b9a2009-01-12 23:05:11 +0000795
796 def CheckNextIncludeOrder(self, header_type):
797 """Returns a non-empty error message if the next header is out of order.
798
799 This function also updates the internal state to be ready to check
800 the next include.
801
802 Args:
803 header_type: One of the _XXX_HEADER constants defined above.
804
805 Returns:
806 The empty string if the header is in the right order, or an
807 error message describing what's wrong.
808
809 """
810 error_message = ('Found %s after %s' %
811 (self._TYPE_NAMES[header_type],
812 self._SECTION_NAMES[self._section]))
813
erg@google.coma868d2d2009-10-09 21:18:45 +0000814 last_section = self._section
815
erg@google.com4e00b9a2009-01-12 23:05:11 +0000816 if header_type == _C_SYS_HEADER:
817 if self._section <= self._C_SECTION:
818 self._section = self._C_SECTION
819 else:
erg@google.coma868d2d2009-10-09 21:18:45 +0000820 self._last_header = ''
erg@google.com4e00b9a2009-01-12 23:05:11 +0000821 return error_message
822 elif header_type == _CPP_SYS_HEADER:
823 if self._section <= self._CPP_SECTION:
824 self._section = self._CPP_SECTION
825 else:
erg@google.coma868d2d2009-10-09 21:18:45 +0000826 self._last_header = ''
erg@google.com4e00b9a2009-01-12 23:05:11 +0000827 return error_message
828 elif header_type == _LIKELY_MY_HEADER:
829 if self._section <= self._MY_H_SECTION:
830 self._section = self._MY_H_SECTION
831 else:
832 self._section = self._OTHER_H_SECTION
833 elif header_type == _POSSIBLE_MY_HEADER:
834 if self._section <= self._MY_H_SECTION:
835 self._section = self._MY_H_SECTION
836 else:
837 # This will always be the fallback because we're not sure
838 # enough that the header is associated with this file.
839 self._section = self._OTHER_H_SECTION
840 else:
841 assert header_type == _OTHER_HEADER
842 self._section = self._OTHER_H_SECTION
843
erg@google.coma868d2d2009-10-09 21:18:45 +0000844 if last_section != self._section:
845 self._last_header = ''
846
erg@google.com4e00b9a2009-01-12 23:05:11 +0000847 return ''
848
849
850class _CppLintState(object):
851 """Maintains module-wide state.."""
852
853 def __init__(self):
854 self.verbose_level = 1 # global setting.
855 self.error_count = 0 # global count of reported errors
erg@google.come35f7652009-06-19 20:52:09 +0000856 # filters to apply when emitting error messages
857 self.filters = _DEFAULT_FILTERS[:]
erg@google.com7430eef2014-07-28 22:33:46 +0000858 # backup of filter list. Used to restore the state after each file.
859 self._filters_backup = self.filters[:]
erg@google.coma868d2d2009-10-09 21:18:45 +0000860 self.counting = 'total' # In what way are we counting errors?
861 self.errors_by_category = {} # string to int dict storing error counts
erg@google.com4e00b9a2009-01-12 23:05:11 +0000862
863 # output format:
864 # "emacs" - format that emacs can parse (default)
865 # "vs7" - format that Microsoft Visual Studio 7 can parse
866 self.output_format = 'emacs'
867
868 def SetOutputFormat(self, output_format):
869 """Sets the output format for errors."""
870 self.output_format = output_format
871
872 def SetVerboseLevel(self, level):
873 """Sets the module's verbosity, and returns the previous setting."""
874 last_verbose_level = self.verbose_level
875 self.verbose_level = level
876 return last_verbose_level
877
erg@google.coma868d2d2009-10-09 21:18:45 +0000878 def SetCountingStyle(self, counting_style):
879 """Sets the module's counting options."""
880 self.counting = counting_style
881
erg@google.com4e00b9a2009-01-12 23:05:11 +0000882 def SetFilters(self, filters):
883 """Sets the error-message filters.
884
885 These filters are applied when deciding whether to emit a given
886 error message.
887
888 Args:
889 filters: A string of comma-separated filters (eg "+whitespace/indent").
890 Each filter should start with + or -; else we die.
erg@google.coma87abb82009-02-24 01:41:01 +0000891
892 Raises:
893 ValueError: The comma-separated filters did not all start with '+' or '-'.
894 E.g. "-,+whitespace,-whitespace/indent,whitespace/badfilter"
erg@google.com4e00b9a2009-01-12 23:05:11 +0000895 """
erg@google.come35f7652009-06-19 20:52:09 +0000896 # Default filters always have less priority than the flag ones.
897 self.filters = _DEFAULT_FILTERS[:]
erg@google.com7430eef2014-07-28 22:33:46 +0000898 self.AddFilters(filters)
899
900 def AddFilters(self, filters):
901 """ Adds more filters to the existing list of error-message filters. """
erg@google.come35f7652009-06-19 20:52:09 +0000902 for filt in filters.split(','):
903 clean_filt = filt.strip()
904 if clean_filt:
905 self.filters.append(clean_filt)
erg@google.com4e00b9a2009-01-12 23:05:11 +0000906 for filt in self.filters:
907 if not (filt.startswith('+') or filt.startswith('-')):
908 raise ValueError('Every filter in --filters must start with + or -'
909 ' (%s does not)' % filt)
910
erg@google.com7430eef2014-07-28 22:33:46 +0000911 def BackupFilters(self):
912 """ Saves the current filter list to backup storage."""
913 self._filters_backup = self.filters[:]
914
915 def RestoreFilters(self):
916 """ Restores filters previously backed up."""
917 self.filters = self._filters_backup[:]
918
erg@google.coma868d2d2009-10-09 21:18:45 +0000919 def ResetErrorCounts(self):
erg@google.com4e00b9a2009-01-12 23:05:11 +0000920 """Sets the module's error statistic back to zero."""
921 self.error_count = 0
erg@google.coma868d2d2009-10-09 21:18:45 +0000922 self.errors_by_category = {}
erg@google.com4e00b9a2009-01-12 23:05:11 +0000923
erg@google.coma868d2d2009-10-09 21:18:45 +0000924 def IncrementErrorCount(self, category):
erg@google.com4e00b9a2009-01-12 23:05:11 +0000925 """Bumps the module's error statistic."""
926 self.error_count += 1
erg@google.coma868d2d2009-10-09 21:18:45 +0000927 if self.counting in ('toplevel', 'detailed'):
928 if self.counting != 'detailed':
929 category = category.split('/')[0]
930 if category not in self.errors_by_category:
931 self.errors_by_category[category] = 0
932 self.errors_by_category[category] += 1
erg@google.com4e00b9a2009-01-12 23:05:11 +0000933
erg@google.coma868d2d2009-10-09 21:18:45 +0000934 def PrintErrorCounts(self):
935 """Print a summary of errors by category, and the total."""
936 for category, count in self.errors_by_category.iteritems():
937 sys.stderr.write('Category \'%s\' errors found: %d\n' %
938 (category, count))
LukeCze09f4782016-09-28 19:13:37 -0500939 sys.stdout.write('Total errors found: %d\n' % self.error_count)
erg@google.com4e00b9a2009-01-12 23:05:11 +0000940
941_cpplint_state = _CppLintState()
942
943
944def _OutputFormat():
945 """Gets the module's output format."""
946 return _cpplint_state.output_format
947
948
949def _SetOutputFormat(output_format):
950 """Sets the module's output format."""
951 _cpplint_state.SetOutputFormat(output_format)
952
953
954def _VerboseLevel():
955 """Returns the module's verbosity setting."""
956 return _cpplint_state.verbose_level
957
958
959def _SetVerboseLevel(level):
960 """Sets the module's verbosity, and returns the previous setting."""
961 return _cpplint_state.SetVerboseLevel(level)
962
963
erg@google.coma868d2d2009-10-09 21:18:45 +0000964def _SetCountingStyle(level):
965 """Sets the module's counting options."""
966 _cpplint_state.SetCountingStyle(level)
967
968
erg@google.com4e00b9a2009-01-12 23:05:11 +0000969def _Filters():
970 """Returns the module's list of output filters, as a list."""
971 return _cpplint_state.filters
972
973
974def _SetFilters(filters):
975 """Sets the module's error-message filters.
976
977 These filters are applied when deciding whether to emit a given
978 error message.
979
980 Args:
981 filters: A string of comma-separated filters (eg "whitespace/indent").
982 Each filter should start with + or -; else we die.
983 """
984 _cpplint_state.SetFilters(filters)
985
erg@google.com7430eef2014-07-28 22:33:46 +0000986def _AddFilters(filters):
987 """Adds more filter overrides.
988
989 Unlike _SetFilters, this function does not reset the current list of filters
990 available.
991
992 Args:
993 filters: A string of comma-separated filters (eg "whitespace/indent").
994 Each filter should start with + or -; else we die.
995 """
996 _cpplint_state.AddFilters(filters)
997
998def _BackupFilters():
999 """ Saves the current filter list to backup storage."""
1000 _cpplint_state.BackupFilters()
1001
1002def _RestoreFilters():
1003 """ Restores filters previously backed up."""
1004 _cpplint_state.RestoreFilters()
erg@google.com4e00b9a2009-01-12 23:05:11 +00001005
1006class _FunctionState(object):
1007 """Tracks current function name and the number of lines in its body."""
1008
1009 _NORMAL_TRIGGER = 250 # for --v=0, 500 for --v=1, etc.
1010 _TEST_TRIGGER = 400 # about 50% more than _NORMAL_TRIGGER.
1011
1012 def __init__(self):
1013 self.in_a_function = False
1014 self.lines_in_function = 0
1015 self.current_function = ''
1016
1017 def Begin(self, function_name):
1018 """Start analyzing function body.
1019
1020 Args:
1021 function_name: The name of the function being tracked.
1022 """
1023 self.in_a_function = True
1024 self.lines_in_function = 0
1025 self.current_function = function_name
1026
1027 def Count(self):
1028 """Count line in current function body."""
1029 if self.in_a_function:
1030 self.lines_in_function += 1
1031
1032 def Check(self, error, filename, linenum):
1033 """Report if too many lines in function body.
1034
1035 Args:
1036 error: The function to call with any errors found.
1037 filename: The name of the current file.
1038 linenum: The number of the line to check.
1039 """
Alex Vakulenko01e47232016-05-06 13:54:15 -07001040 if not self.in_a_function:
1041 return
1042
erg@google.com4e00b9a2009-01-12 23:05:11 +00001043 if Match(r'T(EST|est)', self.current_function):
1044 base_trigger = self._TEST_TRIGGER
1045 else:
1046 base_trigger = self._NORMAL_TRIGGER
1047 trigger = base_trigger * 2**_VerboseLevel()
1048
1049 if self.lines_in_function > trigger:
1050 error_level = int(math.log(self.lines_in_function / base_trigger, 2))
1051 # 50 => 0, 100 => 1, 200 => 2, 400 => 3, 800 => 4, 1600 => 5, ...
1052 if error_level > 5:
1053 error_level = 5
1054 error(filename, linenum, 'readability/fn_size', error_level,
1055 'Small and focused functions are preferred:'
1056 ' %s has %d non-comment lines'
1057 ' (error triggered by exceeding %d lines).' % (
1058 self.current_function, self.lines_in_function, trigger))
1059
1060 def End(self):
erg@google.com8a95ecc2011-09-08 00:45:54 +00001061 """Stop analyzing function body."""
erg@google.com4e00b9a2009-01-12 23:05:11 +00001062 self.in_a_function = False
1063
1064
1065class _IncludeError(Exception):
1066 """Indicates a problem with the include order in a file."""
1067 pass
1068
1069
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00001070class FileInfo(object):
erg@google.com4e00b9a2009-01-12 23:05:11 +00001071 """Provides utility functions for filenames.
1072
1073 FileInfo provides easy access to the components of a file's path
1074 relative to the project root.
1075 """
1076
1077 def __init__(self, filename):
1078 self._filename = filename
1079
1080 def FullName(self):
1081 """Make Windows paths like Unix."""
1082 return os.path.abspath(self._filename).replace('\\', '/')
1083
1084 def RepositoryName(self):
1085 """FullName after removing the local path to the repository.
1086
1087 If we have a real absolute path name here we can try to do something smart:
1088 detecting the root of the checkout and truncating /path/to/checkout from
1089 the name so that we get header guards that don't include things like
1090 "C:\Documents and Settings\..." or "/home/username/..." in them and thus
1091 people on different computers who have checked the source out to different
1092 locations won't see bogus errors.
1093 """
1094 fullname = self.FullName()
1095
1096 if os.path.exists(fullname):
1097 project_dir = os.path.dirname(fullname)
1098
1099 if os.path.exists(os.path.join(project_dir, ".svn")):
1100 # If there's a .svn file in the current directory, we recursively look
1101 # up the directory tree for the top of the SVN checkout
1102 root_dir = project_dir
1103 one_up_dir = os.path.dirname(root_dir)
1104 while os.path.exists(os.path.join(one_up_dir, ".svn")):
1105 root_dir = os.path.dirname(root_dir)
1106 one_up_dir = os.path.dirname(one_up_dir)
1107
1108 prefix = os.path.commonprefix([root_dir, project_dir])
1109 return fullname[len(prefix) + 1:]
1110
erg@google.com3dc74262011-11-30 01:12:00 +00001111 # Not SVN <= 1.6? Try to find a git, hg, or svn top level directory by
1112 # searching up from the current path.
Alex Vakulenko01e47232016-05-06 13:54:15 -07001113 root_dir = current_dir = os.path.dirname(fullname)
1114 while current_dir != os.path.dirname(current_dir):
1115 if (os.path.exists(os.path.join(current_dir, ".git")) or
1116 os.path.exists(os.path.join(current_dir, ".hg")) or
1117 os.path.exists(os.path.join(current_dir, ".svn"))):
1118 root_dir = current_dir
1119 current_dir = os.path.dirname(current_dir)
erg@google.com42e59b02010-10-04 22:18:07 +00001120
1121 if (os.path.exists(os.path.join(root_dir, ".git")) or
erg@google.com3dc74262011-11-30 01:12:00 +00001122 os.path.exists(os.path.join(root_dir, ".hg")) or
1123 os.path.exists(os.path.join(root_dir, ".svn"))):
erg@google.com42e59b02010-10-04 22:18:07 +00001124 prefix = os.path.commonprefix([root_dir, project_dir])
1125 return fullname[len(prefix) + 1:]
erg@google.com4e00b9a2009-01-12 23:05:11 +00001126
1127 # Don't know what to do; header guard warnings may be wrong...
1128 return fullname
1129
1130 def Split(self):
1131 """Splits the file into the directory, basename, and extension.
1132
1133 For 'chrome/browser/browser.cc', Split() would
1134 return ('chrome/browser', 'browser', '.cc')
1135
1136 Returns:
1137 A tuple of (directory, basename, extension).
1138 """
1139
1140 googlename = self.RepositoryName()
1141 project, rest = os.path.split(googlename)
1142 return (project,) + os.path.splitext(rest)
1143
1144 def BaseName(self):
1145 """File base name - text after the final slash, before the final period."""
1146 return self.Split()[1]
1147
1148 def Extension(self):
1149 """File extension - text following the final period."""
1150 return self.Split()[2]
1151
1152 def NoExtension(self):
1153 """File has no source file extension."""
1154 return '/'.join(self.Split()[0:2])
1155
1156 def IsSource(self):
1157 """File has a source file extension."""
Alex Vakulenko01e47232016-05-06 13:54:15 -07001158 return _IsSourceExtension(self.Extension()[1:])
erg@google.com4e00b9a2009-01-12 23:05:11 +00001159
1160
erg+personal@google.com05189642010-04-30 20:43:03 +00001161def _ShouldPrintError(category, confidence, linenum):
erg@google.com8a95ecc2011-09-08 00:45:54 +00001162 """If confidence >= verbose, category passes filter and is not suppressed."""
erg+personal@google.com05189642010-04-30 20:43:03 +00001163
1164 # There are three ways we might decide not to print an error message:
1165 # a "NOLINT(category)" comment appears in the source,
erg@google.com4e00b9a2009-01-12 23:05:11 +00001166 # the verbosity level isn't high enough, or the filters filter it out.
erg+personal@google.com05189642010-04-30 20:43:03 +00001167 if IsErrorSuppressedByNolint(category, linenum):
1168 return False
avakulenko@google.com02af6282014-06-04 18:53:25 +00001169
erg@google.com4e00b9a2009-01-12 23:05:11 +00001170 if confidence < _cpplint_state.verbose_level:
1171 return False
1172
1173 is_filtered = False
1174 for one_filter in _Filters():
1175 if one_filter.startswith('-'):
1176 if category.startswith(one_filter[1:]):
1177 is_filtered = True
1178 elif one_filter.startswith('+'):
1179 if category.startswith(one_filter[1:]):
1180 is_filtered = False
1181 else:
1182 assert False # should have been checked for in SetFilter.
1183 if is_filtered:
1184 return False
1185
1186 return True
1187
1188
1189def Error(filename, linenum, category, confidence, message):
1190 """Logs the fact we've found a lint error.
1191
1192 We log where the error was found, and also our confidence in the error,
1193 that is, how certain we are this is a legitimate style regression, and
1194 not a misidentification or a use that's sometimes justified.
1195
erg+personal@google.com05189642010-04-30 20:43:03 +00001196 False positives can be suppressed by the use of
1197 "cpplint(category)" comments on the offending line. These are
1198 parsed into _error_suppressions.
1199
erg@google.com4e00b9a2009-01-12 23:05:11 +00001200 Args:
1201 filename: The name of the file containing the error.
1202 linenum: The number of the line containing the error.
1203 category: A string used to describe the "category" this bug
1204 falls under: "whitespace", say, or "runtime". Categories
1205 may have a hierarchy separated by slashes: "whitespace/indent".
1206 confidence: A number from 1-5 representing a confidence score for
1207 the error, with 5 meaning that we are certain of the problem,
1208 and 1 meaning that it could be a legitimate construct.
1209 message: The error message.
1210 """
erg+personal@google.com05189642010-04-30 20:43:03 +00001211 if _ShouldPrintError(category, confidence, linenum):
erg@google.coma868d2d2009-10-09 21:18:45 +00001212 _cpplint_state.IncrementErrorCount(category)
erg@google.com4e00b9a2009-01-12 23:05:11 +00001213 if _cpplint_state.output_format == 'vs7':
schoetbi819c5722017-05-03 10:09:12 +02001214 sys.stderr.write('%s(%s): error cpplint: [%s] %s [%d]\n' % (
1215 filename, linenum, category, message, confidence))
erg@google.com02c27fd2013-05-28 21:34:34 +00001216 elif _cpplint_state.output_format == 'eclipse':
1217 sys.stderr.write('%s:%s: warning: %s [%s] [%d]\n' % (
1218 filename, linenum, message, category, confidence))
erg@google.com4e00b9a2009-01-12 23:05:11 +00001219 else:
1220 sys.stderr.write('%s:%s: %s [%s] [%d]\n' % (
1221 filename, linenum, message, category, confidence))
1222
1223
erg@google.com2aa59982013-10-28 19:09:25 +00001224# Matches standard C++ escape sequences per 2.13.2.3 of the C++ standard.
erg@google.com4e00b9a2009-01-12 23:05:11 +00001225_RE_PATTERN_CLEANSE_LINE_ESCAPES = re.compile(
1226 r'\\([abfnrtv?"\\\']|\d+|x[0-9a-fA-F]+)')
avakulenko@google.com02af6282014-06-04 18:53:25 +00001227# Match a single C style comment on the same line.
1228_RE_PATTERN_C_COMMENTS = r'/\*(?:[^*]|\*(?!/))*\*/'
1229# Matches multi-line C style comments.
erg@google.com4e00b9a2009-01-12 23:05:11 +00001230# This RE is a little bit more complicated than one might expect, because we
1231# have to take care of space removals tools so we can handle comments inside
1232# statements better.
1233# The current rule is: We only clear spaces from both sides when we're at the
1234# end of the line. Otherwise, we try to remove spaces from the right side,
1235# if this doesn't work we try on left side but only if there's a non-character
1236# on the right.
1237_RE_PATTERN_CLEANSE_LINE_C_COMMENTS = re.compile(
avakulenko@google.com02af6282014-06-04 18:53:25 +00001238 r'(\s*' + _RE_PATTERN_C_COMMENTS + r'\s*$|' +
1239 _RE_PATTERN_C_COMMENTS + r'\s+|' +
1240 r'\s+' + _RE_PATTERN_C_COMMENTS + r'(?=\W)|' +
1241 _RE_PATTERN_C_COMMENTS + r')')
erg@google.com4e00b9a2009-01-12 23:05:11 +00001242
1243
1244def IsCppString(line):
1245 """Does line terminate so, that the next symbol is in string constant.
1246
1247 This function does not consider single-line nor multi-line comments.
1248
1249 Args:
1250 line: is a partial line of code starting from the 0..n.
1251
1252 Returns:
1253 True, if next character appended to 'line' is inside a
1254 string constant.
1255 """
1256
1257 line = line.replace(r'\\', 'XX') # after this, \\" does not match to \"
1258 return ((line.count('"') - line.count(r'\"') - line.count("'\"'")) & 1) == 1
1259
1260
erg@google.com2aa59982013-10-28 19:09:25 +00001261def CleanseRawStrings(raw_lines):
1262 """Removes C++11 raw strings from lines.
1263
1264 Before:
1265 static const char kData[] = R"(
1266 multi-line string
1267 )";
1268
1269 After:
1270 static const char kData[] = ""
1271 (replaced by blank line)
1272 "";
1273
1274 Args:
1275 raw_lines: list of raw lines.
1276
1277 Returns:
1278 list of lines with C++11 raw strings replaced by empty strings.
1279 """
1280
1281 delimiter = None
1282 lines_without_raw_strings = []
1283 for line in raw_lines:
1284 if delimiter:
1285 # Inside a raw string, look for the end
1286 end = line.find(delimiter)
1287 if end >= 0:
1288 # Found the end of the string, match leading space for this
1289 # line and resume copying the original lines, and also insert
1290 # a "" on the last line.
1291 leading_space = Match(r'^(\s*)\S', line)
1292 line = leading_space.group(1) + '""' + line[end + len(delimiter):]
1293 delimiter = None
1294 else:
1295 # Haven't found the end yet, append a blank line.
avakulenko@google.com02af6282014-06-04 18:53:25 +00001296 line = '""'
erg@google.com2aa59982013-10-28 19:09:25 +00001297
avakulenko@google.com02af6282014-06-04 18:53:25 +00001298 # Look for beginning of a raw string, and replace them with
1299 # empty strings. This is done in a loop to handle multiple raw
1300 # strings on the same line.
1301 while delimiter is None:
erg@google.com2aa59982013-10-28 19:09:25 +00001302 # Look for beginning of a raw string.
1303 # See 2.14.15 [lex.string] for syntax.
Alex Vakulenko01e47232016-05-06 13:54:15 -07001304 #
1305 # Once we have matched a raw string, we check the prefix of the
1306 # line to make sure that the line is not part of a single line
1307 # comment. It's done this way because we remove raw strings
1308 # before removing comments as opposed to removing comments
1309 # before removing raw strings. This is because there are some
1310 # cpplint checks that requires the comments to be preserved, but
1311 # we don't want to check comments that are inside raw strings.
1312 matched = Match(r'^(.*?)\b(?:R|u8R|uR|UR|LR)"([^\s\\()]*)\((.*)$', line)
1313 if (matched and
1314 not Match(r'^([^\'"]|\'(\\.|[^\'])*\'|"(\\.|[^"])*")*//',
1315 matched.group(1))):
erg@google.com2aa59982013-10-28 19:09:25 +00001316 delimiter = ')' + matched.group(2) + '"'
1317
1318 end = matched.group(3).find(delimiter)
1319 if end >= 0:
1320 # Raw string ended on same line
1321 line = (matched.group(1) + '""' +
1322 matched.group(3)[end + len(delimiter):])
1323 delimiter = None
1324 else:
1325 # Start of a multi-line raw string
1326 line = matched.group(1) + '""'
avakulenko@google.com02af6282014-06-04 18:53:25 +00001327 else:
1328 break
erg@google.com2aa59982013-10-28 19:09:25 +00001329
1330 lines_without_raw_strings.append(line)
1331
1332 # TODO(unknown): if delimiter is not None here, we might want to
1333 # emit a warning for unterminated string.
1334 return lines_without_raw_strings
1335
1336
erg@google.com4e00b9a2009-01-12 23:05:11 +00001337def FindNextMultiLineCommentStart(lines, lineix):
1338 """Find the beginning marker for a multiline comment."""
1339 while lineix < len(lines):
1340 if lines[lineix].strip().startswith('/*'):
1341 # Only return this marker if the comment goes beyond this line
1342 if lines[lineix].strip().find('*/', 2) < 0:
1343 return lineix
1344 lineix += 1
1345 return len(lines)
1346
1347
1348def FindNextMultiLineCommentEnd(lines, lineix):
1349 """We are inside a comment, find the end marker."""
1350 while lineix < len(lines):
1351 if lines[lineix].strip().endswith('*/'):
1352 return lineix
1353 lineix += 1
1354 return len(lines)
1355
1356
1357def RemoveMultiLineCommentsFromRange(lines, begin, end):
1358 """Clears a range of lines for multi-line comments."""
1359 # Having // dummy comments makes the lines non-empty, so we will not get
1360 # unnecessary blank line warnings later in the code.
1361 for i in range(begin, end):
avakulenko@google.com554223d2014-12-04 22:00:20 +00001362 lines[i] = '/**/'
erg@google.com4e00b9a2009-01-12 23:05:11 +00001363
1364
1365def RemoveMultiLineComments(filename, lines, error):
1366 """Removes multiline (c-style) comments from lines."""
1367 lineix = 0
1368 while lineix < len(lines):
1369 lineix_begin = FindNextMultiLineCommentStart(lines, lineix)
1370 if lineix_begin >= len(lines):
1371 return
1372 lineix_end = FindNextMultiLineCommentEnd(lines, lineix_begin)
1373 if lineix_end >= len(lines):
1374 error(filename, lineix_begin + 1, 'readability/multiline_comment', 5,
1375 'Could not find end of multi-line comment')
1376 return
1377 RemoveMultiLineCommentsFromRange(lines, lineix_begin, lineix_end + 1)
1378 lineix = lineix_end + 1
1379
1380
1381def CleanseComments(line):
1382 """Removes //-comments and single-line C-style /* */ comments.
1383
1384 Args:
1385 line: A line of C++ source.
1386
1387 Returns:
1388 The line with single-line comments removed.
1389 """
1390 commentpos = line.find('//')
1391 if commentpos != -1 and not IsCppString(line[:commentpos]):
erg@google.comd7d27472011-09-07 17:36:35 +00001392 line = line[:commentpos].rstrip()
erg@google.com4e00b9a2009-01-12 23:05:11 +00001393 # get rid of /* ... */
1394 return _RE_PATTERN_CLEANSE_LINE_C_COMMENTS.sub('', line)
1395
1396
erg@google.coma87abb82009-02-24 01:41:01 +00001397class CleansedLines(object):
avakulenko@google.com554223d2014-12-04 22:00:20 +00001398 """Holds 4 copies of all lines with different preprocessing applied to them.
erg@google.com4e00b9a2009-01-12 23:05:11 +00001399
avakulenko@google.com554223d2014-12-04 22:00:20 +00001400 1) elided member contains lines without strings and comments.
1401 2) lines member contains lines without comments.
erg@google.comd350fe52013-01-14 17:51:48 +00001402 3) raw_lines member contains all the lines without processing.
avakulenko@google.com554223d2014-12-04 22:00:20 +00001403 4) lines_without_raw_strings member is same as raw_lines, but with C++11 raw
1404 strings removed.
1405 All these members are of <type 'list'>, and of the same length.
erg@google.com4e00b9a2009-01-12 23:05:11 +00001406 """
1407
1408 def __init__(self, lines):
1409 self.elided = []
1410 self.lines = []
1411 self.raw_lines = lines
1412 self.num_lines = len(lines)
erg@google.com2aa59982013-10-28 19:09:25 +00001413 self.lines_without_raw_strings = CleanseRawStrings(lines)
1414 for linenum in range(len(self.lines_without_raw_strings)):
1415 self.lines.append(CleanseComments(
1416 self.lines_without_raw_strings[linenum]))
1417 elided = self._CollapseStrings(self.lines_without_raw_strings[linenum])
erg@google.com4e00b9a2009-01-12 23:05:11 +00001418 self.elided.append(CleanseComments(elided))
1419
1420 def NumLines(self):
1421 """Returns the number of lines represented."""
1422 return self.num_lines
1423
1424 @staticmethod
1425 def _CollapseStrings(elided):
1426 """Collapses strings and chars on a line to simple "" or '' blocks.
1427
1428 We nix strings first so we're not fooled by text like '"http://"'
1429
1430 Args:
1431 elided: The line being processed.
1432
1433 Returns:
1434 The line with collapsed strings.
1435 """
avakulenko@google.com02af6282014-06-04 18:53:25 +00001436 if _RE_PATTERN_INCLUDE.match(elided):
1437 return elided
1438
1439 # Remove escaped characters first to make quote/single quote collapsing
1440 # basic. Things that look like escaped characters shouldn't occur
1441 # outside of strings and chars.
1442 elided = _RE_PATTERN_CLEANSE_LINE_ESCAPES.sub('', elided)
1443
1444 # Replace quoted strings and digit separators. Both single quotes
1445 # and double quotes are processed in the same loop, otherwise
1446 # nested quotes wouldn't work.
1447 collapsed = ''
1448 while True:
1449 # Find the first quote character
1450 match = Match(r'^([^\'"]*)([\'"])(.*)$', elided)
1451 if not match:
1452 collapsed += elided
1453 break
1454 head, quote, tail = match.groups()
1455
1456 if quote == '"':
1457 # Collapse double quoted strings
1458 second_quote = tail.find('"')
1459 if second_quote >= 0:
1460 collapsed += head + '""'
1461 elided = tail[second_quote + 1:]
1462 else:
1463 # Unmatched double quote, don't bother processing the rest
1464 # of the line since this is probably a multiline string.
1465 collapsed += elided
1466 break
1467 else:
1468 # Found single quote, check nearby text to eliminate digit separators.
1469 #
1470 # There is no special handling for floating point here, because
1471 # the integer/fractional/exponent parts would all be parsed
1472 # correctly as long as there are digits on both sides of the
1473 # separator. So we are fine as long as we don't see something
1474 # like "0.'3" (gcc 4.9.0 will not allow this literal).
1475 if Search(r'\b(?:0[bBxX]?|[1-9])[0-9a-fA-F]*$', head):
1476 match_literal = Match(r'^((?:\'?[0-9a-zA-Z_])*)(.*)$', "'" + tail)
1477 collapsed += head + match_literal.group(1).replace("'", '')
1478 elided = match_literal.group(2)
1479 else:
1480 second_quote = tail.find('\'')
1481 if second_quote >= 0:
1482 collapsed += head + "''"
1483 elided = tail[second_quote + 1:]
1484 else:
1485 # Unmatched single quote
1486 collapsed += elided
1487 break
1488
1489 return collapsed
erg@google.com4e00b9a2009-01-12 23:05:11 +00001490
1491
avakulenko@google.com02af6282014-06-04 18:53:25 +00001492def FindEndOfExpressionInLine(line, startpos, stack):
1493 """Find the position just after the end of current parenthesized expression.
erg@google.comd350fe52013-01-14 17:51:48 +00001494
1495 Args:
1496 line: a CleansedLines line.
1497 startpos: start searching at this position.
avakulenko@google.com02af6282014-06-04 18:53:25 +00001498 stack: nesting stack at startpos.
erg@google.comd350fe52013-01-14 17:51:48 +00001499
1500 Returns:
avakulenko@google.com02af6282014-06-04 18:53:25 +00001501 On finding matching end: (index just after matching end, None)
1502 On finding an unclosed expression: (-1, None)
1503 Otherwise: (-1, new stack at end of this line)
erg@google.comd350fe52013-01-14 17:51:48 +00001504 """
1505 for i in xrange(startpos, len(line)):
avakulenko@google.com02af6282014-06-04 18:53:25 +00001506 char = line[i]
1507 if char in '([{':
1508 # Found start of parenthesized expression, push to expression stack
1509 stack.append(char)
1510 elif char == '<':
1511 # Found potential start of template argument list
1512 if i > 0 and line[i - 1] == '<':
1513 # Left shift operator
1514 if stack and stack[-1] == '<':
1515 stack.pop()
1516 if not stack:
1517 return (-1, None)
1518 elif i > 0 and Search(r'\boperator\s*$', line[0:i]):
1519 # operator<, don't add to stack
1520 continue
1521 else:
1522 # Tentative start of template argument list
1523 stack.append('<')
1524 elif char in ')]}':
1525 # Found end of parenthesized expression.
1526 #
1527 # If we are currently expecting a matching '>', the pending '<'
1528 # must have been an operator. Remove them from expression stack.
1529 while stack and stack[-1] == '<':
1530 stack.pop()
1531 if not stack:
1532 return (-1, None)
1533 if ((stack[-1] == '(' and char == ')') or
1534 (stack[-1] == '[' and char == ']') or
1535 (stack[-1] == '{' and char == '}')):
1536 stack.pop()
1537 if not stack:
1538 return (i + 1, None)
1539 else:
1540 # Mismatched parentheses
1541 return (-1, None)
1542 elif char == '>':
1543 # Found potential end of template argument list.
1544
1545 # Ignore "->" and operator functions
1546 if (i > 0 and
1547 (line[i - 1] == '-' or Search(r'\boperator\s*$', line[0:i - 1]))):
1548 continue
1549
1550 # Pop the stack if there is a matching '<'. Otherwise, ignore
1551 # this '>' since it must be an operator.
1552 if stack:
1553 if stack[-1] == '<':
1554 stack.pop()
1555 if not stack:
1556 return (i + 1, None)
1557 elif char == ';':
1558 # Found something that look like end of statements. If we are currently
1559 # expecting a '>', the matching '<' must have been an operator, since
1560 # template argument list should not contain statements.
1561 while stack and stack[-1] == '<':
1562 stack.pop()
1563 if not stack:
1564 return (-1, None)
1565
1566 # Did not find end of expression or unbalanced parentheses on this line
1567 return (-1, stack)
erg@google.comd350fe52013-01-14 17:51:48 +00001568
1569
erg@google.com4e00b9a2009-01-12 23:05:11 +00001570def CloseExpression(clean_lines, linenum, pos):
erg@google.com2aa59982013-10-28 19:09:25 +00001571 """If input points to ( or { or [ or <, finds the position that closes it.
erg@google.com4e00b9a2009-01-12 23:05:11 +00001572
erg@google.com2aa59982013-10-28 19:09:25 +00001573 If lines[linenum][pos] points to a '(' or '{' or '[' or '<', finds the
erg@google.com4e00b9a2009-01-12 23:05:11 +00001574 linenum/pos that correspond to the closing of the expression.
1575
avakulenko@google.com02af6282014-06-04 18:53:25 +00001576 TODO(unknown): cpplint spends a fair bit of time matching parentheses.
1577 Ideally we would want to index all opening and closing parentheses once
1578 and have CloseExpression be just a simple lookup, but due to preprocessor
1579 tricks, this is not so easy.
1580
erg@google.com4e00b9a2009-01-12 23:05:11 +00001581 Args:
1582 clean_lines: A CleansedLines instance containing the file.
1583 linenum: The number of the line to check.
1584 pos: A position on the line.
1585
1586 Returns:
1587 A tuple (line, linenum, pos) pointer *past* the closing brace, or
1588 (line, len(lines), -1) if we never find a close. Note we ignore
1589 strings and comments when matching; and the line we return is the
1590 'cleansed' line at linenum.
1591 """
1592
1593 line = clean_lines.elided[linenum]
avakulenko@google.com02af6282014-06-04 18:53:25 +00001594 if (line[pos] not in '({[<') or Match(r'<[<=]', line[pos:]):
erg@google.com4e00b9a2009-01-12 23:05:11 +00001595 return (line, clean_lines.NumLines(), -1)
erg@google.com4e00b9a2009-01-12 23:05:11 +00001596
erg@google.comd350fe52013-01-14 17:51:48 +00001597 # Check first line
avakulenko@google.com02af6282014-06-04 18:53:25 +00001598 (end_pos, stack) = FindEndOfExpressionInLine(line, pos, [])
erg@google.comd350fe52013-01-14 17:51:48 +00001599 if end_pos > -1:
1600 return (line, linenum, end_pos)
erg@google.com2aa59982013-10-28 19:09:25 +00001601
1602 # Continue scanning forward
avakulenko@google.com02af6282014-06-04 18:53:25 +00001603 while stack and linenum < clean_lines.NumLines() - 1:
erg@google.com4e00b9a2009-01-12 23:05:11 +00001604 linenum += 1
1605 line = clean_lines.elided[linenum]
avakulenko@google.com02af6282014-06-04 18:53:25 +00001606 (end_pos, stack) = FindEndOfExpressionInLine(line, 0, stack)
erg@google.com2aa59982013-10-28 19:09:25 +00001607 if end_pos > -1:
1608 return (line, linenum, end_pos)
erg@google.com4e00b9a2009-01-12 23:05:11 +00001609
avakulenko@google.com02af6282014-06-04 18:53:25 +00001610 # Did not find end of expression before end of file, give up
erg@google.comd350fe52013-01-14 17:51:48 +00001611 return (line, clean_lines.NumLines(), -1)
erg@google.com4e00b9a2009-01-12 23:05:11 +00001612
erg@google.com2aa59982013-10-28 19:09:25 +00001613
avakulenko@google.com02af6282014-06-04 18:53:25 +00001614def FindStartOfExpressionInLine(line, endpos, stack):
1615 """Find position at the matching start of current expression.
erg@google.com2aa59982013-10-28 19:09:25 +00001616
1617 This is almost the reverse of FindEndOfExpressionInLine, but note
1618 that the input position and returned position differs by 1.
1619
1620 Args:
1621 line: a CleansedLines line.
1622 endpos: start searching at this position.
avakulenko@google.com02af6282014-06-04 18:53:25 +00001623 stack: nesting stack at endpos.
erg@google.com2aa59982013-10-28 19:09:25 +00001624
1625 Returns:
avakulenko@google.com02af6282014-06-04 18:53:25 +00001626 On finding matching start: (index at matching start, None)
1627 On finding an unclosed expression: (-1, None)
1628 Otherwise: (-1, new stack at beginning of this line)
erg@google.com2aa59982013-10-28 19:09:25 +00001629 """
avakulenko@google.com02af6282014-06-04 18:53:25 +00001630 i = endpos
1631 while i >= 0:
1632 char = line[i]
1633 if char in ')]}':
1634 # Found end of expression, push to expression stack
1635 stack.append(char)
1636 elif char == '>':
1637 # Found potential end of template argument list.
1638 #
1639 # Ignore it if it's a "->" or ">=" or "operator>"
1640 if (i > 0 and
1641 (line[i - 1] == '-' or
1642 Match(r'\s>=\s', line[i - 1:]) or
1643 Search(r'\boperator\s*$', line[0:i]))):
1644 i -= 1
1645 else:
1646 stack.append('>')
1647 elif char == '<':
1648 # Found potential start of template argument list
1649 if i > 0 and line[i - 1] == '<':
1650 # Left shift operator
1651 i -= 1
1652 else:
1653 # If there is a matching '>', we can pop the expression stack.
1654 # Otherwise, ignore this '<' since it must be an operator.
1655 if stack and stack[-1] == '>':
1656 stack.pop()
1657 if not stack:
1658 return (i, None)
1659 elif char in '([{':
1660 # Found start of expression.
1661 #
1662 # If there are any unmatched '>' on the stack, they must be
1663 # operators. Remove those.
1664 while stack and stack[-1] == '>':
1665 stack.pop()
1666 if not stack:
1667 return (-1, None)
1668 if ((char == '(' and stack[-1] == ')') or
1669 (char == '[' and stack[-1] == ']') or
1670 (char == '{' and stack[-1] == '}')):
1671 stack.pop()
1672 if not stack:
1673 return (i, None)
1674 else:
1675 # Mismatched parentheses
1676 return (-1, None)
1677 elif char == ';':
1678 # Found something that look like end of statements. If we are currently
1679 # expecting a '<', the matching '>' must have been an operator, since
1680 # template argument list should not contain statements.
1681 while stack and stack[-1] == '>':
1682 stack.pop()
1683 if not stack:
1684 return (-1, None)
1685
1686 i -= 1
1687
1688 return (-1, stack)
erg@google.com2aa59982013-10-28 19:09:25 +00001689
1690
1691def ReverseCloseExpression(clean_lines, linenum, pos):
1692 """If input points to ) or } or ] or >, finds the position that opens it.
1693
1694 If lines[linenum][pos] points to a ')' or '}' or ']' or '>', finds the
1695 linenum/pos that correspond to the opening of the expression.
1696
1697 Args:
1698 clean_lines: A CleansedLines instance containing the file.
1699 linenum: The number of the line to check.
1700 pos: A position on the line.
1701
1702 Returns:
1703 A tuple (line, linenum, pos) pointer *at* the opening brace, or
1704 (line, 0, -1) if we never find the matching opening brace. Note
1705 we ignore strings and comments when matching; and the line we
1706 return is the 'cleansed' line at linenum.
1707 """
1708 line = clean_lines.elided[linenum]
avakulenko@google.com02af6282014-06-04 18:53:25 +00001709 if line[pos] not in ')}]>':
erg@google.com2aa59982013-10-28 19:09:25 +00001710 return (line, 0, -1)
erg@google.com2aa59982013-10-28 19:09:25 +00001711
1712 # Check last line
avakulenko@google.com02af6282014-06-04 18:53:25 +00001713 (start_pos, stack) = FindStartOfExpressionInLine(line, pos, [])
erg@google.com2aa59982013-10-28 19:09:25 +00001714 if start_pos > -1:
1715 return (line, linenum, start_pos)
1716
1717 # Continue scanning backward
avakulenko@google.com02af6282014-06-04 18:53:25 +00001718 while stack and linenum > 0:
erg@google.com2aa59982013-10-28 19:09:25 +00001719 linenum -= 1
1720 line = clean_lines.elided[linenum]
avakulenko@google.com02af6282014-06-04 18:53:25 +00001721 (start_pos, stack) = FindStartOfExpressionInLine(line, len(line) - 1, stack)
erg@google.com2aa59982013-10-28 19:09:25 +00001722 if start_pos > -1:
1723 return (line, linenum, start_pos)
1724
avakulenko@google.com02af6282014-06-04 18:53:25 +00001725 # Did not find start of expression before beginning of file, give up
erg@google.com2aa59982013-10-28 19:09:25 +00001726 return (line, 0, -1)
1727
1728
erg@google.com4e00b9a2009-01-12 23:05:11 +00001729def CheckForCopyright(filename, lines, error):
1730 """Logs an error if no Copyright message appears at the top of the file."""
1731
1732 # We'll say it should occur by line 10. Don't forget there's a
1733 # dummy line at the front.
1734 for line in xrange(1, min(len(lines), 11)):
1735 if re.search(r'Copyright', lines[line], re.I): break
1736 else: # means no copyright line was found
1737 error(filename, 0, 'legal/copyright', 5,
1738 'No copyright message found. '
1739 'You should have a line: "Copyright [year] <Copyright Owner>"')
1740
1741
avakulenko@google.com02af6282014-06-04 18:53:25 +00001742def GetIndentLevel(line):
1743 """Return the number of leading spaces in line.
1744
1745 Args:
1746 line: A string to check.
1747
1748 Returns:
1749 An integer count of leading spaces, possibly zero.
1750 """
1751 indent = Match(r'^( *)\S', line)
1752 if indent:
1753 return len(indent.group(1))
1754 else:
1755 return 0
1756
1757
erg@google.com4e00b9a2009-01-12 23:05:11 +00001758def GetHeaderGuardCPPVariable(filename):
1759 """Returns the CPP variable that should be used as a header guard.
1760
1761 Args:
1762 filename: The name of a C++ header file.
1763
1764 Returns:
1765 The CPP variable that should be used as a header guard in the
1766 named file.
1767
1768 """
1769
erg+personal@google.com05189642010-04-30 20:43:03 +00001770 # Restores original filename in case that cpplint is invoked from Emacs's
1771 # flymake.
1772 filename = re.sub(r'_flymake\.h$', '.h', filename)
erg@google.comd350fe52013-01-14 17:51:48 +00001773 filename = re.sub(r'/\.flymake/([^/]*)$', r'/\1', filename)
avakulenko@google.com554223d2014-12-04 22:00:20 +00001774 # Replace 'c++' with 'cpp'.
1775 filename = filename.replace('C++', 'cpp').replace('c++', 'cpp')
Alex Vakulenko01e47232016-05-06 13:54:15 -07001776
erg@google.com4e00b9a2009-01-12 23:05:11 +00001777 fileinfo = FileInfo(filename)
erg@google.com4d70a882013-04-16 21:06:32 +00001778 file_path_from_root = fileinfo.RepositoryName()
1779 if _root:
Sergey Sharybin3b0ea892016-05-31 00:21:14 +02001780 suffix = os.sep
1781 # On Windows using directory separator will leave us with
1782 # "bogus escape error" unless we properly escape regex.
1783 if suffix == '\\':
1784 suffix += '\\'
1785 file_path_from_root = re.sub('^' + _root + suffix, '', file_path_from_root)
avakulenko@google.com554223d2014-12-04 22:00:20 +00001786 return re.sub(r'[^a-zA-Z0-9]', '_', file_path_from_root).upper() + '_'
erg@google.com4e00b9a2009-01-12 23:05:11 +00001787
1788
avakulenko@google.com554223d2014-12-04 22:00:20 +00001789def CheckForHeaderGuard(filename, clean_lines, error):
erg@google.com4e00b9a2009-01-12 23:05:11 +00001790 """Checks that the file contains a header guard.
1791
erg@google.coma87abb82009-02-24 01:41:01 +00001792 Logs an error if no #ifndef header guard is present. For other
erg@google.com4e00b9a2009-01-12 23:05:11 +00001793 headers, checks that the full pathname is used.
1794
1795 Args:
1796 filename: The name of the C++ header file.
avakulenko@google.com554223d2014-12-04 22:00:20 +00001797 clean_lines: A CleansedLines instance containing the file.
erg@google.com4e00b9a2009-01-12 23:05:11 +00001798 error: The function to call with any errors found.
1799 """
1800
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00001801 # Don't check for header guards if there are error suppression
1802 # comments somewhere in this file.
1803 #
1804 # Because this is silencing a warning for a nonexistent line, we
1805 # only support the very specific NOLINT(build/header_guard) syntax,
1806 # and not the general NOLINT or NOLINT(*) syntax.
avakulenko@google.com554223d2014-12-04 22:00:20 +00001807 raw_lines = clean_lines.lines_without_raw_strings
1808 for i in raw_lines:
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00001809 if Search(r'//\s*NOLINT\(build/header_guard\)', i):
1810 return
1811
erg@google.com4e00b9a2009-01-12 23:05:11 +00001812 cppvar = GetHeaderGuardCPPVariable(filename)
1813
avakulenko@google.com554223d2014-12-04 22:00:20 +00001814 ifndef = ''
erg@google.com4e00b9a2009-01-12 23:05:11 +00001815 ifndef_linenum = 0
avakulenko@google.com554223d2014-12-04 22:00:20 +00001816 define = ''
1817 endif = ''
erg@google.com4e00b9a2009-01-12 23:05:11 +00001818 endif_linenum = 0
avakulenko@google.com554223d2014-12-04 22:00:20 +00001819 for linenum, line in enumerate(raw_lines):
erg@google.com4e00b9a2009-01-12 23:05:11 +00001820 linesplit = line.split()
1821 if len(linesplit) >= 2:
1822 # find the first occurrence of #ifndef and #define, save arg
1823 if not ifndef and linesplit[0] == '#ifndef':
1824 # set ifndef to the header guard presented on the #ifndef line.
1825 ifndef = linesplit[1]
1826 ifndef_linenum = linenum
1827 if not define and linesplit[0] == '#define':
1828 define = linesplit[1]
1829 # find the last occurrence of #endif, save entire line
1830 if line.startswith('#endif'):
1831 endif = line
1832 endif_linenum = linenum
1833
avakulenko@google.com554223d2014-12-04 22:00:20 +00001834 if not ifndef or not define or ifndef != define:
erg@google.com4e00b9a2009-01-12 23:05:11 +00001835 error(filename, 0, 'build/header_guard', 5,
1836 'No #ifndef header guard found, suggested CPP variable is: %s' %
1837 cppvar)
1838 return
1839
1840 # The guard should be PATH_FILE_H_, but we also allow PATH_FILE_H__
1841 # for backward compatibility.
erg+personal@google.com05189642010-04-30 20:43:03 +00001842 if ifndef != cppvar:
erg@google.com4e00b9a2009-01-12 23:05:11 +00001843 error_level = 0
1844 if ifndef != cppvar + '_':
1845 error_level = 5
1846
avakulenko@google.com554223d2014-12-04 22:00:20 +00001847 ParseNolintSuppressions(filename, raw_lines[ifndef_linenum], ifndef_linenum,
erg+personal@google.com05189642010-04-30 20:43:03 +00001848 error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00001849 error(filename, ifndef_linenum, 'build/header_guard', error_level,
1850 '#ifndef header guard has wrong style, please use: %s' % cppvar)
1851
avakulenko@google.com554223d2014-12-04 22:00:20 +00001852 # Check for "//" comments on endif line.
1853 ParseNolintSuppressions(filename, raw_lines[endif_linenum], endif_linenum,
1854 error)
1855 match = Match(r'#endif\s*//\s*' + cppvar + r'(_)?\b', endif)
1856 if match:
1857 if match.group(1) == '_':
1858 # Issue low severity warning for deprecated double trailing underscore
1859 error(filename, endif_linenum, 'build/header_guard', 0,
1860 '#endif line should be "#endif // %s"' % cppvar)
erg@google.comdc289702012-01-26 20:30:03 +00001861 return
1862
avakulenko@google.com554223d2014-12-04 22:00:20 +00001863 # Didn't find the corresponding "//" comment. If this file does not
1864 # contain any "//" comments at all, it could be that the compiler
1865 # only wants "/**/" comments, look for those instead.
1866 no_single_line_comments = True
1867 for i in xrange(1, len(raw_lines) - 1):
1868 line = raw_lines[i]
1869 if Match(r'^(?:(?:\'(?:\.|[^\'])*\')|(?:"(?:\.|[^"])*")|[^\'"])*//', line):
1870 no_single_line_comments = False
1871 break
erg@google.com4e00b9a2009-01-12 23:05:11 +00001872
avakulenko@google.com554223d2014-12-04 22:00:20 +00001873 if no_single_line_comments:
1874 match = Match(r'#endif\s*/\*\s*' + cppvar + r'(_)?\s*\*/', endif)
1875 if match:
1876 if match.group(1) == '_':
1877 # Low severity warning for double trailing underscore
1878 error(filename, endif_linenum, 'build/header_guard', 0,
1879 '#endif line should be "#endif /* %s */"' % cppvar)
1880 return
1881
1882 # Didn't find anything
1883 error(filename, endif_linenum, 'build/header_guard', 5,
1884 '#endif line should be "#endif // %s"' % cppvar)
1885
1886
1887def CheckHeaderFileIncluded(filename, include_state, error):
1888 """Logs an error if a .cc file does not include its header."""
1889
1890 # Do not check test files
Alex Vakulenko01e47232016-05-06 13:54:15 -07001891 fileinfo = FileInfo(filename)
1892 if Search(_TEST_FILE_SUFFIX, fileinfo.BaseName()):
avakulenko@google.com554223d2014-12-04 22:00:20 +00001893 return
1894
Alex Vakulenko01e47232016-05-06 13:54:15 -07001895 headerfile = filename[0:len(filename) - len(fileinfo.Extension())] + '.h'
avakulenko@google.com554223d2014-12-04 22:00:20 +00001896 if not os.path.exists(headerfile):
1897 return
1898 headername = FileInfo(headerfile).RepositoryName()
1899 first_include = 0
1900 for section_list in include_state.include_list:
1901 for f in section_list:
1902 if headername in f[0] or f[0] in headername:
1903 return
1904 if not first_include:
1905 first_include = f[1]
1906
1907 error(filename, first_include, 'build/include', 5,
1908 '%s should include its header file %s' % (fileinfo.RepositoryName(),
1909 headername))
erg@google.com4e00b9a2009-01-12 23:05:11 +00001910
1911
erg@google.com2aa59982013-10-28 19:09:25 +00001912def CheckForBadCharacters(filename, lines, error):
1913 """Logs an error for each line containing bad characters.
erg@google.com4e00b9a2009-01-12 23:05:11 +00001914
erg@google.com2aa59982013-10-28 19:09:25 +00001915 Two kinds of bad characters:
1916
1917 1. Unicode replacement characters: These indicate that either the file
1918 contained invalid UTF-8 (likely) or Unicode replacement characters (which
1919 it shouldn't). Note that it's possible for this to throw off line
1920 numbering if the invalid UTF-8 occurred adjacent to a newline.
1921
1922 2. NUL bytes. These are problematic for some tools.
erg@google.com4e00b9a2009-01-12 23:05:11 +00001923
1924 Args:
1925 filename: The name of the current file.
1926 lines: An array of strings, each representing a line of the file.
1927 error: The function to call with any errors found.
1928 """
1929 for linenum, line in enumerate(lines):
1930 if u'\ufffd' in line:
1931 error(filename, linenum, 'readability/utf8', 5,
1932 'Line contains invalid UTF-8 (or Unicode replacement character).')
erg@google.com2aa59982013-10-28 19:09:25 +00001933 if '\0' in line:
1934 error(filename, linenum, 'readability/nul', 5, 'Line contains NUL byte.')
erg@google.com4e00b9a2009-01-12 23:05:11 +00001935
1936
1937def CheckForNewlineAtEOF(filename, lines, error):
1938 """Logs an error if there is no newline char at the end of the file.
1939
1940 Args:
1941 filename: The name of the current file.
1942 lines: An array of strings, each representing a line of the file.
1943 error: The function to call with any errors found.
1944 """
1945
1946 # The array lines() was created by adding two newlines to the
1947 # original file (go figure), then splitting on \n.
1948 # To verify that the file ends in \n, we just have to make sure the
1949 # last-but-two element of lines() exists and is empty.
1950 if len(lines) < 3 or lines[-2]:
1951 error(filename, len(lines) - 2, 'whitespace/ending_newline', 5,
1952 'Could not find a newline character at the end of the file.')
1953
1954
1955def CheckForMultilineCommentsAndStrings(filename, clean_lines, linenum, error):
1956 """Logs an error if we see /* ... */ or "..." that extend past one line.
1957
1958 /* ... */ comments are legit inside macros, for one line.
1959 Otherwise, we prefer // comments, so it's ok to warn about the
1960 other. Likewise, it's ok for strings to extend across multiple
1961 lines, as long as a line continuation character (backslash)
1962 terminates each line. Although not currently prohibited by the C++
1963 style guide, it's ugly and unnecessary. We don't do well with either
1964 in this lint program, so we warn about both.
1965
1966 Args:
1967 filename: The name of the current file.
1968 clean_lines: A CleansedLines instance containing the file.
1969 linenum: The number of the line to check.
1970 error: The function to call with any errors found.
1971 """
1972 line = clean_lines.elided[linenum]
1973
1974 # Remove all \\ (escaped backslashes) from the line. They are OK, and the
1975 # second (escaped) slash may trigger later \" detection erroneously.
1976 line = line.replace('\\\\', '')
1977
1978 if line.count('/*') > line.count('*/'):
1979 error(filename, linenum, 'readability/multiline_comment', 5,
1980 'Complex multi-line /*...*/-style comment found. '
1981 'Lint may give bogus warnings. '
1982 'Consider replacing these with //-style comments, '
1983 'with #if 0...#endif, '
1984 'or with more clearly structured multi-line comments.')
1985
1986 if (line.count('"') - line.count('\\"')) % 2:
1987 error(filename, linenum, 'readability/multiline_string', 5,
1988 'Multi-line string ("...") found. This lint script doesn\'t '
erg@google.com2aa59982013-10-28 19:09:25 +00001989 'do well with such strings, and may give bogus warnings. '
1990 'Use C++11 raw strings or concatenation instead.')
erg@google.com4e00b9a2009-01-12 23:05:11 +00001991
1992
avakulenko@google.com02af6282014-06-04 18:53:25 +00001993# (non-threadsafe name, thread-safe alternative, validation pattern)
1994#
1995# The validation pattern is used to eliminate false positives such as:
1996# _rand(); // false positive due to substring match.
1997# ->rand(); // some member function rand().
1998# ACMRandom rand(seed); // some variable named rand.
1999# ISAACRandom rand(); // another variable named rand.
2000#
2001# Basically we require the return value of these functions to be used
2002# in some expression context on the same line by matching on some
2003# operator before the function name. This eliminates constructors and
2004# member function calls.
2005_UNSAFE_FUNC_PREFIX = r'(?:[-+*/=%^&|(<]\s*|>\s+)'
2006_THREADING_LIST = (
2007 ('asctime(', 'asctime_r(', _UNSAFE_FUNC_PREFIX + r'asctime\([^)]+\)'),
2008 ('ctime(', 'ctime_r(', _UNSAFE_FUNC_PREFIX + r'ctime\([^)]+\)'),
2009 ('getgrgid(', 'getgrgid_r(', _UNSAFE_FUNC_PREFIX + r'getgrgid\([^)]+\)'),
2010 ('getgrnam(', 'getgrnam_r(', _UNSAFE_FUNC_PREFIX + r'getgrnam\([^)]+\)'),
2011 ('getlogin(', 'getlogin_r(', _UNSAFE_FUNC_PREFIX + r'getlogin\(\)'),
2012 ('getpwnam(', 'getpwnam_r(', _UNSAFE_FUNC_PREFIX + r'getpwnam\([^)]+\)'),
2013 ('getpwuid(', 'getpwuid_r(', _UNSAFE_FUNC_PREFIX + r'getpwuid\([^)]+\)'),
2014 ('gmtime(', 'gmtime_r(', _UNSAFE_FUNC_PREFIX + r'gmtime\([^)]+\)'),
2015 ('localtime(', 'localtime_r(', _UNSAFE_FUNC_PREFIX + r'localtime\([^)]+\)'),
2016 ('rand(', 'rand_r(', _UNSAFE_FUNC_PREFIX + r'rand\(\)'),
2017 ('strtok(', 'strtok_r(',
2018 _UNSAFE_FUNC_PREFIX + r'strtok\([^)]+\)'),
2019 ('ttyname(', 'ttyname_r(', _UNSAFE_FUNC_PREFIX + r'ttyname\([^)]+\)'),
erg@google.com4e00b9a2009-01-12 23:05:11 +00002020 )
2021
2022
2023def CheckPosixThreading(filename, clean_lines, linenum, error):
2024 """Checks for calls to thread-unsafe functions.
2025
2026 Much code has been originally written without consideration of
2027 multi-threading. Also, engineers are relying on their old experience;
2028 they have learned posix before threading extensions were added. These
2029 tests guide the engineers to use thread-safe functions (when using
2030 posix directly).
2031
2032 Args:
2033 filename: The name of the current file.
2034 clean_lines: A CleansedLines instance containing the file.
2035 linenum: The number of the line to check.
2036 error: The function to call with any errors found.
2037 """
2038 line = clean_lines.elided[linenum]
avakulenko@google.com02af6282014-06-04 18:53:25 +00002039 for single_thread_func, multithread_safe_func, pattern in _THREADING_LIST:
2040 # Additional pattern matching check to confirm that this is the
2041 # function we are looking for
2042 if Search(pattern, line):
erg@google.com4e00b9a2009-01-12 23:05:11 +00002043 error(filename, linenum, 'runtime/threadsafe_fn', 2,
avakulenko@google.com02af6282014-06-04 18:53:25 +00002044 'Consider using ' + multithread_safe_func +
2045 '...) instead of ' + single_thread_func +
erg@google.com4e00b9a2009-01-12 23:05:11 +00002046 '...) for improved thread safety.')
2047
2048
erg@google.com2aa59982013-10-28 19:09:25 +00002049def CheckVlogArguments(filename, clean_lines, linenum, error):
2050 """Checks that VLOG() is only used for defining a logging level.
2051
2052 For example, VLOG(2) is correct. VLOG(INFO), VLOG(WARNING), VLOG(ERROR), and
2053 VLOG(FATAL) are not.
2054
2055 Args:
2056 filename: The name of the current file.
2057 clean_lines: A CleansedLines instance containing the file.
2058 linenum: The number of the line to check.
2059 error: The function to call with any errors found.
2060 """
2061 line = clean_lines.elided[linenum]
2062 if Search(r'\bVLOG\((INFO|ERROR|WARNING|DFATAL|FATAL)\)', line):
2063 error(filename, linenum, 'runtime/vlog', 5,
2064 'VLOG() should be used with numeric verbosity level. '
2065 'Use LOG() if you want symbolic severity levels.')
2066
erg@google.coma868d2d2009-10-09 21:18:45 +00002067# Matches invalid increment: *count++, which moves pointer instead of
erg@google.com36649102009-03-25 21:18:36 +00002068# incrementing a value.
erg@google.coma868d2d2009-10-09 21:18:45 +00002069_RE_PATTERN_INVALID_INCREMENT = re.compile(
erg@google.com36649102009-03-25 21:18:36 +00002070 r'^\s*\*\w+(\+\+|--);')
2071
2072
2073def CheckInvalidIncrement(filename, clean_lines, linenum, error):
erg@google.coma868d2d2009-10-09 21:18:45 +00002074 """Checks for invalid increment *count++.
erg@google.com36649102009-03-25 21:18:36 +00002075
2076 For example following function:
2077 void increment_counter(int* count) {
2078 *count++;
2079 }
2080 is invalid, because it effectively does count++, moving pointer, and should
2081 be replaced with ++*count, (*count)++ or *count += 1.
2082
2083 Args:
2084 filename: The name of the current file.
2085 clean_lines: A CleansedLines instance containing the file.
2086 linenum: The number of the line to check.
2087 error: The function to call with any errors found.
2088 """
2089 line = clean_lines.elided[linenum]
erg@google.coma868d2d2009-10-09 21:18:45 +00002090 if _RE_PATTERN_INVALID_INCREMENT.match(line):
erg@google.com36649102009-03-25 21:18:36 +00002091 error(filename, linenum, 'runtime/invalid_increment', 5,
2092 'Changing pointer instead of value (or unused value of operator*).')
2093
2094
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00002095def IsMacroDefinition(clean_lines, linenum):
2096 if Search(r'^#define', clean_lines[linenum]):
2097 return True
2098
2099 if linenum > 0 and Search(r'\\$', clean_lines[linenum - 1]):
2100 return True
2101
2102 return False
2103
2104
2105def IsForwardClassDeclaration(clean_lines, linenum):
2106 return Match(r'^\s*(\btemplate\b)*.*class\s+\w+;\s*$', clean_lines[linenum])
2107
2108
erg@google.comd350fe52013-01-14 17:51:48 +00002109class _BlockInfo(object):
2110 """Stores information about a generic block of code."""
2111
Alex Vakulenko01e47232016-05-06 13:54:15 -07002112 def __init__(self, linenum, seen_open_brace):
2113 self.starting_linenum = linenum
erg@google.comd350fe52013-01-14 17:51:48 +00002114 self.seen_open_brace = seen_open_brace
2115 self.open_parentheses = 0
2116 self.inline_asm = _NO_ASM
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00002117 self.check_namespace_indentation = False
erg@google.comd350fe52013-01-14 17:51:48 +00002118
2119 def CheckBegin(self, filename, clean_lines, linenum, error):
2120 """Run checks that applies to text up to the opening brace.
2121
2122 This is mostly for checking the text after the class identifier
2123 and the "{", usually where the base class is specified. For other
2124 blocks, there isn't much to check, so we always pass.
2125
2126 Args:
2127 filename: The name of the current file.
2128 clean_lines: A CleansedLines instance containing the file.
2129 linenum: The number of the line to check.
2130 error: The function to call with any errors found.
2131 """
2132 pass
2133
2134 def CheckEnd(self, filename, clean_lines, linenum, error):
2135 """Run checks that applies to text after the closing brace.
2136
2137 This is mostly used for checking end of namespace comments.
2138
2139 Args:
2140 filename: The name of the current file.
2141 clean_lines: A CleansedLines instance containing the file.
2142 linenum: The number of the line to check.
2143 error: The function to call with any errors found.
2144 """
2145 pass
2146
avakulenko@google.com02af6282014-06-04 18:53:25 +00002147 def IsBlockInfo(self):
2148 """Returns true if this block is a _BlockInfo.
2149
2150 This is convenient for verifying that an object is an instance of
2151 a _BlockInfo, but not an instance of any of the derived classes.
2152
2153 Returns:
2154 True for this class, False for derived classes.
2155 """
2156 return self.__class__ == _BlockInfo
2157
2158
2159class _ExternCInfo(_BlockInfo):
2160 """Stores information about an 'extern "C"' block."""
2161
Alex Vakulenko01e47232016-05-06 13:54:15 -07002162 def __init__(self, linenum):
2163 _BlockInfo.__init__(self, linenum, True)
avakulenko@google.com02af6282014-06-04 18:53:25 +00002164
erg@google.comd350fe52013-01-14 17:51:48 +00002165
2166class _ClassInfo(_BlockInfo):
erg@google.com4e00b9a2009-01-12 23:05:11 +00002167 """Stores information about a class."""
2168
erg@google.comd350fe52013-01-14 17:51:48 +00002169 def __init__(self, name, class_or_struct, clean_lines, linenum):
Alex Vakulenko01e47232016-05-06 13:54:15 -07002170 _BlockInfo.__init__(self, linenum, False)
erg@google.com4e00b9a2009-01-12 23:05:11 +00002171 self.name = name
erg@google.com4e00b9a2009-01-12 23:05:11 +00002172 self.is_derived = False
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00002173 self.check_namespace_indentation = True
erg@google.comd350fe52013-01-14 17:51:48 +00002174 if class_or_struct == 'struct':
2175 self.access = 'public'
erg@google.comfd5da632013-10-25 17:39:45 +00002176 self.is_struct = True
erg@google.comd350fe52013-01-14 17:51:48 +00002177 else:
2178 self.access = 'private'
erg@google.comfd5da632013-10-25 17:39:45 +00002179 self.is_struct = False
2180
2181 # Remember initial indentation level for this class. Using raw_lines here
erg@google.comc6671232013-10-25 21:44:03 +00002182 # instead of elided to account for leading comments.
avakulenko@google.com02af6282014-06-04 18:53:25 +00002183 self.class_indent = GetIndentLevel(clean_lines.raw_lines[linenum])
erg@google.com4e00b9a2009-01-12 23:05:11 +00002184
erg@google.com8a95ecc2011-09-08 00:45:54 +00002185 # Try to find the end of the class. This will be confused by things like:
2186 # class A {
2187 # } *x = { ...
2188 #
2189 # But it's still good enough for CheckSectionSpacing.
2190 self.last_line = 0
2191 depth = 0
2192 for i in range(linenum, clean_lines.NumLines()):
erg@google.comd350fe52013-01-14 17:51:48 +00002193 line = clean_lines.elided[i]
erg@google.com8a95ecc2011-09-08 00:45:54 +00002194 depth += line.count('{') - line.count('}')
2195 if not depth:
2196 self.last_line = i
2197 break
2198
erg@google.comd350fe52013-01-14 17:51:48 +00002199 def CheckBegin(self, filename, clean_lines, linenum, error):
2200 # Look for a bare ':'
2201 if Search('(^|[^:]):($|[^:])', clean_lines.elided[linenum]):
2202 self.is_derived = True
erg@google.com4e00b9a2009-01-12 23:05:11 +00002203
erg@google.comfd5da632013-10-25 17:39:45 +00002204 def CheckEnd(self, filename, clean_lines, linenum, error):
avakulenko@google.com554223d2014-12-04 22:00:20 +00002205 # If there is a DISALLOW macro, it should appear near the end of
2206 # the class.
2207 seen_last_thing_in_class = False
2208 for i in xrange(linenum - 1, self.starting_linenum, -1):
2209 match = Search(
2210 r'\b(DISALLOW_COPY_AND_ASSIGN|DISALLOW_IMPLICIT_CONSTRUCTORS)\(' +
2211 self.name + r'\)',
2212 clean_lines.elided[i])
2213 if match:
2214 if seen_last_thing_in_class:
2215 error(filename, i, 'readability/constructors', 3,
2216 match.group(1) + ' should be the last thing in the class')
2217 break
2218
2219 if not Match(r'^\s*$', clean_lines.elided[i]):
2220 seen_last_thing_in_class = True
2221
erg@google.comfd5da632013-10-25 17:39:45 +00002222 # Check that closing brace is aligned with beginning of the class.
2223 # Only do this if the closing brace is indented by only whitespaces.
2224 # This means we will not check single-line class definitions.
2225 indent = Match(r'^( *)\}', clean_lines.elided[linenum])
2226 if indent and len(indent.group(1)) != self.class_indent:
2227 if self.is_struct:
2228 parent = 'struct ' + self.name
2229 else:
2230 parent = 'class ' + self.name
2231 error(filename, linenum, 'whitespace/indent', 3,
2232 'Closing brace should be aligned with beginning of %s' % parent)
2233
erg@google.com4e00b9a2009-01-12 23:05:11 +00002234
erg@google.comd350fe52013-01-14 17:51:48 +00002235class _NamespaceInfo(_BlockInfo):
2236 """Stores information about a namespace."""
2237
2238 def __init__(self, name, linenum):
Alex Vakulenko01e47232016-05-06 13:54:15 -07002239 _BlockInfo.__init__(self, linenum, False)
erg@google.comd350fe52013-01-14 17:51:48 +00002240 self.name = name or ''
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00002241 self.check_namespace_indentation = True
erg@google.comd350fe52013-01-14 17:51:48 +00002242
2243 def CheckEnd(self, filename, clean_lines, linenum, error):
2244 """Check end of namespace comments."""
2245 line = clean_lines.raw_lines[linenum]
2246
2247 # Check how many lines is enclosed in this namespace. Don't issue
2248 # warning for missing namespace comments if there aren't enough
2249 # lines. However, do apply checks if there is already an end of
2250 # namespace comment and it's incorrect.
2251 #
2252 # TODO(unknown): We always want to check end of namespace comments
2253 # if a namespace is large, but sometimes we also want to apply the
2254 # check if a short namespace contained nontrivial things (something
2255 # other than forward declarations). There is currently no logic on
2256 # deciding what these nontrivial things are, so this check is
2257 # triggered by namespace size only, which works most of the time.
2258 if (linenum - self.starting_linenum < 10
Alex Vakulenko01e47232016-05-06 13:54:15 -07002259 and not Match(r'^\s*};*\s*(//|/\*).*\bnamespace\b', line)):
erg@google.comd350fe52013-01-14 17:51:48 +00002260 return
2261
2262 # Look for matching comment at end of namespace.
2263 #
2264 # Note that we accept C style "/* */" comments for terminating
2265 # namespaces, so that code that terminate namespaces inside
erg@google.comc6671232013-10-25 21:44:03 +00002266 # preprocessor macros can be cpplint clean.
erg@google.comd350fe52013-01-14 17:51:48 +00002267 #
2268 # We also accept stuff like "// end of namespace <name>." with the
2269 # period at the end.
2270 #
2271 # Besides these, we don't accept anything else, otherwise we might
2272 # get false negatives when existing comment is a substring of the
erg@google.comc6671232013-10-25 21:44:03 +00002273 # expected namespace.
erg@google.comd350fe52013-01-14 17:51:48 +00002274 if self.name:
2275 # Named namespace
Alex Vakulenko01e47232016-05-06 13:54:15 -07002276 if not Match((r'^\s*};*\s*(//|/\*).*\bnamespace\s+' +
2277 re.escape(self.name) + r'[\*/\.\\\s]*$'),
erg@google.comd350fe52013-01-14 17:51:48 +00002278 line):
2279 error(filename, linenum, 'readability/namespace', 5,
2280 'Namespace should be terminated with "// namespace %s"' %
2281 self.name)
2282 else:
2283 # Anonymous namespace
Alex Vakulenko01e47232016-05-06 13:54:15 -07002284 if not Match(r'^\s*};*\s*(//|/\*).*\bnamespace[\*/\.\\\s]*$', line):
avakulenko@google.com02af6282014-06-04 18:53:25 +00002285 # If "// namespace anonymous" or "// anonymous namespace (more text)",
2286 # mention "// anonymous namespace" as an acceptable form
Alex Vakulenko01e47232016-05-06 13:54:15 -07002287 if Match(r'^\s*}.*\b(namespace anonymous|anonymous namespace)\b', line):
avakulenko@google.com02af6282014-06-04 18:53:25 +00002288 error(filename, linenum, 'readability/namespace', 5,
2289 'Anonymous namespace should be terminated with "// namespace"'
2290 ' or "// anonymous namespace"')
2291 else:
2292 error(filename, linenum, 'readability/namespace', 5,
2293 'Anonymous namespace should be terminated with "// namespace"')
erg@google.comd350fe52013-01-14 17:51:48 +00002294
2295
2296class _PreprocessorInfo(object):
2297 """Stores checkpoints of nesting stacks when #if/#else is seen."""
2298
2299 def __init__(self, stack_before_if):
2300 # The entire nesting stack before #if
2301 self.stack_before_if = stack_before_if
2302
2303 # The entire nesting stack up to #else
2304 self.stack_before_else = []
2305
2306 # Whether we have already seen #else or #elif
2307 self.seen_else = False
2308
2309
avakulenko@google.com02af6282014-06-04 18:53:25 +00002310class NestingState(object):
erg@google.comd350fe52013-01-14 17:51:48 +00002311 """Holds states related to parsing braces."""
erg@google.com4e00b9a2009-01-12 23:05:11 +00002312
2313 def __init__(self):
erg@google.comd350fe52013-01-14 17:51:48 +00002314 # Stack for tracking all braces. An object is pushed whenever we
2315 # see a "{", and popped when we see a "}". Only 3 types of
2316 # objects are possible:
2317 # - _ClassInfo: a class or struct.
2318 # - _NamespaceInfo: a namespace.
2319 # - _BlockInfo: some other type of block.
2320 self.stack = []
erg@google.com4e00b9a2009-01-12 23:05:11 +00002321
avakulenko@google.com02af6282014-06-04 18:53:25 +00002322 # Top of the previous stack before each Update().
2323 #
2324 # Because the nesting_stack is updated at the end of each line, we
2325 # had to do some convoluted checks to find out what is the current
2326 # scope at the beginning of the line. This check is simplified by
2327 # saving the previous top of nesting stack.
2328 #
2329 # We could save the full stack, but we only need the top. Copying
2330 # the full nesting stack would slow down cpplint by ~10%.
2331 self.previous_stack_top = []
2332
erg@google.comd350fe52013-01-14 17:51:48 +00002333 # Stack of _PreprocessorInfo objects.
2334 self.pp_stack = []
2335
2336 def SeenOpenBrace(self):
2337 """Check if we have seen the opening brace for the innermost block.
2338
2339 Returns:
2340 True if we have seen the opening brace, False if the innermost
2341 block is still expecting an opening brace.
2342 """
2343 return (not self.stack) or self.stack[-1].seen_open_brace
2344
2345 def InNamespaceBody(self):
2346 """Check if we are currently one level inside a namespace body.
2347
2348 Returns:
2349 True if top of the stack is a namespace block, False otherwise.
2350 """
2351 return self.stack and isinstance(self.stack[-1], _NamespaceInfo)
2352
avakulenko@google.com02af6282014-06-04 18:53:25 +00002353 def InExternC(self):
2354 """Check if we are currently one level inside an 'extern "C"' block.
2355
2356 Returns:
2357 True if top of the stack is an extern block, False otherwise.
2358 """
2359 return self.stack and isinstance(self.stack[-1], _ExternCInfo)
2360
2361 def InClassDeclaration(self):
2362 """Check if we are currently one level inside a class or struct declaration.
2363
2364 Returns:
2365 True if top of the stack is a class/struct, False otherwise.
2366 """
2367 return self.stack and isinstance(self.stack[-1], _ClassInfo)
2368
2369 def InAsmBlock(self):
2370 """Check if we are currently one level inside an inline ASM block.
2371
2372 Returns:
2373 True if the top of the stack is a block containing inline ASM.
2374 """
2375 return self.stack and self.stack[-1].inline_asm != _NO_ASM
2376
2377 def InTemplateArgumentList(self, clean_lines, linenum, pos):
2378 """Check if current position is inside template argument list.
2379
2380 Args:
2381 clean_lines: A CleansedLines instance containing the file.
2382 linenum: The number of the line to check.
2383 pos: position just after the suspected template argument.
2384 Returns:
2385 True if (linenum, pos) is inside template arguments.
2386 """
2387 while linenum < clean_lines.NumLines():
2388 # Find the earliest character that might indicate a template argument
2389 line = clean_lines.elided[linenum]
2390 match = Match(r'^[^{};=\[\]\.<>]*(.)', line[pos:])
2391 if not match:
2392 linenum += 1
2393 pos = 0
2394 continue
2395 token = match.group(1)
2396 pos += len(match.group(0))
2397
2398 # These things do not look like template argument list:
2399 # class Suspect {
2400 # class Suspect x; }
2401 if token in ('{', '}', ';'): return False
2402
2403 # These things look like template argument list:
2404 # template <class Suspect>
2405 # template <class Suspect = default_value>
2406 # template <class Suspect[]>
2407 # template <class Suspect...>
2408 if token in ('>', '=', '[', ']', '.'): return True
2409
2410 # Check if token is an unmatched '<'.
2411 # If not, move on to the next character.
2412 if token != '<':
2413 pos += 1
2414 if pos >= len(line):
2415 linenum += 1
2416 pos = 0
2417 continue
2418
2419 # We can't be sure if we just find a single '<', and need to
2420 # find the matching '>'.
2421 (_, end_line, end_pos) = CloseExpression(clean_lines, linenum, pos - 1)
2422 if end_pos < 0:
2423 # Not sure if template argument list or syntax error in file
2424 return False
2425 linenum = end_line
2426 pos = end_pos
2427 return False
2428
erg@google.comd350fe52013-01-14 17:51:48 +00002429 def UpdatePreprocessor(self, line):
2430 """Update preprocessor stack.
2431
2432 We need to handle preprocessors due to classes like this:
2433 #ifdef SWIG
2434 struct ResultDetailsPageElementExtensionPoint {
2435 #else
2436 struct ResultDetailsPageElementExtensionPoint : public Extension {
2437 #endif
erg@google.comd350fe52013-01-14 17:51:48 +00002438
2439 We make the following assumptions (good enough for most files):
2440 - Preprocessor condition evaluates to true from #if up to first
2441 #else/#elif/#endif.
2442
2443 - Preprocessor condition evaluates to false from #else/#elif up
2444 to #endif. We still perform lint checks on these lines, but
2445 these do not affect nesting stack.
2446
2447 Args:
2448 line: current line to check.
2449 """
2450 if Match(r'^\s*#\s*(if|ifdef|ifndef)\b', line):
2451 # Beginning of #if block, save the nesting stack here. The saved
2452 # stack will allow us to restore the parsing state in the #else case.
2453 self.pp_stack.append(_PreprocessorInfo(copy.deepcopy(self.stack)))
2454 elif Match(r'^\s*#\s*(else|elif)\b', line):
2455 # Beginning of #else block
2456 if self.pp_stack:
2457 if not self.pp_stack[-1].seen_else:
2458 # This is the first #else or #elif block. Remember the
2459 # whole nesting stack up to this point. This is what we
2460 # keep after the #endif.
2461 self.pp_stack[-1].seen_else = True
2462 self.pp_stack[-1].stack_before_else = copy.deepcopy(self.stack)
2463
2464 # Restore the stack to how it was before the #if
2465 self.stack = copy.deepcopy(self.pp_stack[-1].stack_before_if)
2466 else:
2467 # TODO(unknown): unexpected #else, issue warning?
2468 pass
2469 elif Match(r'^\s*#\s*endif\b', line):
2470 # End of #if or #else blocks.
2471 if self.pp_stack:
2472 # If we saw an #else, we will need to restore the nesting
2473 # stack to its former state before the #else, otherwise we
2474 # will just continue from where we left off.
2475 if self.pp_stack[-1].seen_else:
2476 # Here we can just use a shallow copy since we are the last
2477 # reference to it.
2478 self.stack = self.pp_stack[-1].stack_before_else
2479 # Drop the corresponding #if
2480 self.pp_stack.pop()
2481 else:
2482 # TODO(unknown): unexpected #endif, issue warning?
2483 pass
2484
avakulenko@google.com02af6282014-06-04 18:53:25 +00002485 # TODO(unknown): Update() is too long, but we will refactor later.
erg@google.comd350fe52013-01-14 17:51:48 +00002486 def Update(self, filename, clean_lines, linenum, error):
2487 """Update nesting state with current line.
2488
2489 Args:
2490 filename: The name of the current file.
2491 clean_lines: A CleansedLines instance containing the file.
2492 linenum: The number of the line to check.
2493 error: The function to call with any errors found.
2494 """
2495 line = clean_lines.elided[linenum]
2496
avakulenko@google.com02af6282014-06-04 18:53:25 +00002497 # Remember top of the previous nesting stack.
2498 #
2499 # The stack is always pushed/popped and not modified in place, so
2500 # we can just do a shallow copy instead of copy.deepcopy. Using
2501 # deepcopy would slow down cpplint by ~28%.
2502 if self.stack:
2503 self.previous_stack_top = self.stack[-1]
2504 else:
2505 self.previous_stack_top = None
2506
2507 # Update pp_stack
erg@google.comd350fe52013-01-14 17:51:48 +00002508 self.UpdatePreprocessor(line)
2509
2510 # Count parentheses. This is to avoid adding struct arguments to
2511 # the nesting stack.
2512 if self.stack:
2513 inner_block = self.stack[-1]
2514 depth_change = line.count('(') - line.count(')')
2515 inner_block.open_parentheses += depth_change
2516
2517 # Also check if we are starting or ending an inline assembly block.
2518 if inner_block.inline_asm in (_NO_ASM, _END_ASM):
2519 if (depth_change != 0 and
2520 inner_block.open_parentheses == 1 and
2521 _MATCH_ASM.match(line)):
2522 # Enter assembly block
2523 inner_block.inline_asm = _INSIDE_ASM
2524 else:
2525 # Not entering assembly block. If previous line was _END_ASM,
2526 # we will now shift to _NO_ASM state.
2527 inner_block.inline_asm = _NO_ASM
2528 elif (inner_block.inline_asm == _INSIDE_ASM and
2529 inner_block.open_parentheses == 0):
2530 # Exit assembly block
2531 inner_block.inline_asm = _END_ASM
2532
2533 # Consume namespace declaration at the beginning of the line. Do
2534 # this in a loop so that we catch same line declarations like this:
2535 # namespace proto2 { namespace bridge { class MessageSet; } }
2536 while True:
2537 # Match start of namespace. The "\b\s*" below catches namespace
2538 # declarations even if it weren't followed by a whitespace, this
2539 # is so that we don't confuse our namespace checker. The
2540 # missing spaces will be flagged by CheckSpacing.
2541 namespace_decl_match = Match(r'^\s*namespace\b\s*([:\w]+)?(.*)$', line)
2542 if not namespace_decl_match:
2543 break
2544
2545 new_namespace = _NamespaceInfo(namespace_decl_match.group(1), linenum)
2546 self.stack.append(new_namespace)
2547
2548 line = namespace_decl_match.group(2)
2549 if line.find('{') != -1:
2550 new_namespace.seen_open_brace = True
2551 line = line[line.find('{') + 1:]
2552
2553 # Look for a class declaration in whatever is left of the line
2554 # after parsing namespaces. The regexp accounts for decorated classes
2555 # such as in:
2556 # class LOCKABLE API Object {
2557 # };
erg@google.comd350fe52013-01-14 17:51:48 +00002558 class_decl_match = Match(
avakulenko@google.com02af6282014-06-04 18:53:25 +00002559 r'^(\s*(?:template\s*<[\w\s<>,:]*>\s*)?'
2560 r'(class|struct)\s+(?:[A-Z_]+\s+)*(\w+(?:::\w+)*))'
2561 r'(.*)$', line)
erg@google.comd350fe52013-01-14 17:51:48 +00002562 if (class_decl_match and
2563 (not self.stack or self.stack[-1].open_parentheses == 0)):
avakulenko@google.com02af6282014-06-04 18:53:25 +00002564 # We do not want to accept classes that are actually template arguments:
2565 # template <class Ignore1,
2566 # class Ignore2 = Default<Args>,
2567 # template <Args> class Ignore3>
2568 # void Function() {};
2569 #
2570 # To avoid template argument cases, we scan forward and look for
2571 # an unmatched '>'. If we see one, assume we are inside a
2572 # template argument list.
2573 end_declaration = len(class_decl_match.group(1))
2574 if not self.InTemplateArgumentList(clean_lines, linenum, end_declaration):
2575 self.stack.append(_ClassInfo(
2576 class_decl_match.group(3), class_decl_match.group(2),
2577 clean_lines, linenum))
2578 line = class_decl_match.group(4)
erg@google.comd350fe52013-01-14 17:51:48 +00002579
2580 # If we have not yet seen the opening brace for the innermost block,
2581 # run checks here.
2582 if not self.SeenOpenBrace():
2583 self.stack[-1].CheckBegin(filename, clean_lines, linenum, error)
2584
2585 # Update access control if we are inside a class/struct
2586 if self.stack and isinstance(self.stack[-1], _ClassInfo):
erg@google.comfd5da632013-10-25 17:39:45 +00002587 classinfo = self.stack[-1]
2588 access_match = Match(
2589 r'^(.*)\b(public|private|protected|signals)(\s+(?:slots\s*)?)?'
2590 r':(?:[^:]|$)',
2591 line)
erg@google.comd350fe52013-01-14 17:51:48 +00002592 if access_match:
erg@google.comfd5da632013-10-25 17:39:45 +00002593 classinfo.access = access_match.group(2)
2594
2595 # Check that access keywords are indented +1 space. Skip this
erg@google.comc6671232013-10-25 21:44:03 +00002596 # check if the keywords are not preceded by whitespaces.
erg@google.comfd5da632013-10-25 17:39:45 +00002597 indent = access_match.group(1)
2598 if (len(indent) != classinfo.class_indent + 1 and
2599 Match(r'^\s*$', indent)):
2600 if classinfo.is_struct:
2601 parent = 'struct ' + classinfo.name
2602 else:
2603 parent = 'class ' + classinfo.name
2604 slots = ''
2605 if access_match.group(3):
2606 slots = access_match.group(3)
2607 error(filename, linenum, 'whitespace/indent', 3,
2608 '%s%s: should be indented +1 space inside %s' % (
2609 access_match.group(2), slots, parent))
erg@google.comd350fe52013-01-14 17:51:48 +00002610
2611 # Consume braces or semicolons from what's left of the line
2612 while True:
2613 # Match first brace, semicolon, or closed parenthesis.
2614 matched = Match(r'^[^{;)}]*([{;)}])(.*)$', line)
2615 if not matched:
2616 break
2617
2618 token = matched.group(1)
2619 if token == '{':
2620 # If namespace or class hasn't seen a opening brace yet, mark
2621 # namespace/class head as complete. Push a new block onto the
2622 # stack otherwise.
2623 if not self.SeenOpenBrace():
2624 self.stack[-1].seen_open_brace = True
avakulenko@google.com02af6282014-06-04 18:53:25 +00002625 elif Match(r'^extern\s*"[^"]*"\s*\{', line):
Alex Vakulenko01e47232016-05-06 13:54:15 -07002626 self.stack.append(_ExternCInfo(linenum))
erg@google.comd350fe52013-01-14 17:51:48 +00002627 else:
Alex Vakulenko01e47232016-05-06 13:54:15 -07002628 self.stack.append(_BlockInfo(linenum, True))
erg@google.comd350fe52013-01-14 17:51:48 +00002629 if _MATCH_ASM.match(line):
2630 self.stack[-1].inline_asm = _BLOCK_ASM
avakulenko@google.com02af6282014-06-04 18:53:25 +00002631
erg@google.comd350fe52013-01-14 17:51:48 +00002632 elif token == ';' or token == ')':
2633 # If we haven't seen an opening brace yet, but we already saw
2634 # a semicolon, this is probably a forward declaration. Pop
2635 # the stack for these.
2636 #
2637 # Similarly, if we haven't seen an opening brace yet, but we
2638 # already saw a closing parenthesis, then these are probably
2639 # function arguments with extra "class" or "struct" keywords.
2640 # Also pop these stack for these.
2641 if not self.SeenOpenBrace():
2642 self.stack.pop()
2643 else: # token == '}'
2644 # Perform end of block checks and pop the stack.
2645 if self.stack:
2646 self.stack[-1].CheckEnd(filename, clean_lines, linenum, error)
2647 self.stack.pop()
2648 line = matched.group(2)
2649
2650 def InnermostClass(self):
2651 """Get class info on the top of the stack.
2652
2653 Returns:
2654 A _ClassInfo object if we are inside a class, or None otherwise.
2655 """
2656 for i in range(len(self.stack), 0, -1):
2657 classinfo = self.stack[i - 1]
2658 if isinstance(classinfo, _ClassInfo):
2659 return classinfo
2660 return None
2661
erg@google.com2aa59982013-10-28 19:09:25 +00002662 def CheckCompletedBlocks(self, filename, error):
2663 """Checks that all classes and namespaces have been completely parsed.
erg@google.com4e00b9a2009-01-12 23:05:11 +00002664
2665 Call this when all lines in a file have been processed.
2666 Args:
2667 filename: The name of the current file.
2668 error: The function to call with any errors found.
2669 """
erg@google.comd350fe52013-01-14 17:51:48 +00002670 # Note: This test can result in false positives if #ifdef constructs
2671 # get in the way of brace matching. See the testBuildClass test in
2672 # cpplint_unittest.py for an example of this.
2673 for obj in self.stack:
2674 if isinstance(obj, _ClassInfo):
2675 error(filename, obj.starting_linenum, 'build/class', 5,
2676 'Failed to find complete declaration of class %s' %
2677 obj.name)
erg@google.com2aa59982013-10-28 19:09:25 +00002678 elif isinstance(obj, _NamespaceInfo):
2679 error(filename, obj.starting_linenum, 'build/namespaces', 5,
2680 'Failed to find complete declaration of namespace %s' %
2681 obj.name)
erg@google.com4e00b9a2009-01-12 23:05:11 +00002682
2683
2684def CheckForNonStandardConstructs(filename, clean_lines, linenum,
erg@google.comd350fe52013-01-14 17:51:48 +00002685 nesting_state, error):
erg@google.com2aa59982013-10-28 19:09:25 +00002686 r"""Logs an error if we see certain non-ANSI constructs ignored by gcc-2.
erg@google.com4e00b9a2009-01-12 23:05:11 +00002687
2688 Complain about several constructs which gcc-2 accepts, but which are
2689 not standard C++. Warning about these in lint is one way to ease the
2690 transition to new compilers.
2691 - put storage class first (e.g. "static const" instead of "const static").
2692 - "%lld" instead of %qd" in printf-type functions.
2693 - "%1$d" is non-standard in printf-type functions.
2694 - "\%" is an undefined character escape sequence.
2695 - text after #endif is not allowed.
2696 - invalid inner-style forward declaration.
2697 - >? and <? operators, and their >?= and <?= cousins.
erg@google.com4e00b9a2009-01-12 23:05:11 +00002698
erg@google.coma868d2d2009-10-09 21:18:45 +00002699 Additionally, check for constructor/destructor style violations and reference
2700 members, as it is very convenient to do so while checking for
2701 gcc-2 compliance.
erg@google.com4e00b9a2009-01-12 23:05:11 +00002702
2703 Args:
2704 filename: The name of the current file.
2705 clean_lines: A CleansedLines instance containing the file.
2706 linenum: The number of the line to check.
avakulenko@google.com02af6282014-06-04 18:53:25 +00002707 nesting_state: A NestingState instance which maintains information about
erg@google.comd350fe52013-01-14 17:51:48 +00002708 the current stack of nested blocks being parsed.
erg@google.com4e00b9a2009-01-12 23:05:11 +00002709 error: A callable to which errors are reported, which takes 4 arguments:
2710 filename, line number, error level, and message
2711 """
2712
2713 # Remove comments from the line, but leave in strings for now.
2714 line = clean_lines.lines[linenum]
2715
2716 if Search(r'printf\s*\(.*".*%[-+ ]?\d*q', line):
2717 error(filename, linenum, 'runtime/printf_format', 3,
2718 '%q in format strings is deprecated. Use %ll instead.')
2719
2720 if Search(r'printf\s*\(.*".*%\d+\$', line):
2721 error(filename, linenum, 'runtime/printf_format', 2,
2722 '%N$ formats are unconventional. Try rewriting to avoid them.')
2723
2724 # Remove escaped backslashes before looking for undefined escapes.
2725 line = line.replace('\\\\', '')
2726
2727 if Search(r'("|\').*\\(%|\[|\(|{)', line):
2728 error(filename, linenum, 'build/printf_format', 3,
2729 '%, [, (, and { are undefined character escapes. Unescape them.')
2730
2731 # For the rest, work with both comments and strings removed.
2732 line = clean_lines.elided[linenum]
2733
2734 if Search(r'\b(const|volatile|void|char|short|int|long'
2735 r'|float|double|signed|unsigned'
2736 r'|schar|u?int8|u?int16|u?int32|u?int64)'
erg@google.comd350fe52013-01-14 17:51:48 +00002737 r'\s+(register|static|extern|typedef)\b',
erg@google.com4e00b9a2009-01-12 23:05:11 +00002738 line):
2739 error(filename, linenum, 'build/storage_class', 5,
Alex Vakulenko01e47232016-05-06 13:54:15 -07002740 'Storage-class specifier (static, extern, typedef, etc) should be '
2741 'at the beginning of the declaration.')
erg@google.com4e00b9a2009-01-12 23:05:11 +00002742
Elliot Glaysherae118112016-09-30 15:34:26 -07002743 if Match(r'\s*#\s*endif\s*[^/\s]+', line):
2744 error(filename, linenum, 'build/endif_comment', 5,
2745 'Uncommented text after #endif is non-standard. Use a comment.')
2746
erg@google.com4e00b9a2009-01-12 23:05:11 +00002747 if Match(r'\s*class\s+(\w+\s*::\s*)+\w+\s*;', line):
2748 error(filename, linenum, 'build/forward_decl', 5,
2749 'Inner-style forward declarations are invalid. Remove this line.')
2750
2751 if Search(r'(\w+|[+-]?\d+(\.\d*)?)\s*(<|>)\?=?\s*(\w+|[+-]?\d+)(\.\d*)?',
2752 line):
2753 error(filename, linenum, 'build/deprecated', 3,
2754 '>? and <? (max and min) operators are non-standard and deprecated.')
2755
erg@google.coma868d2d2009-10-09 21:18:45 +00002756 if Search(r'^\s*const\s*string\s*&\s*\w+\s*;', line):
2757 # TODO(unknown): Could it be expanded safely to arbitrary references,
2758 # without triggering too many false positives? The first
2759 # attempt triggered 5 warnings for mostly benign code in the regtest, hence
2760 # the restriction.
2761 # Here's the original regexp, for the reference:
2762 # type_name = r'\w+((\s*::\s*\w+)|(\s*<\s*\w+?\s*>))?'
2763 # r'\s*const\s*' + type_name + '\s*&\s*\w+\s*;'
2764 error(filename, linenum, 'runtime/member_string_references', 2,
2765 'const string& members are dangerous. It is much better to use '
2766 'alternatives, such as pointers or simple constants.')
2767
erg@google.comd350fe52013-01-14 17:51:48 +00002768 # Everything else in this function operates on class declarations.
2769 # Return early if the top of the nesting stack is not a class, or if
2770 # the class head is not completed yet.
2771 classinfo = nesting_state.InnermostClass()
2772 if not classinfo or not classinfo.seen_open_brace:
erg@google.com4e00b9a2009-01-12 23:05:11 +00002773 return
2774
erg@google.com4e00b9a2009-01-12 23:05:11 +00002775 # The class may have been declared with namespace or classname qualifiers.
2776 # The constructor and destructor will not have those qualifiers.
2777 base_classname = classinfo.name.split('::')[-1]
2778
2779 # Look for single-argument constructors that aren't marked explicit.
Alex Vakulenko01e47232016-05-06 13:54:15 -07002780 # Technically a valid construct, but against style.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00002781 explicit_constructor_match = Match(
Dana Jansenscf4071c2017-02-22 12:02:39 -05002782 r'\s+(?:(?:inline|constexpr)\s+)*(explicit\s+)?'
2783 r'(?:(?:inline|constexpr)\s+)*%s\s*'
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00002784 r'\(((?:[^()]|\([^()]*\))*)\)'
2785 % re.escape(base_classname),
2786 line)
2787
2788 if explicit_constructor_match:
2789 is_marked_explicit = explicit_constructor_match.group(1)
2790
2791 if not explicit_constructor_match.group(2):
2792 constructor_args = []
2793 else:
2794 constructor_args = explicit_constructor_match.group(2).split(',')
2795
2796 # collapse arguments so that commas in template parameter lists and function
2797 # argument parameter lists don't split arguments in two
2798 i = 0
2799 while i < len(constructor_args):
2800 constructor_arg = constructor_args[i]
2801 while (constructor_arg.count('<') > constructor_arg.count('>') or
2802 constructor_arg.count('(') > constructor_arg.count(')')):
2803 constructor_arg += ',' + constructor_args[i + 1]
2804 del constructor_args[i + 1]
2805 constructor_args[i] = constructor_arg
2806 i += 1
2807
2808 defaulted_args = [arg for arg in constructor_args if '=' in arg]
2809 noarg_constructor = (not constructor_args or # empty arg list
2810 # 'void' arg specifier
2811 (len(constructor_args) == 1 and
2812 constructor_args[0].strip() == 'void'))
2813 onearg_constructor = ((len(constructor_args) == 1 and # exactly one arg
2814 not noarg_constructor) or
2815 # all but at most one arg defaulted
2816 (len(constructor_args) >= 1 and
2817 not noarg_constructor and
2818 len(defaulted_args) >= len(constructor_args) - 1))
2819 initializer_list_constructor = bool(
2820 onearg_constructor and
2821 Search(r'\bstd\s*::\s*initializer_list\b', constructor_args[0]))
2822 copy_constructor = bool(
2823 onearg_constructor and
2824 Match(r'(const\s+)?%s(\s*<[^>]*>)?(\s+const)?\s*(?:<\w+>\s*)?&'
2825 % re.escape(base_classname), constructor_args[0].strip()))
2826
2827 if (not is_marked_explicit and
2828 onearg_constructor and
2829 not initializer_list_constructor and
2830 not copy_constructor):
2831 if defaulted_args:
2832 error(filename, linenum, 'runtime/explicit', 5,
2833 'Constructors callable with one argument '
2834 'should be marked explicit.')
2835 else:
2836 error(filename, linenum, 'runtime/explicit', 5,
2837 'Single-parameter constructors should be marked explicit.')
2838 elif is_marked_explicit and not onearg_constructor:
2839 if noarg_constructor:
2840 error(filename, linenum, 'runtime/explicit', 5,
2841 'Zero-parameter constructors should not be marked explicit.')
erg@google.com4e00b9a2009-01-12 23:05:11 +00002842
erg@google.com4e00b9a2009-01-12 23:05:11 +00002843
avakulenko@google.com02af6282014-06-04 18:53:25 +00002844def CheckSpacingForFunctionCall(filename, clean_lines, linenum, error):
erg@google.com4e00b9a2009-01-12 23:05:11 +00002845 """Checks for the correctness of various spacing around function calls.
2846
2847 Args:
2848 filename: The name of the current file.
avakulenko@google.com02af6282014-06-04 18:53:25 +00002849 clean_lines: A CleansedLines instance containing the file.
erg@google.com4e00b9a2009-01-12 23:05:11 +00002850 linenum: The number of the line to check.
2851 error: The function to call with any errors found.
2852 """
avakulenko@google.com02af6282014-06-04 18:53:25 +00002853 line = clean_lines.elided[linenum]
erg@google.com4e00b9a2009-01-12 23:05:11 +00002854
2855 # Since function calls often occur inside if/for/while/switch
2856 # expressions - which have their own, more liberal conventions - we
2857 # first see if we should be looking inside such an expression for a
2858 # function call, to which we can apply more strict standards.
2859 fncall = line # if there's no control flow construct, look at whole line
2860 for pattern in (r'\bif\s*\((.*)\)\s*{',
2861 r'\bfor\s*\((.*)\)\s*{',
2862 r'\bwhile\s*\((.*)\)\s*[{;]',
2863 r'\bswitch\s*\((.*)\)\s*{'):
2864 match = Search(pattern, line)
2865 if match:
2866 fncall = match.group(1) # look inside the parens for function calls
2867 break
2868
2869 # Except in if/for/while/switch, there should never be space
2870 # immediately inside parens (eg "f( 3, 4 )"). We make an exception
2871 # for nested parens ( (a+b) + c ). Likewise, there should never be
2872 # a space before a ( when it's a function argument. I assume it's a
2873 # function argument when the char before the whitespace is legal in
2874 # a function name (alnum + _) and we're not starting a macro. Also ignore
2875 # pointers and references to arrays and functions coz they're too tricky:
2876 # we use a very simple way to recognize these:
2877 # " (something)(maybe-something)" or
2878 # " (something)(maybe-something," or
2879 # " (something)[something]"
2880 # Note that we assume the contents of [] to be short enough that
2881 # they'll never need to wrap.
2882 if ( # Ignore control structures.
erg@google.com2aa59982013-10-28 19:09:25 +00002883 not Search(r'\b(if|for|while|switch|return|new|delete|catch|sizeof)\b',
erg@google.comc6671232013-10-25 21:44:03 +00002884 fncall) and
erg@google.com4e00b9a2009-01-12 23:05:11 +00002885 # Ignore pointers/references to functions.
2886 not Search(r' \([^)]+\)\([^)]*(\)|,$)', fncall) and
2887 # Ignore pointers/references to arrays.
2888 not Search(r' \([^)]+\)\[[^\]]+\]', fncall)):
erg@google.com36649102009-03-25 21:18:36 +00002889 if Search(r'\w\s*\(\s(?!\s*\\$)', fncall): # a ( used for a fn call
erg@google.com4e00b9a2009-01-12 23:05:11 +00002890 error(filename, linenum, 'whitespace/parens', 4,
2891 'Extra space after ( in function call')
erg@google.com36649102009-03-25 21:18:36 +00002892 elif Search(r'\(\s+(?!(\s*\\)|\()', fncall):
erg@google.com4e00b9a2009-01-12 23:05:11 +00002893 error(filename, linenum, 'whitespace/parens', 2,
2894 'Extra space after (')
2895 if (Search(r'\w\s+\(', fncall) and
Alex Vakulenko01e47232016-05-06 13:54:15 -07002896 not Search(r'_{0,2}asm_{0,2}\s+_{0,2}volatile_{0,2}\s+\(', fncall) and
avakulenko@google.com02af6282014-06-04 18:53:25 +00002897 not Search(r'#\s*define|typedef|using\s+\w+\s*=', fncall) and
avakulenko@google.com554223d2014-12-04 22:00:20 +00002898 not Search(r'\w\s+\((\w+::)*\*\w+\)\(', fncall) and
2899 not Search(r'\bcase\s+\(', fncall)):
avakulenko@google.com02af6282014-06-04 18:53:25 +00002900 # TODO(unknown): Space after an operator function seem to be a common
2901 # error, silence those for now by restricting them to highest verbosity.
2902 if Search(r'\boperator_*\b', line):
2903 error(filename, linenum, 'whitespace/parens', 0,
2904 'Extra space before ( in function call')
2905 else:
2906 error(filename, linenum, 'whitespace/parens', 4,
2907 'Extra space before ( in function call')
erg@google.com4e00b9a2009-01-12 23:05:11 +00002908 # If the ) is followed only by a newline or a { + newline, assume it's
2909 # part of a control statement (if/while/etc), and don't complain
2910 if Search(r'[^)]\s+\)\s*[^{\s]', fncall):
erg@google.com8a95ecc2011-09-08 00:45:54 +00002911 # If the closing parenthesis is preceded by only whitespaces,
2912 # try to give a more descriptive error message.
2913 if Search(r'^\s+\)', fncall):
2914 error(filename, linenum, 'whitespace/parens', 2,
2915 'Closing ) should be moved to the previous line')
2916 else:
2917 error(filename, linenum, 'whitespace/parens', 2,
2918 'Extra space before )')
erg@google.com4e00b9a2009-01-12 23:05:11 +00002919
2920
2921def IsBlankLine(line):
2922 """Returns true if the given line is blank.
2923
2924 We consider a line to be blank if the line is empty or consists of
2925 only white spaces.
2926
2927 Args:
2928 line: A line of a string.
2929
2930 Returns:
2931 True, if the given line is blank.
2932 """
2933 return not line or line.isspace()
2934
2935
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00002936def CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line,
2937 error):
2938 is_namespace_indent_item = (
2939 len(nesting_state.stack) > 1 and
2940 nesting_state.stack[-1].check_namespace_indentation and
2941 isinstance(nesting_state.previous_stack_top, _NamespaceInfo) and
2942 nesting_state.previous_stack_top == nesting_state.stack[-2])
2943
2944 if ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item,
2945 clean_lines.elided, line):
2946 CheckItemIndentationInNamespace(filename, clean_lines.elided,
2947 line, error)
2948
2949
erg@google.com4e00b9a2009-01-12 23:05:11 +00002950def CheckForFunctionLengths(filename, clean_lines, linenum,
2951 function_state, error):
2952 """Reports for long function bodies.
2953
2954 For an overview why this is done, see:
Ackermann Yuriy79692902016-04-01 21:41:34 +13002955 https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions
erg@google.com4e00b9a2009-01-12 23:05:11 +00002956
2957 Uses a simplistic algorithm assuming other style guidelines
2958 (especially spacing) are followed.
2959 Only checks unindented functions, so class members are unchecked.
2960 Trivial bodies are unchecked, so constructors with huge initializer lists
2961 may be missed.
2962 Blank/comment lines are not counted so as to avoid encouraging the removal
erg@google.com8a95ecc2011-09-08 00:45:54 +00002963 of vertical space and comments just to get through a lint check.
erg@google.com4e00b9a2009-01-12 23:05:11 +00002964 NOLINT *on the last line of a function* disables this check.
2965
2966 Args:
2967 filename: The name of the current file.
2968 clean_lines: A CleansedLines instance containing the file.
2969 linenum: The number of the line to check.
2970 function_state: Current function name and lines in body so far.
2971 error: The function to call with any errors found.
2972 """
2973 lines = clean_lines.lines
2974 line = lines[linenum]
erg@google.com4e00b9a2009-01-12 23:05:11 +00002975 joined_line = ''
2976
2977 starting_func = False
erg@google.coma87abb82009-02-24 01:41:01 +00002978 regexp = r'(\w(\w|::|\*|\&|\s)*)\(' # decls * & space::name( ...
erg@google.com4e00b9a2009-01-12 23:05:11 +00002979 match_result = Match(regexp, line)
2980 if match_result:
2981 # If the name is all caps and underscores, figure it's a macro and
2982 # ignore it, unless it's TEST or TEST_F.
2983 function_name = match_result.group(1).split()[-1]
2984 if function_name == 'TEST' or function_name == 'TEST_F' or (
2985 not Match(r'[A-Z_]+$', function_name)):
2986 starting_func = True
2987
2988 if starting_func:
2989 body_found = False
erg@google.coma87abb82009-02-24 01:41:01 +00002990 for start_linenum in xrange(linenum, clean_lines.NumLines()):
erg@google.com4e00b9a2009-01-12 23:05:11 +00002991 start_line = lines[start_linenum]
2992 joined_line += ' ' + start_line.lstrip()
2993 if Search(r'(;|})', start_line): # Declarations and trivial functions
2994 body_found = True
2995 break # ... ignore
2996 elif Search(r'{', start_line):
2997 body_found = True
2998 function = Search(r'((\w|:)*)\(', line).group(1)
2999 if Match(r'TEST', function): # Handle TEST... macros
3000 parameter_regexp = Search(r'(\(.*\))', joined_line)
3001 if parameter_regexp: # Ignore bad syntax
3002 function += parameter_regexp.group(1)
3003 else:
3004 function += '()'
3005 function_state.Begin(function)
3006 break
3007 if not body_found:
erg@google.coma87abb82009-02-24 01:41:01 +00003008 # No body for the function (or evidence of a non-function) was found.
erg@google.com4e00b9a2009-01-12 23:05:11 +00003009 error(filename, linenum, 'readability/fn_size', 5,
3010 'Lint failed to find start of function body.')
3011 elif Match(r'^\}\s*$', line): # function end
erg+personal@google.com05189642010-04-30 20:43:03 +00003012 function_state.Check(error, filename, linenum)
erg@google.com4e00b9a2009-01-12 23:05:11 +00003013 function_state.End()
3014 elif not Match(r'^\s*$', line):
3015 function_state.Count() # Count non-blank/non-comment lines.
3016
3017
3018_RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?')
3019
3020
avakulenko@google.com02af6282014-06-04 18:53:25 +00003021def CheckComment(line, filename, linenum, next_line_start, error):
3022 """Checks for common mistakes in comments.
erg@google.com4e00b9a2009-01-12 23:05:11 +00003023
3024 Args:
avakulenko@google.com02af6282014-06-04 18:53:25 +00003025 line: The line in question.
erg@google.com4e00b9a2009-01-12 23:05:11 +00003026 filename: The name of the current file.
3027 linenum: The number of the line to check.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003028 next_line_start: The first non-whitespace column of the next line.
erg@google.com4e00b9a2009-01-12 23:05:11 +00003029 error: The function to call with any errors found.
3030 """
avakulenko@google.com02af6282014-06-04 18:53:25 +00003031 commentpos = line.find('//')
3032 if commentpos != -1:
3033 # Check if the // may be in quotes. If so, ignore it
Alex Vakulenko01e47232016-05-06 13:54:15 -07003034 if re.sub(r'\\.', '', line[0:commentpos]).count('"') % 2 == 0:
avakulenko@google.com02af6282014-06-04 18:53:25 +00003035 # Allow one space for new scopes, two spaces otherwise:
3036 if (not (Match(r'^.*{ *//', line) and next_line_start == commentpos) and
3037 ((commentpos >= 1 and
3038 line[commentpos-1] not in string.whitespace) or
3039 (commentpos >= 2 and
3040 line[commentpos-2] not in string.whitespace))):
3041 error(filename, linenum, 'whitespace/comments', 2,
3042 'At least two spaces is best between code and comments')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003043
avakulenko@google.com02af6282014-06-04 18:53:25 +00003044 # Checks for common mistakes in TODO comments.
3045 comment = line[commentpos:]
3046 match = _RE_PATTERN_TODO.match(comment)
3047 if match:
3048 # One whitespace is correct; zero whitespace is handled elsewhere.
3049 leading_whitespace = match.group(1)
3050 if len(leading_whitespace) > 1:
3051 error(filename, linenum, 'whitespace/todo', 2,
3052 'Too many spaces before TODO')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003053
avakulenko@google.com02af6282014-06-04 18:53:25 +00003054 username = match.group(2)
3055 if not username:
3056 error(filename, linenum, 'readability/todo', 2,
3057 'Missing username in TODO; it should look like '
3058 '"// TODO(my_username): Stuff."')
3059
3060 middle_whitespace = match.group(3)
3061 # Comparisons made explicit for correctness -- pylint: disable=g-explicit-bool-comparison
3062 if middle_whitespace != ' ' and middle_whitespace != '':
3063 error(filename, linenum, 'whitespace/todo', 2,
3064 'TODO(my_username) should be followed by a space')
3065
3066 # If the comment contains an alphanumeric character, there
avakulenko@google.com554223d2014-12-04 22:00:20 +00003067 # should be a space somewhere between it and the // unless
3068 # it's a /// or //! Doxygen comment.
3069 if (Match(r'//[^ ]*\w', comment) and
3070 not Match(r'(///|//\!)(\s+|$)', comment)):
avakulenko@google.com02af6282014-06-04 18:53:25 +00003071 error(filename, linenum, 'whitespace/comments', 4,
3072 'Should have a space between // and comment')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003073
avakulenko@google.com554223d2014-12-04 22:00:20 +00003074
erg@google.comd350fe52013-01-14 17:51:48 +00003075def CheckSpacing(filename, clean_lines, linenum, nesting_state, error):
erg@google.com4e00b9a2009-01-12 23:05:11 +00003076 """Checks for the correctness of various spacing issues in the code.
3077
3078 Things we check for: spaces around operators, spaces after
3079 if/for/while/switch, no spaces around parens in function calls, two
3080 spaces between code and comment, don't start a block with a blank
erg@google.com8a95ecc2011-09-08 00:45:54 +00003081 line, don't end a function with a blank line, don't add a blank line
3082 after public/protected/private, don't have too many blank lines in a row.
erg@google.com4e00b9a2009-01-12 23:05:11 +00003083
3084 Args:
3085 filename: The name of the current file.
3086 clean_lines: A CleansedLines instance containing the file.
3087 linenum: The number of the line to check.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003088 nesting_state: A NestingState instance which maintains information about
erg@google.comd350fe52013-01-14 17:51:48 +00003089 the current stack of nested blocks being parsed.
erg@google.com4e00b9a2009-01-12 23:05:11 +00003090 error: The function to call with any errors found.
3091 """
3092
erg@google.com2aa59982013-10-28 19:09:25 +00003093 # Don't use "elided" lines here, otherwise we can't check commented lines.
3094 # Don't want to use "raw" either, because we don't want to check inside C++11
3095 # raw strings,
3096 raw = clean_lines.lines_without_raw_strings
erg@google.com4e00b9a2009-01-12 23:05:11 +00003097 line = raw[linenum]
3098
3099 # Before nixing comments, check if the line is blank for no good
3100 # reason. This includes the first line after a block is opened, and
3101 # blank lines at the end of a function (ie, right before a line like '}'
erg@google.comd350fe52013-01-14 17:51:48 +00003102 #
3103 # Skip all the blank line checks if we are immediately inside a
3104 # namespace body. In other words, don't issue blank line warnings
3105 # for this block:
3106 # namespace {
3107 #
3108 # }
3109 #
3110 # A warning about missing end of namespace comments will be issued instead.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003111 #
3112 # Also skip blank line checks for 'extern "C"' blocks, which are formatted
3113 # like namespaces.
3114 if (IsBlankLine(line) and
3115 not nesting_state.InNamespaceBody() and
3116 not nesting_state.InExternC()):
erg@google.com4e00b9a2009-01-12 23:05:11 +00003117 elided = clean_lines.elided
3118 prev_line = elided[linenum - 1]
3119 prevbrace = prev_line.rfind('{')
3120 # TODO(unknown): Don't complain if line before blank line, and line after,
3121 # both start with alnums and are indented the same amount.
3122 # This ignores whitespace at the start of a namespace block
3123 # because those are not usually indented.
erg@google.comd350fe52013-01-14 17:51:48 +00003124 if prevbrace != -1 and prev_line[prevbrace:].find('}') == -1:
erg@google.com4e00b9a2009-01-12 23:05:11 +00003125 # OK, we have a blank line at the start of a code block. Before we
3126 # complain, we check if it is an exception to the rule: The previous
erg@google.com8a95ecc2011-09-08 00:45:54 +00003127 # non-empty line has the parameters of a function header that are indented
erg@google.com4e00b9a2009-01-12 23:05:11 +00003128 # 4 spaces (because they did not fit in a 80 column line when placed on
3129 # the same line as the function name). We also check for the case where
3130 # the previous line is indented 6 spaces, which may happen when the
3131 # initializers of a constructor do not fit into a 80 column line.
3132 exception = False
3133 if Match(r' {6}\w', prev_line): # Initializer list?
3134 # We are looking for the opening column of initializer list, which
3135 # should be indented 4 spaces to cause 6 space indentation afterwards.
3136 search_position = linenum-2
3137 while (search_position >= 0
3138 and Match(r' {6}\w', elided[search_position])):
3139 search_position -= 1
3140 exception = (search_position >= 0
3141 and elided[search_position][:5] == ' :')
3142 else:
3143 # Search for the function arguments or an initializer list. We use a
3144 # simple heuristic here: If the line is indented 4 spaces; and we have a
3145 # closing paren, without the opening paren, followed by an opening brace
3146 # or colon (for initializer lists) we assume that it is the last line of
3147 # a function header. If we have a colon indented 4 spaces, it is an
3148 # initializer list.
3149 exception = (Match(r' {4}\w[^\(]*\)\s*(const\s*)?(\{\s*$|:)',
3150 prev_line)
3151 or Match(r' {4}:', prev_line))
3152
3153 if not exception:
3154 error(filename, linenum, 'whitespace/blank_line', 2,
erg@google.com2aa59982013-10-28 19:09:25 +00003155 'Redundant blank line at the start of a code block '
3156 'should be deleted.')
erg@google.comd350fe52013-01-14 17:51:48 +00003157 # Ignore blank lines at the end of a block in a long if-else
erg@google.com4e00b9a2009-01-12 23:05:11 +00003158 # chain, like this:
3159 # if (condition1) {
3160 # // Something followed by a blank line
3161 #
3162 # } else if (condition2) {
3163 # // Something else
3164 # }
3165 if linenum + 1 < clean_lines.NumLines():
3166 next_line = raw[linenum + 1]
3167 if (next_line
3168 and Match(r'\s*}', next_line)
erg@google.com4e00b9a2009-01-12 23:05:11 +00003169 and next_line.find('} else ') == -1):
3170 error(filename, linenum, 'whitespace/blank_line', 3,
erg@google.com2aa59982013-10-28 19:09:25 +00003171 'Redundant blank line at the end of a code block '
3172 'should be deleted.')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003173
erg@google.com8a95ecc2011-09-08 00:45:54 +00003174 matched = Match(r'\s*(public|protected|private):', prev_line)
3175 if matched:
3176 error(filename, linenum, 'whitespace/blank_line', 3,
3177 'Do not leave a blank line after "%s:"' % matched.group(1))
3178
avakulenko@google.com02af6282014-06-04 18:53:25 +00003179 # Next, check comments
3180 next_line_start = 0
3181 if linenum + 1 < clean_lines.NumLines():
3182 next_line = raw[linenum + 1]
3183 next_line_start = len(next_line) - len(next_line.lstrip())
3184 CheckComment(line, filename, linenum, next_line_start, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00003185
avakulenko@google.com02af6282014-06-04 18:53:25 +00003186 # get rid of comments and strings
3187 line = clean_lines.elided[linenum]
erg@google.com4e00b9a2009-01-12 23:05:11 +00003188
avakulenko@google.com02af6282014-06-04 18:53:25 +00003189 # You shouldn't have spaces before your brackets, except maybe after
3190 # 'delete []' or 'return []() {};'
3191 if Search(r'\w\s+\[', line) and not Search(r'(?:delete|return)\s+\[', line):
3192 error(filename, linenum, 'whitespace/braces', 5,
3193 'Extra space before [')
3194
3195 # In range-based for, we wanted spaces before and after the colon, but
3196 # not around "::" tokens that might appear.
3197 if (Search(r'for *\(.*[^:]:[^: ]', line) or
3198 Search(r'for *\(.*[^: ]:[^:]', line)):
3199 error(filename, linenum, 'whitespace/forcolon', 2,
3200 'Missing space around colon in range-based for loop')
3201
3202
3203def CheckOperatorSpacing(filename, clean_lines, linenum, error):
3204 """Checks for horizontal spacing around operators.
3205
3206 Args:
3207 filename: The name of the current file.
3208 clean_lines: A CleansedLines instance containing the file.
3209 linenum: The number of the line to check.
3210 error: The function to call with any errors found.
3211 """
3212 line = clean_lines.elided[linenum]
3213
3214 # Don't try to do spacing checks for operator methods. Do this by
3215 # replacing the troublesome characters with something else,
3216 # preserving column position for all other characters.
3217 #
3218 # The replacement is done repeatedly to avoid false positives from
3219 # operators that call operators.
3220 while True:
3221 match = Match(r'^(.*\boperator\b)(\S+)(\s*\(.*)$', line)
3222 if match:
3223 line = match.group(1) + ('_' * len(match.group(2))) + match.group(3)
3224 else:
3225 break
erg@google.com4e00b9a2009-01-12 23:05:11 +00003226
3227 # We allow no-spaces around = within an if: "if ( (a=Foo()) == 0 )".
3228 # Otherwise not. Note we only check for non-spaces on *both* sides;
3229 # sometimes people put non-spaces on one side when aligning ='s among
3230 # many lines (not that this is behavior that I approve of...)
avakulenko@google.com554223d2014-12-04 22:00:20 +00003231 if ((Search(r'[\w.]=', line) or
3232 Search(r'=[\w.]', line))
3233 and not Search(r'\b(if|while|for) ', line)
3234 # Operators taken from [lex.operators] in C++11 standard.
3235 and not Search(r'(>=|<=|==|!=|&=|\^=|\|=|\+=|\*=|\/=|\%=)', line)
3236 and not Search(r'operator=', line)):
erg@google.com4e00b9a2009-01-12 23:05:11 +00003237 error(filename, linenum, 'whitespace/operators', 4,
3238 'Missing spaces around =')
3239
3240 # It's ok not to have spaces around binary operators like + - * /, but if
3241 # there's too little whitespace, we get concerned. It's hard to tell,
3242 # though, so we punt on this one for now. TODO.
3243
3244 # You should always have whitespace around binary operators.
erg@google.comd350fe52013-01-14 17:51:48 +00003245 #
3246 # Check <= and >= first to avoid false positives with < and >, then
3247 # check non-include lines for spacing around < and >.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003248 #
3249 # If the operator is followed by a comma, assume it's be used in a
3250 # macro context and don't do any checks. This avoids false
3251 # positives.
3252 #
Alex Vakulenko01e47232016-05-06 13:54:15 -07003253 # Note that && is not included here. This is because there are too
3254 # many false positives due to RValue references.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003255 match = Search(r'[^<>=!\s](==|!=|<=|>=|\|\|)[^<>=!\s,;\)]', line)
erg@google.com4e00b9a2009-01-12 23:05:11 +00003256 if match:
3257 error(filename, linenum, 'whitespace/operators', 3,
3258 'Missing spaces around %s' % match.group(1))
erg@google.comd350fe52013-01-14 17:51:48 +00003259 elif not Match(r'#.*include', line):
erg@google.comd350fe52013-01-14 17:51:48 +00003260 # Look for < that is not surrounded by spaces. This is only
3261 # triggered if both sides are missing spaces, even though
3262 # technically should should flag if at least one side is missing a
3263 # space. This is done to avoid some false positives with shifts.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003264 match = Match(r'^(.*[^\s<])<[^\s=<,]', line)
3265 if match:
3266 (_, _, end_pos) = CloseExpression(
3267 clean_lines, linenum, len(match.group(1)))
3268 if end_pos <= -1:
3269 error(filename, linenum, 'whitespace/operators', 3,
3270 'Missing spaces around <')
erg@google.comd350fe52013-01-14 17:51:48 +00003271
3272 # Look for > that is not surrounded by spaces. Similar to the
3273 # above, we only trigger if both sides are missing spaces to avoid
3274 # false positives with shifts.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003275 match = Match(r'^(.*[^-\s>])>[^\s=>,]', line)
3276 if match:
3277 (_, _, start_pos) = ReverseCloseExpression(
3278 clean_lines, linenum, len(match.group(1)))
3279 if start_pos <= -1:
3280 error(filename, linenum, 'whitespace/operators', 3,
3281 'Missing spaces around >')
3282
3283 # We allow no-spaces around << when used like this: 10<<20, but
3284 # not otherwise (particularly, not when used as streams)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00003285 #
avakulenko@google.com02af6282014-06-04 18:53:25 +00003286 # We also allow operators following an opening parenthesis, since
3287 # those tend to be macros that deal with operators.
Alex Vakulenko01e47232016-05-06 13:54:15 -07003288 match = Search(r'(operator|[^\s(<])(?:L|UL|LL|ULL|l|ul|ll|ull)?<<([^\s,=<])', line)
avakulenko@google.com554223d2014-12-04 22:00:20 +00003289 if (match and not (match.group(1).isdigit() and match.group(2).isdigit()) and
avakulenko@google.com02af6282014-06-04 18:53:25 +00003290 not (match.group(1) == 'operator' and match.group(2) == ';')):
3291 error(filename, linenum, 'whitespace/operators', 3,
3292 'Missing spaces around <<')
erg@google.comd350fe52013-01-14 17:51:48 +00003293
3294 # We allow no-spaces around >> for almost anything. This is because
3295 # C++11 allows ">>" to close nested templates, which accounts for
3296 # most cases when ">>" is not followed by a space.
3297 #
3298 # We still warn on ">>" followed by alpha character, because that is
3299 # likely due to ">>" being used for right shifts, e.g.:
3300 # value >> alpha
3301 #
3302 # When ">>" is used to close templates, the alphanumeric letter that
3303 # follows would be part of an identifier, and there should still be
3304 # a space separating the template type and the identifier.
3305 # type<type<type>> alpha
3306 match = Search(r'>>[a-zA-Z_]', line)
erg@google.com4e00b9a2009-01-12 23:05:11 +00003307 if match:
3308 error(filename, linenum, 'whitespace/operators', 3,
erg@google.comd350fe52013-01-14 17:51:48 +00003309 'Missing spaces around >>')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003310
3311 # There shouldn't be space around unary operators
3312 match = Search(r'(!\s|~\s|[\s]--[\s;]|[\s]\+\+[\s;])', line)
3313 if match:
3314 error(filename, linenum, 'whitespace/operators', 4,
3315 'Extra space for operator %s' % match.group(1))
3316
avakulenko@google.com02af6282014-06-04 18:53:25 +00003317
3318def CheckParenthesisSpacing(filename, clean_lines, linenum, error):
3319 """Checks for horizontal spacing around parentheses.
3320
3321 Args:
3322 filename: The name of the current file.
3323 clean_lines: A CleansedLines instance containing the file.
3324 linenum: The number of the line to check.
3325 error: The function to call with any errors found.
3326 """
3327 line = clean_lines.elided[linenum]
3328
3329 # No spaces after an if, while, switch, or for
erg@google.com4e00b9a2009-01-12 23:05:11 +00003330 match = Search(r' (if\(|for\(|while\(|switch\()', line)
3331 if match:
3332 error(filename, linenum, 'whitespace/parens', 5,
3333 'Missing space before ( in %s' % match.group(1))
3334
3335 # For if/for/while/switch, the left and right parens should be
3336 # consistent about how many spaces are inside the parens, and
3337 # there should either be zero or one spaces inside the parens.
3338 # We don't want: "if ( foo)" or "if ( foo )".
erg@google.come35f7652009-06-19 20:52:09 +00003339 # Exception: "for ( ; foo; bar)" and "for (foo; bar; )" are allowed.
erg@google.com4e00b9a2009-01-12 23:05:11 +00003340 match = Search(r'\b(if|for|while|switch)\s*'
3341 r'\(([ ]*)(.).*[^ ]+([ ]*)\)\s*{\s*$',
3342 line)
3343 if match:
3344 if len(match.group(2)) != len(match.group(4)):
3345 if not (match.group(3) == ';' and
erg@google.come35f7652009-06-19 20:52:09 +00003346 len(match.group(2)) == 1 + len(match.group(4)) or
3347 not match.group(2) and Search(r'\bfor\s*\(.*; \)', line)):
erg@google.com4e00b9a2009-01-12 23:05:11 +00003348 error(filename, linenum, 'whitespace/parens', 5,
3349 'Mismatching spaces inside () in %s' % match.group(1))
erg@google.comc6671232013-10-25 21:44:03 +00003350 if len(match.group(2)) not in [0, 1]:
erg@google.com4e00b9a2009-01-12 23:05:11 +00003351 error(filename, linenum, 'whitespace/parens', 5,
3352 'Should have zero or one spaces inside ( and ) in %s' %
3353 match.group(1))
3354
avakulenko@google.com02af6282014-06-04 18:53:25 +00003355
3356def CheckCommaSpacing(filename, clean_lines, linenum, error):
3357 """Checks for horizontal spacing near commas and semicolons.
3358
3359 Args:
3360 filename: The name of the current file.
3361 clean_lines: A CleansedLines instance containing the file.
3362 linenum: The number of the line to check.
3363 error: The function to call with any errors found.
3364 """
3365 raw = clean_lines.lines_without_raw_strings
3366 line = clean_lines.elided[linenum]
3367
erg@google.com4e00b9a2009-01-12 23:05:11 +00003368 # You should always have a space after a comma (either as fn arg or operator)
erg@google.comc6671232013-10-25 21:44:03 +00003369 #
3370 # This does not apply when the non-space character following the
3371 # comma is another comma, since the only time when that happens is
3372 # for empty macro arguments.
erg@google.com2aa59982013-10-28 19:09:25 +00003373 #
3374 # We run this check in two passes: first pass on elided lines to
3375 # verify that lines contain missing whitespaces, second pass on raw
3376 # lines to confirm that those missing whitespaces are not due to
3377 # elided comments.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00003378 if (Search(r',[^,\s]', ReplaceAll(r'\boperator\s*,\s*\(', 'F(', line)) and
3379 Search(r',[^,\s]', raw[linenum])):
erg@google.com4e00b9a2009-01-12 23:05:11 +00003380 error(filename, linenum, 'whitespace/comma', 3,
3381 'Missing space after ,')
3382
erg@google.comd7d27472011-09-07 17:36:35 +00003383 # You should always have a space after a semicolon
3384 # except for few corner cases
3385 # TODO(unknown): clarify if 'if (1) { return 1;}' is requires one more
3386 # space after ;
3387 if Search(r';[^\s};\\)/]', line):
3388 error(filename, linenum, 'whitespace/semicolon', 3,
3389 'Missing space after ;')
3390
avakulenko@google.com02af6282014-06-04 18:53:25 +00003391
Alex Vakulenko01e47232016-05-06 13:54:15 -07003392def _IsType(clean_lines, nesting_state, expr):
3393 """Check if expression looks like a type name, returns true if so.
3394
3395 Args:
3396 clean_lines: A CleansedLines instance containing the file.
3397 nesting_state: A NestingState instance which maintains information about
3398 the current stack of nested blocks being parsed.
3399 expr: The expression to check.
3400 Returns:
3401 True, if token looks like a type.
3402 """
3403 # Keep only the last token in the expression
3404 last_word = Match(r'^.*(\b\S+)$', expr)
3405 if last_word:
3406 token = last_word.group(1)
3407 else:
3408 token = expr
3409
3410 # Match native types and stdint types
3411 if _TYPES.match(token):
3412 return True
3413
3414 # Try a bit harder to match templated types. Walk up the nesting
3415 # stack until we find something that resembles a typename
3416 # declaration for what we are looking for.
3417 typename_pattern = (r'\b(?:typename|class|struct)\s+' + re.escape(token) +
3418 r'\b')
3419 block_index = len(nesting_state.stack) - 1
3420 while block_index >= 0:
3421 if isinstance(nesting_state.stack[block_index], _NamespaceInfo):
3422 return False
3423
3424 # Found where the opening brace is. We want to scan from this
3425 # line up to the beginning of the function, minus a few lines.
3426 # template <typename Type1, // stop scanning here
3427 # ...>
3428 # class C
3429 # : public ... { // start scanning here
3430 last_line = nesting_state.stack[block_index].starting_linenum
3431
3432 next_block_start = 0
3433 if block_index > 0:
3434 next_block_start = nesting_state.stack[block_index - 1].starting_linenum
3435 first_line = last_line
3436 while first_line >= next_block_start:
3437 if clean_lines.elided[first_line].find('template') >= 0:
3438 break
3439 first_line -= 1
3440 if first_line < next_block_start:
3441 # Didn't find any "template" keyword before reaching the next block,
3442 # there are probably no template things to check for this block
3443 block_index -= 1
3444 continue
3445
3446 # Look for typename in the specified range
3447 for i in xrange(first_line, last_line + 1, 1):
3448 if Search(typename_pattern, clean_lines.elided[i]):
3449 return True
3450 block_index -= 1
3451
3452 return False
3453
3454
3455def CheckBracesSpacing(filename, clean_lines, linenum, nesting_state, error):
avakulenko@google.com02af6282014-06-04 18:53:25 +00003456 """Checks for horizontal spacing near commas.
3457
3458 Args:
3459 filename: The name of the current file.
3460 clean_lines: A CleansedLines instance containing the file.
3461 linenum: The number of the line to check.
Alex Vakulenko01e47232016-05-06 13:54:15 -07003462 nesting_state: A NestingState instance which maintains information about
3463 the current stack of nested blocks being parsed.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003464 error: The function to call with any errors found.
3465 """
3466 line = clean_lines.elided[linenum]
erg@google.com4e00b9a2009-01-12 23:05:11 +00003467
erg@google.com8a95ecc2011-09-08 00:45:54 +00003468 # Except after an opening paren, or after another opening brace (in case of
3469 # an initializer list, for instance), you should have spaces before your
Alex Vakulenko01e47232016-05-06 13:54:15 -07003470 # braces when they are delimiting blocks, classes, namespaces etc.
3471 # And since you should never have braces at the beginning of a line,
3472 # this is an easy test. Except that braces used for initialization don't
3473 # follow the same rule; we often don't want spaces before those.
avakulenko@google.com554223d2014-12-04 22:00:20 +00003474 match = Match(r'^(.*[^ ({>]){', line)
Alex Vakulenko01e47232016-05-06 13:54:15 -07003475
erg@google.com2aa59982013-10-28 19:09:25 +00003476 if match:
3477 # Try a bit harder to check for brace initialization. This
3478 # happens in one of the following forms:
3479 # Constructor() : initializer_list_{} { ... }
3480 # Constructor{}.MemberFunction()
3481 # Type variable{};
3482 # FunctionCall(type{}, ...);
3483 # LastArgument(..., type{});
3484 # LOG(INFO) << type{} << " ...";
3485 # map_of_type[{...}] = ...;
avakulenko@google.com02af6282014-06-04 18:53:25 +00003486 # ternary = expr ? new type{} : nullptr;
3487 # OuterTemplate<InnerTemplateConstructor<Type>{}>
erg@google.com2aa59982013-10-28 19:09:25 +00003488 #
3489 # We check for the character following the closing brace, and
3490 # silence the warning if it's one of those listed above, i.e.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003491 # "{.;,)<>]:".
erg@google.com2aa59982013-10-28 19:09:25 +00003492 #
3493 # To account for nested initializer list, we allow any number of
3494 # closing braces up to "{;,)<". We can't simply silence the
3495 # warning on first sight of closing brace, because that would
3496 # cause false negatives for things that are not initializer lists.
3497 # Silence this: But not this:
3498 # Outer{ if (...) {
3499 # Inner{...} if (...){ // Missing space before {
3500 # }; }
3501 #
3502 # There is a false negative with this approach if people inserted
3503 # spurious semicolons, e.g. "if (cond){};", but we will catch the
3504 # spurious semicolon with a separate check.
Alex Vakulenko01e47232016-05-06 13:54:15 -07003505 leading_text = match.group(1)
erg@google.com2aa59982013-10-28 19:09:25 +00003506 (endline, endlinenum, endpos) = CloseExpression(
3507 clean_lines, linenum, len(match.group(1)))
3508 trailing_text = ''
3509 if endpos > -1:
3510 trailing_text = endline[endpos:]
3511 for offset in xrange(endlinenum + 1,
3512 min(endlinenum + 3, clean_lines.NumLines() - 1)):
3513 trailing_text += clean_lines.elided[offset]
Alex Vakulenko01e47232016-05-06 13:54:15 -07003514 # We also suppress warnings for `uint64_t{expression}` etc., as the style
3515 # guide recommends brace initialization for integral types to avoid
3516 # overflow/truncation.
3517 if (not Match(r'^[\s}]*[{.;,)<>\]:]', trailing_text)
3518 and not _IsType(clean_lines, nesting_state, leading_text)):
erg@google.com2aa59982013-10-28 19:09:25 +00003519 error(filename, linenum, 'whitespace/braces', 5,
3520 'Missing space before {')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003521
3522 # Make sure '} else {' has spaces.
3523 if Search(r'}else', line):
3524 error(filename, linenum, 'whitespace/braces', 5,
3525 'Missing space before else')
3526
erg@google.com4e00b9a2009-01-12 23:05:11 +00003527 # You shouldn't have a space before a semicolon at the end of the line.
3528 # There's a special case for "for" since the style guide allows space before
3529 # the semicolon there.
3530 if Search(r':\s*;\s*$', line):
3531 error(filename, linenum, 'whitespace/semicolon', 5,
erg@google.comd350fe52013-01-14 17:51:48 +00003532 'Semicolon defining empty statement. Use {} instead.')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003533 elif Search(r'^\s*;\s*$', line):
3534 error(filename, linenum, 'whitespace/semicolon', 5,
3535 'Line contains only semicolon. If this should be an empty statement, '
erg@google.comd350fe52013-01-14 17:51:48 +00003536 'use {} instead.')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003537 elif (Search(r'\s+;\s*$', line) and
3538 not Search(r'\bfor\b', line)):
3539 error(filename, linenum, 'whitespace/semicolon', 5,
3540 'Extra space before last semicolon. If this should be an empty '
erg@google.comd350fe52013-01-14 17:51:48 +00003541 'statement, use {} instead.')
3542
avakulenko@google.com02af6282014-06-04 18:53:25 +00003543
3544def IsDecltype(clean_lines, linenum, column):
3545 """Check if the token ending on (linenum, column) is decltype().
3546
3547 Args:
3548 clean_lines: A CleansedLines instance containing the file.
3549 linenum: the number of the line to check.
3550 column: end column of the token to check.
3551 Returns:
3552 True if this token is decltype() expression, False otherwise.
3553 """
3554 (text, _, start_col) = ReverseCloseExpression(clean_lines, linenum, column)
3555 if start_col < 0:
3556 return False
3557 if Search(r'\bdecltype\s*$', text[0:start_col]):
3558 return True
3559 return False
3560
3561
erg@google.com8a95ecc2011-09-08 00:45:54 +00003562def CheckSectionSpacing(filename, clean_lines, class_info, linenum, error):
3563 """Checks for additional blank line issues related to sections.
3564
3565 Currently the only thing checked here is blank line before protected/private.
3566
3567 Args:
3568 filename: The name of the current file.
3569 clean_lines: A CleansedLines instance containing the file.
3570 class_info: A _ClassInfo objects.
3571 linenum: The number of the line to check.
3572 error: The function to call with any errors found.
3573 """
3574 # Skip checks if the class is small, where small means 25 lines or less.
3575 # 25 lines seems like a good cutoff since that's the usual height of
3576 # terminals, and any class that can't fit in one screen can't really
3577 # be considered "small".
3578 #
3579 # Also skip checks if we are on the first line. This accounts for
3580 # classes that look like
3581 # class Foo { public: ... };
3582 #
3583 # If we didn't find the end of the class, last_line would be zero,
3584 # and the check will be skipped by the first condition.
erg@google.comd350fe52013-01-14 17:51:48 +00003585 if (class_info.last_line - class_info.starting_linenum <= 24 or
3586 linenum <= class_info.starting_linenum):
erg@google.com8a95ecc2011-09-08 00:45:54 +00003587 return
3588
3589 matched = Match(r'\s*(public|protected|private):', clean_lines.lines[linenum])
3590 if matched:
3591 # Issue warning if the line before public/protected/private was
3592 # not a blank line, but don't do this if the previous line contains
3593 # "class" or "struct". This can happen two ways:
3594 # - We are at the beginning of the class.
3595 # - We are forward-declaring an inner class that is semantically
3596 # private, but needed to be public for implementation reasons.
erg@google.comd350fe52013-01-14 17:51:48 +00003597 # Also ignores cases where the previous line ends with a backslash as can be
3598 # common when defining classes in C macros.
erg@google.com8a95ecc2011-09-08 00:45:54 +00003599 prev_line = clean_lines.lines[linenum - 1]
3600 if (not IsBlankLine(prev_line) and
erg@google.comd350fe52013-01-14 17:51:48 +00003601 not Search(r'\b(class|struct)\b', prev_line) and
3602 not Search(r'\\$', prev_line)):
erg@google.com8a95ecc2011-09-08 00:45:54 +00003603 # Try a bit harder to find the beginning of the class. This is to
3604 # account for multi-line base-specifier lists, e.g.:
3605 # class Derived
3606 # : public Base {
erg@google.comd350fe52013-01-14 17:51:48 +00003607 end_class_head = class_info.starting_linenum
3608 for i in range(class_info.starting_linenum, linenum):
erg@google.com8a95ecc2011-09-08 00:45:54 +00003609 if Search(r'\{\s*$', clean_lines.lines[i]):
3610 end_class_head = i
3611 break
3612 if end_class_head < linenum - 1:
3613 error(filename, linenum, 'whitespace/blank_line', 3,
3614 '"%s:" should be preceded by a blank line' % matched.group(1))
3615
3616
erg@google.com4e00b9a2009-01-12 23:05:11 +00003617def GetPreviousNonBlankLine(clean_lines, linenum):
3618 """Return the most recent non-blank line and its line number.
3619
3620 Args:
3621 clean_lines: A CleansedLines instance containing the file contents.
3622 linenum: The number of the line to check.
3623
3624 Returns:
3625 A tuple with two elements. The first element is the contents of the last
3626 non-blank line before the current line, or the empty string if this is the
3627 first non-blank line. The second is the line number of that line, or -1
3628 if this is the first non-blank line.
3629 """
3630
3631 prevlinenum = linenum - 1
3632 while prevlinenum >= 0:
3633 prevline = clean_lines.elided[prevlinenum]
3634 if not IsBlankLine(prevline): # if not a blank line...
3635 return (prevline, prevlinenum)
3636 prevlinenum -= 1
3637 return ('', -1)
3638
3639
3640def CheckBraces(filename, clean_lines, linenum, error):
3641 """Looks for misplaced braces (e.g. at the end of line).
3642
3643 Args:
3644 filename: The name of the current file.
3645 clean_lines: A CleansedLines instance containing the file.
3646 linenum: The number of the line to check.
3647 error: The function to call with any errors found.
3648 """
3649
3650 line = clean_lines.elided[linenum] # get rid of comments and strings
3651
3652 if Match(r'\s*{\s*$', line):
erg@google.com2aa59982013-10-28 19:09:25 +00003653 # We allow an open brace to start a line in the case where someone is using
3654 # braces in a block to explicitly create a new scope, which is commonly used
3655 # to control the lifetime of stack-allocated variables. Braces are also
3656 # used for brace initializers inside function calls. We don't detect this
3657 # perfectly: we just don't complain if the last non-whitespace character on
3658 # the previous non-blank line is ',', ';', ':', '(', '{', or '}', or if the
Alex Vakulenko01e47232016-05-06 13:54:15 -07003659 # previous line starts a preprocessor block. We also allow a brace on the
3660 # following line if it is part of an array initialization and would not fit
3661 # within the 80 character limit of the preceding line.
erg@google.com4e00b9a2009-01-12 23:05:11 +00003662 prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0]
erg@google.com2aa59982013-10-28 19:09:25 +00003663 if (not Search(r'[,;:}{(]\s*$', prevline) and
Alex Vakulenko01e47232016-05-06 13:54:15 -07003664 not Match(r'\s*#', prevline) and
3665 not (GetLineWidth(prevline) > _line_length - 2 and '[]' in prevline)):
erg@google.com4e00b9a2009-01-12 23:05:11 +00003666 error(filename, linenum, 'whitespace/braces', 4,
3667 '{ should almost always be at the end of the previous line')
3668
3669 # An else clause should be on the same line as the preceding closing brace.
avakulenko@google.com02af6282014-06-04 18:53:25 +00003670 if Match(r'\s*else\b\s*(?:if\b|\{|$)', line):
erg@google.com4e00b9a2009-01-12 23:05:11 +00003671 prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0]
3672 if Match(r'\s*}\s*$', prevline):
3673 error(filename, linenum, 'whitespace/newline', 4,
3674 'An else should appear on the same line as the preceding }')
3675
3676 # If braces come on one side of an else, they should be on both.
3677 # However, we have to worry about "else if" that spans multiple lines!
avakulenko@google.com02af6282014-06-04 18:53:25 +00003678 if Search(r'else if\s*\(', line): # could be multi-line if
3679 brace_on_left = bool(Search(r'}\s*else if\s*\(', line))
3680 # find the ( after the if
3681 pos = line.find('else if')
3682 pos = line.find('(', pos)
3683 if pos > 0:
3684 (endline, _, endpos) = CloseExpression(clean_lines, linenum, pos)
3685 brace_on_right = endline[endpos:].find('{') != -1
3686 if brace_on_left != brace_on_right: # must be brace after if
3687 error(filename, linenum, 'readability/braces', 5,
3688 'If an else has a brace on one side, it should have it on both')
3689 elif Search(r'}\s*else[^{]*$', line) or Match(r'[^}]*else\s*{', line):
3690 error(filename, linenum, 'readability/braces', 5,
3691 'If an else has a brace on one side, it should have it on both')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003692
3693 # Likewise, an else should never have the else clause on the same line
3694 if Search(r'\belse [^\s{]', line) and not Search(r'\belse if\b', line):
3695 error(filename, linenum, 'whitespace/newline', 4,
3696 'Else clause should never be on same line as else (use 2 lines)')
3697
3698 # In the same way, a do/while should never be on one line
3699 if Match(r'\s*do [^\s{]', line):
3700 error(filename, linenum, 'whitespace/newline', 4,
3701 'do/while clauses should not be on a single line')
3702
avakulenko@google.com02af6282014-06-04 18:53:25 +00003703 # Check single-line if/else bodies. The style guide says 'curly braces are not
3704 # required for single-line statements'. We additionally allow multi-line,
3705 # single statements, but we reject anything with more than one semicolon in
3706 # it. This means that the first semicolon after the if should be at the end of
3707 # its line, and the line after that should have an indent level equal to or
3708 # lower than the if. We also check for ambiguous if/else nesting without
3709 # braces.
3710 if_else_match = Search(r'\b(if\s*\(|else\b)', line)
3711 if if_else_match and not Match(r'\s*#', line):
3712 if_indent = GetIndentLevel(line)
3713 endline, endlinenum, endpos = line, linenum, if_else_match.end()
3714 if_match = Search(r'\bif\s*\(', line)
3715 if if_match:
3716 # This could be a multiline if condition, so find the end first.
3717 pos = if_match.end() - 1
3718 (endline, endlinenum, endpos) = CloseExpression(clean_lines, linenum, pos)
3719 # Check for an opening brace, either directly after the if or on the next
3720 # line. If found, this isn't a single-statement conditional.
3721 if (not Match(r'\s*{', endline[endpos:])
3722 and not (Match(r'\s*$', endline[endpos:])
3723 and endlinenum < (len(clean_lines.elided) - 1)
3724 and Match(r'\s*{', clean_lines.elided[endlinenum + 1]))):
3725 while (endlinenum < len(clean_lines.elided)
3726 and ';' not in clean_lines.elided[endlinenum][endpos:]):
3727 endlinenum += 1
3728 endpos = 0
3729 if endlinenum < len(clean_lines.elided):
3730 endline = clean_lines.elided[endlinenum]
3731 # We allow a mix of whitespace and closing braces (e.g. for one-liner
3732 # methods) and a single \ after the semicolon (for macros)
3733 endpos = endline.find(';')
3734 if not Match(r';[\s}]*(\\?)$', endline[endpos:]):
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00003735 # Semicolon isn't the last character, there's something trailing.
3736 # Output a warning if the semicolon is not contained inside
3737 # a lambda expression.
3738 if not Match(r'^[^{};]*\[[^\[\]]*\][^{}]*\{[^{}]*\}\s*\)*[;,]\s*$',
3739 endline):
3740 error(filename, linenum, 'readability/braces', 4,
3741 'If/else bodies with multiple statements require braces')
avakulenko@google.com02af6282014-06-04 18:53:25 +00003742 elif endlinenum < len(clean_lines.elided) - 1:
3743 # Make sure the next line is dedented
3744 next_line = clean_lines.elided[endlinenum + 1]
3745 next_indent = GetIndentLevel(next_line)
3746 # With ambiguous nested if statements, this will error out on the
3747 # if that *doesn't* match the else, regardless of whether it's the
3748 # inner one or outer one.
3749 if (if_match and Match(r'\s*else\b', next_line)
3750 and next_indent != if_indent):
3751 error(filename, linenum, 'readability/braces', 4,
3752 'Else clause should be indented at the same level as if. '
3753 'Ambiguous nested if/else chains require braces.')
3754 elif next_indent > if_indent:
3755 error(filename, linenum, 'readability/braces', 4,
3756 'If/else bodies with multiple statements require braces')
3757
3758
3759def CheckTrailingSemicolon(filename, clean_lines, linenum, error):
3760 """Looks for redundant trailing semicolon.
3761
3762 Args:
3763 filename: The name of the current file.
3764 clean_lines: A CleansedLines instance containing the file.
3765 linenum: The number of the line to check.
3766 error: The function to call with any errors found.
3767 """
3768
3769 line = clean_lines.elided[linenum]
3770
erg@google.com2aa59982013-10-28 19:09:25 +00003771 # Block bodies should not be followed by a semicolon. Due to C++11
3772 # brace initialization, there are more places where semicolons are
3773 # required than not, so we use a whitelist approach to check these
3774 # rather than a blacklist. These are the places where "};" should
3775 # be replaced by just "}":
3776 # 1. Some flavor of block following closing parenthesis:
3777 # for (;;) {};
3778 # while (...) {};
3779 # switch (...) {};
3780 # Function(...) {};
3781 # if (...) {};
3782 # if (...) else if (...) {};
3783 #
3784 # 2. else block:
3785 # if (...) else {};
3786 #
3787 # 3. const member function:
3788 # Function(...) const {};
3789 #
3790 # 4. Block following some statement:
3791 # x = 42;
3792 # {};
3793 #
3794 # 5. Block at the beginning of a function:
3795 # Function(...) {
3796 # {};
3797 # }
3798 #
3799 # Note that naively checking for the preceding "{" will also match
3800 # braces inside multi-dimensional arrays, but this is fine since
3801 # that expression will not contain semicolons.
3802 #
3803 # 6. Block following another block:
3804 # while (true) {}
3805 # {};
3806 #
3807 # 7. End of namespaces:
3808 # namespace {};
3809 #
3810 # These semicolons seems far more common than other kinds of
3811 # redundant semicolons, possibly due to people converting classes
3812 # to namespaces. For now we do not warn for this case.
3813 #
3814 # Try matching case 1 first.
3815 match = Match(r'^(.*\)\s*)\{', line)
3816 if match:
3817 # Matched closing parenthesis (case 1). Check the token before the
3818 # matching opening parenthesis, and don't warn if it looks like a
3819 # macro. This avoids these false positives:
3820 # - macro that defines a base class
3821 # - multi-line macro that defines a base class
3822 # - macro that defines the whole class-head
3823 #
3824 # But we still issue warnings for macros that we know are safe to
3825 # warn, specifically:
3826 # - TEST, TEST_F, TEST_P, MATCHER, MATCHER_P
3827 # - TYPED_TEST
3828 # - INTERFACE_DEF
3829 # - EXCLUSIVE_LOCKS_REQUIRED, SHARED_LOCKS_REQUIRED, LOCKS_EXCLUDED:
3830 #
3831 # We implement a whitelist of safe macros instead of a blacklist of
3832 # unsafe macros, even though the latter appears less frequently in
3833 # google code and would have been easier to implement. This is because
3834 # the downside for getting the whitelist wrong means some extra
3835 # semicolons, while the downside for getting the blacklist wrong
3836 # would result in compile errors.
3837 #
avakulenko@google.com554223d2014-12-04 22:00:20 +00003838 # In addition to macros, we also don't want to warn on
3839 # - Compound literals
3840 # - Lambdas
Alex Vakulenko01e47232016-05-06 13:54:15 -07003841 # - alignas specifier with anonymous structs
3842 # - decltype
erg@google.com2aa59982013-10-28 19:09:25 +00003843 closing_brace_pos = match.group(1).rfind(')')
3844 opening_parenthesis = ReverseCloseExpression(
3845 clean_lines, linenum, closing_brace_pos)
3846 if opening_parenthesis[2] > -1:
3847 line_prefix = opening_parenthesis[0][0:opening_parenthesis[2]]
Alex Vakulenko01e47232016-05-06 13:54:15 -07003848 macro = Search(r'\b([A-Z_][A-Z0-9_]*)\s*$', line_prefix)
avakulenko@google.com02af6282014-06-04 18:53:25 +00003849 func = Match(r'^(.*\])\s*$', line_prefix)
erg@google.com2aa59982013-10-28 19:09:25 +00003850 if ((macro and
3851 macro.group(1) not in (
3852 'TEST', 'TEST_F', 'MATCHER', 'MATCHER_P', 'TYPED_TEST',
3853 'EXCLUSIVE_LOCKS_REQUIRED', 'SHARED_LOCKS_REQUIRED',
3854 'LOCKS_EXCLUDED', 'INTERFACE_DEF')) or
avakulenko@google.com02af6282014-06-04 18:53:25 +00003855 (func and not Search(r'\boperator\s*\[\s*\]', func.group(1))) or
avakulenko@google.com554223d2014-12-04 22:00:20 +00003856 Search(r'\b(?:struct|union)\s+alignas\s*$', line_prefix) or
Alex Vakulenko01e47232016-05-06 13:54:15 -07003857 Search(r'\bdecltype$', line_prefix) or
erg@google.com2aa59982013-10-28 19:09:25 +00003858 Search(r'\s+=\s*$', line_prefix)):
3859 match = None
avakulenko@google.com02af6282014-06-04 18:53:25 +00003860 if (match and
3861 opening_parenthesis[1] > 1 and
3862 Search(r'\]\s*$', clean_lines.elided[opening_parenthesis[1] - 1])):
3863 # Multi-line lambda-expression
3864 match = None
erg@google.com2aa59982013-10-28 19:09:25 +00003865
3866 else:
3867 # Try matching cases 2-3.
3868 match = Match(r'^(.*(?:else|\)\s*const)\s*)\{', line)
3869 if not match:
3870 # Try matching cases 4-6. These are always matched on separate lines.
3871 #
3872 # Note that we can't simply concatenate the previous line to the
3873 # current line and do a single match, otherwise we may output
3874 # duplicate warnings for the blank line case:
3875 # if (cond) {
3876 # // blank line
3877 # }
3878 prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0]
3879 if prevline and Search(r'[;{}]\s*$', prevline):
3880 match = Match(r'^(\s*)\{', line)
3881
3882 # Check matching closing brace
3883 if match:
3884 (endline, endlinenum, endpos) = CloseExpression(
3885 clean_lines, linenum, len(match.group(1)))
3886 if endpos > -1 and Match(r'^\s*;', endline[endpos:]):
3887 # Current {} pair is eligible for semicolon check, and we have found
3888 # the redundant semicolon, output warning here.
3889 #
3890 # Note: because we are scanning forward for opening braces, and
3891 # outputting warnings for the matching closing brace, if there are
3892 # nested blocks with trailing semicolons, we will get the error
3893 # messages in reversed order.
Piotr Semenovb7e2ef62016-05-20 18:39:34 +03003894
3895 # We need to check the line forward for NOLINT
3896 raw_lines = clean_lines.raw_lines
3897 ParseNolintSuppressions(filename, raw_lines[endlinenum-1], endlinenum-1,
3898 error)
3899 ParseNolintSuppressions(filename, raw_lines[endlinenum], endlinenum,
3900 error)
3901
erg@google.com2aa59982013-10-28 19:09:25 +00003902 error(filename, endlinenum, 'readability/braces', 4,
3903 "You don't need a ; after a }")
erg@google.com4e00b9a2009-01-12 23:05:11 +00003904
3905
erg@google.comc6671232013-10-25 21:44:03 +00003906def CheckEmptyBlockBody(filename, clean_lines, linenum, error):
3907 """Look for empty loop/conditional body with only a single semicolon.
erg@google.comd350fe52013-01-14 17:51:48 +00003908
3909 Args:
3910 filename: The name of the current file.
3911 clean_lines: A CleansedLines instance containing the file.
3912 linenum: The number of the line to check.
3913 error: The function to call with any errors found.
3914 """
3915
3916 # Search for loop keywords at the beginning of the line. Because only
3917 # whitespaces are allowed before the keywords, this will also ignore most
3918 # do-while-loops, since those lines should start with closing brace.
erg@google.comc6671232013-10-25 21:44:03 +00003919 #
3920 # We also check "if" blocks here, since an empty conditional block
3921 # is likely an error.
erg@google.comd350fe52013-01-14 17:51:48 +00003922 line = clean_lines.elided[linenum]
erg@google.comc6671232013-10-25 21:44:03 +00003923 matched = Match(r'\s*(for|while|if)\s*\(', line)
3924 if matched:
Alex Vakulenko01e47232016-05-06 13:54:15 -07003925 # Find the end of the conditional expression.
erg@google.comd350fe52013-01-14 17:51:48 +00003926 (end_line, end_linenum, end_pos) = CloseExpression(
3927 clean_lines, linenum, line.find('('))
3928
3929 # Output warning if what follows the condition expression is a semicolon.
3930 # No warning for all other cases, including whitespace or newline, since we
3931 # have a separate check for semicolons preceded by whitespace.
3932 if end_pos >= 0 and Match(r';', end_line[end_pos:]):
erg@google.comc6671232013-10-25 21:44:03 +00003933 if matched.group(1) == 'if':
3934 error(filename, end_linenum, 'whitespace/empty_conditional_body', 5,
3935 'Empty conditional bodies should use {}')
3936 else:
3937 error(filename, end_linenum, 'whitespace/empty_loop_body', 5,
3938 'Empty loop bodies should use {} or continue')
erg@google.com4e00b9a2009-01-12 23:05:11 +00003939
Alex Vakulenko01e47232016-05-06 13:54:15 -07003940 # Check for if statements that have completely empty bodies (no comments)
3941 # and no else clauses.
3942 if end_pos >= 0 and matched.group(1) == 'if':
3943 # Find the position of the opening { for the if statement.
3944 # Return without logging an error if it has no brackets.
3945 opening_linenum = end_linenum
3946 opening_line_fragment = end_line[end_pos:]
3947 # Loop until EOF or find anything that's not whitespace or opening {.
3948 while not Search(r'^\s*\{', opening_line_fragment):
3949 if Search(r'^(?!\s*$)', opening_line_fragment):
3950 # Conditional has no brackets.
3951 return
3952 opening_linenum += 1
3953 if opening_linenum == len(clean_lines.elided):
3954 # Couldn't find conditional's opening { or any code before EOF.
3955 return
3956 opening_line_fragment = clean_lines.elided[opening_linenum]
3957 # Set opening_line (opening_line_fragment may not be entire opening line).
3958 opening_line = clean_lines.elided[opening_linenum]
3959
3960 # Find the position of the closing }.
3961 opening_pos = opening_line_fragment.find('{')
3962 if opening_linenum == end_linenum:
3963 # We need to make opening_pos relative to the start of the entire line.
3964 opening_pos += end_pos
3965 (closing_line, closing_linenum, closing_pos) = CloseExpression(
3966 clean_lines, opening_linenum, opening_pos)
3967 if closing_pos < 0:
3968 return
3969
3970 # Now construct the body of the conditional. This consists of the portion
3971 # of the opening line after the {, all lines until the closing line,
3972 # and the portion of the closing line before the }.
3973 if (clean_lines.raw_lines[opening_linenum] !=
3974 CleanseComments(clean_lines.raw_lines[opening_linenum])):
3975 # Opening line ends with a comment, so conditional isn't empty.
3976 return
3977 if closing_linenum > opening_linenum:
3978 # Opening line after the {. Ignore comments here since we checked above.
3979 body = list(opening_line[opening_pos+1:])
3980 # All lines until closing line, excluding closing line, with comments.
3981 body.extend(clean_lines.raw_lines[opening_linenum+1:closing_linenum])
3982 # Closing line before the }. Won't (and can't) have comments.
3983 body.append(clean_lines.elided[closing_linenum][:closing_pos-1])
3984 body = '\n'.join(body)
3985 else:
3986 # If statement has brackets and fits on a single line.
3987 body = opening_line[opening_pos+1:closing_pos-1]
3988
3989 # Check if the body is empty
3990 if not _EMPTY_CONDITIONAL_BODY_PATTERN.search(body):
3991 return
3992 # The body is empty. Now make sure there's not an else clause.
3993 current_linenum = closing_linenum
3994 current_line_fragment = closing_line[closing_pos:]
3995 # Loop until EOF or find anything that's not whitespace or else clause.
3996 while Search(r'^\s*$|^(?=\s*else)', current_line_fragment):
3997 if Search(r'^(?=\s*else)', current_line_fragment):
3998 # Found an else clause, so don't log an error.
3999 return
4000 current_linenum += 1
4001 if current_linenum == len(clean_lines.elided):
4002 break
4003 current_line_fragment = clean_lines.elided[current_linenum]
4004
4005 # The body is empty and there's no else clause until EOF or other code.
4006 error(filename, end_linenum, 'whitespace/empty_if_body', 4,
4007 ('If statement had no body and no else clause'))
4008
erg@google.com4e00b9a2009-01-12 23:05:11 +00004009
avakulenko@google.com02af6282014-06-04 18:53:25 +00004010def FindCheckMacro(line):
4011 """Find a replaceable CHECK-like macro.
4012
4013 Args:
4014 line: line to search on.
4015 Returns:
4016 (macro name, start position), or (None, -1) if no replaceable
4017 macro is found.
4018 """
4019 for macro in _CHECK_MACROS:
4020 i = line.find(macro)
4021 if i >= 0:
4022 # Find opening parenthesis. Do a regular expression match here
4023 # to make sure that we are matching the expected CHECK macro, as
4024 # opposed to some other macro that happens to contain the CHECK
4025 # substring.
4026 matched = Match(r'^(.*\b' + macro + r'\s*)\(', line)
4027 if not matched:
4028 continue
4029 return (macro, len(matched.group(1)))
4030 return (None, -1)
4031
4032
erg@google.com4e00b9a2009-01-12 23:05:11 +00004033def CheckCheck(filename, clean_lines, linenum, error):
4034 """Checks the use of CHECK and EXPECT macros.
4035
4036 Args:
4037 filename: The name of the current file.
4038 clean_lines: A CleansedLines instance containing the file.
4039 linenum: The number of the line to check.
4040 error: The function to call with any errors found.
4041 """
4042
4043 # Decide the set of replacement macros that should be suggested
erg@google.comc6671232013-10-25 21:44:03 +00004044 lines = clean_lines.elided
avakulenko@google.com02af6282014-06-04 18:53:25 +00004045 (check_macro, start_pos) = FindCheckMacro(lines[linenum])
4046 if not check_macro:
erg@google.com4e00b9a2009-01-12 23:05:11 +00004047 return
4048
erg@google.comc6671232013-10-25 21:44:03 +00004049 # Find end of the boolean expression by matching parentheses
4050 (last_line, end_line, end_pos) = CloseExpression(
4051 clean_lines, linenum, start_pos)
4052 if end_pos < 0:
4053 return
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004054
4055 # If the check macro is followed by something other than a
4056 # semicolon, assume users will log their own custom error messages
4057 # and don't suggest any replacements.
4058 if not Match(r'\s*;', last_line[end_pos:]):
4059 return
4060
erg@google.comc6671232013-10-25 21:44:03 +00004061 if linenum == end_line:
4062 expression = lines[linenum][start_pos + 1:end_pos - 1]
4063 else:
4064 expression = lines[linenum][start_pos + 1:]
4065 for i in xrange(linenum + 1, end_line):
4066 expression += lines[i]
4067 expression += last_line[0:end_pos - 1]
erg@google.com4e00b9a2009-01-12 23:05:11 +00004068
erg@google.comc6671232013-10-25 21:44:03 +00004069 # Parse expression so that we can take parentheses into account.
4070 # This avoids false positives for inputs like "CHECK((a < 4) == b)",
4071 # which is not replaceable by CHECK_LE.
4072 lhs = ''
4073 rhs = ''
4074 operator = None
4075 while expression:
4076 matched = Match(r'^\s*(<<|<<=|>>|>>=|->\*|->|&&|\|\||'
4077 r'==|!=|>=|>|<=|<|\()(.*)$', expression)
4078 if matched:
4079 token = matched.group(1)
4080 if token == '(':
4081 # Parenthesized operand
4082 expression = matched.group(2)
avakulenko@google.com02af6282014-06-04 18:53:25 +00004083 (end, _) = FindEndOfExpressionInLine(expression, 0, ['('])
erg@google.comc6671232013-10-25 21:44:03 +00004084 if end < 0:
4085 return # Unmatched parenthesis
4086 lhs += '(' + expression[0:end]
4087 expression = expression[end:]
4088 elif token in ('&&', '||'):
4089 # Logical and/or operators. This means the expression
4090 # contains more than one term, for example:
4091 # CHECK(42 < a && a < b);
4092 #
4093 # These are not replaceable with CHECK_LE, so bail out early.
4094 return
4095 elif token in ('<<', '<<=', '>>', '>>=', '->*', '->'):
4096 # Non-relational operator
4097 lhs += token
4098 expression = matched.group(2)
4099 else:
4100 # Relational operator
4101 operator = token
4102 rhs = matched.group(2)
4103 break
4104 else:
4105 # Unparenthesized operand. Instead of appending to lhs one character
4106 # at a time, we do another regular expression match to consume several
4107 # characters at once if possible. Trivial benchmark shows that this
4108 # is more efficient when the operands are longer than a single
4109 # character, which is generally the case.
4110 matched = Match(r'^([^-=!<>()&|]+)(.*)$', expression)
4111 if not matched:
4112 matched = Match(r'^(\s*\S)(.*)$', expression)
4113 if not matched:
4114 break
4115 lhs += matched.group(1)
4116 expression = matched.group(2)
4117
4118 # Only apply checks if we got all parts of the boolean expression
4119 if not (lhs and operator and rhs):
4120 return
4121
4122 # Check that rhs do not contain logical operators. We already know
4123 # that lhs is fine since the loop above parses out && and ||.
4124 if rhs.find('&&') > -1 or rhs.find('||') > -1:
4125 return
4126
4127 # At least one of the operands must be a constant literal. This is
4128 # to avoid suggesting replacements for unprintable things like
4129 # CHECK(variable != iterator)
4130 #
4131 # The following pattern matches decimal, hex integers, strings, and
4132 # characters (in that order).
4133 lhs = lhs.strip()
4134 rhs = rhs.strip()
4135 match_constant = r'^([-+]?(\d+|0[xX][0-9a-fA-F]+)[lLuU]{0,3}|".*"|\'.*\')$'
4136 if Match(match_constant, lhs) or Match(match_constant, rhs):
4137 # Note: since we know both lhs and rhs, we can provide a more
4138 # descriptive error message like:
4139 # Consider using CHECK_EQ(x, 42) instead of CHECK(x == 42)
4140 # Instead of:
4141 # Consider using CHECK_EQ instead of CHECK(a == b)
4142 #
4143 # We are still keeping the less descriptive message because if lhs
4144 # or rhs gets long, the error message might become unreadable.
4145 error(filename, linenum, 'readability/check', 2,
4146 'Consider using %s instead of %s(a %s b)' % (
4147 _CHECK_REPLACEMENT[check_macro][operator],
4148 check_macro, operator))
erg@google.com4e00b9a2009-01-12 23:05:11 +00004149
4150
erg@google.comd350fe52013-01-14 17:51:48 +00004151def CheckAltTokens(filename, clean_lines, linenum, error):
4152 """Check alternative keywords being used in boolean expressions.
4153
4154 Args:
4155 filename: The name of the current file.
4156 clean_lines: A CleansedLines instance containing the file.
4157 linenum: The number of the line to check.
4158 error: The function to call with any errors found.
4159 """
4160 line = clean_lines.elided[linenum]
4161
4162 # Avoid preprocessor lines
4163 if Match(r'^\s*#', line):
4164 return
4165
4166 # Last ditch effort to avoid multi-line comments. This will not help
4167 # if the comment started before the current line or ended after the
4168 # current line, but it catches most of the false positives. At least,
4169 # it provides a way to workaround this warning for people who use
4170 # multi-line comments in preprocessor macros.
4171 #
4172 # TODO(unknown): remove this once cpplint has better support for
4173 # multi-line comments.
4174 if line.find('/*') >= 0 or line.find('*/') >= 0:
4175 return
4176
4177 for match in _ALT_TOKEN_REPLACEMENT_PATTERN.finditer(line):
4178 error(filename, linenum, 'readability/alt_tokens', 2,
4179 'Use operator %s instead of %s' % (
4180 _ALT_TOKEN_REPLACEMENT[match.group(1)], match.group(1)))
4181
4182
erg@google.com4e00b9a2009-01-12 23:05:11 +00004183def GetLineWidth(line):
4184 """Determines the width of the line in column positions.
4185
4186 Args:
4187 line: A string, which may be a Unicode string.
4188
4189 Returns:
4190 The width of the line in column positions, accounting for Unicode
4191 combining characters and wide characters.
4192 """
4193 if isinstance(line, unicode):
4194 width = 0
erg@google.com8a95ecc2011-09-08 00:45:54 +00004195 for uc in unicodedata.normalize('NFC', line):
4196 if unicodedata.east_asian_width(uc) in ('W', 'F'):
erg@google.com4e00b9a2009-01-12 23:05:11 +00004197 width += 2
erg@google.com8a95ecc2011-09-08 00:45:54 +00004198 elif not unicodedata.combining(uc):
erg@google.com4e00b9a2009-01-12 23:05:11 +00004199 width += 1
4200 return width
4201 else:
4202 return len(line)
4203
4204
erg@google.comd350fe52013-01-14 17:51:48 +00004205def CheckStyle(filename, clean_lines, linenum, file_extension, nesting_state,
erg@google.com8a95ecc2011-09-08 00:45:54 +00004206 error):
erg@google.com4e00b9a2009-01-12 23:05:11 +00004207 """Checks rules from the 'C++ style rules' section of cppguide.html.
4208
4209 Most of these rules are hard to test (naming, comment style), but we
4210 do what we can. In particular we check for 2-space indents, line lengths,
4211 tab usage, spaces inside code, etc.
4212
4213 Args:
4214 filename: The name of the current file.
4215 clean_lines: A CleansedLines instance containing the file.
4216 linenum: The number of the line to check.
4217 file_extension: The extension (without the dot) of the filename.
avakulenko@google.com02af6282014-06-04 18:53:25 +00004218 nesting_state: A NestingState instance which maintains information about
erg@google.comd350fe52013-01-14 17:51:48 +00004219 the current stack of nested blocks being parsed.
erg@google.com4e00b9a2009-01-12 23:05:11 +00004220 error: The function to call with any errors found.
4221 """
4222
erg@google.com2aa59982013-10-28 19:09:25 +00004223 # Don't use "elided" lines here, otherwise we can't check commented lines.
4224 # Don't want to use "raw" either, because we don't want to check inside C++11
4225 # raw strings,
4226 raw_lines = clean_lines.lines_without_raw_strings
erg@google.com4e00b9a2009-01-12 23:05:11 +00004227 line = raw_lines[linenum]
Alex Vakulenko01e47232016-05-06 13:54:15 -07004228 prev = raw_lines[linenum - 1] if linenum > 0 else ''
erg@google.com4e00b9a2009-01-12 23:05:11 +00004229
4230 if line.find('\t') != -1:
4231 error(filename, linenum, 'whitespace/tab', 1,
4232 'Tab found; better to use spaces')
4233
4234 # One or three blank spaces at the beginning of the line is weird; it's
4235 # hard to reconcile that with 2-space indents.
4236 # NOTE: here are the conditions rob pike used for his tests. Mine aren't
4237 # as sophisticated, but it may be worth becoming so: RLENGTH==initial_spaces
4238 # if(RLENGTH > 20) complain = 0;
4239 # if(match($0, " +(error|private|public|protected):")) complain = 0;
4240 # if(match(prev, "&& *$")) complain = 0;
4241 # if(match(prev, "\\|\\| *$")) complain = 0;
4242 # if(match(prev, "[\",=><] *$")) complain = 0;
4243 # if(match($0, " <<")) complain = 0;
4244 # if(match(prev, " +for \\(")) complain = 0;
4245 # if(prevodd && match(prevprev, " +for \\(")) complain = 0;
avakulenko@google.com02af6282014-06-04 18:53:25 +00004246 scope_or_label_pattern = r'\s*\w+\s*:\s*\\?$'
4247 classinfo = nesting_state.InnermostClass()
erg@google.com4e00b9a2009-01-12 23:05:11 +00004248 initial_spaces = 0
4249 cleansed_line = clean_lines.elided[linenum]
4250 while initial_spaces < len(line) and line[initial_spaces] == ' ':
4251 initial_spaces += 1
avakulenko@google.com02af6282014-06-04 18:53:25 +00004252 # There are certain situations we allow one space, notably for
4253 # section labels, and also lines containing multi-line raw strings.
Alex Vakulenko01e47232016-05-06 13:54:15 -07004254 # We also don't check for lines that look like continuation lines
4255 # (of lines ending in double quotes, commas, equals, or angle brackets)
4256 # because the rules for how to indent those are non-trivial.
4257 if (not Search(r'[",=><] *$', prev) and
4258 (initial_spaces == 1 or initial_spaces == 3) and
4259 not Match(scope_or_label_pattern, cleansed_line) and
4260 not (clean_lines.raw_lines[linenum] != line and
4261 Match(r'^\s*""', line))):
erg@google.com4e00b9a2009-01-12 23:05:11 +00004262 error(filename, linenum, 'whitespace/indent', 3,
4263 'Weird number of spaces at line-start. '
4264 'Are you using a 2-space indent?')
erg@google.com4e00b9a2009-01-12 23:05:11 +00004265
Alex Vakulenko01e47232016-05-06 13:54:15 -07004266 if line and line[-1].isspace():
4267 error(filename, linenum, 'whitespace/end_of_line', 4,
4268 'Line ends in whitespace. Consider deleting these extra spaces.')
4269
erg@google.com4e00b9a2009-01-12 23:05:11 +00004270 # Check if the line is a header guard.
4271 is_header_guard = False
LukeCz7197a242016-09-24 13:27:35 -05004272 if IsHeaderExtension(file_extension):
erg@google.com4e00b9a2009-01-12 23:05:11 +00004273 cppvar = GetHeaderGuardCPPVariable(filename)
4274 if (line.startswith('#ifndef %s' % cppvar) or
4275 line.startswith('#define %s' % cppvar) or
4276 line.startswith('#endif // %s' % cppvar)):
4277 is_header_guard = True
4278 # #include lines and header guards can be long, since there's no clean way to
4279 # split them.
erg@google.coma87abb82009-02-24 01:41:01 +00004280 #
4281 # URLs can be long too. It's possible to split these, but it makes them
4282 # harder to cut&paste.
erg@google.comd7d27472011-09-07 17:36:35 +00004283 #
4284 # The "$Id:...$" comment may also get very long without it being the
4285 # developers fault.
erg@google.coma87abb82009-02-24 01:41:01 +00004286 if (not line.startswith('#include') and not is_header_guard and
erg@google.comd7d27472011-09-07 17:36:35 +00004287 not Match(r'^\s*//.*http(s?)://\S*$', line) and
Alex Vakulenko01e47232016-05-06 13:54:15 -07004288 not Match(r'^\s*//\s*[^\s]*$', line) and
erg@google.comd7d27472011-09-07 17:36:35 +00004289 not Match(r'^// \$Id:.*#[0-9]+ \$$', line)):
erg@google.com4e00b9a2009-01-12 23:05:11 +00004290 line_width = GetLineWidth(line)
Alex Vakulenko01e47232016-05-06 13:54:15 -07004291 if line_width > _line_length:
erg@google.com4e00b9a2009-01-12 23:05:11 +00004292 error(filename, linenum, 'whitespace/line_length', 2,
erg@google.comab53edf2013-11-05 22:23:37 +00004293 'Lines should be <= %i characters long' % _line_length)
erg@google.com4e00b9a2009-01-12 23:05:11 +00004294
4295 if (cleansed_line.count(';') > 1 and
4296 # for loops are allowed two ;'s (and may run over two lines).
4297 cleansed_line.find('for') == -1 and
4298 (GetPreviousNonBlankLine(clean_lines, linenum)[0].find('for') == -1 or
4299 GetPreviousNonBlankLine(clean_lines, linenum)[0].find(';') != -1) and
4300 # It's ok to have many commands in a switch case that fits in 1 line
4301 not ((cleansed_line.find('case ') != -1 or
4302 cleansed_line.find('default:') != -1) and
4303 cleansed_line.find('break;') != -1)):
erg@google.comd350fe52013-01-14 17:51:48 +00004304 error(filename, linenum, 'whitespace/newline', 0,
erg@google.com4e00b9a2009-01-12 23:05:11 +00004305 'More than one command on the same line')
4306
4307 # Some more style checks
4308 CheckBraces(filename, clean_lines, linenum, error)
avakulenko@google.com02af6282014-06-04 18:53:25 +00004309 CheckTrailingSemicolon(filename, clean_lines, linenum, error)
erg@google.comc6671232013-10-25 21:44:03 +00004310 CheckEmptyBlockBody(filename, clean_lines, linenum, error)
erg@google.comd350fe52013-01-14 17:51:48 +00004311 CheckSpacing(filename, clean_lines, linenum, nesting_state, error)
avakulenko@google.com02af6282014-06-04 18:53:25 +00004312 CheckOperatorSpacing(filename, clean_lines, linenum, error)
4313 CheckParenthesisSpacing(filename, clean_lines, linenum, error)
4314 CheckCommaSpacing(filename, clean_lines, linenum, error)
Alex Vakulenko01e47232016-05-06 13:54:15 -07004315 CheckBracesSpacing(filename, clean_lines, linenum, nesting_state, error)
avakulenko@google.com02af6282014-06-04 18:53:25 +00004316 CheckSpacingForFunctionCall(filename, clean_lines, linenum, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00004317 CheckCheck(filename, clean_lines, linenum, error)
erg@google.comd350fe52013-01-14 17:51:48 +00004318 CheckAltTokens(filename, clean_lines, linenum, error)
4319 classinfo = nesting_state.InnermostClass()
4320 if classinfo:
4321 CheckSectionSpacing(filename, clean_lines, classinfo, linenum, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00004322
4323
erg@google.com4e00b9a2009-01-12 23:05:11 +00004324_RE_PATTERN_INCLUDE = re.compile(r'^\s*#\s*include\s*([<"])([^>"]*)[>"].*$')
4325# Matches the first component of a filename delimited by -s and _s. That is:
4326# _RE_FIRST_COMPONENT.match('foo').group(0) == 'foo'
4327# _RE_FIRST_COMPONENT.match('foo.cc').group(0) == 'foo'
4328# _RE_FIRST_COMPONENT.match('foo-bar_baz.cc').group(0) == 'foo'
4329# _RE_FIRST_COMPONENT.match('foo_bar-baz.cc').group(0) == 'foo'
4330_RE_FIRST_COMPONENT = re.compile(r'^[^-_.]+')
4331
4332
4333def _DropCommonSuffixes(filename):
4334 """Drops common suffixes like _test.cc or -inl.h from filename.
4335
4336 For example:
4337 >>> _DropCommonSuffixes('foo/foo-inl.h')
4338 'foo/foo'
4339 >>> _DropCommonSuffixes('foo/bar/foo.cc')
4340 'foo/bar/foo'
4341 >>> _DropCommonSuffixes('foo/foo_internal.h')
4342 'foo/foo'
4343 >>> _DropCommonSuffixes('foo/foo_unusualinternal.h')
4344 'foo/foo_unusualinternal'
4345
4346 Args:
4347 filename: The input filename.
4348
4349 Returns:
4350 The filename with the common suffix removed.
4351 """
4352 for suffix in ('test.cc', 'regtest.cc', 'unittest.cc',
4353 'inl.h', 'impl.h', 'internal.h'):
4354 if (filename.endswith(suffix) and len(filename) > len(suffix) and
4355 filename[-len(suffix) - 1] in ('-', '_')):
4356 return filename[:-len(suffix) - 1]
4357 return os.path.splitext(filename)[0]
4358
4359
erg@google.com4e00b9a2009-01-12 23:05:11 +00004360def _ClassifyInclude(fileinfo, include, is_system):
4361 """Figures out what kind of header 'include' is.
4362
4363 Args:
4364 fileinfo: The current file cpplint is running over. A FileInfo instance.
4365 include: The path to a #included file.
4366 is_system: True if the #include used <> rather than "".
4367
4368 Returns:
4369 One of the _XXX_HEADER constants.
4370
4371 For example:
4372 >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'stdio.h', True)
4373 _C_SYS_HEADER
4374 >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'string', True)
4375 _CPP_SYS_HEADER
4376 >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/foo.h', False)
4377 _LIKELY_MY_HEADER
4378 >>> _ClassifyInclude(FileInfo('foo/foo_unknown_extension.cc'),
4379 ... 'bar/foo_other_ext.h', False)
4380 _POSSIBLE_MY_HEADER
4381 >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/bar.h', False)
4382 _OTHER_HEADER
4383 """
4384 # This is a list of all standard c++ header files, except
4385 # those already checked for above.
erg@google.comfd5da632013-10-25 17:39:45 +00004386 is_cpp_h = include in _CPP_HEADERS
erg@google.com4e00b9a2009-01-12 23:05:11 +00004387
4388 if is_system:
4389 if is_cpp_h:
4390 return _CPP_SYS_HEADER
4391 else:
4392 return _C_SYS_HEADER
4393
4394 # If the target file and the include we're checking share a
4395 # basename when we drop common extensions, and the include
4396 # lives in . , then it's likely to be owned by the target file.
4397 target_dir, target_base = (
4398 os.path.split(_DropCommonSuffixes(fileinfo.RepositoryName())))
4399 include_dir, include_base = os.path.split(_DropCommonSuffixes(include))
4400 if target_base == include_base and (
4401 include_dir == target_dir or
4402 include_dir == os.path.normpath(target_dir + '/../public')):
4403 return _LIKELY_MY_HEADER
4404
4405 # If the target and include share some initial basename
4406 # component, it's possible the target is implementing the
4407 # include, so it's allowed to be first, but we'll never
4408 # complain if it's not there.
4409 target_first_component = _RE_FIRST_COMPONENT.match(target_base)
4410 include_first_component = _RE_FIRST_COMPONENT.match(include_base)
4411 if (target_first_component and include_first_component and
4412 target_first_component.group(0) ==
4413 include_first_component.group(0)):
4414 return _POSSIBLE_MY_HEADER
4415
4416 return _OTHER_HEADER
4417
4418
erg@google.coma87abb82009-02-24 01:41:01 +00004419
erg@google.come35f7652009-06-19 20:52:09 +00004420def CheckIncludeLine(filename, clean_lines, linenum, include_state, error):
4421 """Check rules that are applicable to #include lines.
erg@google.com4e00b9a2009-01-12 23:05:11 +00004422
erg@google.come35f7652009-06-19 20:52:09 +00004423 Strings on #include lines are NOT removed from elided line, to make
4424 certain tasks easier. However, to prevent false positives, checks
4425 applicable to #include lines in CheckLanguage must be put here.
erg@google.com4e00b9a2009-01-12 23:05:11 +00004426
4427 Args:
4428 filename: The name of the current file.
4429 clean_lines: A CleansedLines instance containing the file.
4430 linenum: The number of the line to check.
erg@google.com4e00b9a2009-01-12 23:05:11 +00004431 include_state: An _IncludeState instance in which the headers are inserted.
4432 error: The function to call with any errors found.
4433 """
4434 fileinfo = FileInfo(filename)
erg@google.come35f7652009-06-19 20:52:09 +00004435 line = clean_lines.lines[linenum]
erg@google.com4e00b9a2009-01-12 23:05:11 +00004436
4437 # "include" should use the new style "foo/bar.h" instead of just "bar.h"
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004438 # Only do this check if the included header follows google naming
4439 # conventions. If not, assume that it's a 3rd party API that
4440 # requires special include conventions.
4441 #
4442 # We also make an exception for Lua headers, which follow google
4443 # naming convention but not the include convention.
4444 match = Match(r'#include\s*"([^/]+\.h)"', line)
4445 if match and not _THIRD_PARTY_HEADERS_PATTERN.match(match.group(1)):
erg@google.com4e00b9a2009-01-12 23:05:11 +00004446 error(filename, linenum, 'build/include', 4,
4447 'Include the directory when naming .h files')
4448
4449 # we shouldn't include a file more than once. actually, there are a
4450 # handful of instances where doing so is okay, but in general it's
4451 # not.
erg@google.come35f7652009-06-19 20:52:09 +00004452 match = _RE_PATTERN_INCLUDE.search(line)
erg@google.com4e00b9a2009-01-12 23:05:11 +00004453 if match:
4454 include = match.group(2)
4455 is_system = (match.group(1) == '<')
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004456 duplicate_line = include_state.FindHeader(include)
4457 if duplicate_line >= 0:
erg@google.com4e00b9a2009-01-12 23:05:11 +00004458 error(filename, linenum, 'build/include', 4,
4459 '"%s" already included at %s:%s' %
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004460 (include, filename, duplicate_line))
avakulenko@google.com554223d2014-12-04 22:00:20 +00004461 elif (include.endswith('.cc') and
4462 os.path.dirname(fileinfo.RepositoryName()) != os.path.dirname(include)):
4463 error(filename, linenum, 'build/include', 4,
4464 'Do not include .cc files from other packages')
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004465 elif not _THIRD_PARTY_HEADERS_PATTERN.match(include):
4466 include_state.include_list[-1].append((include, linenum))
erg@google.com4e00b9a2009-01-12 23:05:11 +00004467
4468 # We want to ensure that headers appear in the right order:
4469 # 1) for foo.cc, foo.h (preferred location)
4470 # 2) c system files
4471 # 3) cpp system files
4472 # 4) for foo.cc, foo.h (deprecated location)
4473 # 5) other google headers
4474 #
4475 # We classify each include statement as one of those 5 types
4476 # using a number of techniques. The include_state object keeps
4477 # track of the highest type seen, and complains if we see a
4478 # lower type after that.
4479 error_message = include_state.CheckNextIncludeOrder(
4480 _ClassifyInclude(fileinfo, include, is_system))
4481 if error_message:
4482 error(filename, linenum, 'build/include_order', 4,
4483 '%s. Should be: %s.h, c system, c++ system, other.' %
4484 (error_message, fileinfo.BaseName()))
erg@google.comfd5da632013-10-25 17:39:45 +00004485 canonical_include = include_state.CanonicalizeAlphabeticalOrder(include)
4486 if not include_state.IsInAlphabeticalOrder(
4487 clean_lines, linenum, canonical_include):
erg@google.coma868d2d2009-10-09 21:18:45 +00004488 error(filename, linenum, 'build/include_alpha', 4,
4489 'Include "%s" not in alphabetical order' % include)
erg@google.comfd5da632013-10-25 17:39:45 +00004490 include_state.SetLastHeader(canonical_include)
erg@google.com4e00b9a2009-01-12 23:05:11 +00004491
erg@google.come35f7652009-06-19 20:52:09 +00004492
erg@google.com8a95ecc2011-09-08 00:45:54 +00004493
4494def _GetTextInside(text, start_pattern):
erg@google.com2aa59982013-10-28 19:09:25 +00004495 r"""Retrieves all the text between matching open and close parentheses.
erg@google.com8a95ecc2011-09-08 00:45:54 +00004496
4497 Given a string of lines and a regular expression string, retrieve all the text
4498 following the expression and between opening punctuation symbols like
4499 (, [, or {, and the matching close-punctuation symbol. This properly nested
4500 occurrences of the punctuations, so for the text like
4501 printf(a(), b(c()));
4502 a call to _GetTextInside(text, r'printf\(') will return 'a(), b(c())'.
4503 start_pattern must match string having an open punctuation symbol at the end.
4504
4505 Args:
4506 text: The lines to extract text. Its comments and strings must be elided.
4507 It can be single line and can span multiple lines.
4508 start_pattern: The regexp string indicating where to start extracting
4509 the text.
4510 Returns:
4511 The extracted text.
4512 None if either the opening string or ending punctuation could not be found.
4513 """
avakulenko@google.com02af6282014-06-04 18:53:25 +00004514 # TODO(unknown): Audit cpplint.py to see what places could be profitably
erg@google.com8a95ecc2011-09-08 00:45:54 +00004515 # rewritten to use _GetTextInside (and use inferior regexp matching today).
4516
4517 # Give opening punctuations to get the matching close-punctuations.
4518 matching_punctuation = {'(': ')', '{': '}', '[': ']'}
4519 closing_punctuation = set(matching_punctuation.itervalues())
4520
4521 # Find the position to start extracting text.
4522 match = re.search(start_pattern, text, re.M)
4523 if not match: # start_pattern not found in text.
4524 return None
4525 start_position = match.end(0)
4526
4527 assert start_position > 0, (
4528 'start_pattern must ends with an opening punctuation.')
4529 assert text[start_position - 1] in matching_punctuation, (
4530 'start_pattern must ends with an opening punctuation.')
4531 # Stack of closing punctuations we expect to have in text after position.
4532 punctuation_stack = [matching_punctuation[text[start_position - 1]]]
4533 position = start_position
4534 while punctuation_stack and position < len(text):
4535 if text[position] == punctuation_stack[-1]:
4536 punctuation_stack.pop()
4537 elif text[position] in closing_punctuation:
4538 # A closing punctuation without matching opening punctuations.
4539 return None
4540 elif text[position] in matching_punctuation:
4541 punctuation_stack.append(matching_punctuation[text[position]])
4542 position += 1
4543 if punctuation_stack:
4544 # Opening punctuations left without matching close-punctuations.
4545 return None
4546 # punctuations match.
4547 return text[start_position:position - 1]
4548
4549
erg@google.comfd5da632013-10-25 17:39:45 +00004550# Patterns for matching call-by-reference parameters.
erg@google.com2aa59982013-10-28 19:09:25 +00004551#
4552# Supports nested templates up to 2 levels deep using this messy pattern:
4553# < (?: < (?: < [^<>]*
4554# >
4555# | [^<>] )*
4556# >
4557# | [^<>] )*
4558# >
erg@google.comfd5da632013-10-25 17:39:45 +00004559_RE_PATTERN_IDENT = r'[_a-zA-Z]\w*' # =~ [[:alpha:]][[:alnum:]]*
4560_RE_PATTERN_TYPE = (
4561 r'(?:const\s+)?(?:typename\s+|class\s+|struct\s+|union\s+|enum\s+)?'
erg@google.com2aa59982013-10-28 19:09:25 +00004562 r'(?:\w|'
4563 r'\s*<(?:<(?:<[^<>]*>|[^<>])*>|[^<>])*>|'
4564 r'::)+')
erg@google.comfd5da632013-10-25 17:39:45 +00004565# A call-by-reference parameter ends with '& identifier'.
4566_RE_PATTERN_REF_PARAM = re.compile(
4567 r'(' + _RE_PATTERN_TYPE + r'(?:\s*(?:\bconst\b|[*]))*\s*'
4568 r'&\s*' + _RE_PATTERN_IDENT + r')\s*(?:=[^,()]+)?[,)]')
4569# A call-by-const-reference parameter either ends with 'const& identifier'
4570# or looks like 'const type& identifier' when 'type' is atomic.
4571_RE_PATTERN_CONST_REF_PARAM = (
4572 r'(?:.*\s*\bconst\s*&\s*' + _RE_PATTERN_IDENT +
4573 r'|const\s+' + _RE_PATTERN_TYPE + r'\s*&\s*' + _RE_PATTERN_IDENT + r')')
Alex Vakulenko01e47232016-05-06 13:54:15 -07004574# Stream types.
4575_RE_PATTERN_REF_STREAM_PARAM = (
4576 r'(?:.*stream\s*&\s*' + _RE_PATTERN_IDENT + r')')
erg@google.comfd5da632013-10-25 17:39:45 +00004577
4578
4579def CheckLanguage(filename, clean_lines, linenum, file_extension,
4580 include_state, nesting_state, error):
erg@google.come35f7652009-06-19 20:52:09 +00004581 """Checks rules from the 'C++ language rules' section of cppguide.html.
4582
4583 Some of these rules are hard to test (function overloading, using
4584 uint32 inappropriately), but we do the best we can.
4585
4586 Args:
4587 filename: The name of the current file.
4588 clean_lines: A CleansedLines instance containing the file.
4589 linenum: The number of the line to check.
4590 file_extension: The extension (without the dot) of the filename.
4591 include_state: An _IncludeState instance in which the headers are inserted.
avakulenko@google.com02af6282014-06-04 18:53:25 +00004592 nesting_state: A NestingState instance which maintains information about
erg@google.comfd5da632013-10-25 17:39:45 +00004593 the current stack of nested blocks being parsed.
erg@google.come35f7652009-06-19 20:52:09 +00004594 error: The function to call with any errors found.
4595 """
erg@google.com4e00b9a2009-01-12 23:05:11 +00004596 # If the line is empty or consists of entirely a comment, no need to
4597 # check it.
4598 line = clean_lines.elided[linenum]
4599 if not line:
4600 return
4601
erg@google.come35f7652009-06-19 20:52:09 +00004602 match = _RE_PATTERN_INCLUDE.search(line)
4603 if match:
4604 CheckIncludeLine(filename, clean_lines, linenum, include_state, error)
4605 return
4606
erg@google.com2aa59982013-10-28 19:09:25 +00004607 # Reset include state across preprocessor directives. This is meant
4608 # to silence warnings for conditional includes.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004609 match = Match(r'^\s*#\s*(if|ifdef|ifndef|elif|else|endif)\b', line)
4610 if match:
4611 include_state.ResetSection(match.group(1))
erg@google.com2aa59982013-10-28 19:09:25 +00004612
erg@google.com4e00b9a2009-01-12 23:05:11 +00004613 # Make Windows paths like Unix.
4614 fullname = os.path.abspath(filename).replace('\\', '/')
Alex Vakulenko01e47232016-05-06 13:54:15 -07004615
avakulenko@google.com02af6282014-06-04 18:53:25 +00004616 # Perform other checks now that we are sure that this is not an include line
4617 CheckCasts(filename, clean_lines, linenum, error)
4618 CheckGlobalStatic(filename, clean_lines, linenum, error)
4619 CheckPrintf(filename, clean_lines, linenum, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00004620
LukeCz7197a242016-09-24 13:27:35 -05004621 if IsHeaderExtension(file_extension):
erg@google.com4e00b9a2009-01-12 23:05:11 +00004622 # TODO(unknown): check that 1-arg constructors are explicit.
4623 # How to tell it's a constructor?
4624 # (handled in CheckForNonStandardConstructs for now)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004625 # TODO(unknown): check that classes declare or disable copy/assign
erg@google.com4e00b9a2009-01-12 23:05:11 +00004626 # (level 1 error)
4627 pass
4628
4629 # Check if people are using the verboten C basic types. The only exception
4630 # we regularly allow is "unsigned short port" for port.
4631 if Search(r'\bshort port\b', line):
4632 if not Search(r'\bunsigned short port\b', line):
4633 error(filename, linenum, 'runtime/int', 4,
4634 'Use "unsigned short" for ports, not "short"')
4635 else:
4636 match = Search(r'\b(short|long(?! +double)|long long)\b', line)
4637 if match:
4638 error(filename, linenum, 'runtime/int', 4,
4639 'Use int16/int64/etc, rather than the C type %s' % match.group(1))
4640
erg@google.coma868d2d2009-10-09 21:18:45 +00004641 # Check if some verboten operator overloading is going on
4642 # TODO(unknown): catch out-of-line unary operator&:
4643 # class X {};
4644 # int operator&(const X& x) { return 42; } // unary operator&
4645 # The trick is it's hard to tell apart from binary operator&:
4646 # class Y { int operator&(const Y& x) { return 23; } }; // binary operator&
4647 if Search(r'\boperator\s*&\s*\(\s*\)', line):
4648 error(filename, linenum, 'runtime/operator', 4,
4649 'Unary operator& is dangerous. Do not use it.')
4650
erg@google.com4e00b9a2009-01-12 23:05:11 +00004651 # Check for suspicious usage of "if" like
4652 # } if (a == b) {
4653 if Search(r'\}\s*if\s*\(', line):
4654 error(filename, linenum, 'readability/braces', 4,
4655 'Did you mean "else if"? If not, start a new line for "if".')
4656
4657 # Check for potential format string bugs like printf(foo).
4658 # We constrain the pattern not to pick things like DocidForPrintf(foo).
4659 # Not perfect but it can catch printf(foo.c_str()) and printf(foo->c_str())
avakulenko@google.com02af6282014-06-04 18:53:25 +00004660 # TODO(unknown): Catch the following case. Need to change the calling
erg@google.com8a95ecc2011-09-08 00:45:54 +00004661 # convention of the whole function to process multiple line to handle it.
4662 # printf(
4663 # boy_this_is_a_really_long_variable_that_cannot_fit_on_the_prev_line);
4664 printf_args = _GetTextInside(line, r'(?i)\b(string)?printf\s*\(')
4665 if printf_args:
4666 match = Match(r'([\w.\->()]+)$', printf_args)
erg@google.comd350fe52013-01-14 17:51:48 +00004667 if match and match.group(1) != '__VA_ARGS__':
erg@google.com8a95ecc2011-09-08 00:45:54 +00004668 function_name = re.search(r'\b((?:string)?printf)\s*\(',
4669 line, re.I).group(1)
4670 error(filename, linenum, 'runtime/printf', 4,
4671 'Potential format string bug. Do %s("%%s", %s) instead.'
4672 % (function_name, match.group(1)))
erg@google.com4e00b9a2009-01-12 23:05:11 +00004673
4674 # Check for potential memset bugs like memset(buf, sizeof(buf), 0).
4675 match = Search(r'memset\s*\(([^,]*),\s*([^,]*),\s*0\s*\)', line)
4676 if match and not Match(r"^''|-?[0-9]+|0x[0-9A-Fa-f]$", match.group(2)):
4677 error(filename, linenum, 'runtime/memset', 4,
4678 'Did you mean "memset(%s, 0, %s)"?'
4679 % (match.group(1), match.group(2)))
4680
4681 if Search(r'\busing namespace\b', line):
4682 error(filename, linenum, 'build/namespaces', 5,
4683 'Do not use namespace using-directives. '
4684 'Use using-declarations instead.')
4685
4686 # Detect variable-length arrays.
4687 match = Match(r'\s*(.+::)?(\w+) [a-z]\w*\[(.+)];', line)
4688 if (match and match.group(2) != 'return' and match.group(2) != 'delete' and
4689 match.group(3).find(']') == -1):
4690 # Split the size using space and arithmetic operators as delimiters.
4691 # If any of the resulting tokens are not compile time constants then
4692 # report the error.
4693 tokens = re.split(r'\s|\+|\-|\*|\/|<<|>>]', match.group(3))
4694 is_const = True
4695 skip_next = False
4696 for tok in tokens:
4697 if skip_next:
4698 skip_next = False
4699 continue
4700
4701 if Search(r'sizeof\(.+\)', tok): continue
4702 if Search(r'arraysize\(\w+\)', tok): continue
4703
4704 tok = tok.lstrip('(')
4705 tok = tok.rstrip(')')
4706 if not tok: continue
4707 if Match(r'\d+', tok): continue
4708 if Match(r'0[xX][0-9a-fA-F]+', tok): continue
4709 if Match(r'k[A-Z0-9]\w*', tok): continue
4710 if Match(r'(.+::)?k[A-Z0-9]\w*', tok): continue
4711 if Match(r'(.+::)?[A-Z][A-Z0-9_]*', tok): continue
4712 # A catch all for tricky sizeof cases, including 'sizeof expression',
4713 # 'sizeof(*type)', 'sizeof(const type)', 'sizeof(struct StructName)'
erg@google.com8a95ecc2011-09-08 00:45:54 +00004714 # requires skipping the next token because we split on ' ' and '*'.
erg@google.com4e00b9a2009-01-12 23:05:11 +00004715 if tok.startswith('sizeof'):
4716 skip_next = True
4717 continue
4718 is_const = False
4719 break
4720 if not is_const:
4721 error(filename, linenum, 'runtime/arrays', 1,
4722 'Do not use variable-length arrays. Use an appropriately named '
4723 "('k' followed by CamelCase) compile-time constant for the size.")
4724
erg@google.com4e00b9a2009-01-12 23:05:11 +00004725 # Check for use of unnamed namespaces in header files. Registration
4726 # macros are typically OK, so we allow use of "namespace {" on lines
4727 # that end with backslashes.
LukeCz7197a242016-09-24 13:27:35 -05004728 if (IsHeaderExtension(file_extension)
erg@google.com4e00b9a2009-01-12 23:05:11 +00004729 and Search(r'\bnamespace\s*{', line)
4730 and line[-1] != '\\'):
4731 error(filename, linenum, 'build/namespaces', 4,
4732 'Do not use unnamed namespaces in header files. See '
Ackermann Yuriy79692902016-04-01 21:41:34 +13004733 'https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Namespaces'
erg@google.com4e00b9a2009-01-12 23:05:11 +00004734 ' for more information.')
4735
avakulenko@google.com02af6282014-06-04 18:53:25 +00004736
4737def CheckGlobalStatic(filename, clean_lines, linenum, error):
4738 """Check for unsafe global or static objects.
4739
4740 Args:
4741 filename: The name of the current file.
4742 clean_lines: A CleansedLines instance containing the file.
4743 linenum: The number of the line to check.
4744 error: The function to call with any errors found.
4745 """
4746 line = clean_lines.elided[linenum]
4747
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004748 # Match two lines at a time to support multiline declarations
4749 if linenum + 1 < clean_lines.NumLines() and not Search(r'[;({]', line):
4750 line += clean_lines.elided[linenum + 1].strip()
4751
avakulenko@google.com02af6282014-06-04 18:53:25 +00004752 # Check for people declaring static/global STL strings at the top level.
4753 # This is dangerous because the C++ language does not guarantee that
Alex Vakulenko01e47232016-05-06 13:54:15 -07004754 # globals with constructors are initialized before the first access, and
4755 # also because globals can be destroyed when some threads are still running.
4756 # TODO(unknown): Generalize this to also find static unique_ptr instances.
4757 # TODO(unknown): File bugs for clang-tidy to find these.
avakulenko@google.com02af6282014-06-04 18:53:25 +00004758 match = Match(
Alex Vakulenko01e47232016-05-06 13:54:15 -07004759 r'((?:|static +)(?:|const +))(?::*std::)?string( +const)? +'
4760 r'([a-zA-Z0-9_:]+)\b(.*)',
avakulenko@google.com02af6282014-06-04 18:53:25 +00004761 line)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004762
avakulenko@google.com02af6282014-06-04 18:53:25 +00004763 # Remove false positives:
4764 # - String pointers (as opposed to values).
4765 # string *pointer
4766 # const string *pointer
4767 # string const *pointer
4768 # string *const pointer
4769 #
4770 # - Functions and template specializations.
4771 # string Function<Type>(...
4772 # string Class<Type>::Method(...
4773 #
4774 # - Operators. These are matched separately because operator names
4775 # cross non-word boundaries, and trying to match both operators
4776 # and functions at the same time would decrease accuracy of
4777 # matching identifiers.
4778 # string Class::operator*()
4779 if (match and
Alex Vakulenko01e47232016-05-06 13:54:15 -07004780 not Search(r'\bstring\b(\s+const)?\s*[\*\&]\s*(const\s+)?\w', line) and
avakulenko@google.com02af6282014-06-04 18:53:25 +00004781 not Search(r'\boperator\W', line) and
Alex Vakulenko01e47232016-05-06 13:54:15 -07004782 not Match(r'\s*(<.*>)?(::[a-zA-Z0-9_]+)*\s*\(([^"]|$)', match.group(4))):
4783 if Search(r'\bconst\b', line):
4784 error(filename, linenum, 'runtime/string', 4,
4785 'For a static/global string constant, use a C style string '
4786 'instead: "%schar%s %s[]".' %
4787 (match.group(1), match.group(2) or '', match.group(3)))
4788 else:
4789 error(filename, linenum, 'runtime/string', 4,
4790 'Static/global string variables are not permitted.')
avakulenko@google.com02af6282014-06-04 18:53:25 +00004791
Alex Vakulenko01e47232016-05-06 13:54:15 -07004792 if (Search(r'\b([A-Za-z0-9_]*_)\(\1\)', line) or
4793 Search(r'\b([A-Za-z0-9_]*_)\(CHECK_NOTNULL\(\1\)\)', line)):
avakulenko@google.com02af6282014-06-04 18:53:25 +00004794 error(filename, linenum, 'runtime/init', 4,
4795 'You seem to be initializing a member variable with itself.')
4796
4797
4798def CheckPrintf(filename, clean_lines, linenum, error):
4799 """Check for printf related issues.
4800
4801 Args:
4802 filename: The name of the current file.
4803 clean_lines: A CleansedLines instance containing the file.
4804 linenum: The number of the line to check.
4805 error: The function to call with any errors found.
4806 """
4807 line = clean_lines.elided[linenum]
4808
4809 # When snprintf is used, the second argument shouldn't be a literal.
4810 match = Search(r'snprintf\s*\(([^,]*),\s*([0-9]*)\s*,', line)
4811 if match and match.group(2) != '0':
4812 # If 2nd arg is zero, snprintf is used to calculate size.
4813 error(filename, linenum, 'runtime/printf', 3,
4814 'If you can, use sizeof(%s) instead of %s as the 2nd arg '
4815 'to snprintf.' % (match.group(1), match.group(2)))
4816
4817 # Check if some verboten C functions are being used.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004818 if Search(r'\bsprintf\s*\(', line):
avakulenko@google.com02af6282014-06-04 18:53:25 +00004819 error(filename, linenum, 'runtime/printf', 5,
4820 'Never use sprintf. Use snprintf instead.')
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004821 match = Search(r'\b(strcpy|strcat)\s*\(', line)
avakulenko@google.com02af6282014-06-04 18:53:25 +00004822 if match:
4823 error(filename, linenum, 'runtime/printf', 4,
4824 'Almost always, snprintf is better than %s' % match.group(1))
4825
4826
4827def IsDerivedFunction(clean_lines, linenum):
4828 """Check if current line contains an inherited function.
4829
4830 Args:
4831 clean_lines: A CleansedLines instance containing the file.
4832 linenum: The number of the line to check.
4833 Returns:
4834 True if current line contains a function with "override"
4835 virt-specifier.
4836 """
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004837 # Scan back a few lines for start of current function
4838 for i in xrange(linenum, max(-1, linenum - 10), -1):
4839 match = Match(r'^([^()]*\w+)\(', clean_lines.elided[i])
4840 if match:
4841 # Look for "override" after the matching closing parenthesis
4842 line, _, closing_paren = CloseExpression(
4843 clean_lines, i, len(match.group(1)))
4844 return (closing_paren >= 0 and
4845 Search(r'\boverride\b', line[closing_paren:]))
4846 return False
avakulenko@google.com02af6282014-06-04 18:53:25 +00004847
4848
avakulenko@google.com554223d2014-12-04 22:00:20 +00004849def IsOutOfLineMethodDefinition(clean_lines, linenum):
4850 """Check if current line contains an out-of-line method definition.
4851
4852 Args:
4853 clean_lines: A CleansedLines instance containing the file.
4854 linenum: The number of the line to check.
4855 Returns:
4856 True if current line contains an out-of-line method definition.
4857 """
4858 # Scan back a few lines for start of current function
4859 for i in xrange(linenum, max(-1, linenum - 10), -1):
4860 if Match(r'^([^()]*\w+)\(', clean_lines.elided[i]):
4861 return Match(r'^[^()]*\w+::\w+\(', clean_lines.elided[i]) is not None
4862 return False
4863
4864
avakulenko@google.com02af6282014-06-04 18:53:25 +00004865def IsInitializerList(clean_lines, linenum):
4866 """Check if current line is inside constructor initializer list.
4867
4868 Args:
4869 clean_lines: A CleansedLines instance containing the file.
4870 linenum: The number of the line to check.
4871 Returns:
4872 True if current line appears to be inside constructor initializer
4873 list, False otherwise.
4874 """
4875 for i in xrange(linenum, 1, -1):
4876 line = clean_lines.elided[i]
4877 if i == linenum:
4878 remove_function_body = Match(r'^(.*)\{\s*$', line)
4879 if remove_function_body:
4880 line = remove_function_body.group(1)
4881
4882 if Search(r'\s:\s*\w+[({]', line):
4883 # A lone colon tend to indicate the start of a constructor
4884 # initializer list. It could also be a ternary operator, which
4885 # also tend to appear in constructor initializer lists as
4886 # opposed to parameter lists.
4887 return True
4888 if Search(r'\}\s*,\s*$', line):
4889 # A closing brace followed by a comma is probably the end of a
4890 # brace-initialized member in constructor initializer list.
4891 return True
4892 if Search(r'[{};]\s*$', line):
4893 # Found one of the following:
4894 # - A closing brace or semicolon, probably the end of the previous
4895 # function.
4896 # - An opening brace, probably the start of current class or namespace.
4897 #
4898 # Current line is probably not inside an initializer list since
4899 # we saw one of those things without seeing the starting colon.
4900 return False
4901
4902 # Got to the beginning of the file without seeing the start of
4903 # constructor initializer list.
4904 return False
4905
4906
erg@google.comc6671232013-10-25 21:44:03 +00004907def CheckForNonConstReference(filename, clean_lines, linenum,
4908 nesting_state, error):
4909 """Check for non-const references.
4910
4911 Separate from CheckLanguage since it scans backwards from current
4912 line, instead of scanning forward.
4913
4914 Args:
4915 filename: The name of the current file.
4916 clean_lines: A CleansedLines instance containing the file.
4917 linenum: The number of the line to check.
avakulenko@google.com02af6282014-06-04 18:53:25 +00004918 nesting_state: A NestingState instance which maintains information about
erg@google.comc6671232013-10-25 21:44:03 +00004919 the current stack of nested blocks being parsed.
4920 error: The function to call with any errors found.
4921 """
4922 # Do nothing if there is no '&' on current line.
4923 line = clean_lines.elided[linenum]
4924 if '&' not in line:
4925 return
4926
avakulenko@google.com02af6282014-06-04 18:53:25 +00004927 # If a function is inherited, current function doesn't have much of
4928 # a choice, so any non-const references should not be blamed on
4929 # derived function.
4930 if IsDerivedFunction(clean_lines, linenum):
4931 return
4932
avakulenko@google.com554223d2014-12-04 22:00:20 +00004933 # Don't warn on out-of-line method definitions, as we would warn on the
4934 # in-line declaration, if it isn't marked with 'override'.
4935 if IsOutOfLineMethodDefinition(clean_lines, linenum):
4936 return
4937
erg@google.com2aa59982013-10-28 19:09:25 +00004938 # Long type names may be broken across multiple lines, usually in one
4939 # of these forms:
4940 # LongType
4941 # ::LongTypeContinued &identifier
4942 # LongType::
4943 # LongTypeContinued &identifier
4944 # LongType<
4945 # ...>::LongTypeContinued &identifier
4946 #
4947 # If we detected a type split across two lines, join the previous
4948 # line to current line so that we can match const references
4949 # accordingly.
erg@google.comc6671232013-10-25 21:44:03 +00004950 #
4951 # Note that this only scans back one line, since scanning back
4952 # arbitrary number of lines would be expensive. If you have a type
4953 # that spans more than 2 lines, please use a typedef.
4954 if linenum > 1:
4955 previous = None
erg@google.com2aa59982013-10-28 19:09:25 +00004956 if Match(r'\s*::(?:[\w<>]|::)+\s*&\s*\S', line):
erg@google.comc6671232013-10-25 21:44:03 +00004957 # previous_line\n + ::current_line
erg@google.com2aa59982013-10-28 19:09:25 +00004958 previous = Search(r'\b((?:const\s*)?(?:[\w<>]|::)+[\w<>])\s*$',
erg@google.comc6671232013-10-25 21:44:03 +00004959 clean_lines.elided[linenum - 1])
erg@google.com2aa59982013-10-28 19:09:25 +00004960 elif Match(r'\s*[a-zA-Z_]([\w<>]|::)+\s*&\s*\S', line):
erg@google.comc6671232013-10-25 21:44:03 +00004961 # previous_line::\n + current_line
erg@google.com2aa59982013-10-28 19:09:25 +00004962 previous = Search(r'\b((?:const\s*)?(?:[\w<>]|::)+::)\s*$',
erg@google.comc6671232013-10-25 21:44:03 +00004963 clean_lines.elided[linenum - 1])
4964 if previous:
4965 line = previous.group(1) + line.lstrip()
erg@google.com2aa59982013-10-28 19:09:25 +00004966 else:
4967 # Check for templated parameter that is split across multiple lines
4968 endpos = line.rfind('>')
4969 if endpos > -1:
4970 (_, startline, startpos) = ReverseCloseExpression(
4971 clean_lines, linenum, endpos)
4972 if startpos > -1 and startline < linenum:
4973 # Found the matching < on an earlier line, collect all
4974 # pieces up to current line.
4975 line = ''
4976 for i in xrange(startline, linenum + 1):
4977 line += clean_lines.elided[i].strip()
erg@google.comc6671232013-10-25 21:44:03 +00004978
4979 # Check for non-const references in function parameters. A single '&' may
4980 # found in the following places:
4981 # inside expression: binary & for bitwise AND
4982 # inside expression: unary & for taking the address of something
4983 # inside declarators: reference parameter
4984 # We will exclude the first two cases by checking that we are not inside a
4985 # function body, including one that was just introduced by a trailing '{'.
erg@google.comc6671232013-10-25 21:44:03 +00004986 # TODO(unknown): Doesn't account for 'catch(Exception& e)' [rare].
avakulenko@google.com02af6282014-06-04 18:53:25 +00004987 if (nesting_state.previous_stack_top and
4988 not (isinstance(nesting_state.previous_stack_top, _ClassInfo) or
4989 isinstance(nesting_state.previous_stack_top, _NamespaceInfo))):
4990 # Not at toplevel, not within a class, and not within a namespace
4991 return
4992
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00004993 # Avoid initializer lists. We only need to scan back from the
4994 # current line for something that starts with ':'.
4995 #
4996 # We don't need to check the current line, since the '&' would
4997 # appear inside the second set of parentheses on the current line as
4998 # opposed to the first set.
4999 if linenum > 0:
5000 for i in xrange(linenum - 1, max(0, linenum - 10), -1):
5001 previous_line = clean_lines.elided[i]
5002 if not Search(r'[),]\s*$', previous_line):
5003 break
5004 if Match(r'^\s*:\s+\S', previous_line):
5005 return
5006
avakulenko@google.com02af6282014-06-04 18:53:25 +00005007 # Avoid preprocessors
5008 if Search(r'\\\s*$', line):
5009 return
5010
5011 # Avoid constructor initializer lists
5012 if IsInitializerList(clean_lines, linenum):
5013 return
5014
erg@google.comc6671232013-10-25 21:44:03 +00005015 # We allow non-const references in a few standard places, like functions
5016 # called "swap()" or iostream operators like "<<" or ">>". Do not check
5017 # those function parameters.
5018 #
5019 # We also accept & in static_assert, which looks like a function but
5020 # it's actually a declaration expression.
5021 whitelisted_functions = (r'(?:[sS]wap(?:<\w:+>)?|'
5022 r'operator\s*[<>][<>]|'
5023 r'static_assert|COMPILE_ASSERT'
5024 r')\s*\(')
5025 if Search(whitelisted_functions, line):
avakulenko@google.com02af6282014-06-04 18:53:25 +00005026 return
erg@google.comc6671232013-10-25 21:44:03 +00005027 elif not Search(r'\S+\([^)]*$', line):
5028 # Don't see a whitelisted function on this line. Actually we
5029 # didn't see any function name on this line, so this is likely a
5030 # multi-line parameter list. Try a bit harder to catch this case.
5031 for i in xrange(2):
5032 if (linenum > i and
5033 Search(whitelisted_functions, clean_lines.elided[linenum - i - 1])):
avakulenko@google.com02af6282014-06-04 18:53:25 +00005034 return
erg@google.comc6671232013-10-25 21:44:03 +00005035
avakulenko@google.com02af6282014-06-04 18:53:25 +00005036 decls = ReplaceAll(r'{[^}]*}', ' ', line) # exclude function body
5037 for parameter in re.findall(_RE_PATTERN_REF_PARAM, decls):
Alex Vakulenko01e47232016-05-06 13:54:15 -07005038 if (not Match(_RE_PATTERN_CONST_REF_PARAM, parameter) and
5039 not Match(_RE_PATTERN_REF_STREAM_PARAM, parameter)):
avakulenko@google.com02af6282014-06-04 18:53:25 +00005040 error(filename, linenum, 'runtime/references', 2,
5041 'Is this a non-const reference? '
5042 'If so, make const or use a pointer: ' +
5043 ReplaceAll(' *<', '<', parameter))
5044
5045
5046def CheckCasts(filename, clean_lines, linenum, error):
5047 """Various cast related checks.
5048
5049 Args:
5050 filename: The name of the current file.
5051 clean_lines: A CleansedLines instance containing the file.
5052 linenum: The number of the line to check.
5053 error: The function to call with any errors found.
5054 """
5055 line = clean_lines.elided[linenum]
5056
5057 # Check to see if they're using an conversion function cast.
5058 # I just try to capture the most common basic types, though there are more.
5059 # Parameterless conversion functions, such as bool(), are allowed as they are
5060 # probably a member operator declaration or default constructor.
5061 match = Search(
Alex Vakulenko01e47232016-05-06 13:54:15 -07005062 r'(\bnew\s+(?:const\s+)?|\S<\s*(?:const\s+)?)?\b'
avakulenko@google.com02af6282014-06-04 18:53:25 +00005063 r'(int|float|double|bool|char|int32|uint32|int64|uint64)'
5064 r'(\([^)].*)', line)
5065 expecting_function = ExpectingFunctionArgs(clean_lines, linenum)
5066 if match and not expecting_function:
5067 matched_type = match.group(2)
5068
5069 # matched_new_or_template is used to silence two false positives:
5070 # - New operators
5071 # - Template arguments with function types
5072 #
5073 # For template arguments, we match on types immediately following
5074 # an opening bracket without any spaces. This is a fast way to
5075 # silence the common case where the function type is the first
5076 # template argument. False negative with less-than comparison is
5077 # avoided because those operators are usually followed by a space.
5078 #
5079 # function<double(double)> // bracket + no space = false positive
5080 # value < double(42) // bracket + space = true positive
5081 matched_new_or_template = match.group(1)
5082
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005083 # Avoid arrays by looking for brackets that come after the closing
5084 # parenthesis.
5085 if Match(r'\([^()]+\)\s*\[', match.group(3)):
5086 return
5087
avakulenko@google.com02af6282014-06-04 18:53:25 +00005088 # Other things to ignore:
5089 # - Function pointers
5090 # - Casts to pointer types
5091 # - Placement new
5092 # - Alias declarations
5093 matched_funcptr = match.group(3)
5094 if (matched_new_or_template is None and
5095 not (matched_funcptr and
5096 (Match(r'\((?:[^() ]+::\s*\*\s*)?[^() ]+\)\s*\(',
5097 matched_funcptr) or
5098 matched_funcptr.startswith('(*)'))) and
5099 not Match(r'\s*using\s+\S+\s*=\s*' + matched_type, line) and
5100 not Search(r'new\(\S+\)\s*' + matched_type, line)):
5101 error(filename, linenum, 'readability/casting', 4,
5102 'Using deprecated casting style. '
5103 'Use static_cast<%s>(...) instead' %
5104 matched_type)
5105
5106 if not expecting_function:
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005107 CheckCStyleCast(filename, clean_lines, linenum, 'static_cast',
avakulenko@google.com02af6282014-06-04 18:53:25 +00005108 r'\((int|float|double|bool|char|u?int(16|32|64))\)', error)
5109
5110 # This doesn't catch all cases. Consider (const char * const)"hello".
5111 #
5112 # (char *) "foo" should always be a const_cast (reinterpret_cast won't
5113 # compile).
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005114 if CheckCStyleCast(filename, clean_lines, linenum, 'const_cast',
5115 r'\((char\s?\*+\s?)\)\s*"', error):
avakulenko@google.com02af6282014-06-04 18:53:25 +00005116 pass
5117 else:
5118 # Check pointer casts for other than string constants
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005119 CheckCStyleCast(filename, clean_lines, linenum, 'reinterpret_cast',
5120 r'\((\w+\s?\*+\s?)\)', error)
avakulenko@google.com02af6282014-06-04 18:53:25 +00005121
5122 # In addition, we look for people taking the address of a cast. This
5123 # is dangerous -- casts can assign to temporaries, so the pointer doesn't
5124 # point where you think.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005125 #
5126 # Some non-identifier character is required before the '&' for the
5127 # expression to be recognized as a cast. These are casts:
5128 # expression = &static_cast<int*>(temporary());
5129 # function(&(int*)(temporary()));
5130 #
5131 # This is not a cast:
5132 # reference_type&(int* function_param);
avakulenko@google.com02af6282014-06-04 18:53:25 +00005133 match = Search(
avakulenko@google.com554223d2014-12-04 22:00:20 +00005134 r'(?:[^\w]&\(([^)*][^)]*)\)[\w(])|'
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005135 r'(?:[^\w]&(static|dynamic|down|reinterpret)_cast\b)', line)
avakulenko@google.com554223d2014-12-04 22:00:20 +00005136 if match:
avakulenko@google.com02af6282014-06-04 18:53:25 +00005137 # Try a better error message when the & is bound to something
5138 # dereferenced by the casted pointer, as opposed to the casted
5139 # pointer itself.
5140 parenthesis_error = False
5141 match = Match(r'^(.*&(?:static|dynamic|down|reinterpret)_cast\b)<', line)
5142 if match:
5143 _, y1, x1 = CloseExpression(clean_lines, linenum, len(match.group(1)))
5144 if x1 >= 0 and clean_lines.elided[y1][x1] == '(':
5145 _, y2, x2 = CloseExpression(clean_lines, y1, x1)
5146 if x2 >= 0:
5147 extended_line = clean_lines.elided[y2][x2:]
5148 if y2 < clean_lines.NumLines() - 1:
5149 extended_line += clean_lines.elided[y2 + 1]
5150 if Match(r'\s*(?:->|\[)', extended_line):
5151 parenthesis_error = True
5152
5153 if parenthesis_error:
5154 error(filename, linenum, 'readability/casting', 4,
5155 ('Are you taking an address of something dereferenced '
5156 'from a cast? Wrapping the dereferenced expression in '
5157 'parentheses will make the binding more obvious'))
5158 else:
5159 error(filename, linenum, 'runtime/casting', 4,
5160 ('Are you taking an address of a cast? '
5161 'This is dangerous: could be a temp var. '
5162 'Take the address before doing the cast, rather than after'))
erg@google.comc6671232013-10-25 21:44:03 +00005163
erg@google.com4e00b9a2009-01-12 23:05:11 +00005164
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005165def CheckCStyleCast(filename, clean_lines, linenum, cast_type, pattern, error):
erg@google.com4e00b9a2009-01-12 23:05:11 +00005166 """Checks for a C-style cast by looking for the pattern.
5167
erg@google.com4e00b9a2009-01-12 23:05:11 +00005168 Args:
5169 filename: The name of the current file.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005170 clean_lines: A CleansedLines instance containing the file.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005171 linenum: The number of the line to check.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005172 cast_type: The string for the C++ cast to recommend. This is either
erg@google.com8a95ecc2011-09-08 00:45:54 +00005173 reinterpret_cast, static_cast, or const_cast, depending.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005174 pattern: The regular expression used to find C-style casts.
5175 error: The function to call with any errors found.
erg@google.com8a95ecc2011-09-08 00:45:54 +00005176
5177 Returns:
5178 True if an error was emitted.
5179 False otherwise.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005180 """
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005181 line = clean_lines.elided[linenum]
erg@google.com4e00b9a2009-01-12 23:05:11 +00005182 match = Search(pattern, line)
5183 if not match:
erg@google.com8a95ecc2011-09-08 00:45:54 +00005184 return False
erg@google.com4e00b9a2009-01-12 23:05:11 +00005185
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005186 # Exclude lines with keywords that tend to look like casts
5187 context = line[0:match.start(1) - 1]
5188 if Match(r'.*\b(?:sizeof|alignof|alignas|[_A-Z][_A-Z0-9]*)\s*$', context):
5189 return False
5190
5191 # Try expanding current context to see if we one level of
5192 # parentheses inside a macro.
5193 if linenum > 0:
5194 for i in xrange(linenum - 1, max(0, linenum - 5), -1):
5195 context = clean_lines.elided[i] + context
5196 if Match(r'.*\b[_A-Z][_A-Z0-9]*\s*\((?:\([^()]*\)|[^()])*$', context):
erg@google.comfd5da632013-10-25 17:39:45 +00005197 return False
erg@google.com4e00b9a2009-01-12 23:05:11 +00005198
erg@google.comd350fe52013-01-14 17:51:48 +00005199 # operator++(int) and operator--(int)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005200 if context.endswith(' operator++') or context.endswith(' operator--'):
erg@google.comd350fe52013-01-14 17:51:48 +00005201 return False
5202
Alex Vakulenko01e47232016-05-06 13:54:15 -07005203 # A single unnamed argument for a function tends to look like old style cast.
5204 # If we see those, don't issue warnings for deprecated casts.
erg@google.comc6671232013-10-25 21:44:03 +00005205 remainder = line[match.end(0):]
avakulenko@google.com554223d2014-12-04 22:00:20 +00005206 if Match(r'^\s*(?:;|const\b|throw\b|final\b|override\b|[=>{),]|->)',
avakulenko@google.com02af6282014-06-04 18:53:25 +00005207 remainder):
Alex Vakulenko01e47232016-05-06 13:54:15 -07005208 return False
erg@google.com4e00b9a2009-01-12 23:05:11 +00005209
5210 # At this point, all that should be left is actual casts.
5211 error(filename, linenum, 'readability/casting', 4,
5212 'Using C-style cast. Use %s<%s>(...) instead' %
5213 (cast_type, match.group(1)))
5214
erg@google.com8a95ecc2011-09-08 00:45:54 +00005215 return True
5216
erg@google.com4e00b9a2009-01-12 23:05:11 +00005217
avakulenko@google.com02af6282014-06-04 18:53:25 +00005218def ExpectingFunctionArgs(clean_lines, linenum):
5219 """Checks whether where function type arguments are expected.
5220
5221 Args:
5222 clean_lines: A CleansedLines instance containing the file.
5223 linenum: The number of the line to check.
5224
5225 Returns:
5226 True if the line at 'linenum' is inside something that expects arguments
5227 of function types.
5228 """
5229 line = clean_lines.elided[linenum]
5230 return (Match(r'^\s*MOCK_(CONST_)?METHOD\d+(_T)?\(', line) or
5231 (linenum >= 2 and
5232 (Match(r'^\s*MOCK_(?:CONST_)?METHOD\d+(?:_T)?\((?:\S+,)?\s*$',
5233 clean_lines.elided[linenum - 1]) or
5234 Match(r'^\s*MOCK_(?:CONST_)?METHOD\d+(?:_T)?\(\s*$',
5235 clean_lines.elided[linenum - 2]) or
5236 Search(r'\bstd::m?function\s*\<\s*$',
5237 clean_lines.elided[linenum - 1]))))
5238
5239
erg@google.com4e00b9a2009-01-12 23:05:11 +00005240_HEADERS_CONTAINING_TEMPLATES = (
5241 ('<deque>', ('deque',)),
5242 ('<functional>', ('unary_function', 'binary_function',
5243 'plus', 'minus', 'multiplies', 'divides', 'modulus',
5244 'negate',
5245 'equal_to', 'not_equal_to', 'greater', 'less',
5246 'greater_equal', 'less_equal',
5247 'logical_and', 'logical_or', 'logical_not',
5248 'unary_negate', 'not1', 'binary_negate', 'not2',
5249 'bind1st', 'bind2nd',
5250 'pointer_to_unary_function',
5251 'pointer_to_binary_function',
5252 'ptr_fun',
5253 'mem_fun_t', 'mem_fun', 'mem_fun1_t', 'mem_fun1_ref_t',
5254 'mem_fun_ref_t',
5255 'const_mem_fun_t', 'const_mem_fun1_t',
5256 'const_mem_fun_ref_t', 'const_mem_fun1_ref_t',
5257 'mem_fun_ref',
5258 )),
5259 ('<limits>', ('numeric_limits',)),
5260 ('<list>', ('list',)),
5261 ('<map>', ('map', 'multimap',)),
lhchavez2890dff2016-07-11 19:37:29 -07005262 ('<memory>', ('allocator', 'make_shared', 'make_unique', 'shared_ptr',
5263 'unique_ptr', 'weak_ptr')),
erg@google.com4e00b9a2009-01-12 23:05:11 +00005264 ('<queue>', ('queue', 'priority_queue',)),
5265 ('<set>', ('set', 'multiset',)),
5266 ('<stack>', ('stack',)),
5267 ('<string>', ('char_traits', 'basic_string',)),
avakulenko@google.com554223d2014-12-04 22:00:20 +00005268 ('<tuple>', ('tuple',)),
lhchavez2890dff2016-07-11 19:37:29 -07005269 ('<unordered_map>', ('unordered_map', 'unordered_multimap')),
5270 ('<unordered_set>', ('unordered_set', 'unordered_multiset')),
erg@google.com4e00b9a2009-01-12 23:05:11 +00005271 ('<utility>', ('pair',)),
5272 ('<vector>', ('vector',)),
5273
5274 # gcc extensions.
5275 # Note: std::hash is their hash, ::hash is our hash
5276 ('<hash_map>', ('hash_map', 'hash_multimap',)),
5277 ('<hash_set>', ('hash_set', 'hash_multiset',)),
5278 ('<slist>', ('slist',)),
5279 )
5280
Alex Vakulenko01e47232016-05-06 13:54:15 -07005281_HEADERS_MAYBE_TEMPLATES = (
5282 ('<algorithm>', ('copy', 'max', 'min', 'min_element', 'sort',
5283 'transform',
5284 )),
lhchavez2890dff2016-07-11 19:37:29 -07005285 ('<utility>', ('forward', 'make_pair', 'move', 'swap')),
Alex Vakulenko01e47232016-05-06 13:54:15 -07005286 )
5287
erg@google.com4e00b9a2009-01-12 23:05:11 +00005288_RE_PATTERN_STRING = re.compile(r'\bstring\b')
5289
Alex Vakulenko01e47232016-05-06 13:54:15 -07005290_re_pattern_headers_maybe_templates = []
5291for _header, _templates in _HEADERS_MAYBE_TEMPLATES:
5292 for _template in _templates:
5293 # Match max<type>(..., ...), max(..., ...), but not foo->max, foo.max or
5294 # type::max().
5295 _re_pattern_headers_maybe_templates.append(
5296 (re.compile(r'[^>.]\b' + _template + r'(<.*?>)?\([^\)]'),
5297 _template,
5298 _header))
erg@google.com4e00b9a2009-01-12 23:05:11 +00005299
Alex Vakulenko01e47232016-05-06 13:54:15 -07005300# Other scripts may reach in and modify this pattern.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005301_re_pattern_templates = []
5302for _header, _templates in _HEADERS_CONTAINING_TEMPLATES:
5303 for _template in _templates:
5304 _re_pattern_templates.append(
5305 (re.compile(r'(\<|\b)' + _template + r'\s*\<'),
5306 _template + '<>',
5307 _header))
5308
5309
erg@google.come35f7652009-06-19 20:52:09 +00005310def FilesBelongToSameModule(filename_cc, filename_h):
5311 """Check if these two filenames belong to the same module.
5312
5313 The concept of a 'module' here is a as follows:
5314 foo.h, foo-inl.h, foo.cc, foo_test.cc and foo_unittest.cc belong to the
5315 same 'module' if they are in the same directory.
5316 some/path/public/xyzzy and some/path/internal/xyzzy are also considered
5317 to belong to the same module here.
5318
5319 If the filename_cc contains a longer path than the filename_h, for example,
5320 '/absolute/path/to/base/sysinfo.cc', and this file would include
5321 'base/sysinfo.h', this function also produces the prefix needed to open the
5322 header. This is used by the caller of this function to more robustly open the
5323 header file. We don't have access to the real include paths in this context,
5324 so we need this guesswork here.
5325
5326 Known bugs: tools/base/bar.cc and base/bar.h belong to the same module
5327 according to this implementation. Because of this, this function gives
5328 some false positives. This should be sufficiently rare in practice.
5329
5330 Args:
5331 filename_cc: is the path for the .cc file
5332 filename_h: is the path for the header path
5333
5334 Returns:
5335 Tuple with a bool and a string:
5336 bool: True if filename_cc and filename_h belong to the same module.
5337 string: the additional prefix needed to open the header file.
5338 """
5339
Alex Vakulenko01e47232016-05-06 13:54:15 -07005340 fileinfo = FileInfo(filename_cc)
5341 if not fileinfo.IsSource():
erg@google.come35f7652009-06-19 20:52:09 +00005342 return (False, '')
Alex Vakulenko01e47232016-05-06 13:54:15 -07005343 filename_cc = filename_cc[:-len(fileinfo.Extension())]
5344 matched_test_suffix = Search(_TEST_FILE_SUFFIX, fileinfo.BaseName())
5345 if matched_test_suffix:
5346 filename_cc = filename_cc[:-len(matched_test_suffix.group(1))]
erg@google.come35f7652009-06-19 20:52:09 +00005347 filename_cc = filename_cc.replace('/public/', '/')
5348 filename_cc = filename_cc.replace('/internal/', '/')
5349
5350 if not filename_h.endswith('.h'):
5351 return (False, '')
5352 filename_h = filename_h[:-len('.h')]
5353 if filename_h.endswith('-inl'):
5354 filename_h = filename_h[:-len('-inl')]
5355 filename_h = filename_h.replace('/public/', '/')
5356 filename_h = filename_h.replace('/internal/', '/')
5357
5358 files_belong_to_same_module = filename_cc.endswith(filename_h)
5359 common_path = ''
5360 if files_belong_to_same_module:
5361 common_path = filename_cc[:-len(filename_h)]
5362 return files_belong_to_same_module, common_path
5363
5364
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005365def UpdateIncludeState(filename, include_dict, io=codecs):
5366 """Fill up the include_dict with new includes found from the file.
erg@google.come35f7652009-06-19 20:52:09 +00005367
5368 Args:
5369 filename: the name of the header to read.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005370 include_dict: a dictionary in which the headers are inserted.
erg@google.come35f7652009-06-19 20:52:09 +00005371 io: The io factory to use to read the file. Provided for testability.
5372
5373 Returns:
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005374 True if a header was successfully added. False otherwise.
erg@google.come35f7652009-06-19 20:52:09 +00005375 """
5376 headerfile = None
5377 try:
5378 headerfile = io.open(filename, 'r', 'utf8', 'replace')
5379 except IOError:
5380 return False
5381 linenum = 0
5382 for line in headerfile:
5383 linenum += 1
5384 clean_line = CleanseComments(line)
5385 match = _RE_PATTERN_INCLUDE.search(clean_line)
5386 if match:
5387 include = match.group(2)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005388 include_dict.setdefault(include, linenum)
erg@google.come35f7652009-06-19 20:52:09 +00005389 return True
5390
5391
5392def CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error,
5393 io=codecs):
erg@google.com4e00b9a2009-01-12 23:05:11 +00005394 """Reports for missing stl includes.
5395
5396 This function will output warnings to make sure you are including the headers
5397 necessary for the stl containers and functions that you use. We only give one
5398 reason to include a header. For example, if you use both equal_to<> and
5399 less<> in a .h file, only one (the latter in the file) of these will be
5400 reported as a reason to include the <functional>.
5401
erg@google.com4e00b9a2009-01-12 23:05:11 +00005402 Args:
5403 filename: The name of the current file.
5404 clean_lines: A CleansedLines instance containing the file.
5405 include_state: An _IncludeState instance.
5406 error: The function to call with any errors found.
erg@google.come35f7652009-06-19 20:52:09 +00005407 io: The IO factory to use to read the header file. Provided for unittest
5408 injection.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005409 """
erg@google.com4e00b9a2009-01-12 23:05:11 +00005410 required = {} # A map of header name to linenumber and the template entity.
5411 # Example of required: { '<functional>': (1219, 'less<>') }
5412
5413 for linenum in xrange(clean_lines.NumLines()):
5414 line = clean_lines.elided[linenum]
5415 if not line or line[0] == '#':
5416 continue
5417
5418 # String is special -- it is a non-templatized type in STL.
erg@google.com8a95ecc2011-09-08 00:45:54 +00005419 matched = _RE_PATTERN_STRING.search(line)
5420 if matched:
erg+personal@google.com05189642010-04-30 20:43:03 +00005421 # Don't warn about strings in non-STL namespaces:
5422 # (We check only the first match per line; good enough.)
erg@google.com8a95ecc2011-09-08 00:45:54 +00005423 prefix = line[:matched.start()]
erg+personal@google.com05189642010-04-30 20:43:03 +00005424 if prefix.endswith('std::') or not prefix.endswith('::'):
5425 required['<string>'] = (linenum, 'string')
erg@google.com4e00b9a2009-01-12 23:05:11 +00005426
Alex Vakulenko01e47232016-05-06 13:54:15 -07005427 for pattern, template, header in _re_pattern_headers_maybe_templates:
erg@google.com4e00b9a2009-01-12 23:05:11 +00005428 if pattern.search(line):
5429 required[header] = (linenum, template)
5430
5431 # The following function is just a speed up, no semantics are changed.
5432 if not '<' in line: # Reduces the cpu time usage by skipping lines.
5433 continue
5434
5435 for pattern, template, header in _re_pattern_templates:
lhchavez3ae81f12016-07-11 19:00:34 -07005436 matched = pattern.search(line)
5437 if matched:
5438 # Don't warn about IWYU in non-STL namespaces:
5439 # (We check only the first match per line; good enough.)
5440 prefix = line[:matched.start()]
5441 if prefix.endswith('std::') or not prefix.endswith('::'):
5442 required[header] = (linenum, template)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005443
erg@google.come35f7652009-06-19 20:52:09 +00005444 # The policy is that if you #include something in foo.h you don't need to
5445 # include it again in foo.cc. Here, we will look at possible includes.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005446 # Let's flatten the include_state include_list and copy it into a dictionary.
5447 include_dict = dict([item for sublist in include_state.include_list
5448 for item in sublist])
erg@google.come35f7652009-06-19 20:52:09 +00005449
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005450 # Did we find the header for this file (if any) and successfully load it?
erg@google.come35f7652009-06-19 20:52:09 +00005451 header_found = False
5452
5453 # Use the absolute path so that matching works properly.
erg@google.com90ecb622012-01-30 19:34:23 +00005454 abs_filename = FileInfo(filename).FullName()
erg@google.come35f7652009-06-19 20:52:09 +00005455
5456 # For Emacs's flymake.
5457 # If cpplint is invoked from Emacs's flymake, a temporary file is generated
5458 # by flymake and that file name might end with '_flymake.cc'. In that case,
5459 # restore original file name here so that the corresponding header file can be
5460 # found.
5461 # e.g. If the file name is 'foo_flymake.cc', we should search for 'foo.h'
5462 # instead of 'foo_flymake.h'
erg+personal@google.com05189642010-04-30 20:43:03 +00005463 abs_filename = re.sub(r'_flymake\.cc$', '.cc', abs_filename)
erg@google.come35f7652009-06-19 20:52:09 +00005464
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005465 # include_dict is modified during iteration, so we iterate over a copy of
erg@google.come35f7652009-06-19 20:52:09 +00005466 # the keys.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005467 header_keys = include_dict.keys()
erg@google.com8a95ecc2011-09-08 00:45:54 +00005468 for header in header_keys:
erg@google.come35f7652009-06-19 20:52:09 +00005469 (same_module, common_path) = FilesBelongToSameModule(abs_filename, header)
5470 fullpath = common_path + header
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005471 if same_module and UpdateIncludeState(fullpath, include_dict, io):
erg@google.come35f7652009-06-19 20:52:09 +00005472 header_found = True
5473
5474 # If we can't find the header file for a .cc, assume it's because we don't
5475 # know where to look. In that case we'll give up as we're not sure they
5476 # didn't include it in the .h file.
5477 # TODO(unknown): Do a better job of finding .h files so we are confident that
5478 # not having the .h file means there isn't one.
5479 if filename.endswith('.cc') and not header_found:
5480 return
5481
erg@google.com4e00b9a2009-01-12 23:05:11 +00005482 # All the lines have been processed, report the errors found.
5483 for required_header_unstripped in required:
5484 template = required[required_header_unstripped][1]
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005485 if required_header_unstripped.strip('<>"') not in include_dict:
erg@google.com4e00b9a2009-01-12 23:05:11 +00005486 error(filename, required[required_header_unstripped][0],
5487 'build/include_what_you_use', 4,
5488 'Add #include ' + required_header_unstripped + ' for ' + template)
5489
5490
erg@google.com8a95ecc2011-09-08 00:45:54 +00005491_RE_PATTERN_EXPLICIT_MAKEPAIR = re.compile(r'\bmake_pair\s*<')
5492
5493
5494def CheckMakePairUsesDeduction(filename, clean_lines, linenum, error):
5495 """Check that make_pair's template arguments are deduced.
5496
avakulenko@google.com02af6282014-06-04 18:53:25 +00005497 G++ 4.6 in C++11 mode fails badly if make_pair's template arguments are
erg@google.com8a95ecc2011-09-08 00:45:54 +00005498 specified explicitly, and such use isn't intended in any case.
5499
5500 Args:
5501 filename: The name of the current file.
5502 clean_lines: A CleansedLines instance containing the file.
5503 linenum: The number of the line to check.
5504 error: The function to call with any errors found.
5505 """
erg@google.com2aa59982013-10-28 19:09:25 +00005506 line = clean_lines.elided[linenum]
erg@google.com8a95ecc2011-09-08 00:45:54 +00005507 match = _RE_PATTERN_EXPLICIT_MAKEPAIR.search(line)
5508 if match:
5509 error(filename, linenum, 'build/explicit_make_pair',
5510 4, # 4 = high confidence
erg@google.comd350fe52013-01-14 17:51:48 +00005511 'For C++11-compatibility, omit template arguments from make_pair'
5512 ' OR use pair directly OR if appropriate, construct a pair directly')
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005513
5514
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005515def CheckRedundantVirtual(filename, clean_lines, linenum, error):
5516 """Check if line contains a redundant "virtual" function-specifier.
5517
5518 Args:
5519 filename: The name of the current file.
5520 clean_lines: A CleansedLines instance containing the file.
5521 linenum: The number of the line to check.
5522 error: The function to call with any errors found.
5523 """
5524 # Look for "virtual" on current line.
5525 line = clean_lines.elided[linenum]
avakulenko@google.com554223d2014-12-04 22:00:20 +00005526 virtual = Match(r'^(.*)(\bvirtual\b)(.*)$', line)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005527 if not virtual: return
5528
avakulenko@google.com554223d2014-12-04 22:00:20 +00005529 # Ignore "virtual" keywords that are near access-specifiers. These
5530 # are only used in class base-specifier and do not apply to member
5531 # functions.
5532 if (Search(r'\b(public|protected|private)\s+$', virtual.group(1)) or
5533 Match(r'^\s+(public|protected|private)\b', virtual.group(3))):
5534 return
5535
5536 # Ignore the "virtual" keyword from virtual base classes. Usually
5537 # there is a column on the same line in these cases (virtual base
5538 # classes are rare in google3 because multiple inheritance is rare).
5539 if Match(r'^.*[^:]:[^:].*$', line): return
5540
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005541 # Look for the next opening parenthesis. This is the start of the
5542 # parameter list (possibly on the next line shortly after virtual).
5543 # TODO(unknown): doesn't work if there are virtual functions with
5544 # decltype() or other things that use parentheses, but csearch suggests
5545 # that this is rare.
5546 end_col = -1
5547 end_line = -1
avakulenko@google.com554223d2014-12-04 22:00:20 +00005548 start_col = len(virtual.group(2))
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005549 for start_line in xrange(linenum, min(linenum + 3, clean_lines.NumLines())):
5550 line = clean_lines.elided[start_line][start_col:]
5551 parameter_list = Match(r'^([^(]*)\(', line)
5552 if parameter_list:
5553 # Match parentheses to find the end of the parameter list
5554 (_, end_line, end_col) = CloseExpression(
5555 clean_lines, start_line, start_col + len(parameter_list.group(1)))
5556 break
5557 start_col = 0
5558
5559 if end_col < 0:
5560 return # Couldn't find end of parameter list, give up
5561
5562 # Look for "override" or "final" after the parameter list
5563 # (possibly on the next few lines).
5564 for i in xrange(end_line, min(end_line + 3, clean_lines.NumLines())):
5565 line = clean_lines.elided[i][end_col:]
5566 match = Search(r'\b(override|final)\b', line)
5567 if match:
5568 error(filename, linenum, 'readability/inheritance', 4,
5569 ('"virtual" is redundant since function is '
5570 'already declared as "%s"' % match.group(1)))
5571
5572 # Set end_col to check whole lines after we are done with the
5573 # first line.
5574 end_col = 0
5575 if Search(r'[^\w]\s*$', line):
5576 break
5577
5578
5579def CheckRedundantOverrideOrFinal(filename, clean_lines, linenum, error):
5580 """Check if line contains a redundant "override" or "final" virt-specifier.
5581
5582 Args:
5583 filename: The name of the current file.
5584 clean_lines: A CleansedLines instance containing the file.
5585 linenum: The number of the line to check.
5586 error: The function to call with any errors found.
5587 """
avakulenko@google.com554223d2014-12-04 22:00:20 +00005588 # Look for closing parenthesis nearby. We need one to confirm where
5589 # the declarator ends and where the virt-specifier starts to avoid
5590 # false positives.
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005591 line = clean_lines.elided[linenum]
avakulenko@google.com554223d2014-12-04 22:00:20 +00005592 declarator_end = line.rfind(')')
5593 if declarator_end >= 0:
5594 fragment = line[declarator_end:]
5595 else:
5596 if linenum > 1 and clean_lines.elided[linenum - 1].rfind(')') >= 0:
5597 fragment = line
5598 else:
5599 return
5600
5601 # Check that at most one of "override" or "final" is present, not both
5602 if Search(r'\boverride\b', fragment) and Search(r'\bfinal\b', fragment):
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005603 error(filename, linenum, 'readability/inheritance', 4,
5604 ('"override" is redundant since function is '
5605 'already declared as "final"'))
5606
5607
5608
5609
5610# Returns true if we are at a new block, and it is directly
5611# inside of a namespace.
5612def IsBlockInNameSpace(nesting_state, is_forward_declaration):
5613 """Checks that the new block is directly in a namespace.
5614
5615 Args:
5616 nesting_state: The _NestingState object that contains info about our state.
5617 is_forward_declaration: If the class is a forward declared class.
5618 Returns:
5619 Whether or not the new block is directly in a namespace.
5620 """
5621 if is_forward_declaration:
5622 if len(nesting_state.stack) >= 1 and (
5623 isinstance(nesting_state.stack[-1], _NamespaceInfo)):
5624 return True
5625 else:
5626 return False
5627
5628 return (len(nesting_state.stack) > 1 and
5629 nesting_state.stack[-1].check_namespace_indentation and
5630 isinstance(nesting_state.stack[-2], _NamespaceInfo))
5631
5632
5633def ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item,
5634 raw_lines_no_comments, linenum):
5635 """This method determines if we should apply our namespace indentation check.
5636
5637 Args:
5638 nesting_state: The current nesting state.
5639 is_namespace_indent_item: If we just put a new class on the stack, True.
5640 If the top of the stack is not a class, or we did not recently
5641 add the class, False.
5642 raw_lines_no_comments: The lines without the comments.
5643 linenum: The current line number we are processing.
5644
5645 Returns:
5646 True if we should apply our namespace indentation check. Currently, it
5647 only works for classes and namespaces inside of a namespace.
5648 """
5649
5650 is_forward_declaration = IsForwardClassDeclaration(raw_lines_no_comments,
5651 linenum)
5652
5653 if not (is_namespace_indent_item or is_forward_declaration):
5654 return False
5655
5656 # If we are in a macro, we do not want to check the namespace indentation.
5657 if IsMacroDefinition(raw_lines_no_comments, linenum):
5658 return False
5659
5660 return IsBlockInNameSpace(nesting_state, is_forward_declaration)
5661
5662
5663# Call this method if the line is directly inside of a namespace.
5664# If the line above is blank (excluding comments) or the start of
5665# an inner namespace, it cannot be indented.
5666def CheckItemIndentationInNamespace(filename, raw_lines_no_comments, linenum,
5667 error):
5668 line = raw_lines_no_comments[linenum]
5669 if Match(r'^\s+', line):
5670 error(filename, linenum, 'runtime/indentation_namespace', 4,
5671 'Do not indent within a namespace')
erg@google.com8a95ecc2011-09-08 00:45:54 +00005672
5673
erg@google.comd350fe52013-01-14 17:51:48 +00005674def ProcessLine(filename, file_extension, clean_lines, line,
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005675 include_state, function_state, nesting_state, error,
5676 extra_check_functions=[]):
erg@google.com4e00b9a2009-01-12 23:05:11 +00005677 """Processes a single line in the file.
5678
5679 Args:
5680 filename: Filename of the file that is being processed.
5681 file_extension: The extension (dot not included) of the file.
5682 clean_lines: An array of strings, each representing a line of the file,
5683 with comments stripped.
5684 line: Number of line being processed.
5685 include_state: An _IncludeState instance in which the headers are inserted.
5686 function_state: A _FunctionState instance which counts function lines, etc.
avakulenko@google.com02af6282014-06-04 18:53:25 +00005687 nesting_state: A NestingState instance which maintains information about
erg@google.comd350fe52013-01-14 17:51:48 +00005688 the current stack of nested blocks being parsed.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005689 error: A callable to which errors are reported, which takes 4 arguments:
5690 filename, line number, error level, and message
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005691 extra_check_functions: An array of additional check functions that will be
5692 run on each source line. Each function takes 4
5693 arguments: filename, clean_lines, line, error
erg@google.com4e00b9a2009-01-12 23:05:11 +00005694 """
5695 raw_lines = clean_lines.raw_lines
erg+personal@google.com05189642010-04-30 20:43:03 +00005696 ParseNolintSuppressions(filename, raw_lines[line], line, error)
erg@google.comd350fe52013-01-14 17:51:48 +00005697 nesting_state.Update(filename, clean_lines, line, error)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005698 CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line,
5699 error)
avakulenko@google.com02af6282014-06-04 18:53:25 +00005700 if nesting_state.InAsmBlock(): return
erg@google.com4e00b9a2009-01-12 23:05:11 +00005701 CheckForFunctionLengths(filename, clean_lines, line, function_state, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005702 CheckForMultilineCommentsAndStrings(filename, clean_lines, line, error)
erg@google.comd350fe52013-01-14 17:51:48 +00005703 CheckStyle(filename, clean_lines, line, file_extension, nesting_state, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005704 CheckLanguage(filename, clean_lines, line, file_extension, include_state,
erg@google.comfd5da632013-10-25 17:39:45 +00005705 nesting_state, error)
erg@google.comc6671232013-10-25 21:44:03 +00005706 CheckForNonConstReference(filename, clean_lines, line, nesting_state, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005707 CheckForNonStandardConstructs(filename, clean_lines, line,
erg@google.comd350fe52013-01-14 17:51:48 +00005708 nesting_state, error)
erg@google.com2aa59982013-10-28 19:09:25 +00005709 CheckVlogArguments(filename, clean_lines, line, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005710 CheckPosixThreading(filename, clean_lines, line, error)
erg@google.com36649102009-03-25 21:18:36 +00005711 CheckInvalidIncrement(filename, clean_lines, line, error)
erg@google.com8a95ecc2011-09-08 00:45:54 +00005712 CheckMakePairUsesDeduction(filename, clean_lines, line, error)
avakulenko@google.coma8ee7ea2014-08-11 19:41:35 +00005713 CheckRedundantVirtual(filename, clean_lines, line, error)
5714 CheckRedundantOverrideOrFinal(filename, clean_lines, line, error)
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005715 for check_fn in extra_check_functions:
5716 check_fn(filename, clean_lines, line, error)
erg@google.com7430eef2014-07-28 22:33:46 +00005717
avakulenko@google.com02af6282014-06-04 18:53:25 +00005718def FlagCxx11Features(filename, clean_lines, linenum, error):
5719 """Flag those c++11 features that we only allow in certain places.
5720
5721 Args:
5722 filename: The name of the current file.
5723 clean_lines: A CleansedLines instance containing the file.
5724 linenum: The number of the line to check.
5725 error: The function to call with any errors found.
5726 """
5727 line = clean_lines.elided[linenum]
5728
avakulenko@google.com02af6282014-06-04 18:53:25 +00005729 include = Match(r'\s*#\s*include\s+[<"]([^<"]+)[">]', line)
Alex Vakulenko01e47232016-05-06 13:54:15 -07005730
5731 # Flag unapproved C++ TR1 headers.
5732 if include and include.group(1).startswith('tr1/'):
5733 error(filename, linenum, 'build/c++tr1', 5,
5734 ('C++ TR1 headers such as <%s> are unapproved.') % include.group(1))
5735
5736 # Flag unapproved C++11 headers.
avakulenko@google.com02af6282014-06-04 18:53:25 +00005737 if include and include.group(1) in ('cfenv',
5738 'condition_variable',
5739 'fenv.h',
5740 'future',
5741 'mutex',
5742 'thread',
5743 'chrono',
5744 'ratio',
5745 'regex',
5746 'system_error',
5747 ):
5748 error(filename, linenum, 'build/c++11', 5,
5749 ('<%s> is an unapproved C++11 header.') % include.group(1))
5750
5751 # The only place where we need to worry about C++11 keywords and library
5752 # features in preprocessor directives is in macro definitions.
5753 if Match(r'\s*#', line) and not Match(r'\s*#\s*define\b', line): return
5754
5755 # These are classes and free functions. The classes are always
5756 # mentioned as std::*, but we only catch the free functions if
5757 # they're not found by ADL. They're alphabetical by header.
5758 for top_name in (
5759 # type_traits
5760 'alignment_of',
5761 'aligned_union',
avakulenko@google.com02af6282014-06-04 18:53:25 +00005762 ):
5763 if Search(r'\bstd::%s\b' % top_name, line):
5764 error(filename, linenum, 'build/c++11', 5,
5765 ('std::%s is an unapproved C++11 class or function. Send c-style '
5766 'an example of where it would make your code more readable, and '
5767 'they may let you use it.') % top_name)
5768
5769
Alex Vakulenko01e47232016-05-06 13:54:15 -07005770def FlagCxx14Features(filename, clean_lines, linenum, error):
5771 """Flag those C++14 features that we restrict.
5772
5773 Args:
5774 filename: The name of the current file.
5775 clean_lines: A CleansedLines instance containing the file.
5776 linenum: The number of the line to check.
5777 error: The function to call with any errors found.
5778 """
5779 line = clean_lines.elided[linenum]
5780
5781 include = Match(r'\s*#\s*include\s+[<"]([^<"]+)[">]', line)
5782
5783 # Flag unapproved C++14 headers.
5784 if include and include.group(1) in ('scoped_allocator', 'shared_mutex'):
5785 error(filename, linenum, 'build/c++14', 5,
5786 ('<%s> is an unapproved C++14 header.') % include.group(1))
5787
5788
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005789def ProcessFileData(filename, file_extension, lines, error,
5790 extra_check_functions=[]):
erg@google.com4e00b9a2009-01-12 23:05:11 +00005791 """Performs lint checks and reports any errors to the given error function.
5792
5793 Args:
5794 filename: Filename of the file that is being processed.
5795 file_extension: The extension (dot not included) of the file.
5796 lines: An array of strings, each representing a line of the file, with the
erg@google.com8a95ecc2011-09-08 00:45:54 +00005797 last element being empty if the file is terminated with a newline.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005798 error: A callable to which errors are reported, which takes 4 arguments:
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005799 filename, line number, error level, and message
5800 extra_check_functions: An array of additional check functions that will be
5801 run on each source line. Each function takes 4
5802 arguments: filename, clean_lines, line, error
erg@google.com4e00b9a2009-01-12 23:05:11 +00005803 """
5804 lines = (['// marker so line numbers and indices both start at 1'] + lines +
5805 ['// marker so line numbers end in a known way'])
5806
5807 include_state = _IncludeState()
5808 function_state = _FunctionState()
avakulenko@google.com02af6282014-06-04 18:53:25 +00005809 nesting_state = NestingState()
erg@google.com4e00b9a2009-01-12 23:05:11 +00005810
erg+personal@google.com05189642010-04-30 20:43:03 +00005811 ResetNolintSuppressions()
5812
erg@google.com4e00b9a2009-01-12 23:05:11 +00005813 CheckForCopyright(filename, lines, error)
Alex Vakulenko01e47232016-05-06 13:54:15 -07005814 ProcessGlobalSuppresions(lines)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005815 RemoveMultiLineComments(filename, lines, error)
5816 clean_lines = CleansedLines(lines)
avakulenko@google.com554223d2014-12-04 22:00:20 +00005817
LukeCz7197a242016-09-24 13:27:35 -05005818 if IsHeaderExtension(file_extension):
avakulenko@google.com554223d2014-12-04 22:00:20 +00005819 CheckForHeaderGuard(filename, clean_lines, error)
5820
erg@google.com4e00b9a2009-01-12 23:05:11 +00005821 for line in xrange(clean_lines.NumLines()):
5822 ProcessLine(filename, file_extension, clean_lines, line,
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005823 include_state, function_state, nesting_state, error,
5824 extra_check_functions)
avakulenko@google.com02af6282014-06-04 18:53:25 +00005825 FlagCxx11Features(filename, clean_lines, line, error)
erg@google.com2aa59982013-10-28 19:09:25 +00005826 nesting_state.CheckCompletedBlocks(filename, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005827
5828 CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error)
Alex Vakulenko01e47232016-05-06 13:54:15 -07005829
avakulenko@google.com554223d2014-12-04 22:00:20 +00005830 # Check that the .cc file has included its header if it exists.
Alex Vakulenko01e47232016-05-06 13:54:15 -07005831 if _IsSourceExtension(file_extension):
avakulenko@google.com554223d2014-12-04 22:00:20 +00005832 CheckHeaderFileIncluded(filename, include_state, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005833
5834 # We check here rather than inside ProcessLine so that we see raw
5835 # lines rather than "cleaned" lines.
erg@google.com2aa59982013-10-28 19:09:25 +00005836 CheckForBadCharacters(filename, lines, error)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005837
5838 CheckForNewlineAtEOF(filename, lines, error)
5839
erg@google.com7430eef2014-07-28 22:33:46 +00005840def ProcessConfigOverrides(filename):
5841 """ Loads the configuration files and processes the config overrides.
5842
5843 Args:
5844 filename: The name of the file being processed by the linter.
5845
5846 Returns:
5847 False if the current |filename| should not be processed further.
5848 """
5849
5850 abs_filename = os.path.abspath(filename)
5851 cfg_filters = []
5852 keep_looking = True
5853 while keep_looking:
5854 abs_path, base_name = os.path.split(abs_filename)
5855 if not base_name:
5856 break # Reached the root directory.
5857
5858 cfg_file = os.path.join(abs_path, "CPPLINT.cfg")
5859 abs_filename = abs_path
5860 if not os.path.isfile(cfg_file):
5861 continue
5862
5863 try:
5864 with open(cfg_file) as file_handle:
5865 for line in file_handle:
5866 line, _, _ = line.partition('#') # Remove comments.
5867 if not line.strip():
5868 continue
5869
5870 name, _, val = line.partition('=')
5871 name = name.strip()
5872 val = val.strip()
5873 if name == 'set noparent':
5874 keep_looking = False
5875 elif name == 'filter':
5876 cfg_filters.append(val)
5877 elif name == 'exclude_files':
5878 # When matching exclude_files pattern, use the base_name of
5879 # the current file name or the directory name we are processing.
5880 # For example, if we are checking for lint errors in /foo/bar/baz.cc
5881 # and we found the .cfg file at /foo/CPPLINT.cfg, then the config
5882 # file's "exclude_files" filter is meant to be checked against "bar"
5883 # and not "baz" nor "bar/baz.cc".
5884 if base_name:
5885 pattern = re.compile(val)
5886 if pattern.match(base_name):
5887 sys.stderr.write('Ignoring "%s": file excluded by "%s". '
5888 'File path component "%s" matches '
5889 'pattern "%s"\n' %
5890 (filename, cfg_file, base_name, val))
5891 return False
avakulenko@google.com310681b2014-08-22 19:38:55 +00005892 elif name == 'linelength':
5893 global _line_length
5894 try:
5895 _line_length = int(val)
5896 except ValueError:
5897 sys.stderr.write('Line length must be numeric.')
Fabian Guera2322e4f2016-05-01 17:36:30 +02005898 elif name == 'root':
5899 global _root
5900 _root = val
LukeCz7197a242016-09-24 13:27:35 -05005901 elif name == 'headers':
5902 ProcessHppHeadersOption(val)
erg@google.com7430eef2014-07-28 22:33:46 +00005903 else:
5904 sys.stderr.write(
5905 'Invalid configuration option (%s) in file %s\n' %
5906 (name, cfg_file))
5907
5908 except IOError:
5909 sys.stderr.write(
5910 "Skipping config file '%s': Can't open for reading\n" % cfg_file)
5911 keep_looking = False
5912
5913 # Apply all the accumulated filters in reverse order (top-level directory
5914 # config options having the least priority).
5915 for filter in reversed(cfg_filters):
5916 _AddFilters(filter)
5917
5918 return True
5919
avakulenko@google.com02af6282014-06-04 18:53:25 +00005920
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005921def ProcessFile(filename, vlevel, extra_check_functions=[]):
erg@google.com4e00b9a2009-01-12 23:05:11 +00005922 """Does google-lint on a single file.
5923
5924 Args:
5925 filename: The name of the file to parse.
5926
5927 vlevel: The level of errors to report. Every error of confidence
5928 >= verbose_level will be reported. 0 is a good default.
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005929
5930 extra_check_functions: An array of additional check functions that will be
5931 run on each source line. Each function takes 4
5932 arguments: filename, clean_lines, line, error
erg@google.com4e00b9a2009-01-12 23:05:11 +00005933 """
5934
5935 _SetVerboseLevel(vlevel)
erg@google.com7430eef2014-07-28 22:33:46 +00005936 _BackupFilters()
5937
5938 if not ProcessConfigOverrides(filename):
5939 _RestoreFilters()
5940 return
erg@google.com4e00b9a2009-01-12 23:05:11 +00005941
avakulenko@google.com02af6282014-06-04 18:53:25 +00005942 lf_lines = []
5943 crlf_lines = []
erg@google.com4e00b9a2009-01-12 23:05:11 +00005944 try:
5945 # Support the UNIX convention of using "-" for stdin. Note that
5946 # we are not opening the file with universal newline support
5947 # (which codecs doesn't support anyway), so the resulting lines do
5948 # contain trailing '\r' characters if we are reading a file that
5949 # has CRLF endings.
5950 # If after the split a trailing '\r' is present, it is removed
avakulenko@google.com02af6282014-06-04 18:53:25 +00005951 # below.
erg@google.com4e00b9a2009-01-12 23:05:11 +00005952 if filename == '-':
5953 lines = codecs.StreamReaderWriter(sys.stdin,
5954 codecs.getreader('utf8'),
5955 codecs.getwriter('utf8'),
5956 'replace').read().split('\n')
5957 else:
5958 lines = codecs.open(filename, 'r', 'utf8', 'replace').read().split('\n')
5959
erg@google.com4e00b9a2009-01-12 23:05:11 +00005960 # Remove trailing '\r'.
avakulenko@google.com02af6282014-06-04 18:53:25 +00005961 # The -1 accounts for the extra trailing blank line we get from split()
5962 for linenum in range(len(lines) - 1):
erg@google.com4e00b9a2009-01-12 23:05:11 +00005963 if lines[linenum].endswith('\r'):
5964 lines[linenum] = lines[linenum].rstrip('\r')
avakulenko@google.com02af6282014-06-04 18:53:25 +00005965 crlf_lines.append(linenum + 1)
5966 else:
5967 lf_lines.append(linenum + 1)
erg@google.com4e00b9a2009-01-12 23:05:11 +00005968
5969 except IOError:
5970 sys.stderr.write(
5971 "Skipping input '%s': Can't open for reading\n" % filename)
erg@google.com7430eef2014-07-28 22:33:46 +00005972 _RestoreFilters()
erg@google.com4e00b9a2009-01-12 23:05:11 +00005973 return
5974
5975 # Note, if no dot is found, this will give the entire filename as the ext.
5976 file_extension = filename[filename.rfind('.') + 1:]
5977
5978 # When reading from stdin, the extension is unknown, so no cpplint tests
5979 # should rely on the extension.
erg@google.com19680272013-12-16 22:48:54 +00005980 if filename != '-' and file_extension not in _valid_extensions:
erg@google.com2aa59982013-10-28 19:09:25 +00005981 sys.stderr.write('Ignoring %s; not a valid file name '
erg@google.com19680272013-12-16 22:48:54 +00005982 '(%s)\n' % (filename, ', '.join(_valid_extensions)))
erg@google.com4e00b9a2009-01-12 23:05:11 +00005983 else:
avakulenko@google.com4b957b22014-06-04 22:48:14 +00005984 ProcessFileData(filename, file_extension, lines, Error,
5985 extra_check_functions)
avakulenko@google.com02af6282014-06-04 18:53:25 +00005986
5987 # If end-of-line sequences are a mix of LF and CR-LF, issue
5988 # warnings on the lines with CR.
5989 #
5990 # Don't issue any warnings if all lines are uniformly LF or CR-LF,
5991 # since critique can handle these just fine, and the style guide
5992 # doesn't dictate a particular end of line sequence.
5993 #
5994 # We can't depend on os.linesep to determine what the desired
5995 # end-of-line sequence should be, since that will return the
5996 # server-side end-of-line sequence.
5997 if lf_lines and crlf_lines:
5998 # Warn on every line with CR. An alternative approach might be to
5999 # check whether the file is mostly CRLF or just LF, and warn on the
6000 # minority, we bias toward LF here since most tools prefer LF.
6001 for linenum in crlf_lines:
6002 Error(filename, linenum, 'whitespace/newline', 1,
6003 'Unexpected \\r (^M) found; better to use only \\n')
erg@google.com4e00b9a2009-01-12 23:05:11 +00006004
LukeCze09f4782016-09-28 19:13:37 -05006005 sys.stdout.write('Done processing %s\n' % filename)
erg@google.com7430eef2014-07-28 22:33:46 +00006006 _RestoreFilters()
erg@google.com4e00b9a2009-01-12 23:05:11 +00006007
6008
6009def PrintUsage(message):
6010 """Prints a brief usage string and exits, optionally with an error message.
6011
6012 Args:
6013 message: The optional error message.
6014 """
6015 sys.stderr.write(_USAGE)
6016 if message:
6017 sys.exit('\nFATAL ERROR: ' + message)
6018 else:
6019 sys.exit(1)
6020
6021
6022def PrintCategories():
6023 """Prints a list of all the error-categories used by error messages.
6024
6025 These are the categories used to filter messages via --filter.
6026 """
erg+personal@google.com05189642010-04-30 20:43:03 +00006027 sys.stderr.write(''.join(' %s\n' % cat for cat in _ERROR_CATEGORIES))
erg@google.com4e00b9a2009-01-12 23:05:11 +00006028 sys.exit(0)
6029
6030
6031def ParseArguments(args):
6032 """Parses the command line arguments.
6033
6034 This may set the output format and verbosity level as side-effects.
6035
6036 Args:
6037 args: The command line arguments:
6038
6039 Returns:
6040 The list of filenames to lint.
6041 """
6042 try:
6043 (opts, filenames) = getopt.getopt(args, '', ['help', 'output=', 'verbose=',
erg@google.coma868d2d2009-10-09 21:18:45 +00006044 'counting=',
erg@google.com4d70a882013-04-16 21:06:32 +00006045 'filter=',
erg@google.comab53edf2013-11-05 22:23:37 +00006046 'root=',
erg@google.com19680272013-12-16 22:48:54 +00006047 'linelength=',
LukeCz7197a242016-09-24 13:27:35 -05006048 'extensions=',
6049 'headers='])
erg@google.com4e00b9a2009-01-12 23:05:11 +00006050 except getopt.GetoptError:
6051 PrintUsage('Invalid arguments.')
6052
6053 verbosity = _VerboseLevel()
6054 output_format = _OutputFormat()
6055 filters = ''
erg@google.coma868d2d2009-10-09 21:18:45 +00006056 counting_style = ''
erg@google.com4e00b9a2009-01-12 23:05:11 +00006057
6058 for (opt, val) in opts:
6059 if opt == '--help':
6060 PrintUsage(None)
6061 elif opt == '--output':
erg@google.comc6671232013-10-25 21:44:03 +00006062 if val not in ('emacs', 'vs7', 'eclipse'):
erg@google.com02c27fd2013-05-28 21:34:34 +00006063 PrintUsage('The only allowed output formats are emacs, vs7 and eclipse.')
erg@google.com4e00b9a2009-01-12 23:05:11 +00006064 output_format = val
6065 elif opt == '--verbose':
6066 verbosity = int(val)
6067 elif opt == '--filter':
6068 filters = val
erg@google.coma87abb82009-02-24 01:41:01 +00006069 if not filters:
erg@google.com4e00b9a2009-01-12 23:05:11 +00006070 PrintCategories()
erg@google.coma868d2d2009-10-09 21:18:45 +00006071 elif opt == '--counting':
6072 if val not in ('total', 'toplevel', 'detailed'):
6073 PrintUsage('Valid counting options are total, toplevel, and detailed')
6074 counting_style = val
erg@google.com4d70a882013-04-16 21:06:32 +00006075 elif opt == '--root':
6076 global _root
6077 _root = val
erg@google.comab53edf2013-11-05 22:23:37 +00006078 elif opt == '--linelength':
6079 global _line_length
6080 try:
6081 _line_length = int(val)
6082 except ValueError:
6083 PrintUsage('Line length must be digits.')
erg@google.com19680272013-12-16 22:48:54 +00006084 elif opt == '--extensions':
6085 global _valid_extensions
6086 try:
6087 _valid_extensions = set(val.split(','))
6088 except ValueError:
6089 PrintUsage('Extensions must be comma seperated list.')
LukeCz7197a242016-09-24 13:27:35 -05006090 elif opt == '--headers':
6091 ProcessHppHeadersOption(val)
erg@google.com4e00b9a2009-01-12 23:05:11 +00006092
6093 if not filenames:
6094 PrintUsage('No files were specified.')
6095
6096 _SetOutputFormat(output_format)
6097 _SetVerboseLevel(verbosity)
6098 _SetFilters(filters)
erg@google.coma868d2d2009-10-09 21:18:45 +00006099 _SetCountingStyle(counting_style)
erg@google.com4e00b9a2009-01-12 23:05:11 +00006100
6101 return filenames
6102
6103
6104def main():
6105 filenames = ParseArguments(sys.argv[1:])
6106
6107 # Change stderr to write with replacement characters so we don't die
6108 # if we try to print something containing non-ASCII characters.
6109 sys.stderr = codecs.StreamReaderWriter(sys.stderr,
6110 codecs.getreader('utf8'),
6111 codecs.getwriter('utf8'),
6112 'replace')
6113
erg@google.coma868d2d2009-10-09 21:18:45 +00006114 _cpplint_state.ResetErrorCounts()
erg@google.com4e00b9a2009-01-12 23:05:11 +00006115 for filename in filenames:
6116 ProcessFile(filename, _cpplint_state.verbose_level)
erg@google.coma868d2d2009-10-09 21:18:45 +00006117 _cpplint_state.PrintErrorCounts()
6118
erg@google.com4e00b9a2009-01-12 23:05:11 +00006119 sys.exit(_cpplint_state.error_count > 0)
6120
6121
6122if __name__ == '__main__':
6123 main()