]>
Commit | Line | Data |
---|---|---|
1d09f67e TL |
1 | #!/usr/bin/env python3 |
2 | # | |
3 | # Copyright (c) 2009 Google Inc. All rights reserved. | |
4 | # | |
5 | # Redistribution and use in source and binary forms, with or without | |
6 | # modification, are permitted provided that the following conditions are | |
7 | # met: | |
8 | # | |
9 | # * Redistributions of source code must retain the above copyright | |
10 | # notice, this list of conditions and the following disclaimer. | |
11 | # * Redistributions in binary form must reproduce the above | |
12 | # copyright notice, this list of conditions and the following disclaimer | |
13 | # in the documentation and/or other materials provided with the | |
14 | # distribution. | |
15 | # * Neither the name of Google Inc. nor the names of its | |
16 | # contributors may be used to endorse or promote products derived from | |
17 | # this software without specific prior written permission. | |
18 | # | |
19 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS | |
20 | # "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT | |
21 | # LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR | |
22 | # A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT | |
23 | # OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, | |
24 | # SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT | |
25 | # LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, | |
26 | # DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY | |
27 | # THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT | |
28 | # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
29 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
30 | ||
31 | """Does google-lint on c++ files. | |
32 | ||
33 | The goal of this script is to identify places in the code that *may* | |
34 | be in non-compliance with google style. It does not attempt to fix | |
35 | up these problems -- the point is to educate. It does also not | |
36 | attempt to find all problems, or to ensure that everything it does | |
37 | find is legitimately a problem. | |
38 | ||
39 | In particular, we can get very confused by /* and // inside strings! | |
40 | We do a small hack, which is to ignore //'s with "'s after them on the | |
41 | same line, but it is far from perfect (in either direction). | |
42 | """ | |
43 | ||
44 | import codecs | |
45 | import copy | |
46 | import getopt | |
47 | import glob | |
48 | import itertools | |
49 | import math # for log | |
50 | import os | |
51 | import re | |
52 | import sre_compile | |
53 | import string | |
54 | import sys | |
55 | import unicodedata | |
56 | import xml.etree.ElementTree | |
57 | ||
58 | # if empty, use defaults | |
59 | _header_extensions = set([]) | |
60 | ||
61 | # if empty, use defaults | |
62 | _valid_extensions = set([]) | |
63 | ||
64 | ||
65 | # Files with any of these extensions are considered to be | |
66 | # header files (and will undergo different style checks). | |
67 | # This set can be extended by using the --headers | |
68 | # option (also supported in CPPLINT.cfg) | |
69 | def GetHeaderExtensions(): | |
70 | if not _header_extensions: | |
71 | return set(['h', 'hpp', 'hxx', 'h++', 'cuh']) | |
72 | return _header_extensions | |
73 | ||
74 | # The allowed extensions for file names | |
75 | # This is set by --extensions flag | |
76 | def GetAllExtensions(): | |
77 | if not _valid_extensions: | |
78 | return GetHeaderExtensions().union(set(['c', 'cc', 'cpp', 'cxx', 'c++', 'cu'])) | |
79 | return _valid_extensions | |
80 | ||
81 | def GetNonHeaderExtensions(): | |
82 | return GetAllExtensions().difference(GetHeaderExtensions()) | |
83 | ||
84 | ||
85 | _USAGE = """ | |
86 | Syntax: cpplint.py [--verbose=#] [--output=emacs|eclipse|vs7|junit] | |
87 | [--filter=-x,+y,...] | |
88 | [--counting=total|toplevel|detailed] [--repository=path] | |
89 | [--root=subdir] [--linelength=digits] [--recursive] | |
90 | [--exclude=path] | |
91 | [--headers=ext1,ext2] | |
92 | [--extensions=hpp,cpp,...] | |
93 | <file> [file] ... | |
94 | ||
95 | The style guidelines this tries to follow are those in | |
96 | https://google.github.io/styleguide/cppguide.html | |
97 | ||
98 | Every problem is given a confidence score from 1-5, with 5 meaning we are | |
99 | certain of the problem, and 1 meaning it could be a legitimate construct. | |
100 | This will miss some errors, and is not a substitute for a code review. | |
101 | ||
102 | To suppress false-positive errors of a certain category, add a | |
103 | 'NOLINT(category)' comment to the line. NOLINT or NOLINT(*) | |
104 | suppresses errors of all categories on that line. | |
105 | ||
106 | The files passed in will be linted; at least one file must be provided. | |
107 | Default linted extensions are %s. | |
108 | Other file types will be ignored. | |
109 | Change the extensions with the --extensions flag. | |
110 | ||
111 | Flags: | |
112 | ||
113 | output=emacs|eclipse|vs7|junit | |
114 | By default, the output is formatted to ease emacs parsing. Output | |
115 | compatible with eclipse (eclipse), Visual Studio (vs7), and JUnit | |
116 | XML parsers such as those used in Jenkins and Bamboo may also be | |
117 | used. Other formats are unsupported. | |
118 | ||
119 | verbose=# | |
120 | Specify a number 0-5 to restrict errors to certain verbosity levels. | |
121 | Errors with lower verbosity levels have lower confidence and are more | |
122 | likely to be false positives. | |
123 | ||
124 | quiet | |
125 | Suppress output other than linting errors, such as information about | |
126 | which files have been processed and excluded. | |
127 | ||
128 | filter=-x,+y,... | |
129 | Specify a comma-separated list of category-filters to apply: only | |
130 | error messages whose category names pass the filters will be printed. | |
131 | (Category names are printed with the message and look like | |
132 | "[whitespace/indent]".) Filters are evaluated left to right. | |
133 | "-FOO" and "FOO" means "do not print categories that start with FOO". | |
134 | "+FOO" means "do print categories that start with FOO". | |
135 | ||
136 | Examples: --filter=-whitespace,+whitespace/braces | |
137 | --filter=whitespace,runtime/printf,+runtime/printf_format | |
138 | --filter=-,+build/include_what_you_use | |
139 | ||
140 | To see a list of all the categories used in cpplint, pass no arg: | |
141 | --filter= | |
142 | ||
143 | counting=total|toplevel|detailed | |
144 | The total number of errors found is always printed. If | |
145 | 'toplevel' is provided, then the count of errors in each of | |
146 | the top-level categories like 'build' and 'whitespace' will | |
147 | also be printed. If 'detailed' is provided, then a count | |
148 | is provided for each category like 'build/class'. | |
149 | ||
150 | repository=path | |
151 | The top level directory of the repository, used to derive the header | |
152 | guard CPP variable. By default, this is determined by searching for a | |
153 | path that contains .git, .hg, or .svn. When this flag is specified, the | |
154 | given path is used instead. This option allows the header guard CPP | |
155 | variable to remain consistent even if members of a team have different | |
156 | repository root directories (such as when checking out a subdirectory | |
157 | with SVN). In addition, users of non-mainstream version control systems | |
158 | can use this flag to ensure readable header guard CPP variables. | |
159 | ||
160 | Examples: | |
161 | Assuming that Alice checks out ProjectName and Bob checks out | |
162 | ProjectName/trunk and trunk contains src/chrome/ui/browser.h, then | |
163 | with no --repository flag, the header guard CPP variable will be: | |
164 | ||
165 | Alice => TRUNK_SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
166 | Bob => SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
167 | ||
168 | If Alice uses the --repository=trunk flag and Bob omits the flag or | |
169 | uses --repository=. then the header guard CPP variable will be: | |
170 | ||
171 | Alice => SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
172 | Bob => SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
173 | ||
174 | root=subdir | |
175 | The root directory used for deriving header guard CPP variables. This | |
176 | directory is relative to the top level directory of the repository which | |
177 | by default is determined by searching for a directory that contains .git, | |
178 | .hg, or .svn but can also be controlled with the --repository flag. If | |
179 | the specified directory does not exist, this flag is ignored. | |
180 | ||
181 | Examples: | |
182 | Assuming that src is the top level directory of the repository, the | |
183 | header guard CPP variables for src/chrome/browser/ui/browser.h are: | |
184 | ||
185 | No flag => CHROME_BROWSER_UI_BROWSER_H_ | |
186 | --root=chrome => BROWSER_UI_BROWSER_H_ | |
187 | --root=chrome/browser => UI_BROWSER_H_ | |
188 | ||
189 | linelength=digits | |
190 | This is the allowed line length for the project. The default value is | |
191 | 80 characters. | |
192 | ||
193 | Examples: | |
194 | --linelength=120 | |
195 | ||
196 | recursive | |
197 | Search for files to lint recursively. Each directory given in the list | |
198 | of files to be linted is replaced by all files that descend from that | |
199 | directory. Files with extensions not in the valid extensions list are | |
200 | excluded. | |
201 | ||
202 | exclude=path | |
203 | Exclude the given path from the list of files to be linted. Relative | |
204 | paths are evaluated relative to the current directory and shell globbing | |
205 | is performed. This flag can be provided multiple times to exclude | |
206 | multiple files. | |
207 | ||
208 | Examples: | |
209 | --exclude=one.cc | |
210 | --exclude=src/*.cc | |
211 | --exclude=src/*.cc --exclude=test/*.cc | |
212 | ||
213 | extensions=extension,extension,... | |
214 | The allowed file extensions that cpplint will check | |
215 | ||
216 | Examples: | |
217 | --extensions=%s | |
218 | ||
219 | headers=extension,extension,... | |
220 | The allowed header extensions that cpplint will consider to be header files | |
221 | (by default, only files with extensions %s | |
222 | will be assumed to be headers) | |
223 | ||
224 | Examples: | |
225 | --headers=%s | |
226 | ||
227 | cpplint.py supports per-directory configurations specified in CPPLINT.cfg | |
228 | files. CPPLINT.cfg file can contain a number of key=value pairs. | |
229 | Currently the following options are supported: | |
230 | ||
231 | set noparent | |
232 | filter=+filter1,-filter2,... | |
233 | exclude_files=regex | |
234 | linelength=80 | |
235 | root=subdir | |
236 | ||
237 | "set noparent" option prevents cpplint from traversing directory tree | |
238 | upwards looking for more .cfg files in parent directories. This option | |
239 | is usually placed in the top-level project directory. | |
240 | ||
241 | The "filter" option is similar in function to --filter flag. It specifies | |
242 | message filters in addition to the |_DEFAULT_FILTERS| and those specified | |
243 | through --filter command-line flag. | |
244 | ||
245 | "exclude_files" allows to specify a regular expression to be matched against | |
246 | a file name. If the expression matches, the file is skipped and not run | |
247 | through the linter. | |
248 | ||
249 | "linelength" specifies the allowed line length for the project. | |
250 | ||
251 | The "root" option is similar in function to the --root flag (see example | |
252 | above). | |
253 | ||
254 | CPPLINT.cfg has an effect on files in the same directory and all | |
255 | subdirectories, unless overridden by a nested configuration file. | |
256 | ||
257 | Example file: | |
258 | filter=-build/include_order,+build/include_alpha | |
259 | exclude_files=.*\\.cc | |
260 | ||
261 | The above example disables build/include_order warning and enables | |
262 | build/include_alpha as well as excludes all .cc from being | |
263 | processed by linter, in the current directory (where the .cfg | |
264 | file is located) and all subdirectories. | |
265 | """ % (list(GetAllExtensions()), | |
266 | ','.join(list(GetAllExtensions())), | |
267 | GetHeaderExtensions(), | |
268 | ','.join(GetHeaderExtensions())) | |
269 | ||
270 | # We categorize each error message we print. Here are the categories. | |
271 | # We want an explicit list so we can list them all in cpplint --filter=. | |
272 | # If you add a new error message with a new category, add it to the list | |
273 | # here! cpplint_unittest.py should tell you if you forget to do this. | |
274 | _ERROR_CATEGORIES = [ | |
275 | 'build/class', | |
276 | 'build/c++11', | |
277 | 'build/c++14', | |
278 | 'build/c++tr1', | |
279 | 'build/deprecated', | |
280 | 'build/endif_comment', | |
281 | 'build/explicit_make_pair', | |
282 | 'build/forward_decl', | |
283 | 'build/header_guard', | |
284 | 'build/include', | |
285 | 'build/include_subdir', | |
286 | 'build/include_alpha', | |
287 | 'build/include_order', | |
288 | 'build/include_what_you_use', | |
289 | 'build/namespaces_literals', | |
290 | 'build/namespaces', | |
291 | 'build/printf_format', | |
292 | 'build/storage_class', | |
293 | 'legal/copyright', | |
294 | 'readability/alt_tokens', | |
295 | 'readability/braces', | |
296 | 'readability/casting', | |
297 | 'readability/check', | |
298 | 'readability/constructors', | |
299 | 'readability/fn_size', | |
300 | 'readability/inheritance', | |
301 | 'readability/multiline_comment', | |
302 | 'readability/multiline_string', | |
303 | 'readability/namespace', | |
304 | 'readability/nolint', | |
305 | 'readability/nul', | |
306 | 'readability/strings', | |
307 | 'readability/todo', | |
308 | 'readability/utf8', | |
309 | 'runtime/arrays', | |
310 | 'runtime/casting', | |
311 | 'runtime/explicit', | |
312 | 'runtime/int', | |
313 | 'runtime/init', | |
314 | 'runtime/invalid_increment', | |
315 | 'runtime/member_string_references', | |
316 | 'runtime/memset', | |
317 | 'runtime/indentation_namespace', | |
318 | 'runtime/operator', | |
319 | 'runtime/printf', | |
320 | 'runtime/printf_format', | |
321 | 'runtime/references', | |
322 | 'runtime/string', | |
323 | 'runtime/threadsafe_fn', | |
324 | 'runtime/vlog', | |
325 | 'whitespace/blank_line', | |
326 | 'whitespace/braces', | |
327 | 'whitespace/comma', | |
328 | 'whitespace/comments', | |
329 | 'whitespace/empty_conditional_body', | |
330 | 'whitespace/empty_if_body', | |
331 | 'whitespace/empty_loop_body', | |
332 | 'whitespace/end_of_line', | |
333 | 'whitespace/ending_newline', | |
334 | 'whitespace/forcolon', | |
335 | 'whitespace/indent', | |
336 | 'whitespace/line_length', | |
337 | 'whitespace/newline', | |
338 | 'whitespace/operators', | |
339 | 'whitespace/parens', | |
340 | 'whitespace/semicolon', | |
341 | 'whitespace/tab', | |
342 | 'whitespace/todo', | |
343 | ] | |
344 | ||
345 | # These error categories are no longer enforced by cpplint, but for backwards- | |
346 | # compatibility they may still appear in NOLINT comments. | |
347 | _LEGACY_ERROR_CATEGORIES = [ | |
348 | 'readability/streams', | |
349 | 'readability/function', | |
350 | ] | |
351 | ||
352 | # The default state of the category filter. This is overridden by the --filter= | |
353 | # flag. By default all errors are on, so only add here categories that should be | |
354 | # off by default (i.e., categories that must be enabled by the --filter= flags). | |
355 | # All entries here should start with a '-' or '+', as in the --filter= flag. | |
356 | _DEFAULT_FILTERS = ['-build/include_alpha'] | |
357 | ||
358 | # The default list of categories suppressed for C (not C++) files. | |
359 | _DEFAULT_C_SUPPRESSED_CATEGORIES = [ | |
360 | 'readability/casting', | |
361 | ] | |
362 | ||
363 | # The default list of categories suppressed for Linux Kernel files. | |
364 | _DEFAULT_KERNEL_SUPPRESSED_CATEGORIES = [ | |
365 | 'whitespace/tab', | |
366 | ] | |
367 | ||
368 | # We used to check for high-bit characters, but after much discussion we | |
369 | # decided those were OK, as long as they were in UTF-8 and didn't represent | |
370 | # hard-coded international strings, which belong in a separate i18n file. | |
371 | ||
372 | # C++ headers | |
373 | _CPP_HEADERS = frozenset([ | |
374 | # Legacy | |
375 | 'algobase.h', | |
376 | 'algo.h', | |
377 | 'alloc.h', | |
378 | 'builtinbuf.h', | |
379 | 'bvector.h', | |
380 | 'complex.h', | |
381 | 'defalloc.h', | |
382 | 'deque.h', | |
383 | 'editbuf.h', | |
384 | 'fstream.h', | |
385 | 'function.h', | |
386 | 'hash_map', | |
387 | 'hash_map.h', | |
388 | 'hash_set', | |
389 | 'hash_set.h', | |
390 | 'hashtable.h', | |
391 | 'heap.h', | |
392 | 'indstream.h', | |
393 | 'iomanip.h', | |
394 | 'iostream.h', | |
395 | 'istream.h', | |
396 | 'iterator.h', | |
397 | 'list.h', | |
398 | 'map.h', | |
399 | 'multimap.h', | |
400 | 'multiset.h', | |
401 | 'ostream.h', | |
402 | 'pair.h', | |
403 | 'parsestream.h', | |
404 | 'pfstream.h', | |
405 | 'procbuf.h', | |
406 | 'pthread_alloc', | |
407 | 'pthread_alloc.h', | |
408 | 'rope', | |
409 | 'rope.h', | |
410 | 'ropeimpl.h', | |
411 | 'set.h', | |
412 | 'slist', | |
413 | 'slist.h', | |
414 | 'stack.h', | |
415 | 'stdiostream.h', | |
416 | 'stl_alloc.h', | |
417 | 'stl_relops.h', | |
418 | 'streambuf.h', | |
419 | 'stream.h', | |
420 | 'strfile.h', | |
421 | 'strstream.h', | |
422 | 'tempbuf.h', | |
423 | 'tree.h', | |
424 | 'type_traits.h', | |
425 | 'vector.h', | |
426 | # 17.6.1.2 C++ library headers | |
427 | 'algorithm', | |
428 | 'array', | |
429 | 'atomic', | |
430 | 'bitset', | |
431 | 'chrono', | |
432 | 'codecvt', | |
433 | 'complex', | |
434 | 'condition_variable', | |
435 | 'deque', | |
436 | 'exception', | |
437 | 'forward_list', | |
438 | 'fstream', | |
439 | 'functional', | |
440 | 'future', | |
441 | 'initializer_list', | |
442 | 'iomanip', | |
443 | 'ios', | |
444 | 'iosfwd', | |
445 | 'iostream', | |
446 | 'istream', | |
447 | 'iterator', | |
448 | 'limits', | |
449 | 'list', | |
450 | 'locale', | |
451 | 'map', | |
452 | 'memory', | |
453 | 'mutex', | |
454 | 'new', | |
455 | 'numeric', | |
456 | 'ostream', | |
457 | 'queue', | |
458 | 'random', | |
459 | 'ratio', | |
460 | 'regex', | |
461 | 'scoped_allocator', | |
462 | 'set', | |
463 | 'sstream', | |
464 | 'stack', | |
465 | 'stdexcept', | |
466 | 'streambuf', | |
467 | 'string', | |
468 | 'strstream', | |
469 | 'system_error', | |
470 | 'thread', | |
471 | 'tuple', | |
472 | 'typeindex', | |
473 | 'typeinfo', | |
474 | 'type_traits', | |
475 | 'unordered_map', | |
476 | 'unordered_set', | |
477 | 'utility', | |
478 | 'valarray', | |
479 | 'vector', | |
480 | # 17.6.1.2 C++ headers for C library facilities | |
481 | 'cassert', | |
482 | 'ccomplex', | |
483 | 'cctype', | |
484 | 'cerrno', | |
485 | 'cfenv', | |
486 | 'cfloat', | |
487 | 'cinttypes', | |
488 | 'ciso646', | |
489 | 'climits', | |
490 | 'clocale', | |
491 | 'cmath', | |
492 | 'csetjmp', | |
493 | 'csignal', | |
494 | 'cstdalign', | |
495 | 'cstdarg', | |
496 | 'cstdbool', | |
497 | 'cstddef', | |
498 | 'cstdint', | |
499 | 'cstdio', | |
500 | 'cstdlib', | |
501 | 'cstring', | |
502 | 'ctgmath', | |
503 | 'ctime', | |
504 | 'cuchar', | |
505 | 'cwchar', | |
506 | 'cwctype', | |
507 | ]) | |
508 | ||
509 | # Type names | |
510 | _TYPES = re.compile( | |
511 | r'^(?:' | |
512 | # [dcl.type.simple] | |
513 | r'(char(16_t|32_t)?)|wchar_t|' | |
514 | r'bool|short|int|long|signed|unsigned|float|double|' | |
515 | # [support.types] | |
516 | r'(ptrdiff_t|size_t|max_align_t|nullptr_t)|' | |
517 | # [cstdint.syn] | |
518 | r'(u?int(_fast|_least)?(8|16|32|64)_t)|' | |
519 | r'(u?int(max|ptr)_t)|' | |
520 | r')$') | |
521 | ||
522 | ||
523 | # These headers are excluded from [build/include] and [build/include_order] | |
524 | # checks: | |
525 | # - Anything not following google file name conventions (containing an | |
526 | # uppercase character, such as Python.h or nsStringAPI.h, for example). | |
527 | # - Lua headers. | |
528 | _THIRD_PARTY_HEADERS_PATTERN = re.compile( | |
529 | r'^(?:[^/]*[A-Z][^/]*\.h|lua\.h|lauxlib\.h|lualib\.h)$') | |
530 | ||
531 | # Pattern for matching FileInfo.BaseName() against test file name | |
532 | _test_suffixes = ['_test', '_regtest', '_unittest'] | |
533 | _TEST_FILE_SUFFIX = '(' + '|'.join(_test_suffixes) + r')$' | |
534 | ||
535 | # Pattern that matches only complete whitespace, possibly across multiple lines. | |
536 | _EMPTY_CONDITIONAL_BODY_PATTERN = re.compile(r'^\s*$', re.DOTALL) | |
537 | ||
538 | # Assertion macros. These are defined in base/logging.h and | |
539 | # testing/base/public/gunit.h. | |
540 | _CHECK_MACROS = [ | |
541 | 'DCHECK', 'CHECK', | |
542 | 'EXPECT_TRUE', 'ASSERT_TRUE', | |
543 | 'EXPECT_FALSE', 'ASSERT_FALSE', | |
544 | ] | |
545 | ||
546 | # Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE | |
547 | _CHECK_REPLACEMENT = dict([(macro_var, {}) for macro_var in _CHECK_MACROS]) | |
548 | ||
549 | for op, replacement in [('==', 'EQ'), ('!=', 'NE'), | |
550 | ('>=', 'GE'), ('>', 'GT'), | |
551 | ('<=', 'LE'), ('<', 'LT')]: | |
552 | _CHECK_REPLACEMENT['DCHECK'][op] = 'DCHECK_%s' % replacement | |
553 | _CHECK_REPLACEMENT['CHECK'][op] = 'CHECK_%s' % replacement | |
554 | _CHECK_REPLACEMENT['EXPECT_TRUE'][op] = 'EXPECT_%s' % replacement | |
555 | _CHECK_REPLACEMENT['ASSERT_TRUE'][op] = 'ASSERT_%s' % replacement | |
556 | ||
557 | for op, inv_replacement in [('==', 'NE'), ('!=', 'EQ'), | |
558 | ('>=', 'LT'), ('>', 'LE'), | |
559 | ('<=', 'GT'), ('<', 'GE')]: | |
560 | _CHECK_REPLACEMENT['EXPECT_FALSE'][op] = 'EXPECT_%s' % inv_replacement | |
561 | _CHECK_REPLACEMENT['ASSERT_FALSE'][op] = 'ASSERT_%s' % inv_replacement | |
562 | ||
563 | # Alternative tokens and their replacements. For full list, see section 2.5 | |
564 | # Alternative tokens [lex.digraph] in the C++ standard. | |
565 | # | |
566 | # Digraphs (such as '%:') are not included here since it's a mess to | |
567 | # match those on a word boundary. | |
568 | _ALT_TOKEN_REPLACEMENT = { | |
569 | 'and': '&&', | |
570 | 'bitor': '|', | |
571 | 'or': '||', | |
572 | 'xor': '^', | |
573 | 'compl': '~', | |
574 | 'bitand': '&', | |
575 | 'and_eq': '&=', | |
576 | 'or_eq': '|=', | |
577 | 'xor_eq': '^=', | |
578 | 'not': '!', | |
579 | 'not_eq': '!=' | |
580 | } | |
581 | ||
582 | # Compile regular expression that matches all the above keywords. The "[ =()]" | |
583 | # bit is meant to avoid matching these keywords outside of boolean expressions. | |
584 | # | |
585 | # False positives include C-style multi-line comments and multi-line strings | |
586 | # but those have always been troublesome for cpplint. | |
587 | _ALT_TOKEN_REPLACEMENT_PATTERN = re.compile( | |
588 | r'[ =()](' + ('|'.join(_ALT_TOKEN_REPLACEMENT.keys())) + r')(?=[ (]|$)') | |
589 | ||
590 | ||
591 | # These constants define types of headers for use with | |
592 | # _IncludeState.CheckNextIncludeOrder(). | |
593 | _C_SYS_HEADER = 1 | |
594 | _CPP_SYS_HEADER = 2 | |
595 | _LIKELY_MY_HEADER = 3 | |
596 | _POSSIBLE_MY_HEADER = 4 | |
597 | _OTHER_HEADER = 5 | |
598 | ||
599 | # These constants define the current inline assembly state | |
600 | _NO_ASM = 0 # Outside of inline assembly block | |
601 | _INSIDE_ASM = 1 # Inside inline assembly block | |
602 | _END_ASM = 2 # Last line of inline assembly block | |
603 | _BLOCK_ASM = 3 # The whole block is an inline assembly block | |
604 | ||
605 | # Match start of assembly blocks | |
606 | _MATCH_ASM = re.compile(r'^\s*(?:asm|_asm|__asm|__asm__)' | |
607 | r'(?:\s+(volatile|__volatile__))?' | |
608 | r'\s*[{(]') | |
609 | ||
610 | # Match strings that indicate we're working on a C (not C++) file. | |
611 | _SEARCH_C_FILE = re.compile(r'\b(?:LINT_C_FILE|' | |
612 | r'vim?:\s*.*(\s*|:)filetype=c(\s*|:|$))') | |
613 | ||
614 | # Match string that indicates we're working on a Linux Kernel file. | |
615 | _SEARCH_KERNEL_FILE = re.compile(r'\b(?:LINT_KERNEL_FILE)') | |
616 | ||
617 | _regexp_compile_cache = {} | |
618 | ||
619 | # {str, set(int)}: a map from error categories to sets of linenumbers | |
620 | # on which those errors are expected and should be suppressed. | |
621 | _error_suppressions = {} | |
622 | ||
623 | # The root directory used for deriving header guard CPP variable. | |
624 | # This is set by --root flag. | |
625 | _root = None | |
626 | ||
627 | # The top level repository directory. If set, _root is calculated relative to | |
628 | # this directory instead of the directory containing version control artifacts. | |
629 | # This is set by the --repository flag. | |
630 | _repository = None | |
631 | ||
632 | # Files to exclude from linting. This is set by the --exclude flag. | |
633 | _excludes = None | |
634 | ||
635 | # Whether to suppress PrintInfo messages | |
636 | _quiet = False | |
637 | ||
638 | # The allowed line length of files. | |
639 | # This is set by --linelength flag. | |
640 | _line_length = 80 | |
641 | ||
642 | try: | |
643 | xrange(1, 0) | |
644 | except NameError: | |
645 | # -- pylint: disable=redefined-builtin | |
646 | xrange = range | |
647 | ||
648 | try: | |
649 | unicode | |
650 | except NameError: | |
651 | # -- pylint: disable=redefined-builtin | |
652 | basestring = unicode = str | |
653 | ||
654 | try: | |
655 | long(2) | |
656 | except NameError: | |
657 | # -- pylint: disable=redefined-builtin | |
658 | long = int | |
659 | ||
660 | if sys.version_info < (3,): | |
661 | # -- pylint: disable=no-member | |
662 | # BINARY_TYPE = str | |
663 | itervalues = dict.itervalues | |
664 | iteritems = dict.iteritems | |
665 | else: | |
666 | # BINARY_TYPE = bytes | |
667 | itervalues = dict.values | |
668 | iteritems = dict.items | |
669 | ||
670 | def unicode_escape_decode(x): | |
671 | if sys.version_info < (3,): | |
672 | return codecs.unicode_escape_decode(x)[0] | |
673 | else: | |
674 | return x | |
675 | ||
676 | # {str, bool}: a map from error categories to booleans which indicate if the | |
677 | # category should be suppressed for every line. | |
678 | _global_error_suppressions = {} | |
679 | ||
680 | ||
681 | ||
682 | ||
683 | def ParseNolintSuppressions(filename, raw_line, linenum, error): | |
684 | """Updates the global list of line error-suppressions. | |
685 | ||
686 | Parses any NOLINT comments on the current line, updating the global | |
687 | error_suppressions store. Reports an error if the NOLINT comment | |
688 | was malformed. | |
689 | ||
690 | Args: | |
691 | filename: str, the name of the input file. | |
692 | raw_line: str, the line of input text, with comments. | |
693 | linenum: int, the number of the current line. | |
694 | error: function, an error handler. | |
695 | """ | |
696 | matched = Search(r'\bNOLINT(NEXTLINE)?\b(\([^)]+\))?', raw_line) | |
697 | if matched: | |
698 | if matched.group(1): | |
699 | suppressed_line = linenum + 1 | |
700 | else: | |
701 | suppressed_line = linenum | |
702 | category = matched.group(2) | |
703 | if category in (None, '(*)'): # => "suppress all" | |
704 | _error_suppressions.setdefault(None, set()).add(suppressed_line) | |
705 | else: | |
706 | if category.startswith('(') and category.endswith(')'): | |
707 | category = category[1:-1] | |
708 | if category in _ERROR_CATEGORIES: | |
709 | _error_suppressions.setdefault(category, set()).add(suppressed_line) | |
710 | elif category not in _LEGACY_ERROR_CATEGORIES: | |
711 | error(filename, linenum, 'readability/nolint', 5, | |
712 | 'Unknown NOLINT error category: %s' % category) | |
713 | ||
714 | ||
715 | def ProcessGlobalSuppresions(lines): | |
716 | """Updates the list of global error suppressions. | |
717 | ||
718 | Parses any lint directives in the file that have global effect. | |
719 | ||
720 | Args: | |
721 | lines: An array of strings, each representing a line of the file, with the | |
722 | last element being empty if the file is terminated with a newline. | |
723 | """ | |
724 | for line in lines: | |
725 | if _SEARCH_C_FILE.search(line): | |
726 | for category in _DEFAULT_C_SUPPRESSED_CATEGORIES: | |
727 | _global_error_suppressions[category] = True | |
728 | if _SEARCH_KERNEL_FILE.search(line): | |
729 | for category in _DEFAULT_KERNEL_SUPPRESSED_CATEGORIES: | |
730 | _global_error_suppressions[category] = True | |
731 | ||
732 | ||
733 | def ResetNolintSuppressions(): | |
734 | """Resets the set of NOLINT suppressions to empty.""" | |
735 | _error_suppressions.clear() | |
736 | _global_error_suppressions.clear() | |
737 | ||
738 | ||
739 | def IsErrorSuppressedByNolint(category, linenum): | |
740 | """Returns true if the specified error category is suppressed on this line. | |
741 | ||
742 | Consults the global error_suppressions map populated by | |
743 | ParseNolintSuppressions/ProcessGlobalSuppresions/ResetNolintSuppressions. | |
744 | ||
745 | Args: | |
746 | category: str, the category of the error. | |
747 | linenum: int, the current line number. | |
748 | Returns: | |
749 | bool, True iff the error should be suppressed due to a NOLINT comment or | |
750 | global suppression. | |
751 | """ | |
752 | return (_global_error_suppressions.get(category, False) or | |
753 | linenum in _error_suppressions.get(category, set()) or | |
754 | linenum in _error_suppressions.get(None, set())) | |
755 | ||
756 | ||
757 | def Match(pattern, s): | |
758 | """Matches the string with the pattern, caching the compiled regexp.""" | |
759 | # The regexp compilation caching is inlined in both Match and Search for | |
760 | # performance reasons; factoring it out into a separate function turns out | |
761 | # to be noticeably expensive. | |
762 | if pattern not in _regexp_compile_cache: | |
763 | _regexp_compile_cache[pattern] = sre_compile.compile(pattern) | |
764 | return _regexp_compile_cache[pattern].match(s) | |
765 | ||
766 | ||
767 | def ReplaceAll(pattern, rep, s): | |
768 | """Replaces instances of pattern in a string with a replacement. | |
769 | ||
770 | The compiled regex is kept in a cache shared by Match and Search. | |
771 | ||
772 | Args: | |
773 | pattern: regex pattern | |
774 | rep: replacement text | |
775 | s: search string | |
776 | ||
777 | Returns: | |
778 | string with replacements made (or original string if no replacements) | |
779 | """ | |
780 | if pattern not in _regexp_compile_cache: | |
781 | _regexp_compile_cache[pattern] = sre_compile.compile(pattern) | |
782 | return _regexp_compile_cache[pattern].sub(rep, s) | |
783 | ||
784 | ||
785 | def Search(pattern, s): | |
786 | """Searches the string for the pattern, caching the compiled regexp.""" | |
787 | if pattern not in _regexp_compile_cache: | |
788 | _regexp_compile_cache[pattern] = sre_compile.compile(pattern) | |
789 | return _regexp_compile_cache[pattern].search(s) | |
790 | ||
791 | ||
792 | def _IsSourceExtension(s): | |
793 | """File extension (excluding dot) matches a source file extension.""" | |
794 | return s in GetNonHeaderExtensions() | |
795 | ||
796 | ||
797 | class _IncludeState(object): | |
798 | """Tracks line numbers for includes, and the order in which includes appear. | |
799 | ||
800 | include_list contains list of lists of (header, line number) pairs. | |
801 | It's a lists of lists rather than just one flat list to make it | |
802 | easier to update across preprocessor boundaries. | |
803 | ||
804 | Call CheckNextIncludeOrder() once for each header in the file, passing | |
805 | in the type constants defined above. Calls in an illegal order will | |
806 | raise an _IncludeError with an appropriate error message. | |
807 | ||
808 | """ | |
809 | # self._section will move monotonically through this set. If it ever | |
810 | # needs to move backwards, CheckNextIncludeOrder will raise an error. | |
811 | _INITIAL_SECTION = 0 | |
812 | _MY_H_SECTION = 1 | |
813 | _C_SECTION = 2 | |
814 | _CPP_SECTION = 3 | |
815 | _OTHER_H_SECTION = 4 | |
816 | ||
817 | _TYPE_NAMES = { | |
818 | _C_SYS_HEADER: 'C system header', | |
819 | _CPP_SYS_HEADER: 'C++ system header', | |
820 | _LIKELY_MY_HEADER: 'header this file implements', | |
821 | _POSSIBLE_MY_HEADER: 'header this file may implement', | |
822 | _OTHER_HEADER: 'other header', | |
823 | } | |
824 | _SECTION_NAMES = { | |
825 | _INITIAL_SECTION: "... nothing. (This can't be an error.)", | |
826 | _MY_H_SECTION: 'a header this file implements', | |
827 | _C_SECTION: 'C system header', | |
828 | _CPP_SECTION: 'C++ system header', | |
829 | _OTHER_H_SECTION: 'other header', | |
830 | } | |
831 | ||
832 | def __init__(self): | |
833 | self.include_list = [[]] | |
834 | self._section = None | |
835 | self._last_header = None | |
836 | self.ResetSection('') | |
837 | ||
838 | def FindHeader(self, header): | |
839 | """Check if a header has already been included. | |
840 | ||
841 | Args: | |
842 | header: header to check. | |
843 | Returns: | |
844 | Line number of previous occurrence, or -1 if the header has not | |
845 | been seen before. | |
846 | """ | |
847 | for section_list in self.include_list: | |
848 | for f in section_list: | |
849 | if f[0] == header: | |
850 | return f[1] | |
851 | return -1 | |
852 | ||
853 | def ResetSection(self, directive): | |
854 | """Reset section checking for preprocessor directive. | |
855 | ||
856 | Args: | |
857 | directive: preprocessor directive (e.g. "if", "else"). | |
858 | """ | |
859 | # The name of the current section. | |
860 | self._section = self._INITIAL_SECTION | |
861 | # The path of last found header. | |
862 | self._last_header = '' | |
863 | ||
864 | # Update list of includes. Note that we never pop from the | |
865 | # include list. | |
866 | if directive in ('if', 'ifdef', 'ifndef'): | |
867 | self.include_list.append([]) | |
868 | elif directive in ('else', 'elif'): | |
869 | self.include_list[-1] = [] | |
870 | ||
871 | def SetLastHeader(self, header_path): | |
872 | self._last_header = header_path | |
873 | ||
874 | def CanonicalizeAlphabeticalOrder(self, header_path): | |
875 | """Returns a path canonicalized for alphabetical comparison. | |
876 | ||
877 | - replaces "-" with "_" so they both cmp the same. | |
878 | - removes '-inl' since we don't require them to be after the main header. | |
879 | - lowercase everything, just in case. | |
880 | ||
881 | Args: | |
882 | header_path: Path to be canonicalized. | |
883 | ||
884 | Returns: | |
885 | Canonicalized path. | |
886 | """ | |
887 | return header_path.replace('-inl.h', '.h').replace('-', '_').lower() | |
888 | ||
889 | def IsInAlphabeticalOrder(self, clean_lines, linenum, header_path): | |
890 | """Check if a header is in alphabetical order with the previous header. | |
891 | ||
892 | Args: | |
893 | clean_lines: A CleansedLines instance containing the file. | |
894 | linenum: The number of the line to check. | |
895 | header_path: Canonicalized header to be checked. | |
896 | ||
897 | Returns: | |
898 | Returns true if the header is in alphabetical order. | |
899 | """ | |
900 | # If previous section is different from current section, _last_header will | |
901 | # be reset to empty string, so it's always less than current header. | |
902 | # | |
903 | # If previous line was a blank line, assume that the headers are | |
904 | # intentionally sorted the way they are. | |
905 | if (self._last_header > header_path and | |
906 | Match(r'^\s*#\s*include\b', clean_lines.elided[linenum - 1])): | |
907 | return False | |
908 | return True | |
909 | ||
910 | def CheckNextIncludeOrder(self, header_type): | |
911 | """Returns a non-empty error message if the next header is out of order. | |
912 | ||
913 | This function also updates the internal state to be ready to check | |
914 | the next include. | |
915 | ||
916 | Args: | |
917 | header_type: One of the _XXX_HEADER constants defined above. | |
918 | ||
919 | Returns: | |
920 | The empty string if the header is in the right order, or an | |
921 | error message describing what's wrong. | |
922 | ||
923 | """ | |
924 | error_message = ('Found %s after %s' % | |
925 | (self._TYPE_NAMES[header_type], | |
926 | self._SECTION_NAMES[self._section])) | |
927 | ||
928 | last_section = self._section | |
929 | ||
930 | if header_type == _C_SYS_HEADER: | |
931 | if self._section <= self._C_SECTION: | |
932 | self._section = self._C_SECTION | |
933 | else: | |
934 | self._last_header = '' | |
935 | return error_message | |
936 | elif header_type == _CPP_SYS_HEADER: | |
937 | if self._section <= self._CPP_SECTION: | |
938 | self._section = self._CPP_SECTION | |
939 | else: | |
940 | self._last_header = '' | |
941 | return error_message | |
942 | elif header_type == _LIKELY_MY_HEADER: | |
943 | if self._section <= self._MY_H_SECTION: | |
944 | self._section = self._MY_H_SECTION | |
945 | else: | |
946 | self._section = self._OTHER_H_SECTION | |
947 | elif header_type == _POSSIBLE_MY_HEADER: | |
948 | if self._section <= self._MY_H_SECTION: | |
949 | self._section = self._MY_H_SECTION | |
950 | else: | |
951 | # This will always be the fallback because we're not sure | |
952 | # enough that the header is associated with this file. | |
953 | self._section = self._OTHER_H_SECTION | |
954 | else: | |
955 | assert header_type == _OTHER_HEADER | |
956 | self._section = self._OTHER_H_SECTION | |
957 | ||
958 | if last_section != self._section: | |
959 | self._last_header = '' | |
960 | ||
961 | return '' | |
962 | ||
963 | ||
964 | class _CppLintState(object): | |
965 | """Maintains module-wide state..""" | |
966 | ||
967 | def __init__(self): | |
968 | self.verbose_level = 1 # global setting. | |
969 | self.error_count = 0 # global count of reported errors | |
970 | # filters to apply when emitting error messages | |
971 | self.filters = _DEFAULT_FILTERS[:] | |
972 | # backup of filter list. Used to restore the state after each file. | |
973 | self._filters_backup = self.filters[:] | |
974 | self.counting = 'total' # In what way are we counting errors? | |
975 | self.errors_by_category = {} # string to int dict storing error counts | |
976 | ||
977 | # output format: | |
978 | # "emacs" - format that emacs can parse (default) | |
979 | # "eclipse" - format that eclipse can parse | |
980 | # "vs7" - format that Microsoft Visual Studio 7 can parse | |
981 | # "junit" - format that Jenkins, Bamboo, etc can parse | |
982 | self.output_format = 'emacs' | |
983 | ||
984 | # For JUnit output, save errors and failures until the end so that they | |
985 | # can be written into the XML | |
986 | self._junit_errors = [] | |
987 | self._junit_failures = [] | |
988 | ||
989 | def SetOutputFormat(self, output_format): | |
990 | """Sets the output format for errors.""" | |
991 | self.output_format = output_format | |
992 | ||
993 | def SetVerboseLevel(self, level): | |
994 | """Sets the module's verbosity, and returns the previous setting.""" | |
995 | last_verbose_level = self.verbose_level | |
996 | self.verbose_level = level | |
997 | return last_verbose_level | |
998 | ||
999 | def SetCountingStyle(self, counting_style): | |
1000 | """Sets the module's counting options.""" | |
1001 | self.counting = counting_style | |
1002 | ||
1003 | def SetFilters(self, filters): | |
1004 | """Sets the error-message filters. | |
1005 | ||
1006 | These filters are applied when deciding whether to emit a given | |
1007 | error message. | |
1008 | ||
1009 | Args: | |
1010 | filters: A string of comma-separated filters (eg "+whitespace/indent"). | |
1011 | Each filter should start with + or -; else we die. | |
1012 | ||
1013 | Raises: | |
1014 | ValueError: The comma-separated filters did not all start with '+' or '-'. | |
1015 | E.g. "-,+whitespace,-whitespace/indent,whitespace/badfilter" | |
1016 | """ | |
1017 | # Default filters always have less priority than the flag ones. | |
1018 | self.filters = _DEFAULT_FILTERS[:] | |
1019 | self.AddFilters(filters) | |
1020 | ||
1021 | def AddFilters(self, filters): | |
1022 | """ Adds more filters to the existing list of error-message filters. """ | |
1023 | for filt in filters.split(','): | |
1024 | clean_filt = filt.strip() | |
1025 | if clean_filt: | |
1026 | self.filters.append(clean_filt) | |
1027 | for filt in self.filters: | |
1028 | if not (filt.startswith('+') or filt.startswith('-')): | |
1029 | raise ValueError('Every filter in --filters must start with + or -' | |
1030 | ' (%s does not)' % filt) | |
1031 | ||
1032 | def BackupFilters(self): | |
1033 | """ Saves the current filter list to backup storage.""" | |
1034 | self._filters_backup = self.filters[:] | |
1035 | ||
1036 | def RestoreFilters(self): | |
1037 | """ Restores filters previously backed up.""" | |
1038 | self.filters = self._filters_backup[:] | |
1039 | ||
1040 | def ResetErrorCounts(self): | |
1041 | """Sets the module's error statistic back to zero.""" | |
1042 | self.error_count = 0 | |
1043 | self.errors_by_category = {} | |
1044 | ||
1045 | def IncrementErrorCount(self, category): | |
1046 | """Bumps the module's error statistic.""" | |
1047 | self.error_count += 1 | |
1048 | if self.counting in ('toplevel', 'detailed'): | |
1049 | if self.counting != 'detailed': | |
1050 | category = category.split('/')[0] | |
1051 | if category not in self.errors_by_category: | |
1052 | self.errors_by_category[category] = 0 | |
1053 | self.errors_by_category[category] += 1 | |
1054 | ||
1055 | def PrintErrorCounts(self): | |
1056 | """Print a summary of errors by category, and the total.""" | |
1057 | for category, count in sorted(iteritems(self.errors_by_category)): | |
1058 | self.PrintInfo('Category \'%s\' errors found: %d\n' % | |
1059 | (category, count)) | |
1060 | if self.error_count > 0: | |
1061 | self.PrintInfo('Total errors found: %d\n' % self.error_count) | |
1062 | ||
1063 | def PrintInfo(self, message): | |
1064 | if not _quiet and self.output_format != 'junit': | |
1065 | sys.stderr.write(message) | |
1066 | ||
1067 | def PrintError(self, message): | |
1068 | if self.output_format == 'junit': | |
1069 | self._junit_errors.append(message) | |
1070 | else: | |
1071 | sys.stderr.write(message) | |
1072 | ||
1073 | def AddJUnitFailure(self, filename, linenum, message, category, confidence): | |
1074 | self._junit_failures.append((filename, linenum, message, category, | |
1075 | confidence)) | |
1076 | ||
1077 | def FormatJUnitXML(self): | |
1078 | num_errors = len(self._junit_errors) | |
1079 | num_failures = len(self._junit_failures) | |
1080 | ||
1081 | testsuite = xml.etree.ElementTree.Element('testsuite') | |
1082 | testsuite.attrib['name'] = 'cpplint' | |
1083 | testsuite.attrib['errors'] = str(num_errors) | |
1084 | testsuite.attrib['failures'] = str(num_failures) | |
1085 | ||
1086 | if num_errors == 0 and num_failures == 0: | |
1087 | testsuite.attrib['tests'] = str(1) | |
1088 | xml.etree.ElementTree.SubElement(testsuite, 'testcase', name='passed') | |
1089 | ||
1090 | else: | |
1091 | testsuite.attrib['tests'] = str(num_errors + num_failures) | |
1092 | if num_errors > 0: | |
1093 | testcase = xml.etree.ElementTree.SubElement(testsuite, 'testcase') | |
1094 | testcase.attrib['name'] = 'errors' | |
1095 | error = xml.etree.ElementTree.SubElement(testcase, 'error') | |
1096 | error.text = '\n'.join(self._junit_errors) | |
1097 | if num_failures > 0: | |
1098 | # Group failures by file | |
1099 | failed_file_order = [] | |
1100 | failures_by_file = {} | |
1101 | for failure in self._junit_failures: | |
1102 | failed_file = failure[0] | |
1103 | if failed_file not in failed_file_order: | |
1104 | failed_file_order.append(failed_file) | |
1105 | failures_by_file[failed_file] = [] | |
1106 | failures_by_file[failed_file].append(failure) | |
1107 | # Create a testcase for each file | |
1108 | for failed_file in failed_file_order: | |
1109 | failures = failures_by_file[failed_file] | |
1110 | testcase = xml.etree.ElementTree.SubElement(testsuite, 'testcase') | |
1111 | testcase.attrib['name'] = failed_file | |
1112 | failure = xml.etree.ElementTree.SubElement(testcase, 'failure') | |
1113 | template = '{0}: {1} [{2}] [{3}]' | |
1114 | texts = [template.format(f[1], f[2], f[3], f[4]) for f in failures] | |
1115 | failure.text = '\n'.join(texts) | |
1116 | ||
1117 | xml_decl = '<?xml version="1.0" encoding="UTF-8" ?>\n' | |
1118 | return xml_decl + xml.etree.ElementTree.tostring(testsuite, 'utf-8').decode('utf-8') | |
1119 | ||
1120 | ||
1121 | _cpplint_state = _CppLintState() | |
1122 | ||
1123 | ||
1124 | def _OutputFormat(): | |
1125 | """Gets the module's output format.""" | |
1126 | return _cpplint_state.output_format | |
1127 | ||
1128 | ||
1129 | def _SetOutputFormat(output_format): | |
1130 | """Sets the module's output format.""" | |
1131 | _cpplint_state.SetOutputFormat(output_format) | |
1132 | ||
1133 | ||
1134 | def _VerboseLevel(): | |
1135 | """Returns the module's verbosity setting.""" | |
1136 | return _cpplint_state.verbose_level | |
1137 | ||
1138 | ||
1139 | def _SetVerboseLevel(level): | |
1140 | """Sets the module's verbosity, and returns the previous setting.""" | |
1141 | return _cpplint_state.SetVerboseLevel(level) | |
1142 | ||
1143 | ||
1144 | def _SetCountingStyle(level): | |
1145 | """Sets the module's counting options.""" | |
1146 | _cpplint_state.SetCountingStyle(level) | |
1147 | ||
1148 | ||
1149 | def _Filters(): | |
1150 | """Returns the module's list of output filters, as a list.""" | |
1151 | return _cpplint_state.filters | |
1152 | ||
1153 | ||
1154 | def _SetFilters(filters): | |
1155 | """Sets the module's error-message filters. | |
1156 | ||
1157 | These filters are applied when deciding whether to emit a given | |
1158 | error message. | |
1159 | ||
1160 | Args: | |
1161 | filters: A string of comma-separated filters (eg "whitespace/indent"). | |
1162 | Each filter should start with + or -; else we die. | |
1163 | """ | |
1164 | _cpplint_state.SetFilters(filters) | |
1165 | ||
1166 | def _AddFilters(filters): | |
1167 | """Adds more filter overrides. | |
1168 | ||
1169 | Unlike _SetFilters, this function does not reset the current list of filters | |
1170 | available. | |
1171 | ||
1172 | Args: | |
1173 | filters: A string of comma-separated filters (eg "whitespace/indent"). | |
1174 | Each filter should start with + or -; else we die. | |
1175 | """ | |
1176 | _cpplint_state.AddFilters(filters) | |
1177 | ||
1178 | def _BackupFilters(): | |
1179 | """ Saves the current filter list to backup storage.""" | |
1180 | _cpplint_state.BackupFilters() | |
1181 | ||
1182 | def _RestoreFilters(): | |
1183 | """ Restores filters previously backed up.""" | |
1184 | _cpplint_state.RestoreFilters() | |
1185 | ||
1186 | class _FunctionState(object): | |
1187 | """Tracks current function name and the number of lines in its body.""" | |
1188 | ||
1189 | _NORMAL_TRIGGER = 250 # for --v=0, 500 for --v=1, etc. | |
1190 | _TEST_TRIGGER = 400 # about 50% more than _NORMAL_TRIGGER. | |
1191 | ||
1192 | def __init__(self): | |
1193 | self.in_a_function = False | |
1194 | self.lines_in_function = 0 | |
1195 | self.current_function = '' | |
1196 | ||
1197 | def Begin(self, function_name): | |
1198 | """Start analyzing function body. | |
1199 | ||
1200 | Args: | |
1201 | function_name: The name of the function being tracked. | |
1202 | """ | |
1203 | self.in_a_function = True | |
1204 | self.lines_in_function = 0 | |
1205 | self.current_function = function_name | |
1206 | ||
1207 | def Count(self): | |
1208 | """Count line in current function body.""" | |
1209 | if self.in_a_function: | |
1210 | self.lines_in_function += 1 | |
1211 | ||
1212 | def Check(self, error, filename, linenum): | |
1213 | """Report if too many lines in function body. | |
1214 | ||
1215 | Args: | |
1216 | error: The function to call with any errors found. | |
1217 | filename: The name of the current file. | |
1218 | linenum: The number of the line to check. | |
1219 | """ | |
1220 | if not self.in_a_function: | |
1221 | return | |
1222 | ||
1223 | if Match(r'T(EST|est)', self.current_function): | |
1224 | base_trigger = self._TEST_TRIGGER | |
1225 | else: | |
1226 | base_trigger = self._NORMAL_TRIGGER | |
1227 | trigger = base_trigger * 2**_VerboseLevel() | |
1228 | ||
1229 | if self.lines_in_function > trigger: | |
1230 | error_level = int(math.log(self.lines_in_function / base_trigger, 2)) | |
1231 | # 50 => 0, 100 => 1, 200 => 2, 400 => 3, 800 => 4, 1600 => 5, ... | |
1232 | if error_level > 5: | |
1233 | error_level = 5 | |
1234 | error(filename, linenum, 'readability/fn_size', error_level, | |
1235 | 'Small and focused functions are preferred:' | |
1236 | ' %s has %d non-comment lines' | |
1237 | ' (error triggered by exceeding %d lines).' % ( | |
1238 | self.current_function, self.lines_in_function, trigger)) | |
1239 | ||
1240 | def End(self): | |
1241 | """Stop analyzing function body.""" | |
1242 | self.in_a_function = False | |
1243 | ||
1244 | ||
1245 | class _IncludeError(Exception): | |
1246 | """Indicates a problem with the include order in a file.""" | |
1247 | pass | |
1248 | ||
1249 | ||
1250 | class FileInfo(object): | |
1251 | """Provides utility functions for filenames. | |
1252 | ||
1253 | FileInfo provides easy access to the components of a file's path | |
1254 | relative to the project root. | |
1255 | """ | |
1256 | ||
1257 | def __init__(self, filename): | |
1258 | self._filename = filename | |
1259 | ||
1260 | def FullName(self): | |
1261 | """Make Windows paths like Unix.""" | |
1262 | return os.path.abspath(self._filename).replace('\\', '/') | |
1263 | ||
1264 | def RepositoryName(self): | |
1265 | r"""FullName after removing the local path to the repository. | |
1266 | ||
1267 | If we have a real absolute path name here we can try to do something smart: | |
1268 | detecting the root of the checkout and truncating /path/to/checkout from | |
1269 | the name so that we get header guards that don't include things like | |
1270 | "C:\Documents and Settings\..." or "/home/username/..." in them and thus | |
1271 | people on different computers who have checked the source out to different | |
1272 | locations won't see bogus errors. | |
1273 | """ | |
1274 | fullname = self.FullName() | |
1275 | ||
1276 | if os.path.exists(fullname): | |
1277 | project_dir = os.path.dirname(fullname) | |
1278 | ||
1279 | # If the user specified a repository path, it exists, and the file is | |
1280 | # contained in it, use the specified repository path | |
1281 | if _repository: | |
1282 | repo = FileInfo(_repository).FullName() | |
1283 | root_dir = project_dir | |
1284 | while os.path.exists(root_dir): | |
1285 | # allow case insensitive compare on Windows | |
1286 | if os.path.normcase(root_dir) == os.path.normcase(repo): | |
1287 | return os.path.relpath(fullname, root_dir).replace('\\', '/') | |
1288 | one_up_dir = os.path.dirname(root_dir) | |
1289 | if one_up_dir == root_dir: | |
1290 | break | |
1291 | root_dir = one_up_dir | |
1292 | ||
1293 | if os.path.exists(os.path.join(project_dir, ".svn")): | |
1294 | # If there's a .svn file in the current directory, we recursively look | |
1295 | # up the directory tree for the top of the SVN checkout | |
1296 | root_dir = project_dir | |
1297 | one_up_dir = os.path.dirname(root_dir) | |
1298 | while os.path.exists(os.path.join(one_up_dir, ".svn")): | |
1299 | root_dir = os.path.dirname(root_dir) | |
1300 | one_up_dir = os.path.dirname(one_up_dir) | |
1301 | ||
1302 | prefix = os.path.commonprefix([root_dir, project_dir]) | |
1303 | return fullname[len(prefix) + 1:] | |
1304 | ||
1305 | # Not SVN <= 1.6? Try to find a git, hg, or svn top level directory by | |
1306 | # searching up from the current path. | |
1307 | root_dir = current_dir = os.path.dirname(fullname) | |
1308 | while current_dir != os.path.dirname(current_dir): | |
1309 | if (os.path.exists(os.path.join(current_dir, ".git")) or | |
1310 | os.path.exists(os.path.join(current_dir, ".hg")) or | |
1311 | os.path.exists(os.path.join(current_dir, ".svn"))): | |
1312 | root_dir = current_dir | |
1313 | current_dir = os.path.dirname(current_dir) | |
1314 | ||
1315 | if (os.path.exists(os.path.join(root_dir, ".git")) or | |
1316 | os.path.exists(os.path.join(root_dir, ".hg")) or | |
1317 | os.path.exists(os.path.join(root_dir, ".svn"))): | |
1318 | prefix = os.path.commonprefix([root_dir, project_dir]) | |
1319 | return fullname[len(prefix) + 1:] | |
1320 | ||
1321 | # Don't know what to do; header guard warnings may be wrong... | |
1322 | return fullname | |
1323 | ||
1324 | def Split(self): | |
1325 | """Splits the file into the directory, basename, and extension. | |
1326 | ||
1327 | For 'chrome/browser/browser.cc', Split() would | |
1328 | return ('chrome/browser', 'browser', '.cc') | |
1329 | ||
1330 | Returns: | |
1331 | A tuple of (directory, basename, extension). | |
1332 | """ | |
1333 | ||
1334 | googlename = self.RepositoryName() | |
1335 | project, rest = os.path.split(googlename) | |
1336 | return (project,) + os.path.splitext(rest) | |
1337 | ||
1338 | def BaseName(self): | |
1339 | """File base name - text after the final slash, before the final period.""" | |
1340 | return self.Split()[1] | |
1341 | ||
1342 | def Extension(self): | |
1343 | """File extension - text following the final period, includes that period.""" | |
1344 | return self.Split()[2] | |
1345 | ||
1346 | def NoExtension(self): | |
1347 | """File has no source file extension.""" | |
1348 | return '/'.join(self.Split()[0:2]) | |
1349 | ||
1350 | def IsSource(self): | |
1351 | """File has a source file extension.""" | |
1352 | return _IsSourceExtension(self.Extension()[1:]) | |
1353 | ||
1354 | ||
1355 | def _ShouldPrintError(category, confidence, linenum): | |
1356 | """If confidence >= verbose, category passes filter and is not suppressed.""" | |
1357 | ||
1358 | # There are three ways we might decide not to print an error message: | |
1359 | # a "NOLINT(category)" comment appears in the source, | |
1360 | # the verbosity level isn't high enough, or the filters filter it out. | |
1361 | if IsErrorSuppressedByNolint(category, linenum): | |
1362 | return False | |
1363 | ||
1364 | if confidence < _cpplint_state.verbose_level: | |
1365 | return False | |
1366 | ||
1367 | is_filtered = False | |
1368 | for one_filter in _Filters(): | |
1369 | if one_filter.startswith('-'): | |
1370 | if category.startswith(one_filter[1:]): | |
1371 | is_filtered = True | |
1372 | elif one_filter.startswith('+'): | |
1373 | if category.startswith(one_filter[1:]): | |
1374 | is_filtered = False | |
1375 | else: | |
1376 | assert False # should have been checked for in SetFilter. | |
1377 | if is_filtered: | |
1378 | return False | |
1379 | ||
1380 | return True | |
1381 | ||
1382 | ||
1383 | def Error(filename, linenum, category, confidence, message): | |
1384 | """Logs the fact we've found a lint error. | |
1385 | ||
1386 | We log where the error was found, and also our confidence in the error, | |
1387 | that is, how certain we are this is a legitimate style regression, and | |
1388 | not a misidentification or a use that's sometimes justified. | |
1389 | ||
1390 | False positives can be suppressed by the use of | |
1391 | "cpplint(category)" comments on the offending line. These are | |
1392 | parsed into _error_suppressions. | |
1393 | ||
1394 | Args: | |
1395 | filename: The name of the file containing the error. | |
1396 | linenum: The number of the line containing the error. | |
1397 | category: A string used to describe the "category" this bug | |
1398 | falls under: "whitespace", say, or "runtime". Categories | |
1399 | may have a hierarchy separated by slashes: "whitespace/indent". | |
1400 | confidence: A number from 1-5 representing a confidence score for | |
1401 | the error, with 5 meaning that we are certain of the problem, | |
1402 | and 1 meaning that it could be a legitimate construct. | |
1403 | message: The error message. | |
1404 | """ | |
1405 | if _ShouldPrintError(category, confidence, linenum): | |
1406 | _cpplint_state.IncrementErrorCount(category) | |
1407 | if _cpplint_state.output_format == 'vs7': | |
1408 | _cpplint_state.PrintError('%s(%s): warning: %s [%s] [%d]\n' % ( | |
1409 | filename, linenum, message, category, confidence)) | |
1410 | elif _cpplint_state.output_format == 'eclipse': | |
1411 | sys.stderr.write('%s:%s: warning: %s [%s] [%d]\n' % ( | |
1412 | filename, linenum, message, category, confidence)) | |
1413 | elif _cpplint_state.output_format == 'junit': | |
1414 | _cpplint_state.AddJUnitFailure(filename, linenum, message, category, | |
1415 | confidence) | |
1416 | else: | |
1417 | final_message = '%s:%s: %s [%s] [%d]\n' % ( | |
1418 | filename, linenum, message, category, confidence) | |
1419 | sys.stderr.write(final_message) | |
1420 | ||
1421 | # Matches standard C++ escape sequences per 2.13.2.3 of the C++ standard. | |
1422 | _RE_PATTERN_CLEANSE_LINE_ESCAPES = re.compile( | |
1423 | r'\\([abfnrtv?"\\\']|\d+|x[0-9a-fA-F]+)') | |
1424 | # Match a single C style comment on the same line. | |
1425 | _RE_PATTERN_C_COMMENTS = r'/\*(?:[^*]|\*(?!/))*\*/' | |
1426 | # Matches multi-line C style comments. | |
1427 | # This RE is a little bit more complicated than one might expect, because we | |
1428 | # have to take care of space removals tools so we can handle comments inside | |
1429 | # statements better. | |
1430 | # The current rule is: We only clear spaces from both sides when we're at the | |
1431 | # end of the line. Otherwise, we try to remove spaces from the right side, | |
1432 | # if this doesn't work we try on left side but only if there's a non-character | |
1433 | # on the right. | |
1434 | _RE_PATTERN_CLEANSE_LINE_C_COMMENTS = re.compile( | |
1435 | r'(\s*' + _RE_PATTERN_C_COMMENTS + r'\s*$|' + | |
1436 | _RE_PATTERN_C_COMMENTS + r'\s+|' + | |
1437 | r'\s+' + _RE_PATTERN_C_COMMENTS + r'(?=\W)|' + | |
1438 | _RE_PATTERN_C_COMMENTS + r')') | |
1439 | ||
1440 | ||
1441 | def IsCppString(line): | |
1442 | """Does line terminate so, that the next symbol is in string constant. | |
1443 | ||
1444 | This function does not consider single-line nor multi-line comments. | |
1445 | ||
1446 | Args: | |
1447 | line: is a partial line of code starting from the 0..n. | |
1448 | ||
1449 | Returns: | |
1450 | True, if next character appended to 'line' is inside a | |
1451 | string constant. | |
1452 | """ | |
1453 | ||
1454 | line = line.replace(r'\\', 'XX') # after this, \\" does not match to \" | |
1455 | return ((line.count('"') - line.count(r'\"') - line.count("'\"'")) & 1) == 1 | |
1456 | ||
1457 | ||
1458 | def CleanseRawStrings(raw_lines): | |
1459 | """Removes C++11 raw strings from lines. | |
1460 | ||
1461 | Before: | |
1462 | static const char kData[] = R"( | |
1463 | multi-line string | |
1464 | )"; | |
1465 | ||
1466 | After: | |
1467 | static const char kData[] = "" | |
1468 | (replaced by blank line) | |
1469 | ""; | |
1470 | ||
1471 | Args: | |
1472 | raw_lines: list of raw lines. | |
1473 | ||
1474 | Returns: | |
1475 | list of lines with C++11 raw strings replaced by empty strings. | |
1476 | """ | |
1477 | ||
1478 | delimiter = None | |
1479 | lines_without_raw_strings = [] | |
1480 | for line in raw_lines: | |
1481 | if delimiter: | |
1482 | # Inside a raw string, look for the end | |
1483 | end = line.find(delimiter) | |
1484 | if end >= 0: | |
1485 | # Found the end of the string, match leading space for this | |
1486 | # line and resume copying the original lines, and also insert | |
1487 | # a "" on the last line. | |
1488 | leading_space = Match(r'^(\s*)\S', line) | |
1489 | line = leading_space.group(1) + '""' + line[end + len(delimiter):] | |
1490 | delimiter = None | |
1491 | else: | |
1492 | # Haven't found the end yet, append a blank line. | |
1493 | line = '""' | |
1494 | ||
1495 | # Look for beginning of a raw string, and replace them with | |
1496 | # empty strings. This is done in a loop to handle multiple raw | |
1497 | # strings on the same line. | |
1498 | while delimiter is None: | |
1499 | # Look for beginning of a raw string. | |
1500 | # See 2.14.15 [lex.string] for syntax. | |
1501 | # | |
1502 | # Once we have matched a raw string, we check the prefix of the | |
1503 | # line to make sure that the line is not part of a single line | |
1504 | # comment. It's done this way because we remove raw strings | |
1505 | # before removing comments as opposed to removing comments | |
1506 | # before removing raw strings. This is because there are some | |
1507 | # cpplint checks that requires the comments to be preserved, but | |
1508 | # we don't want to check comments that are inside raw strings. | |
1509 | matched = Match(r'^(.*?)\b(?:R|u8R|uR|UR|LR)"([^\s\\()]*)\((.*)$', line) | |
1510 | if (matched and | |
1511 | not Match(r'^([^\'"]|\'(\\.|[^\'])*\'|"(\\.|[^"])*")*//', | |
1512 | matched.group(1))): | |
1513 | delimiter = ')' + matched.group(2) + '"' | |
1514 | ||
1515 | end = matched.group(3).find(delimiter) | |
1516 | if end >= 0: | |
1517 | # Raw string ended on same line | |
1518 | line = (matched.group(1) + '""' + | |
1519 | matched.group(3)[end + len(delimiter):]) | |
1520 | delimiter = None | |
1521 | else: | |
1522 | # Start of a multi-line raw string | |
1523 | line = matched.group(1) + '""' | |
1524 | else: | |
1525 | break | |
1526 | ||
1527 | lines_without_raw_strings.append(line) | |
1528 | ||
1529 | # TODO(unknown): if delimiter is not None here, we might want to | |
1530 | # emit a warning for unterminated string. | |
1531 | return lines_without_raw_strings | |
1532 | ||
1533 | ||
1534 | def FindNextMultiLineCommentStart(lines, lineix): | |
1535 | """Find the beginning marker for a multiline comment.""" | |
1536 | while lineix < len(lines): | |
1537 | if lines[lineix].strip().startswith('/*'): | |
1538 | # Only return this marker if the comment goes beyond this line | |
1539 | if lines[lineix].strip().find('*/', 2) < 0: | |
1540 | return lineix | |
1541 | lineix += 1 | |
1542 | return len(lines) | |
1543 | ||
1544 | ||
1545 | def FindNextMultiLineCommentEnd(lines, lineix): | |
1546 | """We are inside a comment, find the end marker.""" | |
1547 | while lineix < len(lines): | |
1548 | if lines[lineix].strip().endswith('*/'): | |
1549 | return lineix | |
1550 | lineix += 1 | |
1551 | return len(lines) | |
1552 | ||
1553 | ||
1554 | def RemoveMultiLineCommentsFromRange(lines, begin, end): | |
1555 | """Clears a range of lines for multi-line comments.""" | |
1556 | # Having // dummy comments makes the lines non-empty, so we will not get | |
1557 | # unnecessary blank line warnings later in the code. | |
1558 | for i in range(begin, end): | |
1559 | lines[i] = '/**/' | |
1560 | ||
1561 | ||
1562 | def RemoveMultiLineComments(filename, lines, error): | |
1563 | """Removes multiline (c-style) comments from lines.""" | |
1564 | lineix = 0 | |
1565 | while lineix < len(lines): | |
1566 | lineix_begin = FindNextMultiLineCommentStart(lines, lineix) | |
1567 | if lineix_begin >= len(lines): | |
1568 | return | |
1569 | lineix_end = FindNextMultiLineCommentEnd(lines, lineix_begin) | |
1570 | if lineix_end >= len(lines): | |
1571 | error(filename, lineix_begin + 1, 'readability/multiline_comment', 5, | |
1572 | 'Could not find end of multi-line comment') | |
1573 | return | |
1574 | RemoveMultiLineCommentsFromRange(lines, lineix_begin, lineix_end + 1) | |
1575 | lineix = lineix_end + 1 | |
1576 | ||
1577 | ||
1578 | def CleanseComments(line): | |
1579 | """Removes //-comments and single-line C-style /* */ comments. | |
1580 | ||
1581 | Args: | |
1582 | line: A line of C++ source. | |
1583 | ||
1584 | Returns: | |
1585 | The line with single-line comments removed. | |
1586 | """ | |
1587 | commentpos = line.find('//') | |
1588 | if commentpos != -1 and not IsCppString(line[:commentpos]): | |
1589 | line = line[:commentpos].rstrip() | |
1590 | # get rid of /* ... */ | |
1591 | return _RE_PATTERN_CLEANSE_LINE_C_COMMENTS.sub('', line) | |
1592 | ||
1593 | ||
1594 | class CleansedLines(object): | |
1595 | """Holds 4 copies of all lines with different preprocessing applied to them. | |
1596 | ||
1597 | 1) elided member contains lines without strings and comments. | |
1598 | 2) lines member contains lines without comments. | |
1599 | 3) raw_lines member contains all the lines without processing. | |
1600 | 4) lines_without_raw_strings member is same as raw_lines, but with C++11 raw | |
1601 | strings removed. | |
1602 | All these members are of <type 'list'>, and of the same length. | |
1603 | """ | |
1604 | ||
1605 | def __init__(self, lines): | |
1606 | self.elided = [] | |
1607 | self.lines = [] | |
1608 | self.raw_lines = lines | |
1609 | self.num_lines = len(lines) | |
1610 | self.lines_without_raw_strings = CleanseRawStrings(lines) | |
1611 | for linenum in range(len(self.lines_without_raw_strings)): | |
1612 | self.lines.append(CleanseComments( | |
1613 | self.lines_without_raw_strings[linenum])) | |
1614 | elided = self._CollapseStrings(self.lines_without_raw_strings[linenum]) | |
1615 | self.elided.append(CleanseComments(elided)) | |
1616 | ||
1617 | def NumLines(self): | |
1618 | """Returns the number of lines represented.""" | |
1619 | return self.num_lines | |
1620 | ||
1621 | @staticmethod | |
1622 | def _CollapseStrings(elided): | |
1623 | """Collapses strings and chars on a line to simple "" or '' blocks. | |
1624 | ||
1625 | We nix strings first so we're not fooled by text like '"http://"' | |
1626 | ||
1627 | Args: | |
1628 | elided: The line being processed. | |
1629 | ||
1630 | Returns: | |
1631 | The line with collapsed strings. | |
1632 | """ | |
1633 | if _RE_PATTERN_INCLUDE.match(elided): | |
1634 | return elided | |
1635 | ||
1636 | # Remove escaped characters first to make quote/single quote collapsing | |
1637 | # basic. Things that look like escaped characters shouldn't occur | |
1638 | # outside of strings and chars. | |
1639 | elided = _RE_PATTERN_CLEANSE_LINE_ESCAPES.sub('', elided) | |
1640 | ||
1641 | # Replace quoted strings and digit separators. Both single quotes | |
1642 | # and double quotes are processed in the same loop, otherwise | |
1643 | # nested quotes wouldn't work. | |
1644 | collapsed = '' | |
1645 | while True: | |
1646 | # Find the first quote character | |
1647 | match = Match(r'^([^\'"]*)([\'"])(.*)$', elided) | |
1648 | if not match: | |
1649 | collapsed += elided | |
1650 | break | |
1651 | head, quote, tail = match.groups() | |
1652 | ||
1653 | if quote == '"': | |
1654 | # Collapse double quoted strings | |
1655 | second_quote = tail.find('"') | |
1656 | if second_quote >= 0: | |
1657 | collapsed += head + '""' | |
1658 | elided = tail[second_quote + 1:] | |
1659 | else: | |
1660 | # Unmatched double quote, don't bother processing the rest | |
1661 | # of the line since this is probably a multiline string. | |
1662 | collapsed += elided | |
1663 | break | |
1664 | else: | |
1665 | # Found single quote, check nearby text to eliminate digit separators. | |
1666 | # | |
1667 | # There is no special handling for floating point here, because | |
1668 | # the integer/fractional/exponent parts would all be parsed | |
1669 | # correctly as long as there are digits on both sides of the | |
1670 | # separator. So we are fine as long as we don't see something | |
1671 | # like "0.'3" (gcc 4.9.0 will not allow this literal). | |
1672 | if Search(r'\b(?:0[bBxX]?|[1-9])[0-9a-fA-F]*$', head): | |
1673 | match_literal = Match(r'^((?:\'?[0-9a-zA-Z_])*)(.*)$', "'" + tail) | |
1674 | collapsed += head + match_literal.group(1).replace("'", '') | |
1675 | elided = match_literal.group(2) | |
1676 | else: | |
1677 | second_quote = tail.find('\'') | |
1678 | if second_quote >= 0: | |
1679 | collapsed += head + "''" | |
1680 | elided = tail[second_quote + 1:] | |
1681 | else: | |
1682 | # Unmatched single quote | |
1683 | collapsed += elided | |
1684 | break | |
1685 | ||
1686 | return collapsed | |
1687 | ||
1688 | ||
1689 | def FindEndOfExpressionInLine(line, startpos, stack): | |
1690 | """Find the position just after the end of current parenthesized expression. | |
1691 | ||
1692 | Args: | |
1693 | line: a CleansedLines line. | |
1694 | startpos: start searching at this position. | |
1695 | stack: nesting stack at startpos. | |
1696 | ||
1697 | Returns: | |
1698 | On finding matching end: (index just after matching end, None) | |
1699 | On finding an unclosed expression: (-1, None) | |
1700 | Otherwise: (-1, new stack at end of this line) | |
1701 | """ | |
1702 | for i in xrange(startpos, len(line)): | |
1703 | char = line[i] | |
1704 | if char in '([{': | |
1705 | # Found start of parenthesized expression, push to expression stack | |
1706 | stack.append(char) | |
1707 | elif char == '<': | |
1708 | # Found potential start of template argument list | |
1709 | if i > 0 and line[i - 1] == '<': | |
1710 | # Left shift operator | |
1711 | if stack and stack[-1] == '<': | |
1712 | stack.pop() | |
1713 | if not stack: | |
1714 | return (-1, None) | |
1715 | elif i > 0 and Search(r'\boperator\s*$', line[0:i]): | |
1716 | # operator<, don't add to stack | |
1717 | continue | |
1718 | else: | |
1719 | # Tentative start of template argument list | |
1720 | stack.append('<') | |
1721 | elif char in ')]}': | |
1722 | # Found end of parenthesized expression. | |
1723 | # | |
1724 | # If we are currently expecting a matching '>', the pending '<' | |
1725 | # must have been an operator. Remove them from expression stack. | |
1726 | while stack and stack[-1] == '<': | |
1727 | stack.pop() | |
1728 | if not stack: | |
1729 | return (-1, None) | |
1730 | if ((stack[-1] == '(' and char == ')') or | |
1731 | (stack[-1] == '[' and char == ']') or | |
1732 | (stack[-1] == '{' and char == '}')): | |
1733 | stack.pop() | |
1734 | if not stack: | |
1735 | return (i + 1, None) | |
1736 | else: | |
1737 | # Mismatched parentheses | |
1738 | return (-1, None) | |
1739 | elif char == '>': | |
1740 | # Found potential end of template argument list. | |
1741 | ||
1742 | # Ignore "->" and operator functions | |
1743 | if (i > 0 and | |
1744 | (line[i - 1] == '-' or Search(r'\boperator\s*$', line[0:i - 1]))): | |
1745 | continue | |
1746 | ||
1747 | # Pop the stack if there is a matching '<'. Otherwise, ignore | |
1748 | # this '>' since it must be an operator. | |
1749 | if stack: | |
1750 | if stack[-1] == '<': | |
1751 | stack.pop() | |
1752 | if not stack: | |
1753 | return (i + 1, None) | |
1754 | elif char == ';': | |
1755 | # Found something that look like end of statements. If we are currently | |
1756 | # expecting a '>', the matching '<' must have been an operator, since | |
1757 | # template argument list should not contain statements. | |
1758 | while stack and stack[-1] == '<': | |
1759 | stack.pop() | |
1760 | if not stack: | |
1761 | return (-1, None) | |
1762 | ||
1763 | # Did not find end of expression or unbalanced parentheses on this line | |
1764 | return (-1, stack) | |
1765 | ||
1766 | ||
1767 | def CloseExpression(clean_lines, linenum, pos): | |
1768 | """If input points to ( or { or [ or <, finds the position that closes it. | |
1769 | ||
1770 | If lines[linenum][pos] points to a '(' or '{' or '[' or '<', finds the | |
1771 | linenum/pos that correspond to the closing of the expression. | |
1772 | ||
1773 | TODO(unknown): cpplint spends a fair bit of time matching parentheses. | |
1774 | Ideally we would want to index all opening and closing parentheses once | |
1775 | and have CloseExpression be just a simple lookup, but due to preprocessor | |
1776 | tricks, this is not so easy. | |
1777 | ||
1778 | Args: | |
1779 | clean_lines: A CleansedLines instance containing the file. | |
1780 | linenum: The number of the line to check. | |
1781 | pos: A position on the line. | |
1782 | ||
1783 | Returns: | |
1784 | A tuple (line, linenum, pos) pointer *past* the closing brace, or | |
1785 | (line, len(lines), -1) if we never find a close. Note we ignore | |
1786 | strings and comments when matching; and the line we return is the | |
1787 | 'cleansed' line at linenum. | |
1788 | """ | |
1789 | ||
1790 | line = clean_lines.elided[linenum] | |
1791 | if (line[pos] not in '({[<') or Match(r'<[<=]', line[pos:]): | |
1792 | return (line, clean_lines.NumLines(), -1) | |
1793 | ||
1794 | # Check first line | |
1795 | (end_pos, stack) = FindEndOfExpressionInLine(line, pos, []) | |
1796 | if end_pos > -1: | |
1797 | return (line, linenum, end_pos) | |
1798 | ||
1799 | # Continue scanning forward | |
1800 | while stack and linenum < clean_lines.NumLines() - 1: | |
1801 | linenum += 1 | |
1802 | line = clean_lines.elided[linenum] | |
1803 | (end_pos, stack) = FindEndOfExpressionInLine(line, 0, stack) | |
1804 | if end_pos > -1: | |
1805 | return (line, linenum, end_pos) | |
1806 | ||
1807 | # Did not find end of expression before end of file, give up | |
1808 | return (line, clean_lines.NumLines(), -1) | |
1809 | ||
1810 | ||
1811 | def FindStartOfExpressionInLine(line, endpos, stack): | |
1812 | """Find position at the matching start of current expression. | |
1813 | ||
1814 | This is almost the reverse of FindEndOfExpressionInLine, but note | |
1815 | that the input position and returned position differs by 1. | |
1816 | ||
1817 | Args: | |
1818 | line: a CleansedLines line. | |
1819 | endpos: start searching at this position. | |
1820 | stack: nesting stack at endpos. | |
1821 | ||
1822 | Returns: | |
1823 | On finding matching start: (index at matching start, None) | |
1824 | On finding an unclosed expression: (-1, None) | |
1825 | Otherwise: (-1, new stack at beginning of this line) | |
1826 | """ | |
1827 | i = endpos | |
1828 | while i >= 0: | |
1829 | char = line[i] | |
1830 | if char in ')]}': | |
1831 | # Found end of expression, push to expression stack | |
1832 | stack.append(char) | |
1833 | elif char == '>': | |
1834 | # Found potential end of template argument list. | |
1835 | # | |
1836 | # Ignore it if it's a "->" or ">=" or "operator>" | |
1837 | if (i > 0 and | |
1838 | (line[i - 1] == '-' or | |
1839 | Match(r'\s>=\s', line[i - 1:]) or | |
1840 | Search(r'\boperator\s*$', line[0:i]))): | |
1841 | i -= 1 | |
1842 | else: | |
1843 | stack.append('>') | |
1844 | elif char == '<': | |
1845 | # Found potential start of template argument list | |
1846 | if i > 0 and line[i - 1] == '<': | |
1847 | # Left shift operator | |
1848 | i -= 1 | |
1849 | else: | |
1850 | # If there is a matching '>', we can pop the expression stack. | |
1851 | # Otherwise, ignore this '<' since it must be an operator. | |
1852 | if stack and stack[-1] == '>': | |
1853 | stack.pop() | |
1854 | if not stack: | |
1855 | return (i, None) | |
1856 | elif char in '([{': | |
1857 | # Found start of expression. | |
1858 | # | |
1859 | # If there are any unmatched '>' on the stack, they must be | |
1860 | # operators. Remove those. | |
1861 | while stack and stack[-1] == '>': | |
1862 | stack.pop() | |
1863 | if not stack: | |
1864 | return (-1, None) | |
1865 | if ((char == '(' and stack[-1] == ')') or | |
1866 | (char == '[' and stack[-1] == ']') or | |
1867 | (char == '{' and stack[-1] == '}')): | |
1868 | stack.pop() | |
1869 | if not stack: | |
1870 | return (i, None) | |
1871 | else: | |
1872 | # Mismatched parentheses | |
1873 | return (-1, None) | |
1874 | elif char == ';': | |
1875 | # Found something that look like end of statements. If we are currently | |
1876 | # expecting a '<', the matching '>' must have been an operator, since | |
1877 | # template argument list should not contain statements. | |
1878 | while stack and stack[-1] == '>': | |
1879 | stack.pop() | |
1880 | if not stack: | |
1881 | return (-1, None) | |
1882 | ||
1883 | i -= 1 | |
1884 | ||
1885 | return (-1, stack) | |
1886 | ||
1887 | ||
1888 | def ReverseCloseExpression(clean_lines, linenum, pos): | |
1889 | """If input points to ) or } or ] or >, finds the position that opens it. | |
1890 | ||
1891 | If lines[linenum][pos] points to a ')' or '}' or ']' or '>', finds the | |
1892 | linenum/pos that correspond to the opening of the expression. | |
1893 | ||
1894 | Args: | |
1895 | clean_lines: A CleansedLines instance containing the file. | |
1896 | linenum: The number of the line to check. | |
1897 | pos: A position on the line. | |
1898 | ||
1899 | Returns: | |
1900 | A tuple (line, linenum, pos) pointer *at* the opening brace, or | |
1901 | (line, 0, -1) if we never find the matching opening brace. Note | |
1902 | we ignore strings and comments when matching; and the line we | |
1903 | return is the 'cleansed' line at linenum. | |
1904 | """ | |
1905 | line = clean_lines.elided[linenum] | |
1906 | if line[pos] not in ')}]>': | |
1907 | return (line, 0, -1) | |
1908 | ||
1909 | # Check last line | |
1910 | (start_pos, stack) = FindStartOfExpressionInLine(line, pos, []) | |
1911 | if start_pos > -1: | |
1912 | return (line, linenum, start_pos) | |
1913 | ||
1914 | # Continue scanning backward | |
1915 | while stack and linenum > 0: | |
1916 | linenum -= 1 | |
1917 | line = clean_lines.elided[linenum] | |
1918 | (start_pos, stack) = FindStartOfExpressionInLine(line, len(line) - 1, stack) | |
1919 | if start_pos > -1: | |
1920 | return (line, linenum, start_pos) | |
1921 | ||
1922 | # Did not find start of expression before beginning of file, give up | |
1923 | return (line, 0, -1) | |
1924 | ||
1925 | ||
1926 | def CheckForCopyright(filename, lines, error): | |
1927 | """Logs an error if no Copyright message appears at the top of the file.""" | |
1928 | ||
1929 | # We'll say it should occur by line 10. Don't forget there's a | |
1930 | # dummy line at the front. | |
1931 | for line in range(1, min(len(lines), 11)): | |
1932 | if re.search(r'Copyright', lines[line], re.I): break | |
1933 | else: # means no copyright line was found | |
1934 | error(filename, 0, 'legal/copyright', 5, | |
1935 | 'No copyright message found. ' | |
1936 | 'You should have a line: "Copyright [year] <Copyright Owner>"') | |
1937 | ||
1938 | ||
1939 | def GetIndentLevel(line): | |
1940 | """Return the number of leading spaces in line. | |
1941 | ||
1942 | Args: | |
1943 | line: A string to check. | |
1944 | ||
1945 | Returns: | |
1946 | An integer count of leading spaces, possibly zero. | |
1947 | """ | |
1948 | indent = Match(r'^( *)\S', line) | |
1949 | if indent: | |
1950 | return len(indent.group(1)) | |
1951 | else: | |
1952 | return 0 | |
1953 | ||
1954 | ||
1955 | def GetHeaderGuardCPPVariable(filename): | |
1956 | """Returns the CPP variable that should be used as a header guard. | |
1957 | ||
1958 | Args: | |
1959 | filename: The name of a C++ header file. | |
1960 | ||
1961 | Returns: | |
1962 | The CPP variable that should be used as a header guard in the | |
1963 | named file. | |
1964 | ||
1965 | """ | |
1966 | ||
1967 | # Restores original filename in case that cpplint is invoked from Emacs's | |
1968 | # flymake. | |
1969 | filename = re.sub(r'_flymake\.h$', '.h', filename) | |
1970 | filename = re.sub(r'/\.flymake/([^/]*)$', r'/\1', filename) | |
1971 | # Replace 'c++' with 'cpp'. | |
1972 | filename = filename.replace('C++', 'cpp').replace('c++', 'cpp') | |
1973 | ||
1974 | fileinfo = FileInfo(filename) | |
1975 | file_path_from_root = fileinfo.RepositoryName() | |
1976 | if _root: | |
1977 | suffix = os.sep | |
1978 | # On Windows using directory separator will leave us with | |
1979 | # "bogus escape error" unless we properly escape regex. | |
1980 | if suffix == '\\': | |
1981 | suffix += '\\' | |
1982 | file_path_from_root = re.sub('^' + _root + suffix, '', file_path_from_root) | |
1983 | return re.sub(r'[^a-zA-Z0-9]', '_', file_path_from_root).upper() + '_' | |
1984 | ||
1985 | ||
1986 | def CheckForHeaderGuard(filename, clean_lines, error): | |
1987 | """Checks that the file contains a header guard. | |
1988 | ||
1989 | Logs an error if no #ifndef header guard is present. For other | |
1990 | headers, checks that the full pathname is used. | |
1991 | ||
1992 | Args: | |
1993 | filename: The name of the C++ header file. | |
1994 | clean_lines: A CleansedLines instance containing the file. | |
1995 | error: The function to call with any errors found. | |
1996 | """ | |
1997 | ||
1998 | # Don't check for header guards if there are error suppression | |
1999 | # comments somewhere in this file. | |
2000 | # | |
2001 | # Because this is silencing a warning for a nonexistent line, we | |
2002 | # only support the very specific NOLINT(build/header_guard) syntax, | |
2003 | # and not the general NOLINT or NOLINT(*) syntax. | |
2004 | raw_lines = clean_lines.lines_without_raw_strings | |
2005 | for i in raw_lines: | |
2006 | if Search(r'//\s*NOLINT\(build/header_guard\)', i): | |
2007 | return | |
2008 | ||
2009 | # Allow pragma once instead of header guards | |
2010 | for i in raw_lines: | |
2011 | if Search(r'^\s*#pragma\s+once', i): | |
2012 | return | |
2013 | ||
2014 | cppvar = GetHeaderGuardCPPVariable(filename) | |
2015 | ||
2016 | ifndef = '' | |
2017 | ifndef_linenum = 0 | |
2018 | define = '' | |
2019 | endif = '' | |
2020 | endif_linenum = 0 | |
2021 | for linenum, line in enumerate(raw_lines): | |
2022 | linesplit = line.split() | |
2023 | if len(linesplit) >= 2: | |
2024 | # find the first occurrence of #ifndef and #define, save arg | |
2025 | if not ifndef and linesplit[0] == '#ifndef': | |
2026 | # set ifndef to the header guard presented on the #ifndef line. | |
2027 | ifndef = linesplit[1] | |
2028 | ifndef_linenum = linenum | |
2029 | if not define and linesplit[0] == '#define': | |
2030 | define = linesplit[1] | |
2031 | # find the last occurrence of #endif, save entire line | |
2032 | if line.startswith('#endif'): | |
2033 | endif = line | |
2034 | endif_linenum = linenum | |
2035 | ||
2036 | if not ifndef or not define or ifndef != define: | |
2037 | error(filename, 0, 'build/header_guard', 5, | |
2038 | 'No #ifndef header guard found, suggested CPP variable is: %s' % | |
2039 | cppvar) | |
2040 | return | |
2041 | ||
2042 | # The guard should be PATH_FILE_H_, but we also allow PATH_FILE_H__ | |
2043 | # for backward compatibility. | |
2044 | if ifndef != cppvar: | |
2045 | error_level = 0 | |
2046 | if ifndef != cppvar + '_': | |
2047 | error_level = 5 | |
2048 | ||
2049 | ParseNolintSuppressions(filename, raw_lines[ifndef_linenum], ifndef_linenum, | |
2050 | error) | |
2051 | error(filename, ifndef_linenum, 'build/header_guard', error_level, | |
2052 | '#ifndef header guard has wrong style, please use: %s' % cppvar) | |
2053 | ||
2054 | # Check for "//" comments on endif line. | |
2055 | ParseNolintSuppressions(filename, raw_lines[endif_linenum], endif_linenum, | |
2056 | error) | |
2057 | match = Match(r'#endif\s*//\s*' + cppvar + r'(_)?\b', endif) | |
2058 | if match: | |
2059 | if match.group(1) == '_': | |
2060 | # Issue low severity warning for deprecated double trailing underscore | |
2061 | error(filename, endif_linenum, 'build/header_guard', 0, | |
2062 | '#endif line should be "#endif // %s"' % cppvar) | |
2063 | return | |
2064 | ||
2065 | # Didn't find the corresponding "//" comment. If this file does not | |
2066 | # contain any "//" comments at all, it could be that the compiler | |
2067 | # only wants "/**/" comments, look for those instead. | |
2068 | no_single_line_comments = True | |
2069 | for i in xrange(1, len(raw_lines) - 1): | |
2070 | line = raw_lines[i] | |
2071 | if Match(r'^(?:(?:\'(?:\.|[^\'])*\')|(?:"(?:\.|[^"])*")|[^\'"])*//', line): | |
2072 | no_single_line_comments = False | |
2073 | break | |
2074 | ||
2075 | if no_single_line_comments: | |
2076 | match = Match(r'#endif\s*/\*\s*' + cppvar + r'(_)?\s*\*/', endif) | |
2077 | if match: | |
2078 | if match.group(1) == '_': | |
2079 | # Low severity warning for double trailing underscore | |
2080 | error(filename, endif_linenum, 'build/header_guard', 0, | |
2081 | '#endif line should be "#endif /* %s */"' % cppvar) | |
2082 | return | |
2083 | ||
2084 | # Didn't find anything | |
2085 | error(filename, endif_linenum, 'build/header_guard', 5, | |
2086 | '#endif line should be "#endif // %s"' % cppvar) | |
2087 | ||
2088 | ||
2089 | def CheckHeaderFileIncluded(filename, include_state, error): | |
2090 | """Logs an error if a source file does not include its header.""" | |
2091 | ||
2092 | # Do not check test files | |
2093 | fileinfo = FileInfo(filename) | |
2094 | if Search(_TEST_FILE_SUFFIX, fileinfo.BaseName()): | |
2095 | return | |
2096 | ||
2097 | for ext in GetHeaderExtensions(): | |
2098 | basefilename = filename[0:len(filename) - len(fileinfo.Extension())] | |
2099 | headerfile = basefilename + '.' + ext | |
2100 | if not os.path.exists(headerfile): | |
2101 | continue | |
2102 | headername = FileInfo(headerfile).RepositoryName() | |
2103 | first_include = None | |
2104 | for section_list in include_state.include_list: | |
2105 | for f in section_list: | |
2106 | if headername in f[0] or f[0] in headername: | |
2107 | return | |
2108 | if not first_include: | |
2109 | first_include = f[1] | |
2110 | ||
2111 | error(filename, first_include, 'build/include', 5, | |
2112 | '%s should include its header file %s' % (fileinfo.RepositoryName(), | |
2113 | headername)) | |
2114 | ||
2115 | ||
2116 | def CheckForBadCharacters(filename, lines, error): | |
2117 | """Logs an error for each line containing bad characters. | |
2118 | ||
2119 | Two kinds of bad characters: | |
2120 | ||
2121 | 1. Unicode replacement characters: These indicate that either the file | |
2122 | contained invalid UTF-8 (likely) or Unicode replacement characters (which | |
2123 | it shouldn't). Note that it's possible for this to throw off line | |
2124 | numbering if the invalid UTF-8 occurred adjacent to a newline. | |
2125 | ||
2126 | 2. NUL bytes. These are problematic for some tools. | |
2127 | ||
2128 | Args: | |
2129 | filename: The name of the current file. | |
2130 | lines: An array of strings, each representing a line of the file. | |
2131 | error: The function to call with any errors found. | |
2132 | """ | |
2133 | for linenum, line in enumerate(lines): | |
2134 | if unicode_escape_decode('\ufffd') in line: | |
2135 | error(filename, linenum, 'readability/utf8', 5, | |
2136 | 'Line contains invalid UTF-8 (or Unicode replacement character).') | |
2137 | if '\0' in line: | |
2138 | error(filename, linenum, 'readability/nul', 5, 'Line contains NUL byte.') | |
2139 | ||
2140 | ||
2141 | def CheckForNewlineAtEOF(filename, lines, error): | |
2142 | """Logs an error if there is no newline char at the end of the file. | |
2143 | ||
2144 | Args: | |
2145 | filename: The name of the current file. | |
2146 | lines: An array of strings, each representing a line of the file. | |
2147 | error: The function to call with any errors found. | |
2148 | """ | |
2149 | ||
2150 | # The array lines() was created by adding two newlines to the | |
2151 | # original file (go figure), then splitting on \n. | |
2152 | # To verify that the file ends in \n, we just have to make sure the | |
2153 | # last-but-two element of lines() exists and is empty. | |
2154 | if len(lines) < 3 or lines[-2]: | |
2155 | error(filename, len(lines) - 2, 'whitespace/ending_newline', 5, | |
2156 | 'Could not find a newline character at the end of the file.') | |
2157 | ||
2158 | ||
2159 | def CheckForMultilineCommentsAndStrings(filename, clean_lines, linenum, error): | |
2160 | """Logs an error if we see /* ... */ or "..." that extend past one line. | |
2161 | ||
2162 | /* ... */ comments are legit inside macros, for one line. | |
2163 | Otherwise, we prefer // comments, so it's ok to warn about the | |
2164 | other. Likewise, it's ok for strings to extend across multiple | |
2165 | lines, as long as a line continuation character (backslash) | |
2166 | terminates each line. Although not currently prohibited by the C++ | |
2167 | style guide, it's ugly and unnecessary. We don't do well with either | |
2168 | in this lint program, so we warn about both. | |
2169 | ||
2170 | Args: | |
2171 | filename: The name of the current file. | |
2172 | clean_lines: A CleansedLines instance containing the file. | |
2173 | linenum: The number of the line to check. | |
2174 | error: The function to call with any errors found. | |
2175 | """ | |
2176 | line = clean_lines.elided[linenum] | |
2177 | ||
2178 | # Remove all \\ (escaped backslashes) from the line. They are OK, and the | |
2179 | # second (escaped) slash may trigger later \" detection erroneously. | |
2180 | line = line.replace('\\\\', '') | |
2181 | ||
2182 | if line.count('/*') > line.count('*/'): | |
2183 | error(filename, linenum, 'readability/multiline_comment', 5, | |
2184 | 'Complex multi-line /*...*/-style comment found. ' | |
2185 | 'Lint may give bogus warnings. ' | |
2186 | 'Consider replacing these with //-style comments, ' | |
2187 | 'with #if 0...#endif, ' | |
2188 | 'or with more clearly structured multi-line comments.') | |
2189 | ||
2190 | if (line.count('"') - line.count('\\"')) % 2: | |
2191 | error(filename, linenum, 'readability/multiline_string', 5, | |
2192 | 'Multi-line string ("...") found. This lint script doesn\'t ' | |
2193 | 'do well with such strings, and may give bogus warnings. ' | |
2194 | 'Use C++11 raw strings or concatenation instead.') | |
2195 | ||
2196 | ||
2197 | # (non-threadsafe name, thread-safe alternative, validation pattern) | |
2198 | # | |
2199 | # The validation pattern is used to eliminate false positives such as: | |
2200 | # _rand(); // false positive due to substring match. | |
2201 | # ->rand(); // some member function rand(). | |
2202 | # ACMRandom rand(seed); // some variable named rand. | |
2203 | # ISAACRandom rand(); // another variable named rand. | |
2204 | # | |
2205 | # Basically we require the return value of these functions to be used | |
2206 | # in some expression context on the same line by matching on some | |
2207 | # operator before the function name. This eliminates constructors and | |
2208 | # member function calls. | |
2209 | _UNSAFE_FUNC_PREFIX = r'(?:[-+*/=%^&|(<]\s*|>\s+)' | |
2210 | _THREADING_LIST = ( | |
2211 | ('asctime(', 'asctime_r(', _UNSAFE_FUNC_PREFIX + r'asctime\([^)]+\)'), | |
2212 | ('ctime(', 'ctime_r(', _UNSAFE_FUNC_PREFIX + r'ctime\([^)]+\)'), | |
2213 | ('getgrgid(', 'getgrgid_r(', _UNSAFE_FUNC_PREFIX + r'getgrgid\([^)]+\)'), | |
2214 | ('getgrnam(', 'getgrnam_r(', _UNSAFE_FUNC_PREFIX + r'getgrnam\([^)]+\)'), | |
2215 | ('getlogin(', 'getlogin_r(', _UNSAFE_FUNC_PREFIX + r'getlogin\(\)'), | |
2216 | ('getpwnam(', 'getpwnam_r(', _UNSAFE_FUNC_PREFIX + r'getpwnam\([^)]+\)'), | |
2217 | ('getpwuid(', 'getpwuid_r(', _UNSAFE_FUNC_PREFIX + r'getpwuid\([^)]+\)'), | |
2218 | ('gmtime(', 'gmtime_r(', _UNSAFE_FUNC_PREFIX + r'gmtime\([^)]+\)'), | |
2219 | ('localtime(', 'localtime_r(', _UNSAFE_FUNC_PREFIX + r'localtime\([^)]+\)'), | |
2220 | ('rand(', 'rand_r(', _UNSAFE_FUNC_PREFIX + r'rand\(\)'), | |
2221 | ('strtok(', 'strtok_r(', | |
2222 | _UNSAFE_FUNC_PREFIX + r'strtok\([^)]+\)'), | |
2223 | ('ttyname(', 'ttyname_r(', _UNSAFE_FUNC_PREFIX + r'ttyname\([^)]+\)'), | |
2224 | ) | |
2225 | ||
2226 | ||
2227 | def CheckPosixThreading(filename, clean_lines, linenum, error): | |
2228 | """Checks for calls to thread-unsafe functions. | |
2229 | ||
2230 | Much code has been originally written without consideration of | |
2231 | multi-threading. Also, engineers are relying on their old experience; | |
2232 | they have learned posix before threading extensions were added. These | |
2233 | tests guide the engineers to use thread-safe functions (when using | |
2234 | posix directly). | |
2235 | ||
2236 | Args: | |
2237 | filename: The name of the current file. | |
2238 | clean_lines: A CleansedLines instance containing the file. | |
2239 | linenum: The number of the line to check. | |
2240 | error: The function to call with any errors found. | |
2241 | """ | |
2242 | line = clean_lines.elided[linenum] | |
2243 | for single_thread_func, multithread_safe_func, pattern in _THREADING_LIST: | |
2244 | # Additional pattern matching check to confirm that this is the | |
2245 | # function we are looking for | |
2246 | if Search(pattern, line): | |
2247 | error(filename, linenum, 'runtime/threadsafe_fn', 2, | |
2248 | 'Consider using ' + multithread_safe_func + | |
2249 | '...) instead of ' + single_thread_func + | |
2250 | '...) for improved thread safety.') | |
2251 | ||
2252 | ||
2253 | def CheckVlogArguments(filename, clean_lines, linenum, error): | |
2254 | """Checks that VLOG() is only used for defining a logging level. | |
2255 | ||
2256 | For example, VLOG(2) is correct. VLOG(INFO), VLOG(WARNING), VLOG(ERROR), and | |
2257 | VLOG(FATAL) are not. | |
2258 | ||
2259 | Args: | |
2260 | filename: The name of the current file. | |
2261 | clean_lines: A CleansedLines instance containing the file. | |
2262 | linenum: The number of the line to check. | |
2263 | error: The function to call with any errors found. | |
2264 | """ | |
2265 | line = clean_lines.elided[linenum] | |
2266 | if Search(r'\bVLOG\((INFO|ERROR|WARNING|DFATAL|FATAL)\)', line): | |
2267 | error(filename, linenum, 'runtime/vlog', 5, | |
2268 | 'VLOG() should be used with numeric verbosity level. ' | |
2269 | 'Use LOG() if you want symbolic severity levels.') | |
2270 | ||
2271 | # Matches invalid increment: *count++, which moves pointer instead of | |
2272 | # incrementing a value. | |
2273 | _RE_PATTERN_INVALID_INCREMENT = re.compile( | |
2274 | r'^\s*\*\w+(\+\+|--);') | |
2275 | ||
2276 | ||
2277 | def CheckInvalidIncrement(filename, clean_lines, linenum, error): | |
2278 | """Checks for invalid increment *count++. | |
2279 | ||
2280 | For example following function: | |
2281 | void increment_counter(int* count) { | |
2282 | *count++; | |
2283 | } | |
2284 | is invalid, because it effectively does count++, moving pointer, and should | |
2285 | be replaced with ++*count, (*count)++ or *count += 1. | |
2286 | ||
2287 | Args: | |
2288 | filename: The name of the current file. | |
2289 | clean_lines: A CleansedLines instance containing the file. | |
2290 | linenum: The number of the line to check. | |
2291 | error: The function to call with any errors found. | |
2292 | """ | |
2293 | line = clean_lines.elided[linenum] | |
2294 | if _RE_PATTERN_INVALID_INCREMENT.match(line): | |
2295 | error(filename, linenum, 'runtime/invalid_increment', 5, | |
2296 | 'Changing pointer instead of value (or unused value of operator*).') | |
2297 | ||
2298 | ||
2299 | def IsMacroDefinition(clean_lines, linenum): | |
2300 | if Search(r'^#define', clean_lines[linenum]): | |
2301 | return True | |
2302 | ||
2303 | if linenum > 0 and Search(r'\\$', clean_lines[linenum - 1]): | |
2304 | return True | |
2305 | ||
2306 | return False | |
2307 | ||
2308 | ||
2309 | def IsForwardClassDeclaration(clean_lines, linenum): | |
2310 | return Match(r'^\s*(\btemplate\b)*.*class\s+\w+;\s*$', clean_lines[linenum]) | |
2311 | ||
2312 | ||
2313 | class _BlockInfo(object): | |
2314 | """Stores information about a generic block of code.""" | |
2315 | ||
2316 | def __init__(self, linenum, seen_open_brace): | |
2317 | self.starting_linenum = linenum | |
2318 | self.seen_open_brace = seen_open_brace | |
2319 | self.open_parentheses = 0 | |
2320 | self.inline_asm = _NO_ASM | |
2321 | self.check_namespace_indentation = False | |
2322 | ||
2323 | def CheckBegin(self, filename, clean_lines, linenum, error): | |
2324 | """Run checks that applies to text up to the opening brace. | |
2325 | ||
2326 | This is mostly for checking the text after the class identifier | |
2327 | and the "{", usually where the base class is specified. For other | |
2328 | blocks, there isn't much to check, so we always pass. | |
2329 | ||
2330 | Args: | |
2331 | filename: The name of the current file. | |
2332 | clean_lines: A CleansedLines instance containing the file. | |
2333 | linenum: The number of the line to check. | |
2334 | error: The function to call with any errors found. | |
2335 | """ | |
2336 | pass | |
2337 | ||
2338 | def CheckEnd(self, filename, clean_lines, linenum, error): | |
2339 | """Run checks that applies to text after the closing brace. | |
2340 | ||
2341 | This is mostly used for checking end of namespace comments. | |
2342 | ||
2343 | Args: | |
2344 | filename: The name of the current file. | |
2345 | clean_lines: A CleansedLines instance containing the file. | |
2346 | linenum: The number of the line to check. | |
2347 | error: The function to call with any errors found. | |
2348 | """ | |
2349 | pass | |
2350 | ||
2351 | def IsBlockInfo(self): | |
2352 | """Returns true if this block is a _BlockInfo. | |
2353 | ||
2354 | This is convenient for verifying that an object is an instance of | |
2355 | a _BlockInfo, but not an instance of any of the derived classes. | |
2356 | ||
2357 | Returns: | |
2358 | True for this class, False for derived classes. | |
2359 | """ | |
2360 | return self.__class__ == _BlockInfo | |
2361 | ||
2362 | ||
2363 | class _ExternCInfo(_BlockInfo): | |
2364 | """Stores information about an 'extern "C"' block.""" | |
2365 | ||
2366 | def __init__(self, linenum): | |
2367 | _BlockInfo.__init__(self, linenum, True) | |
2368 | ||
2369 | ||
2370 | class _ClassInfo(_BlockInfo): | |
2371 | """Stores information about a class.""" | |
2372 | ||
2373 | def __init__(self, name, class_or_struct, clean_lines, linenum): | |
2374 | _BlockInfo.__init__(self, linenum, False) | |
2375 | self.name = name | |
2376 | self.is_derived = False | |
2377 | self.check_namespace_indentation = True | |
2378 | if class_or_struct == 'struct': | |
2379 | self.access = 'public' | |
2380 | self.is_struct = True | |
2381 | else: | |
2382 | self.access = 'private' | |
2383 | self.is_struct = False | |
2384 | ||
2385 | # Remember initial indentation level for this class. Using raw_lines here | |
2386 | # instead of elided to account for leading comments. | |
2387 | self.class_indent = GetIndentLevel(clean_lines.raw_lines[linenum]) | |
2388 | ||
2389 | # Try to find the end of the class. This will be confused by things like: | |
2390 | # class A { | |
2391 | # } *x = { ... | |
2392 | # | |
2393 | # But it's still good enough for CheckSectionSpacing. | |
2394 | self.last_line = 0 | |
2395 | depth = 0 | |
2396 | for i in range(linenum, clean_lines.NumLines()): | |
2397 | line = clean_lines.elided[i] | |
2398 | depth += line.count('{') - line.count('}') | |
2399 | if not depth: | |
2400 | self.last_line = i | |
2401 | break | |
2402 | ||
2403 | def CheckBegin(self, filename, clean_lines, linenum, error): | |
2404 | # Look for a bare ':' | |
2405 | if Search('(^|[^:]):($|[^:])', clean_lines.elided[linenum]): | |
2406 | self.is_derived = True | |
2407 | ||
2408 | def CheckEnd(self, filename, clean_lines, linenum, error): | |
2409 | # If there is a DISALLOW macro, it should appear near the end of | |
2410 | # the class. | |
2411 | seen_last_thing_in_class = False | |
2412 | for i in xrange(linenum - 1, self.starting_linenum, -1): | |
2413 | match = Search( | |
2414 | r'\b(DISALLOW_COPY_AND_ASSIGN|DISALLOW_IMPLICIT_CONSTRUCTORS)\(' + | |
2415 | self.name + r'\)', | |
2416 | clean_lines.elided[i]) | |
2417 | if match: | |
2418 | if seen_last_thing_in_class: | |
2419 | error(filename, i, 'readability/constructors', 3, | |
2420 | match.group(1) + ' should be the last thing in the class') | |
2421 | break | |
2422 | ||
2423 | if not Match(r'^\s*$', clean_lines.elided[i]): | |
2424 | seen_last_thing_in_class = True | |
2425 | ||
2426 | # Check that closing brace is aligned with beginning of the class. | |
2427 | # Only do this if the closing brace is indented by only whitespaces. | |
2428 | # This means we will not check single-line class definitions. | |
2429 | indent = Match(r'^( *)\}', clean_lines.elided[linenum]) | |
2430 | if indent and len(indent.group(1)) != self.class_indent: | |
2431 | if self.is_struct: | |
2432 | parent = 'struct ' + self.name | |
2433 | else: | |
2434 | parent = 'class ' + self.name | |
2435 | error(filename, linenum, 'whitespace/indent', 3, | |
2436 | 'Closing brace should be aligned with beginning of %s' % parent) | |
2437 | ||
2438 | ||
2439 | class _NamespaceInfo(_BlockInfo): | |
2440 | """Stores information about a namespace.""" | |
2441 | ||
2442 | def __init__(self, name, linenum): | |
2443 | _BlockInfo.__init__(self, linenum, False) | |
2444 | self.name = name or '' | |
2445 | self.check_namespace_indentation = True | |
2446 | ||
2447 | def CheckEnd(self, filename, clean_lines, linenum, error): | |
2448 | """Check end of namespace comments.""" | |
2449 | line = clean_lines.raw_lines[linenum] | |
2450 | ||
2451 | # Check how many lines is enclosed in this namespace. Don't issue | |
2452 | # warning for missing namespace comments if there aren't enough | |
2453 | # lines. However, do apply checks if there is already an end of | |
2454 | # namespace comment and it's incorrect. | |
2455 | # | |
2456 | # TODO(unknown): We always want to check end of namespace comments | |
2457 | # if a namespace is large, but sometimes we also want to apply the | |
2458 | # check if a short namespace contained nontrivial things (something | |
2459 | # other than forward declarations). There is currently no logic on | |
2460 | # deciding what these nontrivial things are, so this check is | |
2461 | # triggered by namespace size only, which works most of the time. | |
2462 | if (linenum - self.starting_linenum < 10 | |
2463 | and not Match(r'^\s*};*\s*(//|/\*).*\bnamespace\b', line)): | |
2464 | return | |
2465 | ||
2466 | # Look for matching comment at end of namespace. | |
2467 | # | |
2468 | # Note that we accept C style "/* */" comments for terminating | |
2469 | # namespaces, so that code that terminate namespaces inside | |
2470 | # preprocessor macros can be cpplint clean. | |
2471 | # | |
2472 | # We also accept stuff like "// end of namespace <name>." with the | |
2473 | # period at the end. | |
2474 | # | |
2475 | # Besides these, we don't accept anything else, otherwise we might | |
2476 | # get false negatives when existing comment is a substring of the | |
2477 | # expected namespace. | |
2478 | if self.name: | |
2479 | # Named namespace | |
2480 | if not Match((r'^\s*};*\s*(//|/\*).*\bnamespace\s+' + | |
2481 | re.escape(self.name) + r'[\*/\.\\\s]*$'), | |
2482 | line): | |
2483 | error(filename, linenum, 'readability/namespace', 5, | |
2484 | 'Namespace should be terminated with "// namespace %s"' % | |
2485 | self.name) | |
2486 | else: | |
2487 | # Anonymous namespace | |
2488 | if not Match(r'^\s*};*\s*(//|/\*).*\bnamespace[\*/\.\\\s]*$', line): | |
2489 | # If "// namespace anonymous" or "// anonymous namespace (more text)", | |
2490 | # mention "// anonymous namespace" as an acceptable form | |
2491 | if Match(r'^\s*}.*\b(namespace anonymous|anonymous namespace)\b', line): | |
2492 | error(filename, linenum, 'readability/namespace', 5, | |
2493 | 'Anonymous namespace should be terminated with "// namespace"' | |
2494 | ' or "// anonymous namespace"') | |
2495 | else: | |
2496 | error(filename, linenum, 'readability/namespace', 5, | |
2497 | 'Anonymous namespace should be terminated with "// namespace"') | |
2498 | ||
2499 | ||
2500 | class _PreprocessorInfo(object): | |
2501 | """Stores checkpoints of nesting stacks when #if/#else is seen.""" | |
2502 | ||
2503 | def __init__(self, stack_before_if): | |
2504 | # The entire nesting stack before #if | |
2505 | self.stack_before_if = stack_before_if | |
2506 | ||
2507 | # The entire nesting stack up to #else | |
2508 | self.stack_before_else = [] | |
2509 | ||
2510 | # Whether we have already seen #else or #elif | |
2511 | self.seen_else = False | |
2512 | ||
2513 | ||
2514 | class NestingState(object): | |
2515 | """Holds states related to parsing braces.""" | |
2516 | ||
2517 | def __init__(self): | |
2518 | # Stack for tracking all braces. An object is pushed whenever we | |
2519 | # see a "{", and popped when we see a "}". Only 3 types of | |
2520 | # objects are possible: | |
2521 | # - _ClassInfo: a class or struct. | |
2522 | # - _NamespaceInfo: a namespace. | |
2523 | # - _BlockInfo: some other type of block. | |
2524 | self.stack = [] | |
2525 | ||
2526 | # Top of the previous stack before each Update(). | |
2527 | # | |
2528 | # Because the nesting_stack is updated at the end of each line, we | |
2529 | # had to do some convoluted checks to find out what is the current | |
2530 | # scope at the beginning of the line. This check is simplified by | |
2531 | # saving the previous top of nesting stack. | |
2532 | # | |
2533 | # We could save the full stack, but we only need the top. Copying | |
2534 | # the full nesting stack would slow down cpplint by ~10%. | |
2535 | self.previous_stack_top = [] | |
2536 | ||
2537 | # Stack of _PreprocessorInfo objects. | |
2538 | self.pp_stack = [] | |
2539 | ||
2540 | def SeenOpenBrace(self): | |
2541 | """Check if we have seen the opening brace for the innermost block. | |
2542 | ||
2543 | Returns: | |
2544 | True if we have seen the opening brace, False if the innermost | |
2545 | block is still expecting an opening brace. | |
2546 | """ | |
2547 | return (not self.stack) or self.stack[-1].seen_open_brace | |
2548 | ||
2549 | def InNamespaceBody(self): | |
2550 | """Check if we are currently one level inside a namespace body. | |
2551 | ||
2552 | Returns: | |
2553 | True if top of the stack is a namespace block, False otherwise. | |
2554 | """ | |
2555 | return self.stack and isinstance(self.stack[-1], _NamespaceInfo) | |
2556 | ||
2557 | def InExternC(self): | |
2558 | """Check if we are currently one level inside an 'extern "C"' block. | |
2559 | ||
2560 | Returns: | |
2561 | True if top of the stack is an extern block, False otherwise. | |
2562 | """ | |
2563 | return self.stack and isinstance(self.stack[-1], _ExternCInfo) | |
2564 | ||
2565 | def InClassDeclaration(self): | |
2566 | """Check if we are currently one level inside a class or struct declaration. | |
2567 | ||
2568 | Returns: | |
2569 | True if top of the stack is a class/struct, False otherwise. | |
2570 | """ | |
2571 | return self.stack and isinstance(self.stack[-1], _ClassInfo) | |
2572 | ||
2573 | def InAsmBlock(self): | |
2574 | """Check if we are currently one level inside an inline ASM block. | |
2575 | ||
2576 | Returns: | |
2577 | True if the top of the stack is a block containing inline ASM. | |
2578 | """ | |
2579 | return self.stack and self.stack[-1].inline_asm != _NO_ASM | |
2580 | ||
2581 | def InTemplateArgumentList(self, clean_lines, linenum, pos): | |
2582 | """Check if current position is inside template argument list. | |
2583 | ||
2584 | Args: | |
2585 | clean_lines: A CleansedLines instance containing the file. | |
2586 | linenum: The number of the line to check. | |
2587 | pos: position just after the suspected template argument. | |
2588 | Returns: | |
2589 | True if (linenum, pos) is inside template arguments. | |
2590 | """ | |
2591 | while linenum < clean_lines.NumLines(): | |
2592 | # Find the earliest character that might indicate a template argument | |
2593 | line = clean_lines.elided[linenum] | |
2594 | match = Match(r'^[^{};=\[\]\.<>]*(.)', line[pos:]) | |
2595 | if not match: | |
2596 | linenum += 1 | |
2597 | pos = 0 | |
2598 | continue | |
2599 | token = match.group(1) | |
2600 | pos += len(match.group(0)) | |
2601 | ||
2602 | # These things do not look like template argument list: | |
2603 | # class Suspect { | |
2604 | # class Suspect x; } | |
2605 | if token in ('{', '}', ';'): return False | |
2606 | ||
2607 | # These things look like template argument list: | |
2608 | # template <class Suspect> | |
2609 | # template <class Suspect = default_value> | |
2610 | # template <class Suspect[]> | |
2611 | # template <class Suspect...> | |
2612 | if token in ('>', '=', '[', ']', '.'): return True | |
2613 | ||
2614 | # Check if token is an unmatched '<'. | |
2615 | # If not, move on to the next character. | |
2616 | if token != '<': | |
2617 | pos += 1 | |
2618 | if pos >= len(line): | |
2619 | linenum += 1 | |
2620 | pos = 0 | |
2621 | continue | |
2622 | ||
2623 | # We can't be sure if we just find a single '<', and need to | |
2624 | # find the matching '>'. | |
2625 | (_, end_line, end_pos) = CloseExpression(clean_lines, linenum, pos - 1) | |
2626 | if end_pos < 0: | |
2627 | # Not sure if template argument list or syntax error in file | |
2628 | return False | |
2629 | linenum = end_line | |
2630 | pos = end_pos | |
2631 | return False | |
2632 | ||
2633 | def UpdatePreprocessor(self, line): | |
2634 | """Update preprocessor stack. | |
2635 | ||
2636 | We need to handle preprocessors due to classes like this: | |
2637 | #ifdef SWIG | |
2638 | struct ResultDetailsPageElementExtensionPoint { | |
2639 | #else | |
2640 | struct ResultDetailsPageElementExtensionPoint : public Extension { | |
2641 | #endif | |
2642 | ||
2643 | We make the following assumptions (good enough for most files): | |
2644 | - Preprocessor condition evaluates to true from #if up to first | |
2645 | #else/#elif/#endif. | |
2646 | ||
2647 | - Preprocessor condition evaluates to false from #else/#elif up | |
2648 | to #endif. We still perform lint checks on these lines, but | |
2649 | these do not affect nesting stack. | |
2650 | ||
2651 | Args: | |
2652 | line: current line to check. | |
2653 | """ | |
2654 | if Match(r'^\s*#\s*(if|ifdef|ifndef)\b', line): | |
2655 | # Beginning of #if block, save the nesting stack here. The saved | |
2656 | # stack will allow us to restore the parsing state in the #else case. | |
2657 | self.pp_stack.append(_PreprocessorInfo(copy.deepcopy(self.stack))) | |
2658 | elif Match(r'^\s*#\s*(else|elif)\b', line): | |
2659 | # Beginning of #else block | |
2660 | if self.pp_stack: | |
2661 | if not self.pp_stack[-1].seen_else: | |
2662 | # This is the first #else or #elif block. Remember the | |
2663 | # whole nesting stack up to this point. This is what we | |
2664 | # keep after the #endif. | |
2665 | self.pp_stack[-1].seen_else = True | |
2666 | self.pp_stack[-1].stack_before_else = copy.deepcopy(self.stack) | |
2667 | ||
2668 | # Restore the stack to how it was before the #if | |
2669 | self.stack = copy.deepcopy(self.pp_stack[-1].stack_before_if) | |
2670 | else: | |
2671 | # TODO(unknown): unexpected #else, issue warning? | |
2672 | pass | |
2673 | elif Match(r'^\s*#\s*endif\b', line): | |
2674 | # End of #if or #else blocks. | |
2675 | if self.pp_stack: | |
2676 | # If we saw an #else, we will need to restore the nesting | |
2677 | # stack to its former state before the #else, otherwise we | |
2678 | # will just continue from where we left off. | |
2679 | if self.pp_stack[-1].seen_else: | |
2680 | # Here we can just use a shallow copy since we are the last | |
2681 | # reference to it. | |
2682 | self.stack = self.pp_stack[-1].stack_before_else | |
2683 | # Drop the corresponding #if | |
2684 | self.pp_stack.pop() | |
2685 | else: | |
2686 | # TODO(unknown): unexpected #endif, issue warning? | |
2687 | pass | |
2688 | ||
2689 | # TODO(unknown): Update() is too long, but we will refactor later. | |
2690 | def Update(self, filename, clean_lines, linenum, error): | |
2691 | """Update nesting state with current line. | |
2692 | ||
2693 | Args: | |
2694 | filename: The name of the current file. | |
2695 | clean_lines: A CleansedLines instance containing the file. | |
2696 | linenum: The number of the line to check. | |
2697 | error: The function to call with any errors found. | |
2698 | """ | |
2699 | line = clean_lines.elided[linenum] | |
2700 | ||
2701 | # Remember top of the previous nesting stack. | |
2702 | # | |
2703 | # The stack is always pushed/popped and not modified in place, so | |
2704 | # we can just do a shallow copy instead of copy.deepcopy. Using | |
2705 | # deepcopy would slow down cpplint by ~28%. | |
2706 | if self.stack: | |
2707 | self.previous_stack_top = self.stack[-1] | |
2708 | else: | |
2709 | self.previous_stack_top = None | |
2710 | ||
2711 | # Update pp_stack | |
2712 | self.UpdatePreprocessor(line) | |
2713 | ||
2714 | # Count parentheses. This is to avoid adding struct arguments to | |
2715 | # the nesting stack. | |
2716 | if self.stack: | |
2717 | inner_block = self.stack[-1] | |
2718 | depth_change = line.count('(') - line.count(')') | |
2719 | inner_block.open_parentheses += depth_change | |
2720 | ||
2721 | # Also check if we are starting or ending an inline assembly block. | |
2722 | if inner_block.inline_asm in (_NO_ASM, _END_ASM): | |
2723 | if (depth_change != 0 and | |
2724 | inner_block.open_parentheses == 1 and | |
2725 | _MATCH_ASM.match(line)): | |
2726 | # Enter assembly block | |
2727 | inner_block.inline_asm = _INSIDE_ASM | |
2728 | else: | |
2729 | # Not entering assembly block. If previous line was _END_ASM, | |
2730 | # we will now shift to _NO_ASM state. | |
2731 | inner_block.inline_asm = _NO_ASM | |
2732 | elif (inner_block.inline_asm == _INSIDE_ASM and | |
2733 | inner_block.open_parentheses == 0): | |
2734 | # Exit assembly block | |
2735 | inner_block.inline_asm = _END_ASM | |
2736 | ||
2737 | # Consume namespace declaration at the beginning of the line. Do | |
2738 | # this in a loop so that we catch same line declarations like this: | |
2739 | # namespace proto2 { namespace bridge { class MessageSet; } } | |
2740 | while True: | |
2741 | # Match start of namespace. The "\b\s*" below catches namespace | |
2742 | # declarations even if it weren't followed by a whitespace, this | |
2743 | # is so that we don't confuse our namespace checker. The | |
2744 | # missing spaces will be flagged by CheckSpacing. | |
2745 | namespace_decl_match = Match(r'^\s*namespace\b\s*([:\w]+)?(.*)$', line) | |
2746 | if not namespace_decl_match: | |
2747 | break | |
2748 | ||
2749 | new_namespace = _NamespaceInfo(namespace_decl_match.group(1), linenum) | |
2750 | self.stack.append(new_namespace) | |
2751 | ||
2752 | line = namespace_decl_match.group(2) | |
2753 | if line.find('{') != -1: | |
2754 | new_namespace.seen_open_brace = True | |
2755 | line = line[line.find('{') + 1:] | |
2756 | ||
2757 | # Look for a class declaration in whatever is left of the line | |
2758 | # after parsing namespaces. The regexp accounts for decorated classes | |
2759 | # such as in: | |
2760 | # class LOCKABLE API Object { | |
2761 | # }; | |
2762 | class_decl_match = Match( | |
2763 | r'^(\s*(?:template\s*<[\w\s<>,:=]*>\s*)?' | |
2764 | r'(class|struct)\s+(?:[A-Z_]+\s+)*(\w+(?:::\w+)*))' | |
2765 | r'(.*)$', line) | |
2766 | if (class_decl_match and | |
2767 | (not self.stack or self.stack[-1].open_parentheses == 0)): | |
2768 | # We do not want to accept classes that are actually template arguments: | |
2769 | # template <class Ignore1, | |
2770 | # class Ignore2 = Default<Args>, | |
2771 | # template <Args> class Ignore3> | |
2772 | # void Function() {}; | |
2773 | # | |
2774 | # To avoid template argument cases, we scan forward and look for | |
2775 | # an unmatched '>'. If we see one, assume we are inside a | |
2776 | # template argument list. | |
2777 | end_declaration = len(class_decl_match.group(1)) | |
2778 | if not self.InTemplateArgumentList(clean_lines, linenum, end_declaration): | |
2779 | self.stack.append(_ClassInfo( | |
2780 | class_decl_match.group(3), class_decl_match.group(2), | |
2781 | clean_lines, linenum)) | |
2782 | line = class_decl_match.group(4) | |
2783 | ||
2784 | # If we have not yet seen the opening brace for the innermost block, | |
2785 | # run checks here. | |
2786 | if not self.SeenOpenBrace(): | |
2787 | self.stack[-1].CheckBegin(filename, clean_lines, linenum, error) | |
2788 | ||
2789 | # Update access control if we are inside a class/struct | |
2790 | if self.stack and isinstance(self.stack[-1], _ClassInfo): | |
2791 | classinfo = self.stack[-1] | |
2792 | access_match = Match( | |
2793 | r'^(.*)\b(public|private|protected|signals)(\s+(?:slots\s*)?)?' | |
2794 | r':(?:[^:]|$)', | |
2795 | line) | |
2796 | if access_match: | |
2797 | classinfo.access = access_match.group(2) | |
2798 | ||
2799 | # Check that access keywords are indented +1 space. Skip this | |
2800 | # check if the keywords are not preceded by whitespaces. | |
2801 | indent = access_match.group(1) | |
2802 | if (len(indent) != classinfo.class_indent + 1 and | |
2803 | Match(r'^\s*$', indent)): | |
2804 | if classinfo.is_struct: | |
2805 | parent = 'struct ' + classinfo.name | |
2806 | else: | |
2807 | parent = 'class ' + classinfo.name | |
2808 | slots = '' | |
2809 | if access_match.group(3): | |
2810 | slots = access_match.group(3) | |
2811 | error(filename, linenum, 'whitespace/indent', 3, | |
2812 | '%s%s: should be indented +1 space inside %s' % ( | |
2813 | access_match.group(2), slots, parent)) | |
2814 | ||
2815 | # Consume braces or semicolons from what's left of the line | |
2816 | while True: | |
2817 | # Match first brace, semicolon, or closed parenthesis. | |
2818 | matched = Match(r'^[^{;)}]*([{;)}])(.*)$', line) | |
2819 | if not matched: | |
2820 | break | |
2821 | ||
2822 | token = matched.group(1) | |
2823 | if token == '{': | |
2824 | # If namespace or class hasn't seen a opening brace yet, mark | |
2825 | # namespace/class head as complete. Push a new block onto the | |
2826 | # stack otherwise. | |
2827 | if not self.SeenOpenBrace(): | |
2828 | self.stack[-1].seen_open_brace = True | |
2829 | elif Match(r'^extern\s*"[^"]*"\s*\{', line): | |
2830 | self.stack.append(_ExternCInfo(linenum)) | |
2831 | else: | |
2832 | self.stack.append(_BlockInfo(linenum, True)) | |
2833 | if _MATCH_ASM.match(line): | |
2834 | self.stack[-1].inline_asm = _BLOCK_ASM | |
2835 | ||
2836 | elif token == ';' or token == ')': | |
2837 | # If we haven't seen an opening brace yet, but we already saw | |
2838 | # a semicolon, this is probably a forward declaration. Pop | |
2839 | # the stack for these. | |
2840 | # | |
2841 | # Similarly, if we haven't seen an opening brace yet, but we | |
2842 | # already saw a closing parenthesis, then these are probably | |
2843 | # function arguments with extra "class" or "struct" keywords. | |
2844 | # Also pop these stack for these. | |
2845 | if not self.SeenOpenBrace(): | |
2846 | self.stack.pop() | |
2847 | else: # token == '}' | |
2848 | # Perform end of block checks and pop the stack. | |
2849 | if self.stack: | |
2850 | self.stack[-1].CheckEnd(filename, clean_lines, linenum, error) | |
2851 | self.stack.pop() | |
2852 | line = matched.group(2) | |
2853 | ||
2854 | def InnermostClass(self): | |
2855 | """Get class info on the top of the stack. | |
2856 | ||
2857 | Returns: | |
2858 | A _ClassInfo object if we are inside a class, or None otherwise. | |
2859 | """ | |
2860 | for i in range(len(self.stack), 0, -1): | |
2861 | classinfo = self.stack[i - 1] | |
2862 | if isinstance(classinfo, _ClassInfo): | |
2863 | return classinfo | |
2864 | return None | |
2865 | ||
2866 | def CheckCompletedBlocks(self, filename, error): | |
2867 | """Checks that all classes and namespaces have been completely parsed. | |
2868 | ||
2869 | Call this when all lines in a file have been processed. | |
2870 | Args: | |
2871 | filename: The name of the current file. | |
2872 | error: The function to call with any errors found. | |
2873 | """ | |
2874 | # Note: This test can result in false positives if #ifdef constructs | |
2875 | # get in the way of brace matching. See the testBuildClass test in | |
2876 | # cpplint_unittest.py for an example of this. | |
2877 | for obj in self.stack: | |
2878 | if isinstance(obj, _ClassInfo): | |
2879 | error(filename, obj.starting_linenum, 'build/class', 5, | |
2880 | 'Failed to find complete declaration of class %s' % | |
2881 | obj.name) | |
2882 | elif isinstance(obj, _NamespaceInfo): | |
2883 | error(filename, obj.starting_linenum, 'build/namespaces', 5, | |
2884 | 'Failed to find complete declaration of namespace %s' % | |
2885 | obj.name) | |
2886 | ||
2887 | ||
2888 | def CheckForNonStandardConstructs(filename, clean_lines, linenum, | |
2889 | nesting_state, error): | |
2890 | r"""Logs an error if we see certain non-ANSI constructs ignored by gcc-2. | |
2891 | ||
2892 | Complain about several constructs which gcc-2 accepts, but which are | |
2893 | not standard C++. Warning about these in lint is one way to ease the | |
2894 | transition to new compilers. | |
2895 | - put storage class first (e.g. "static const" instead of "const static"). | |
2896 | - "%lld" instead of %qd" in printf-type functions. | |
2897 | - "%1$d" is non-standard in printf-type functions. | |
2898 | - "\%" is an undefined character escape sequence. | |
2899 | - text after #endif is not allowed. | |
2900 | - invalid inner-style forward declaration. | |
2901 | - >? and <? operators, and their >?= and <?= cousins. | |
2902 | ||
2903 | Additionally, check for constructor/destructor style violations and reference | |
2904 | members, as it is very convenient to do so while checking for | |
2905 | gcc-2 compliance. | |
2906 | ||
2907 | Args: | |
2908 | filename: The name of the current file. | |
2909 | clean_lines: A CleansedLines instance containing the file. | |
2910 | linenum: The number of the line to check. | |
2911 | nesting_state: A NestingState instance which maintains information about | |
2912 | the current stack of nested blocks being parsed. | |
2913 | error: A callable to which errors are reported, which takes 4 arguments: | |
2914 | filename, line number, error level, and message | |
2915 | """ | |
2916 | ||
2917 | # Remove comments from the line, but leave in strings for now. | |
2918 | line = clean_lines.lines[linenum] | |
2919 | ||
2920 | if Search(r'printf\s*\(.*".*%[-+ ]?\d*q', line): | |
2921 | error(filename, linenum, 'runtime/printf_format', 3, | |
2922 | '%q in format strings is deprecated. Use %ll instead.') | |
2923 | ||
2924 | if Search(r'printf\s*\(.*".*%\d+\$', line): | |
2925 | error(filename, linenum, 'runtime/printf_format', 2, | |
2926 | '%N$ formats are unconventional. Try rewriting to avoid them.') | |
2927 | ||
2928 | # Remove escaped backslashes before looking for undefined escapes. | |
2929 | line = line.replace('\\\\', '') | |
2930 | ||
2931 | if Search(r'("|\').*\\(%|\[|\(|{)', line): | |
2932 | error(filename, linenum, 'build/printf_format', 3, | |
2933 | '%, [, (, and { are undefined character escapes. Unescape them.') | |
2934 | ||
2935 | # For the rest, work with both comments and strings removed. | |
2936 | line = clean_lines.elided[linenum] | |
2937 | ||
2938 | if Search(r'\b(const|volatile|void|char|short|int|long' | |
2939 | r'|float|double|signed|unsigned' | |
2940 | r'|schar|u?int8|u?int16|u?int32|u?int64)' | |
2941 | r'\s+(register|static|extern|typedef)\b', | |
2942 | line): | |
2943 | error(filename, linenum, 'build/storage_class', 5, | |
2944 | 'Storage-class specifier (static, extern, typedef, etc) should be ' | |
2945 | 'at the beginning of the declaration.') | |
2946 | ||
2947 | if Match(r'\s*#\s*endif\s*[^/\s]+', line): | |
2948 | error(filename, linenum, 'build/endif_comment', 5, | |
2949 | 'Uncommented text after #endif is non-standard. Use a comment.') | |
2950 | ||
2951 | if Match(r'\s*class\s+(\w+\s*::\s*)+\w+\s*;', line): | |
2952 | error(filename, linenum, 'build/forward_decl', 5, | |
2953 | 'Inner-style forward declarations are invalid. Remove this line.') | |
2954 | ||
2955 | if Search(r'(\w+|[+-]?\d+(\.\d*)?)\s*(<|>)\?=?\s*(\w+|[+-]?\d+)(\.\d*)?', | |
2956 | line): | |
2957 | error(filename, linenum, 'build/deprecated', 3, | |
2958 | '>? and <? (max and min) operators are non-standard and deprecated.') | |
2959 | ||
2960 | if Search(r'^\s*const\s*string\s*&\s*\w+\s*;', line): | |
2961 | # TODO(unknown): Could it be expanded safely to arbitrary references, | |
2962 | # without triggering too many false positives? The first | |
2963 | # attempt triggered 5 warnings for mostly benign code in the regtest, hence | |
2964 | # the restriction. | |
2965 | # Here's the original regexp, for the reference: | |
2966 | # type_name = r'\w+((\s*::\s*\w+)|(\s*<\s*\w+?\s*>))?' | |
2967 | # r'\s*const\s*' + type_name + '\s*&\s*\w+\s*;' | |
2968 | error(filename, linenum, 'runtime/member_string_references', 2, | |
2969 | 'const string& members are dangerous. It is much better to use ' | |
2970 | 'alternatives, such as pointers or simple constants.') | |
2971 | ||
2972 | # Everything else in this function operates on class declarations. | |
2973 | # Return early if the top of the nesting stack is not a class, or if | |
2974 | # the class head is not completed yet. | |
2975 | classinfo = nesting_state.InnermostClass() | |
2976 | if not classinfo or not classinfo.seen_open_brace: | |
2977 | return | |
2978 | ||
2979 | # The class may have been declared with namespace or classname qualifiers. | |
2980 | # The constructor and destructor will not have those qualifiers. | |
2981 | base_classname = classinfo.name.split('::')[-1] | |
2982 | ||
2983 | # Look for single-argument constructors that aren't marked explicit. | |
2984 | # Technically a valid construct, but against style. | |
2985 | explicit_constructor_match = Match( | |
2986 | r'\s+(?:inline\s+)?(explicit\s+)?(?:inline\s+)?%s\s*' | |
2987 | r'\(((?:[^()]|\([^()]*\))*)\)' | |
2988 | % re.escape(base_classname), | |
2989 | line) | |
2990 | ||
2991 | if explicit_constructor_match: | |
2992 | is_marked_explicit = explicit_constructor_match.group(1) | |
2993 | ||
2994 | if not explicit_constructor_match.group(2): | |
2995 | constructor_args = [] | |
2996 | else: | |
2997 | constructor_args = explicit_constructor_match.group(2).split(',') | |
2998 | ||
2999 | # collapse arguments so that commas in template parameter lists and function | |
3000 | # argument parameter lists don't split arguments in two | |
3001 | i = 0 | |
3002 | while i < len(constructor_args): | |
3003 | constructor_arg = constructor_args[i] | |
3004 | while (constructor_arg.count('<') > constructor_arg.count('>') or | |
3005 | constructor_arg.count('(') > constructor_arg.count(')')): | |
3006 | constructor_arg += ',' + constructor_args[i + 1] | |
3007 | del constructor_args[i + 1] | |
3008 | constructor_args[i] = constructor_arg | |
3009 | i += 1 | |
3010 | ||
3011 | variadic_args = [arg for arg in constructor_args if '&&...' in arg] | |
3012 | defaulted_args = [arg for arg in constructor_args if '=' in arg] | |
3013 | noarg_constructor = (not constructor_args or # empty arg list | |
3014 | # 'void' arg specifier | |
3015 | (len(constructor_args) == 1 and | |
3016 | constructor_args[0].strip() == 'void')) | |
3017 | onearg_constructor = ((len(constructor_args) == 1 and # exactly one arg | |
3018 | not noarg_constructor) or | |
3019 | # all but at most one arg defaulted | |
3020 | (len(constructor_args) >= 1 and | |
3021 | not noarg_constructor and | |
3022 | len(defaulted_args) >= len(constructor_args) - 1) or | |
3023 | # variadic arguments with zero or one argument | |
3024 | (len(constructor_args) <= 2 and | |
3025 | len(variadic_args) >= 1)) | |
3026 | initializer_list_constructor = bool( | |
3027 | onearg_constructor and | |
3028 | Search(r'\bstd\s*::\s*initializer_list\b', constructor_args[0])) | |
3029 | copy_constructor = bool( | |
3030 | onearg_constructor and | |
3031 | Match(r'(const\s+)?%s(\s*<[^>]*>)?(\s+const)?\s*(?:<\w+>\s*)?&' | |
3032 | % re.escape(base_classname), constructor_args[0].strip())) | |
3033 | ||
3034 | if (not is_marked_explicit and | |
3035 | onearg_constructor and | |
3036 | not initializer_list_constructor and | |
3037 | not copy_constructor): | |
3038 | if defaulted_args or variadic_args: | |
3039 | error(filename, linenum, 'runtime/explicit', 5, | |
3040 | 'Constructors callable with one argument ' | |
3041 | 'should be marked explicit.') | |
3042 | else: | |
3043 | error(filename, linenum, 'runtime/explicit', 5, | |
3044 | 'Single-parameter constructors should be marked explicit.') | |
3045 | elif is_marked_explicit and not onearg_constructor: | |
3046 | if noarg_constructor: | |
3047 | error(filename, linenum, 'runtime/explicit', 5, | |
3048 | 'Zero-parameter constructors should not be marked explicit.') | |
3049 | ||
3050 | ||
3051 | def CheckSpacingForFunctionCall(filename, clean_lines, linenum, error): | |
3052 | """Checks for the correctness of various spacing around function calls. | |
3053 | ||
3054 | Args: | |
3055 | filename: The name of the current file. | |
3056 | clean_lines: A CleansedLines instance containing the file. | |
3057 | linenum: The number of the line to check. | |
3058 | error: The function to call with any errors found. | |
3059 | """ | |
3060 | line = clean_lines.elided[linenum] | |
3061 | ||
3062 | # Since function calls often occur inside if/for/while/switch | |
3063 | # expressions - which have their own, more liberal conventions - we | |
3064 | # first see if we should be looking inside such an expression for a | |
3065 | # function call, to which we can apply more strict standards. | |
3066 | fncall = line # if there's no control flow construct, look at whole line | |
3067 | for pattern in (r'\bif\s*\((.*)\)\s*{', | |
3068 | r'\bfor\s*\((.*)\)\s*{', | |
3069 | r'\bwhile\s*\((.*)\)\s*[{;]', | |
3070 | r'\bswitch\s*\((.*)\)\s*{'): | |
3071 | match = Search(pattern, line) | |
3072 | if match: | |
3073 | fncall = match.group(1) # look inside the parens for function calls | |
3074 | break | |
3075 | ||
3076 | # Except in if/for/while/switch, there should never be space | |
3077 | # immediately inside parens (eg "f( 3, 4 )"). We make an exception | |
3078 | # for nested parens ( (a+b) + c ). Likewise, there should never be | |
3079 | # a space before a ( when it's a function argument. I assume it's a | |
3080 | # function argument when the char before the whitespace is legal in | |
3081 | # a function name (alnum + _) and we're not starting a macro. Also ignore | |
3082 | # pointers and references to arrays and functions coz they're too tricky: | |
3083 | # we use a very simple way to recognize these: | |
3084 | # " (something)(maybe-something)" or | |
3085 | # " (something)(maybe-something," or | |
3086 | # " (something)[something]" | |
3087 | # Note that we assume the contents of [] to be short enough that | |
3088 | # they'll never need to wrap. | |
3089 | if ( # Ignore control structures. | |
3090 | not Search(r'\b(if|for|while|switch|return|new|delete|catch|sizeof)\b', | |
3091 | fncall) and | |
3092 | # Ignore pointers/references to functions. | |
3093 | not Search(r' \([^)]+\)\([^)]*(\)|,$)', fncall) and | |
3094 | # Ignore pointers/references to arrays. | |
3095 | not Search(r' \([^)]+\)\[[^\]]+\]', fncall)): | |
3096 | if Search(r'\w\s*\(\s(?!\s*\\$)', fncall): # a ( used for a fn call | |
3097 | error(filename, linenum, 'whitespace/parens', 4, | |
3098 | 'Extra space after ( in function call') | |
3099 | elif Search(r'\(\s+(?!(\s*\\)|\()', fncall): | |
3100 | error(filename, linenum, 'whitespace/parens', 2, | |
3101 | 'Extra space after (') | |
3102 | if (Search(r'\w\s+\(', fncall) and | |
3103 | not Search(r'_{0,2}asm_{0,2}\s+_{0,2}volatile_{0,2}\s+\(', fncall) and | |
3104 | not Search(r'#\s*define|typedef|using\s+\w+\s*=', fncall) and | |
3105 | not Search(r'\w\s+\((\w+::)*\*\w+\)\(', fncall) and | |
3106 | not Search(r'\b(' + '|'.join(_ALT_TOKEN_REPLACEMENT.keys()) + r')\b\s+\(', | |
3107 | fncall) and | |
3108 | not Search(r'\bcase\s+\(', fncall)): | |
3109 | # TODO(unknown): Space after an operator function seem to be a common | |
3110 | # error, silence those for now by restricting them to highest verbosity. | |
3111 | if Search(r'\boperator_*\b', line): | |
3112 | error(filename, linenum, 'whitespace/parens', 0, | |
3113 | 'Extra space before ( in function call') | |
3114 | else: | |
3115 | error(filename, linenum, 'whitespace/parens', 4, | |
3116 | 'Extra space before ( in function call') | |
3117 | # If the ) is followed only by a newline or a { + newline, assume it's | |
3118 | # part of a control statement (if/while/etc), and don't complain | |
3119 | if Search(r'[^)]\s+\)\s*[^{\s]', fncall): | |
3120 | # If the closing parenthesis is preceded by only whitespaces, | |
3121 | # try to give a more descriptive error message. | |
3122 | if Search(r'^\s+\)', fncall): | |
3123 | error(filename, linenum, 'whitespace/parens', 2, | |
3124 | 'Closing ) should be moved to the previous line') | |
3125 | else: | |
3126 | error(filename, linenum, 'whitespace/parens', 2, | |
3127 | 'Extra space before )') | |
3128 | ||
3129 | ||
3130 | def IsBlankLine(line): | |
3131 | """Returns true if the given line is blank. | |
3132 | ||
3133 | We consider a line to be blank if the line is empty or consists of | |
3134 | only white spaces. | |
3135 | ||
3136 | Args: | |
3137 | line: A line of a string. | |
3138 | ||
3139 | Returns: | |
3140 | True, if the given line is blank. | |
3141 | """ | |
3142 | return not line or line.isspace() | |
3143 | ||
3144 | ||
3145 | def CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line, | |
3146 | error): | |
3147 | is_namespace_indent_item = ( | |
3148 | len(nesting_state.stack) > 1 and | |
3149 | nesting_state.stack[-1].check_namespace_indentation and | |
3150 | isinstance(nesting_state.previous_stack_top, _NamespaceInfo) and | |
3151 | nesting_state.previous_stack_top == nesting_state.stack[-2]) | |
3152 | ||
3153 | if ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item, | |
3154 | clean_lines.elided, line): | |
3155 | CheckItemIndentationInNamespace(filename, clean_lines.elided, | |
3156 | line, error) | |
3157 | ||
3158 | ||
3159 | def CheckForFunctionLengths(filename, clean_lines, linenum, | |
3160 | function_state, error): | |
3161 | """Reports for long function bodies. | |
3162 | ||
3163 | For an overview why this is done, see: | |
3164 | https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions | |
3165 | ||
3166 | Uses a simplistic algorithm assuming other style guidelines | |
3167 | (especially spacing) are followed. | |
3168 | Only checks unindented functions, so class members are unchecked. | |
3169 | Trivial bodies are unchecked, so constructors with huge initializer lists | |
3170 | may be missed. | |
3171 | Blank/comment lines are not counted so as to avoid encouraging the removal | |
3172 | of vertical space and comments just to get through a lint check. | |
3173 | NOLINT *on the last line of a function* disables this check. | |
3174 | ||
3175 | Args: | |
3176 | filename: The name of the current file. | |
3177 | clean_lines: A CleansedLines instance containing the file. | |
3178 | linenum: The number of the line to check. | |
3179 | function_state: Current function name and lines in body so far. | |
3180 | error: The function to call with any errors found. | |
3181 | """ | |
3182 | lines = clean_lines.lines | |
3183 | line = lines[linenum] | |
3184 | joined_line = '' | |
3185 | ||
3186 | starting_func = False | |
3187 | regexp = r'(\w(\w|::|\*|\&|\s)*)\(' # decls * & space::name( ... | |
3188 | match_result = Match(regexp, line) | |
3189 | if match_result: | |
3190 | # If the name is all caps and underscores, figure it's a macro and | |
3191 | # ignore it, unless it's TEST or TEST_F. | |
3192 | function_name = match_result.group(1).split()[-1] | |
3193 | if function_name == 'TEST' or function_name == 'TEST_F' or ( | |
3194 | not Match(r'[A-Z_]+$', function_name)): | |
3195 | starting_func = True | |
3196 | ||
3197 | if starting_func: | |
3198 | body_found = False | |
3199 | for start_linenum in range(linenum, clean_lines.NumLines()): | |
3200 | start_line = lines[start_linenum] | |
3201 | joined_line += ' ' + start_line.lstrip() | |
3202 | if Search(r'(;|})', start_line): # Declarations and trivial functions | |
3203 | body_found = True | |
3204 | break # ... ignore | |
3205 | elif Search(r'{', start_line): | |
3206 | body_found = True | |
3207 | function = Search(r'((\w|:)*)\(', line).group(1) | |
3208 | if Match(r'TEST', function): # Handle TEST... macros | |
3209 | parameter_regexp = Search(r'(\(.*\))', joined_line) | |
3210 | if parameter_regexp: # Ignore bad syntax | |
3211 | function += parameter_regexp.group(1) | |
3212 | else: | |
3213 | function += '()' | |
3214 | function_state.Begin(function) | |
3215 | break | |
3216 | if not body_found: | |
3217 | # No body for the function (or evidence of a non-function) was found. | |
3218 | error(filename, linenum, 'readability/fn_size', 5, | |
3219 | 'Lint failed to find start of function body.') | |
3220 | elif Match(r'^\}\s*$', line): # function end | |
3221 | function_state.Check(error, filename, linenum) | |
3222 | function_state.End() | |
3223 | elif not Match(r'^\s*$', line): | |
3224 | function_state.Count() # Count non-blank/non-comment lines. | |
3225 | ||
3226 | ||
3227 | _RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?') | |
3228 | ||
3229 | ||
3230 | def CheckComment(line, filename, linenum, next_line_start, error): | |
3231 | """Checks for common mistakes in comments. | |
3232 | ||
3233 | Args: | |
3234 | line: The line in question. | |
3235 | filename: The name of the current file. | |
3236 | linenum: The number of the line to check. | |
3237 | next_line_start: The first non-whitespace column of the next line. | |
3238 | error: The function to call with any errors found. | |
3239 | """ | |
3240 | commentpos = line.find('//') | |
3241 | if commentpos != -1: | |
3242 | # Check if the // may be in quotes. If so, ignore it | |
3243 | if re.sub(r'\\.', '', line[0:commentpos]).count('"') % 2 == 0: | |
3244 | # Allow one space for new scopes, two spaces otherwise: | |
3245 | if (not (Match(r'^.*{ *//', line) and next_line_start == commentpos) and | |
3246 | ((commentpos >= 1 and | |
3247 | line[commentpos-1] not in string.whitespace) or | |
3248 | (commentpos >= 2 and | |
3249 | line[commentpos-2] not in string.whitespace))): | |
3250 | error(filename, linenum, 'whitespace/comments', 2, | |
3251 | 'At least two spaces is best between code and comments') | |
3252 | ||
3253 | # Checks for common mistakes in TODO comments. | |
3254 | comment = line[commentpos:] | |
3255 | match = _RE_PATTERN_TODO.match(comment) | |
3256 | if match: | |
3257 | # One whitespace is correct; zero whitespace is handled elsewhere. | |
3258 | leading_whitespace = match.group(1) | |
3259 | if len(leading_whitespace) > 1: | |
3260 | error(filename, linenum, 'whitespace/todo', 2, | |
3261 | 'Too many spaces before TODO') | |
3262 | ||
3263 | username = match.group(2) | |
3264 | if not username: | |
3265 | error(filename, linenum, 'readability/todo', 2, | |
3266 | 'Missing username in TODO; it should look like ' | |
3267 | '"// TODO(my_username): Stuff."') | |
3268 | ||
3269 | middle_whitespace = match.group(3) | |
3270 | # Comparisons made explicit for correctness -- pylint: disable=g-explicit-bool-comparison | |
3271 | if middle_whitespace != ' ' and middle_whitespace != '': | |
3272 | error(filename, linenum, 'whitespace/todo', 2, | |
3273 | 'TODO(my_username) should be followed by a space') | |
3274 | ||
3275 | # If the comment contains an alphanumeric character, there | |
3276 | # should be a space somewhere between it and the // unless | |
3277 | # it's a /// or //! Doxygen comment. | |
3278 | if (Match(r'//[^ ]*\w', comment) and | |
3279 | not Match(r'(///|//\!)(\s+|$)', comment)): | |
3280 | error(filename, linenum, 'whitespace/comments', 4, | |
3281 | 'Should have a space between // and comment') | |
3282 | ||
3283 | ||
3284 | def CheckAccess(filename, clean_lines, linenum, nesting_state, error): | |
3285 | """Checks for improper use of DISALLOW* macros. | |
3286 | ||
3287 | Args: | |
3288 | filename: The name of the current file. | |
3289 | clean_lines: A CleansedLines instance containing the file. | |
3290 | linenum: The number of the line to check. | |
3291 | nesting_state: A NestingState instance which maintains information about | |
3292 | the current stack of nested blocks being parsed. | |
3293 | error: The function to call with any errors found. | |
3294 | """ | |
3295 | line = clean_lines.elided[linenum] # get rid of comments and strings | |
3296 | ||
3297 | matched = Match((r'\s*(DISALLOW_COPY_AND_ASSIGN|' | |
3298 | r'DISALLOW_IMPLICIT_CONSTRUCTORS)'), line) | |
3299 | if not matched: | |
3300 | return | |
3301 | if nesting_state.stack and isinstance(nesting_state.stack[-1], _ClassInfo): | |
3302 | if nesting_state.stack[-1].access != 'private': | |
3303 | error(filename, linenum, 'readability/constructors', 3, | |
3304 | '%s must be in the private: section' % matched.group(1)) | |
3305 | ||
3306 | else: | |
3307 | # Found DISALLOW* macro outside a class declaration, or perhaps it | |
3308 | # was used inside a function when it should have been part of the | |
3309 | # class declaration. We could issue a warning here, but it | |
3310 | # probably resulted in a compiler error already. | |
3311 | pass | |
3312 | ||
3313 | ||
3314 | def CheckSpacing(filename, clean_lines, linenum, nesting_state, error): | |
3315 | """Checks for the correctness of various spacing issues in the code. | |
3316 | ||
3317 | Things we check for: spaces around operators, spaces after | |
3318 | if/for/while/switch, no spaces around parens in function calls, two | |
3319 | spaces between code and comment, don't start a block with a blank | |
3320 | line, don't end a function with a blank line, don't add a blank line | |
3321 | after public/protected/private, don't have too many blank lines in a row. | |
3322 | ||
3323 | Args: | |
3324 | filename: The name of the current file. | |
3325 | clean_lines: A CleansedLines instance containing the file. | |
3326 | linenum: The number of the line to check. | |
3327 | nesting_state: A NestingState instance which maintains information about | |
3328 | the current stack of nested blocks being parsed. | |
3329 | error: The function to call with any errors found. | |
3330 | """ | |
3331 | ||
3332 | # Don't use "elided" lines here, otherwise we can't check commented lines. | |
3333 | # Don't want to use "raw" either, because we don't want to check inside C++11 | |
3334 | # raw strings, | |
3335 | raw = clean_lines.lines_without_raw_strings | |
3336 | line = raw[linenum] | |
3337 | ||
3338 | # Before nixing comments, check if the line is blank for no good | |
3339 | # reason. This includes the first line after a block is opened, and | |
3340 | # blank lines at the end of a function (ie, right before a line like '}' | |
3341 | # | |
3342 | # Skip all the blank line checks if we are immediately inside a | |
3343 | # namespace body. In other words, don't issue blank line warnings | |
3344 | # for this block: | |
3345 | # namespace { | |
3346 | # | |
3347 | # } | |
3348 | # | |
3349 | # A warning about missing end of namespace comments will be issued instead. | |
3350 | # | |
3351 | # Also skip blank line checks for 'extern "C"' blocks, which are formatted | |
3352 | # like namespaces. | |
3353 | if (IsBlankLine(line) and | |
3354 | not nesting_state.InNamespaceBody() and | |
3355 | not nesting_state.InExternC()): | |
3356 | elided = clean_lines.elided | |
3357 | prev_line = elided[linenum - 1] | |
3358 | prevbrace = prev_line.rfind('{') | |
3359 | # TODO(unknown): Don't complain if line before blank line, and line after, | |
3360 | # both start with alnums and are indented the same amount. | |
3361 | # This ignores whitespace at the start of a namespace block | |
3362 | # because those are not usually indented. | |
3363 | if prevbrace != -1 and prev_line[prevbrace:].find('}') == -1: | |
3364 | # OK, we have a blank line at the start of a code block. Before we | |
3365 | # complain, we check if it is an exception to the rule: The previous | |
3366 | # non-empty line has the parameters of a function header that are indented | |
3367 | # 4 spaces (because they did not fit in a 80 column line when placed on | |
3368 | # the same line as the function name). We also check for the case where | |
3369 | # the previous line is indented 6 spaces, which may happen when the | |
3370 | # initializers of a constructor do not fit into a 80 column line. | |
3371 | exception = False | |
3372 | if Match(r' {6}\w', prev_line): # Initializer list? | |
3373 | # We are looking for the opening column of initializer list, which | |
3374 | # should be indented 4 spaces to cause 6 space indentation afterwards. | |
3375 | search_position = linenum-2 | |
3376 | while (search_position >= 0 | |
3377 | and Match(r' {6}\w', elided[search_position])): | |
3378 | search_position -= 1 | |
3379 | exception = (search_position >= 0 | |
3380 | and elided[search_position][:5] == ' :') | |
3381 | else: | |
3382 | # Search for the function arguments or an initializer list. We use a | |
3383 | # simple heuristic here: If the line is indented 4 spaces; and we have a | |
3384 | # closing paren, without the opening paren, followed by an opening brace | |
3385 | # or colon (for initializer lists) we assume that it is the last line of | |
3386 | # a function header. If we have a colon indented 4 spaces, it is an | |
3387 | # initializer list. | |
3388 | exception = (Match(r' {4}\w[^\(]*\)\s*(const\s*)?(\{\s*$|:)', | |
3389 | prev_line) | |
3390 | or Match(r' {4}:', prev_line)) | |
3391 | ||
3392 | if not exception: | |
3393 | error(filename, linenum, 'whitespace/blank_line', 2, | |
3394 | 'Redundant blank line at the start of a code block ' | |
3395 | 'should be deleted.') | |
3396 | # Ignore blank lines at the end of a block in a long if-else | |
3397 | # chain, like this: | |
3398 | # if (condition1) { | |
3399 | # // Something followed by a blank line | |
3400 | # | |
3401 | # } else if (condition2) { | |
3402 | # // Something else | |
3403 | # } | |
3404 | if linenum + 1 < clean_lines.NumLines(): | |
3405 | next_line = raw[linenum + 1] | |
3406 | if (next_line | |
3407 | and Match(r'\s*}', next_line) | |
3408 | and next_line.find('} else ') == -1): | |
3409 | error(filename, linenum, 'whitespace/blank_line', 3, | |
3410 | 'Redundant blank line at the end of a code block ' | |
3411 | 'should be deleted.') | |
3412 | ||
3413 | matched = Match(r'\s*(public|protected|private):', prev_line) | |
3414 | if matched: | |
3415 | error(filename, linenum, 'whitespace/blank_line', 3, | |
3416 | 'Do not leave a blank line after "%s:"' % matched.group(1)) | |
3417 | ||
3418 | # Next, check comments | |
3419 | next_line_start = 0 | |
3420 | if linenum + 1 < clean_lines.NumLines(): | |
3421 | next_line = raw[linenum + 1] | |
3422 | next_line_start = len(next_line) - len(next_line.lstrip()) | |
3423 | CheckComment(line, filename, linenum, next_line_start, error) | |
3424 | ||
3425 | # get rid of comments and strings | |
3426 | line = clean_lines.elided[linenum] | |
3427 | ||
3428 | # You shouldn't have spaces before your brackets, except maybe after | |
3429 | # 'delete []' or 'return []() {};' | |
3430 | if Search(r'\w\s+\[', line) and not Search(r'(?:delete|return)\s+\[', line): | |
3431 | error(filename, linenum, 'whitespace/braces', 5, | |
3432 | 'Extra space before [') | |
3433 | ||
3434 | # In range-based for, we wanted spaces before and after the colon, but | |
3435 | # not around "::" tokens that might appear. | |
3436 | if (Search(r'for *\(.*[^:]:[^: ]', line) or | |
3437 | Search(r'for *\(.*[^: ]:[^:]', line)): | |
3438 | error(filename, linenum, 'whitespace/forcolon', 2, | |
3439 | 'Missing space around colon in range-based for loop') | |
3440 | ||
3441 | ||
3442 | def CheckOperatorSpacing(filename, clean_lines, linenum, error): | |
3443 | """Checks for horizontal spacing around operators. | |
3444 | ||
3445 | Args: | |
3446 | filename: The name of the current file. | |
3447 | clean_lines: A CleansedLines instance containing the file. | |
3448 | linenum: The number of the line to check. | |
3449 | error: The function to call with any errors found. | |
3450 | """ | |
3451 | line = clean_lines.elided[linenum] | |
3452 | ||
3453 | # Don't try to do spacing checks for operator methods. Do this by | |
3454 | # replacing the troublesome characters with something else, | |
3455 | # preserving column position for all other characters. | |
3456 | # | |
3457 | # The replacement is done repeatedly to avoid false positives from | |
3458 | # operators that call operators. | |
3459 | while True: | |
3460 | match = Match(r'^(.*\boperator\b)(\S+)(\s*\(.*)$', line) | |
3461 | if match: | |
3462 | line = match.group(1) + ('_' * len(match.group(2))) + match.group(3) | |
3463 | else: | |
3464 | break | |
3465 | ||
3466 | # We allow no-spaces around = within an if: "if ( (a=Foo()) == 0 )". | |
3467 | # Otherwise not. Note we only check for non-spaces on *both* sides; | |
3468 | # sometimes people put non-spaces on one side when aligning ='s among | |
3469 | # many lines (not that this is behavior that I approve of...) | |
3470 | if ((Search(r'[\w.]=', line) or | |
3471 | Search(r'=[\w.]', line)) | |
3472 | and not Search(r'\b(if|while|for) ', line) | |
3473 | # Operators taken from [lex.operators] in C++11 standard. | |
3474 | and not Search(r'(>=|<=|==|!=|&=|\^=|\|=|\+=|\*=|\/=|\%=)', line) | |
3475 | and not Search(r'operator=', line)): | |
3476 | error(filename, linenum, 'whitespace/operators', 4, | |
3477 | 'Missing spaces around =') | |
3478 | ||
3479 | # It's ok not to have spaces around binary operators like + - * /, but if | |
3480 | # there's too little whitespace, we get concerned. It's hard to tell, | |
3481 | # though, so we punt on this one for now. TODO. | |
3482 | ||
3483 | # You should always have whitespace around binary operators. | |
3484 | # | |
3485 | # Check <= and >= first to avoid false positives with < and >, then | |
3486 | # check non-include lines for spacing around < and >. | |
3487 | # | |
3488 | # If the operator is followed by a comma, assume it's be used in a | |
3489 | # macro context and don't do any checks. This avoids false | |
3490 | # positives. | |
3491 | # | |
3492 | # Note that && is not included here. This is because there are too | |
3493 | # many false positives due to RValue references. | |
3494 | match = Search(r'[^<>=!\s](==|!=|<=|>=|\|\|)[^<>=!\s,;\)]', line) | |
3495 | if match: | |
3496 | error(filename, linenum, 'whitespace/operators', 3, | |
3497 | 'Missing spaces around %s' % match.group(1)) | |
3498 | elif not Match(r'#.*include', line): | |
3499 | # Look for < that is not surrounded by spaces. This is only | |
3500 | # triggered if both sides are missing spaces, even though | |
3501 | # technically should should flag if at least one side is missing a | |
3502 | # space. This is done to avoid some false positives with shifts. | |
3503 | match = Match(r'^(.*[^\s<])<[^\s=<,]', line) | |
3504 | if match: | |
3505 | (_, _, end_pos) = CloseExpression( | |
3506 | clean_lines, linenum, len(match.group(1))) | |
3507 | if end_pos <= -1: | |
3508 | error(filename, linenum, 'whitespace/operators', 3, | |
3509 | 'Missing spaces around <') | |
3510 | ||
3511 | # Look for > that is not surrounded by spaces. Similar to the | |
3512 | # above, we only trigger if both sides are missing spaces to avoid | |
3513 | # false positives with shifts. | |
3514 | match = Match(r'^(.*[^-\s>])>[^\s=>,]', line) | |
3515 | if match: | |
3516 | (_, _, start_pos) = ReverseCloseExpression( | |
3517 | clean_lines, linenum, len(match.group(1))) | |
3518 | if start_pos <= -1: | |
3519 | error(filename, linenum, 'whitespace/operators', 3, | |
3520 | 'Missing spaces around >') | |
3521 | ||
3522 | # We allow no-spaces around << when used like this: 10<<20, but | |
3523 | # not otherwise (particularly, not when used as streams) | |
3524 | # | |
3525 | # We also allow operators following an opening parenthesis, since | |
3526 | # those tend to be macros that deal with operators. | |
3527 | match = Search(r'(operator|[^\s(<])(?:L|UL|LL|ULL|l|ul|ll|ull)?<<([^\s,=<])', line) | |
3528 | if (match and not (match.group(1).isdigit() and match.group(2).isdigit()) and | |
3529 | not (match.group(1) == 'operator' and match.group(2) == ';')): | |
3530 | error(filename, linenum, 'whitespace/operators', 3, | |
3531 | 'Missing spaces around <<') | |
3532 | ||
3533 | # We allow no-spaces around >> for almost anything. This is because | |
3534 | # C++11 allows ">>" to close nested templates, which accounts for | |
3535 | # most cases when ">>" is not followed by a space. | |
3536 | # | |
3537 | # We still warn on ">>" followed by alpha character, because that is | |
3538 | # likely due to ">>" being used for right shifts, e.g.: | |
3539 | # value >> alpha | |
3540 | # | |
3541 | # When ">>" is used to close templates, the alphanumeric letter that | |
3542 | # follows would be part of an identifier, and there should still be | |
3543 | # a space separating the template type and the identifier. | |
3544 | # type<type<type>> alpha | |
3545 | match = Search(r'>>[a-zA-Z_]', line) | |
3546 | if match: | |
3547 | error(filename, linenum, 'whitespace/operators', 3, | |
3548 | 'Missing spaces around >>') | |
3549 | ||
3550 | # There shouldn't be space around unary operators | |
3551 | match = Search(r'(!\s|~\s|[\s]--[\s;]|[\s]\+\+[\s;])', line) | |
3552 | if match: | |
3553 | error(filename, linenum, 'whitespace/operators', 4, | |
3554 | 'Extra space for operator %s' % match.group(1)) | |
3555 | ||
3556 | ||
3557 | def CheckParenthesisSpacing(filename, clean_lines, linenum, error): | |
3558 | """Checks for horizontal spacing around parentheses. | |
3559 | ||
3560 | Args: | |
3561 | filename: The name of the current file. | |
3562 | clean_lines: A CleansedLines instance containing the file. | |
3563 | linenum: The number of the line to check. | |
3564 | error: The function to call with any errors found. | |
3565 | """ | |
3566 | line = clean_lines.elided[linenum] | |
3567 | ||
3568 | # No spaces after an if, while, switch, or for | |
3569 | match = Search(r' (if\(|for\(|while\(|switch\()', line) | |
3570 | if match: | |
3571 | error(filename, linenum, 'whitespace/parens', 5, | |
3572 | 'Missing space before ( in %s' % match.group(1)) | |
3573 | ||
3574 | # For if/for/while/switch, the left and right parens should be | |
3575 | # consistent about how many spaces are inside the parens, and | |
3576 | # there should either be zero or one spaces inside the parens. | |
3577 | # We don't want: "if ( foo)" or "if ( foo )". | |
3578 | # Exception: "for ( ; foo; bar)" and "for (foo; bar; )" are allowed. | |
3579 | match = Search(r'\b(if|for|while|switch)\s*' | |
3580 | r'\(([ ]*)(.).*[^ ]+([ ]*)\)\s*{\s*$', | |
3581 | line) | |
3582 | if match: | |
3583 | if len(match.group(2)) != len(match.group(4)): | |
3584 | if not (match.group(3) == ';' and | |
3585 | len(match.group(2)) == 1 + len(match.group(4)) or | |
3586 | not match.group(2) and Search(r'\bfor\s*\(.*; \)', line)): | |
3587 | error(filename, linenum, 'whitespace/parens', 5, | |
3588 | 'Mismatching spaces inside () in %s' % match.group(1)) | |
3589 | if len(match.group(2)) not in [0, 1]: | |
3590 | error(filename, linenum, 'whitespace/parens', 5, | |
3591 | 'Should have zero or one spaces inside ( and ) in %s' % | |
3592 | match.group(1)) | |
3593 | ||
3594 | ||
3595 | def CheckCommaSpacing(filename, clean_lines, linenum, error): | |
3596 | """Checks for horizontal spacing near commas and semicolons. | |
3597 | ||
3598 | Args: | |
3599 | filename: The name of the current file. | |
3600 | clean_lines: A CleansedLines instance containing the file. | |
3601 | linenum: The number of the line to check. | |
3602 | error: The function to call with any errors found. | |
3603 | """ | |
3604 | raw = clean_lines.lines_without_raw_strings | |
3605 | line = clean_lines.elided[linenum] | |
3606 | ||
3607 | # You should always have a space after a comma (either as fn arg or operator) | |
3608 | # | |
3609 | # This does not apply when the non-space character following the | |
3610 | # comma is another comma, since the only time when that happens is | |
3611 | # for empty macro arguments. | |
3612 | # | |
3613 | # We run this check in two passes: first pass on elided lines to | |
3614 | # verify that lines contain missing whitespaces, second pass on raw | |
3615 | # lines to confirm that those missing whitespaces are not due to | |
3616 | # elided comments. | |
3617 | if (Search(r',[^,\s]', ReplaceAll(r'\boperator\s*,\s*\(', 'F(', line)) and | |
3618 | Search(r',[^,\s]', raw[linenum])): | |
3619 | error(filename, linenum, 'whitespace/comma', 3, | |
3620 | 'Missing space after ,') | |
3621 | ||
3622 | # You should always have a space after a semicolon | |
3623 | # except for few corner cases | |
3624 | # TODO(unknown): clarify if 'if (1) { return 1;}' is requires one more | |
3625 | # space after ; | |
3626 | if Search(r';[^\s};\\)/]', line): | |
3627 | error(filename, linenum, 'whitespace/semicolon', 3, | |
3628 | 'Missing space after ;') | |
3629 | ||
3630 | ||
3631 | def _IsType(clean_lines, nesting_state, expr): | |
3632 | """Check if expression looks like a type name, returns true if so. | |
3633 | ||
3634 | Args: | |
3635 | clean_lines: A CleansedLines instance containing the file. | |
3636 | nesting_state: A NestingState instance which maintains information about | |
3637 | the current stack of nested blocks being parsed. | |
3638 | expr: The expression to check. | |
3639 | Returns: | |
3640 | True, if token looks like a type. | |
3641 | """ | |
3642 | # Keep only the last token in the expression | |
3643 | last_word = Match(r'^.*(\b\S+)$', expr) | |
3644 | if last_word: | |
3645 | token = last_word.group(1) | |
3646 | else: | |
3647 | token = expr | |
3648 | ||
3649 | # Match native types and stdint types | |
3650 | if _TYPES.match(token): | |
3651 | return True | |
3652 | ||
3653 | # Try a bit harder to match templated types. Walk up the nesting | |
3654 | # stack until we find something that resembles a typename | |
3655 | # declaration for what we are looking for. | |
3656 | typename_pattern = (r'\b(?:typename|class|struct)\s+' + re.escape(token) + | |
3657 | r'\b') | |
3658 | block_index = len(nesting_state.stack) - 1 | |
3659 | while block_index >= 0: | |
3660 | if isinstance(nesting_state.stack[block_index], _NamespaceInfo): | |
3661 | return False | |
3662 | ||
3663 | # Found where the opening brace is. We want to scan from this | |
3664 | # line up to the beginning of the function, minus a few lines. | |
3665 | # template <typename Type1, // stop scanning here | |
3666 | # ...> | |
3667 | # class C | |
3668 | # : public ... { // start scanning here | |
3669 | last_line = nesting_state.stack[block_index].starting_linenum | |
3670 | ||
3671 | next_block_start = 0 | |
3672 | if block_index > 0: | |
3673 | next_block_start = nesting_state.stack[block_index - 1].starting_linenum | |
3674 | first_line = last_line | |
3675 | while first_line >= next_block_start: | |
3676 | if clean_lines.elided[first_line].find('template') >= 0: | |
3677 | break | |
3678 | first_line -= 1 | |
3679 | if first_line < next_block_start: | |
3680 | # Didn't find any "template" keyword before reaching the next block, | |
3681 | # there are probably no template things to check for this block | |
3682 | block_index -= 1 | |
3683 | continue | |
3684 | ||
3685 | # Look for typename in the specified range | |
3686 | for i in xrange(first_line, last_line + 1, 1): | |
3687 | if Search(typename_pattern, clean_lines.elided[i]): | |
3688 | return True | |
3689 | block_index -= 1 | |
3690 | ||
3691 | return False | |
3692 | ||
3693 | ||
3694 | def CheckBracesSpacing(filename, clean_lines, linenum, nesting_state, error): | |
3695 | """Checks for horizontal spacing near commas. | |
3696 | ||
3697 | Args: | |
3698 | filename: The name of the current file. | |
3699 | clean_lines: A CleansedLines instance containing the file. | |
3700 | linenum: The number of the line to check. | |
3701 | nesting_state: A NestingState instance which maintains information about | |
3702 | the current stack of nested blocks being parsed. | |
3703 | error: The function to call with any errors found. | |
3704 | """ | |
3705 | line = clean_lines.elided[linenum] | |
3706 | ||
3707 | # Except after an opening paren, or after another opening brace (in case of | |
3708 | # an initializer list, for instance), you should have spaces before your | |
3709 | # braces when they are delimiting blocks, classes, namespaces etc. | |
3710 | # And since you should never have braces at the beginning of a line, | |
3711 | # this is an easy test. Except that braces used for initialization don't | |
3712 | # follow the same rule; we often don't want spaces before those. | |
3713 | match = Match(r'^(.*[^ ({>]){', line) | |
3714 | ||
3715 | if match: | |
3716 | # Try a bit harder to check for brace initialization. This | |
3717 | # happens in one of the following forms: | |
3718 | # Constructor() : initializer_list_{} { ... } | |
3719 | # Constructor{}.MemberFunction() | |
3720 | # Type variable{}; | |
3721 | # FunctionCall(type{}, ...); | |
3722 | # LastArgument(..., type{}); | |
3723 | # LOG(INFO) << type{} << " ..."; | |
3724 | # map_of_type[{...}] = ...; | |
3725 | # ternary = expr ? new type{} : nullptr; | |
3726 | # OuterTemplate<InnerTemplateConstructor<Type>{}> | |
3727 | # | |
3728 | # We check for the character following the closing brace, and | |
3729 | # silence the warning if it's one of those listed above, i.e. | |
3730 | # "{.;,)<>]:". | |
3731 | # | |
3732 | # To account for nested initializer list, we allow any number of | |
3733 | # closing braces up to "{;,)<". We can't simply silence the | |
3734 | # warning on first sight of closing brace, because that would | |
3735 | # cause false negatives for things that are not initializer lists. | |
3736 | # Silence this: But not this: | |
3737 | # Outer{ if (...) { | |
3738 | # Inner{...} if (...){ // Missing space before { | |
3739 | # }; } | |
3740 | # | |
3741 | # There is a false negative with this approach if people inserted | |
3742 | # spurious semicolons, e.g. "if (cond){};", but we will catch the | |
3743 | # spurious semicolon with a separate check. | |
3744 | leading_text = match.group(1) | |
3745 | (endline, endlinenum, endpos) = CloseExpression( | |
3746 | clean_lines, linenum, len(match.group(1))) | |
3747 | trailing_text = '' | |
3748 | if endpos > -1: | |
3749 | trailing_text = endline[endpos:] | |
3750 | for offset in xrange(endlinenum + 1, | |
3751 | min(endlinenum + 3, clean_lines.NumLines() - 1)): | |
3752 | trailing_text += clean_lines.elided[offset] | |
3753 | # We also suppress warnings for `uint64_t{expression}` etc., as the style | |
3754 | # guide recommends brace initialization for integral types to avoid | |
3755 | # overflow/truncation. | |
3756 | if (not Match(r'^[\s}]*[{.;,)<>\]:]', trailing_text) | |
3757 | and not _IsType(clean_lines, nesting_state, leading_text)): | |
3758 | error(filename, linenum, 'whitespace/braces', 5, | |
3759 | 'Missing space before {') | |
3760 | ||
3761 | # Make sure '} else {' has spaces. | |
3762 | if Search(r'}else', line): | |
3763 | error(filename, linenum, 'whitespace/braces', 5, | |
3764 | 'Missing space before else') | |
3765 | ||
3766 | # You shouldn't have a space before a semicolon at the end of the line. | |
3767 | # There's a special case for "for" since the style guide allows space before | |
3768 | # the semicolon there. | |
3769 | if Search(r':\s*;\s*$', line): | |
3770 | error(filename, linenum, 'whitespace/semicolon', 5, | |
3771 | 'Semicolon defining empty statement. Use {} instead.') | |
3772 | elif Search(r'^\s*;\s*$', line): | |
3773 | error(filename, linenum, 'whitespace/semicolon', 5, | |
3774 | 'Line contains only semicolon. If this should be an empty statement, ' | |
3775 | 'use {} instead.') | |
3776 | elif (Search(r'\s+;\s*$', line) and | |
3777 | not Search(r'\bfor\b', line)): | |
3778 | error(filename, linenum, 'whitespace/semicolon', 5, | |
3779 | 'Extra space before last semicolon. If this should be an empty ' | |
3780 | 'statement, use {} instead.') | |
3781 | ||
3782 | ||
3783 | def IsDecltype(clean_lines, linenum, column): | |
3784 | """Check if the token ending on (linenum, column) is decltype(). | |
3785 | ||
3786 | Args: | |
3787 | clean_lines: A CleansedLines instance containing the file. | |
3788 | linenum: the number of the line to check. | |
3789 | column: end column of the token to check. | |
3790 | Returns: | |
3791 | True if this token is decltype() expression, False otherwise. | |
3792 | """ | |
3793 | (text, _, start_col) = ReverseCloseExpression(clean_lines, linenum, column) | |
3794 | if start_col < 0: | |
3795 | return False | |
3796 | if Search(r'\bdecltype\s*$', text[0:start_col]): | |
3797 | return True | |
3798 | return False | |
3799 | ||
3800 | def CheckSectionSpacing(filename, clean_lines, class_info, linenum, error): | |
3801 | """Checks for additional blank line issues related to sections. | |
3802 | ||
3803 | Currently the only thing checked here is blank line before protected/private. | |
3804 | ||
3805 | Args: | |
3806 | filename: The name of the current file. | |
3807 | clean_lines: A CleansedLines instance containing the file. | |
3808 | class_info: A _ClassInfo objects. | |
3809 | linenum: The number of the line to check. | |
3810 | error: The function to call with any errors found. | |
3811 | """ | |
3812 | # Skip checks if the class is small, where small means 25 lines or less. | |
3813 | # 25 lines seems like a good cutoff since that's the usual height of | |
3814 | # terminals, and any class that can't fit in one screen can't really | |
3815 | # be considered "small". | |
3816 | # | |
3817 | # Also skip checks if we are on the first line. This accounts for | |
3818 | # classes that look like | |
3819 | # class Foo { public: ... }; | |
3820 | # | |
3821 | # If we didn't find the end of the class, last_line would be zero, | |
3822 | # and the check will be skipped by the first condition. | |
3823 | if (class_info.last_line - class_info.starting_linenum <= 24 or | |
3824 | linenum <= class_info.starting_linenum): | |
3825 | return | |
3826 | ||
3827 | matched = Match(r'\s*(public|protected|private):', clean_lines.lines[linenum]) | |
3828 | if matched: | |
3829 | # Issue warning if the line before public/protected/private was | |
3830 | # not a blank line, but don't do this if the previous line contains | |
3831 | # "class" or "struct". This can happen two ways: | |
3832 | # - We are at the beginning of the class. | |
3833 | # - We are forward-declaring an inner class that is semantically | |
3834 | # private, but needed to be public for implementation reasons. | |
3835 | # Also ignores cases where the previous line ends with a backslash as can be | |
3836 | # common when defining classes in C macros. | |
3837 | prev_line = clean_lines.lines[linenum - 1] | |
3838 | if (not IsBlankLine(prev_line) and | |
3839 | not Search(r'\b(class|struct)\b', prev_line) and | |
3840 | not Search(r'\\$', prev_line)): | |
3841 | # Try a bit harder to find the beginning of the class. This is to | |
3842 | # account for multi-line base-specifier lists, e.g.: | |
3843 | # class Derived | |
3844 | # : public Base { | |
3845 | end_class_head = class_info.starting_linenum | |
3846 | for i in range(class_info.starting_linenum, linenum): | |
3847 | if Search(r'\{\s*$', clean_lines.lines[i]): | |
3848 | end_class_head = i | |
3849 | break | |
3850 | if end_class_head < linenum - 1: | |
3851 | error(filename, linenum, 'whitespace/blank_line', 3, | |
3852 | '"%s:" should be preceded by a blank line' % matched.group(1)) | |
3853 | ||
3854 | ||
3855 | def GetPreviousNonBlankLine(clean_lines, linenum): | |
3856 | """Return the most recent non-blank line and its line number. | |
3857 | ||
3858 | Args: | |
3859 | clean_lines: A CleansedLines instance containing the file contents. | |
3860 | linenum: The number of the line to check. | |
3861 | ||
3862 | Returns: | |
3863 | A tuple with two elements. The first element is the contents of the last | |
3864 | non-blank line before the current line, or the empty string if this is the | |
3865 | first non-blank line. The second is the line number of that line, or -1 | |
3866 | if this is the first non-blank line. | |
3867 | """ | |
3868 | ||
3869 | prevlinenum = linenum - 1 | |
3870 | while prevlinenum >= 0: | |
3871 | prevline = clean_lines.elided[prevlinenum] | |
3872 | if not IsBlankLine(prevline): # if not a blank line... | |
3873 | return (prevline, prevlinenum) | |
3874 | prevlinenum -= 1 | |
3875 | return ('', -1) | |
3876 | ||
3877 | ||
3878 | def CheckBraces(filename, clean_lines, linenum, error): | |
3879 | """Looks for misplaced braces (e.g. at the end of line). | |
3880 | ||
3881 | Args: | |
3882 | filename: The name of the current file. | |
3883 | clean_lines: A CleansedLines instance containing the file. | |
3884 | linenum: The number of the line to check. | |
3885 | error: The function to call with any errors found. | |
3886 | """ | |
3887 | ||
3888 | line = clean_lines.elided[linenum] # get rid of comments and strings | |
3889 | ||
3890 | if Match(r'\s*{\s*$', line): | |
3891 | # We allow an open brace to start a line in the case where someone is using | |
3892 | # braces in a block to explicitly create a new scope, which is commonly used | |
3893 | # to control the lifetime of stack-allocated variables. Braces are also | |
3894 | # used for brace initializers inside function calls. We don't detect this | |
3895 | # perfectly: we just don't complain if the last non-whitespace character on | |
3896 | # the previous non-blank line is ',', ';', ':', '(', '{', or '}', or if the | |
3897 | # previous line starts a preprocessor block. We also allow a brace on the | |
3898 | # following line if it is part of an array initialization and would not fit | |
3899 | # within the 80 character limit of the preceding line. | |
3900 | prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] | |
3901 | if (not Search(r'[,;:}{(]\s*$', prevline) and | |
3902 | not Match(r'\s*#', prevline) and | |
3903 | not (GetLineWidth(prevline) > _line_length - 2 and '[]' in prevline)): | |
3904 | error(filename, linenum, 'whitespace/braces', 4, | |
3905 | '{ should almost always be at the end of the previous line') | |
3906 | ||
3907 | # An else clause should be on the same line as the preceding closing brace. | |
3908 | if Match(r'\s*else\b\s*(?:if\b|\{|$)', line): | |
3909 | prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] | |
3910 | if Match(r'\s*}\s*$', prevline): | |
3911 | error(filename, linenum, 'whitespace/newline', 4, | |
3912 | 'An else should appear on the same line as the preceding }') | |
3913 | ||
3914 | # If braces come on one side of an else, they should be on both. | |
3915 | # However, we have to worry about "else if" that spans multiple lines! | |
3916 | if Search(r'else if\s*\(', line): # could be multi-line if | |
3917 | brace_on_left = bool(Search(r'}\s*else if\s*\(', line)) | |
3918 | # find the ( after the if | |
3919 | pos = line.find('else if') | |
3920 | pos = line.find('(', pos) | |
3921 | if pos > 0: | |
3922 | (endline, _, endpos) = CloseExpression(clean_lines, linenum, pos) | |
3923 | brace_on_right = endline[endpos:].find('{') != -1 | |
3924 | if brace_on_left != brace_on_right: # must be brace after if | |
3925 | error(filename, linenum, 'readability/braces', 5, | |
3926 | 'If an else has a brace on one side, it should have it on both') | |
3927 | elif Search(r'}\s*else[^{]*$', line) or Match(r'[^}]*else\s*{', line): | |
3928 | error(filename, linenum, 'readability/braces', 5, | |
3929 | 'If an else has a brace on one side, it should have it on both') | |
3930 | ||
3931 | # Likewise, an else should never have the else clause on the same line | |
3932 | if Search(r'\belse [^\s{]', line) and not Search(r'\belse if\b', line): | |
3933 | error(filename, linenum, 'whitespace/newline', 4, | |
3934 | 'Else clause should never be on same line as else (use 2 lines)') | |
3935 | ||
3936 | # In the same way, a do/while should never be on one line | |
3937 | if Match(r'\s*do [^\s{]', line): | |
3938 | error(filename, linenum, 'whitespace/newline', 4, | |
3939 | 'do/while clauses should not be on a single line') | |
3940 | ||
3941 | # Check single-line if/else bodies. The style guide says 'curly braces are not | |
3942 | # required for single-line statements'. We additionally allow multi-line, | |
3943 | # single statements, but we reject anything with more than one semicolon in | |
3944 | # it. This means that the first semicolon after the if should be at the end of | |
3945 | # its line, and the line after that should have an indent level equal to or | |
3946 | # lower than the if. We also check for ambiguous if/else nesting without | |
3947 | # braces. | |
3948 | if_else_match = Search(r'\b(if\s*\(|else\b)', line) | |
3949 | if if_else_match and not Match(r'\s*#', line): | |
3950 | if_indent = GetIndentLevel(line) | |
3951 | endline, endlinenum, endpos = line, linenum, if_else_match.end() | |
3952 | if_match = Search(r'\bif\s*\(', line) | |
3953 | if if_match: | |
3954 | # This could be a multiline if condition, so find the end first. | |
3955 | pos = if_match.end() - 1 | |
3956 | (endline, endlinenum, endpos) = CloseExpression(clean_lines, linenum, pos) | |
3957 | # Check for an opening brace, either directly after the if or on the next | |
3958 | # line. If found, this isn't a single-statement conditional. | |
3959 | if (not Match(r'\s*{', endline[endpos:]) | |
3960 | and not (Match(r'\s*$', endline[endpos:]) | |
3961 | and endlinenum < (len(clean_lines.elided) - 1) | |
3962 | and Match(r'\s*{', clean_lines.elided[endlinenum + 1]))): | |
3963 | while (endlinenum < len(clean_lines.elided) | |
3964 | and ';' not in clean_lines.elided[endlinenum][endpos:]): | |
3965 | endlinenum += 1 | |
3966 | endpos = 0 | |
3967 | if endlinenum < len(clean_lines.elided): | |
3968 | endline = clean_lines.elided[endlinenum] | |
3969 | # We allow a mix of whitespace and closing braces (e.g. for one-liner | |
3970 | # methods) and a single \ after the semicolon (for macros) | |
3971 | endpos = endline.find(';') | |
3972 | if not Match(r';[\s}]*(\\?)$', endline[endpos:]): | |
3973 | # Semicolon isn't the last character, there's something trailing. | |
3974 | # Output a warning if the semicolon is not contained inside | |
3975 | # a lambda expression. | |
3976 | if not Match(r'^[^{};]*\[[^\[\]]*\][^{}]*\{[^{}]*\}\s*\)*[;,]\s*$', | |
3977 | endline): | |
3978 | error(filename, linenum, 'readability/braces', 4, | |
3979 | 'If/else bodies with multiple statements require braces') | |
3980 | elif endlinenum < len(clean_lines.elided) - 1: | |
3981 | # Make sure the next line is dedented | |
3982 | next_line = clean_lines.elided[endlinenum + 1] | |
3983 | next_indent = GetIndentLevel(next_line) | |
3984 | # With ambiguous nested if statements, this will error out on the | |
3985 | # if that *doesn't* match the else, regardless of whether it's the | |
3986 | # inner one or outer one. | |
3987 | if (if_match and Match(r'\s*else\b', next_line) | |
3988 | and next_indent != if_indent): | |
3989 | error(filename, linenum, 'readability/braces', 4, | |
3990 | 'Else clause should be indented at the same level as if. ' | |
3991 | 'Ambiguous nested if/else chains require braces.') | |
3992 | elif next_indent > if_indent: | |
3993 | error(filename, linenum, 'readability/braces', 4, | |
3994 | 'If/else bodies with multiple statements require braces') | |
3995 | ||
3996 | ||
3997 | def CheckTrailingSemicolon(filename, clean_lines, linenum, error): | |
3998 | """Looks for redundant trailing semicolon. | |
3999 | ||
4000 | Args: | |
4001 | filename: The name of the current file. | |
4002 | clean_lines: A CleansedLines instance containing the file. | |
4003 | linenum: The number of the line to check. | |
4004 | error: The function to call with any errors found. | |
4005 | """ | |
4006 | ||
4007 | line = clean_lines.elided[linenum] | |
4008 | ||
4009 | # Block bodies should not be followed by a semicolon. Due to C++11 | |
4010 | # brace initialization, there are more places where semicolons are | |
4011 | # required than not, so we use explicitly list the allowed rules | |
4012 | # rather than listing the disallowed ones. These are the places | |
4013 | # where "};" should be replaced by just "}": | |
4014 | # 1. Some flavor of block following closing parenthesis: | |
4015 | # for (;;) {}; | |
4016 | # while (...) {}; | |
4017 | # switch (...) {}; | |
4018 | # Function(...) {}; | |
4019 | # if (...) {}; | |
4020 | # if (...) else if (...) {}; | |
4021 | # | |
4022 | # 2. else block: | |
4023 | # if (...) else {}; | |
4024 | # | |
4025 | # 3. const member function: | |
4026 | # Function(...) const {}; | |
4027 | # | |
4028 | # 4. Block following some statement: | |
4029 | # x = 42; | |
4030 | # {}; | |
4031 | # | |
4032 | # 5. Block at the beginning of a function: | |
4033 | # Function(...) { | |
4034 | # {}; | |
4035 | # } | |
4036 | # | |
4037 | # Note that naively checking for the preceding "{" will also match | |
4038 | # braces inside multi-dimensional arrays, but this is fine since | |
4039 | # that expression will not contain semicolons. | |
4040 | # | |
4041 | # 6. Block following another block: | |
4042 | # while (true) {} | |
4043 | # {}; | |
4044 | # | |
4045 | # 7. End of namespaces: | |
4046 | # namespace {}; | |
4047 | # | |
4048 | # These semicolons seems far more common than other kinds of | |
4049 | # redundant semicolons, possibly due to people converting classes | |
4050 | # to namespaces. For now we do not warn for this case. | |
4051 | # | |
4052 | # Try matching case 1 first. | |
4053 | match = Match(r'^(.*\)\s*)\{', line) | |
4054 | if match: | |
4055 | # Matched closing parenthesis (case 1). Check the token before the | |
4056 | # matching opening parenthesis, and don't warn if it looks like a | |
4057 | # macro. This avoids these false positives: | |
4058 | # - macro that defines a base class | |
4059 | # - multi-line macro that defines a base class | |
4060 | # - macro that defines the whole class-head | |
4061 | # | |
4062 | # But we still issue warnings for macros that we know are safe to | |
4063 | # warn, specifically: | |
4064 | # - TEST, TEST_F, TEST_P, MATCHER, MATCHER_P | |
4065 | # - TYPED_TEST | |
4066 | # - INTERFACE_DEF | |
4067 | # - EXCLUSIVE_LOCKS_REQUIRED, SHARED_LOCKS_REQUIRED, LOCKS_EXCLUDED: | |
4068 | # | |
4069 | # We implement a list of safe macros instead of a list of | |
4070 | # unsafe macros, even though the latter appears less frequently in | |
4071 | # google code and would have been easier to implement. This is because | |
4072 | # the downside for getting the allowed checks wrong means some extra | |
4073 | # semicolons, while the downside for getting disallowed checks wrong | |
4074 | # would result in compile errors. | |
4075 | # | |
4076 | # In addition to macros, we also don't want to warn on | |
4077 | # - Compound literals | |
4078 | # - Lambdas | |
4079 | # - alignas specifier with anonymous structs | |
4080 | # - decltype | |
4081 | closing_brace_pos = match.group(1).rfind(')') | |
4082 | opening_parenthesis = ReverseCloseExpression( | |
4083 | clean_lines, linenum, closing_brace_pos) | |
4084 | if opening_parenthesis[2] > -1: | |
4085 | line_prefix = opening_parenthesis[0][0:opening_parenthesis[2]] | |
4086 | macro = Search(r'\b([A-Z_][A-Z0-9_]*)\s*$', line_prefix) | |
4087 | func = Match(r'^(.*\])\s*$', line_prefix) | |
4088 | if ((macro and | |
4089 | macro.group(1) not in ( | |
4090 | 'TEST', 'TEST_F', 'MATCHER', 'MATCHER_P', 'TYPED_TEST', | |
4091 | 'EXCLUSIVE_LOCKS_REQUIRED', 'SHARED_LOCKS_REQUIRED', | |
4092 | 'LOCKS_EXCLUDED', 'INTERFACE_DEF')) or | |
4093 | (func and not Search(r'\boperator\s*\[\s*\]', func.group(1))) or | |
4094 | Search(r'\b(?:struct|union)\s+alignas\s*$', line_prefix) or | |
4095 | Search(r'\bdecltype$', line_prefix) or | |
4096 | Search(r'\s+=\s*$', line_prefix)): | |
4097 | match = None | |
4098 | if (match and | |
4099 | opening_parenthesis[1] > 1 and | |
4100 | Search(r'\]\s*$', clean_lines.elided[opening_parenthesis[1] - 1])): | |
4101 | # Multi-line lambda-expression | |
4102 | match = None | |
4103 | ||
4104 | else: | |
4105 | # Try matching cases 2-3. | |
4106 | match = Match(r'^(.*(?:else|\)\s*const)\s*)\{', line) | |
4107 | if not match: | |
4108 | # Try matching cases 4-6. These are always matched on separate lines. | |
4109 | # | |
4110 | # Note that we can't simply concatenate the previous line to the | |
4111 | # current line and do a single match, otherwise we may output | |
4112 | # duplicate warnings for the blank line case: | |
4113 | # if (cond) { | |
4114 | # // blank line | |
4115 | # } | |
4116 | prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] | |
4117 | if prevline and Search(r'[;{}]\s*$', prevline): | |
4118 | match = Match(r'^(\s*)\{', line) | |
4119 | ||
4120 | # Check matching closing brace | |
4121 | if match: | |
4122 | (endline, endlinenum, endpos) = CloseExpression( | |
4123 | clean_lines, linenum, len(match.group(1))) | |
4124 | if endpos > -1 and Match(r'^\s*;', endline[endpos:]): | |
4125 | # Current {} pair is eligible for semicolon check, and we have found | |
4126 | # the redundant semicolon, output warning here. | |
4127 | # | |
4128 | # Note: because we are scanning forward for opening braces, and | |
4129 | # outputting warnings for the matching closing brace, if there are | |
4130 | # nested blocks with trailing semicolons, we will get the error | |
4131 | # messages in reversed order. | |
4132 | ||
4133 | # We need to check the line forward for NOLINT | |
4134 | raw_lines = clean_lines.raw_lines | |
4135 | ParseNolintSuppressions(filename, raw_lines[endlinenum-1], endlinenum-1, | |
4136 | error) | |
4137 | ParseNolintSuppressions(filename, raw_lines[endlinenum], endlinenum, | |
4138 | error) | |
4139 | ||
4140 | error(filename, endlinenum, 'readability/braces', 4, | |
4141 | "You don't need a ; after a }") | |
4142 | ||
4143 | ||
4144 | def CheckEmptyBlockBody(filename, clean_lines, linenum, error): | |
4145 | """Look for empty loop/conditional body with only a single semicolon. | |
4146 | ||
4147 | Args: | |
4148 | filename: The name of the current file. | |
4149 | clean_lines: A CleansedLines instance containing the file. | |
4150 | linenum: The number of the line to check. | |
4151 | error: The function to call with any errors found. | |
4152 | """ | |
4153 | ||
4154 | # Search for loop keywords at the beginning of the line. Because only | |
4155 | # whitespaces are allowed before the keywords, this will also ignore most | |
4156 | # do-while-loops, since those lines should start with closing brace. | |
4157 | # | |
4158 | # We also check "if" blocks here, since an empty conditional block | |
4159 | # is likely an error. | |
4160 | line = clean_lines.elided[linenum] | |
4161 | matched = Match(r'\s*(for|while|if)\s*\(', line) | |
4162 | if matched: | |
4163 | # Find the end of the conditional expression. | |
4164 | (end_line, end_linenum, end_pos) = CloseExpression( | |
4165 | clean_lines, linenum, line.find('(')) | |
4166 | ||
4167 | # Output warning if what follows the condition expression is a semicolon. | |
4168 | # No warning for all other cases, including whitespace or newline, since we | |
4169 | # have a separate check for semicolons preceded by whitespace. | |
4170 | if end_pos >= 0 and Match(r';', end_line[end_pos:]): | |
4171 | if matched.group(1) == 'if': | |
4172 | error(filename, end_linenum, 'whitespace/empty_conditional_body', 5, | |
4173 | 'Empty conditional bodies should use {}') | |
4174 | else: | |
4175 | error(filename, end_linenum, 'whitespace/empty_loop_body', 5, | |
4176 | 'Empty loop bodies should use {} or continue') | |
4177 | ||
4178 | # Check for if statements that have completely empty bodies (no comments) | |
4179 | # and no else clauses. | |
4180 | if end_pos >= 0 and matched.group(1) == 'if': | |
4181 | # Find the position of the opening { for the if statement. | |
4182 | # Return without logging an error if it has no brackets. | |
4183 | opening_linenum = end_linenum | |
4184 | opening_line_fragment = end_line[end_pos:] | |
4185 | # Loop until EOF or find anything that's not whitespace or opening {. | |
4186 | while not Search(r'^\s*\{', opening_line_fragment): | |
4187 | if Search(r'^(?!\s*$)', opening_line_fragment): | |
4188 | # Conditional has no brackets. | |
4189 | return | |
4190 | opening_linenum += 1 | |
4191 | if opening_linenum == len(clean_lines.elided): | |
4192 | # Couldn't find conditional's opening { or any code before EOF. | |
4193 | return | |
4194 | opening_line_fragment = clean_lines.elided[opening_linenum] | |
4195 | # Set opening_line (opening_line_fragment may not be entire opening line). | |
4196 | opening_line = clean_lines.elided[opening_linenum] | |
4197 | ||
4198 | # Find the position of the closing }. | |
4199 | opening_pos = opening_line_fragment.find('{') | |
4200 | if opening_linenum == end_linenum: | |
4201 | # We need to make opening_pos relative to the start of the entire line. | |
4202 | opening_pos += end_pos | |
4203 | (closing_line, closing_linenum, closing_pos) = CloseExpression( | |
4204 | clean_lines, opening_linenum, opening_pos) | |
4205 | if closing_pos < 0: | |
4206 | return | |
4207 | ||
4208 | # Now construct the body of the conditional. This consists of the portion | |
4209 | # of the opening line after the {, all lines until the closing line, | |
4210 | # and the portion of the closing line before the }. | |
4211 | if (clean_lines.raw_lines[opening_linenum] != | |
4212 | CleanseComments(clean_lines.raw_lines[opening_linenum])): | |
4213 | # Opening line ends with a comment, so conditional isn't empty. | |
4214 | return | |
4215 | if closing_linenum > opening_linenum: | |
4216 | # Opening line after the {. Ignore comments here since we checked above. | |
4217 | bodylist = list(opening_line[opening_pos+1:]) | |
4218 | # All lines until closing line, excluding closing line, with comments. | |
4219 | bodylist.extend(clean_lines.raw_lines[opening_linenum+1:closing_linenum]) | |
4220 | # Closing line before the }. Won't (and can't) have comments. | |
4221 | bodylist.append(clean_lines.elided[closing_linenum][:closing_pos-1]) | |
4222 | body = '\n'.join(bodylist) | |
4223 | else: | |
4224 | # If statement has brackets and fits on a single line. | |
4225 | body = opening_line[opening_pos+1:closing_pos-1] | |
4226 | ||
4227 | # Check if the body is empty | |
4228 | if not _EMPTY_CONDITIONAL_BODY_PATTERN.search(body): | |
4229 | return | |
4230 | # The body is empty. Now make sure there's not an else clause. | |
4231 | current_linenum = closing_linenum | |
4232 | current_line_fragment = closing_line[closing_pos:] | |
4233 | # Loop until EOF or find anything that's not whitespace or else clause. | |
4234 | while Search(r'^\s*$|^(?=\s*else)', current_line_fragment): | |
4235 | if Search(r'^(?=\s*else)', current_line_fragment): | |
4236 | # Found an else clause, so don't log an error. | |
4237 | return | |
4238 | current_linenum += 1 | |
4239 | if current_linenum == len(clean_lines.elided): | |
4240 | break | |
4241 | current_line_fragment = clean_lines.elided[current_linenum] | |
4242 | ||
4243 | # The body is empty and there's no else clause until EOF or other code. | |
4244 | error(filename, end_linenum, 'whitespace/empty_if_body', 4, | |
4245 | ('If statement had no body and no else clause')) | |
4246 | ||
4247 | ||
4248 | def FindCheckMacro(line): | |
4249 | """Find a replaceable CHECK-like macro. | |
4250 | ||
4251 | Args: | |
4252 | line: line to search on. | |
4253 | Returns: | |
4254 | (macro name, start position), or (None, -1) if no replaceable | |
4255 | macro is found. | |
4256 | """ | |
4257 | for macro in _CHECK_MACROS: | |
4258 | i = line.find(macro) | |
4259 | if i >= 0: | |
4260 | # Find opening parenthesis. Do a regular expression match here | |
4261 | # to make sure that we are matching the expected CHECK macro, as | |
4262 | # opposed to some other macro that happens to contain the CHECK | |
4263 | # substring. | |
4264 | matched = Match(r'^(.*\b' + macro + r'\s*)\(', line) | |
4265 | if not matched: | |
4266 | continue | |
4267 | return (macro, len(matched.group(1))) | |
4268 | return (None, -1) | |
4269 | ||
4270 | ||
4271 | def CheckCheck(filename, clean_lines, linenum, error): | |
4272 | """Checks the use of CHECK and EXPECT macros. | |
4273 | ||
4274 | Args: | |
4275 | filename: The name of the current file. | |
4276 | clean_lines: A CleansedLines instance containing the file. | |
4277 | linenum: The number of the line to check. | |
4278 | error: The function to call with any errors found. | |
4279 | """ | |
4280 | ||
4281 | # Decide the set of replacement macros that should be suggested | |
4282 | lines = clean_lines.elided | |
4283 | (check_macro, start_pos) = FindCheckMacro(lines[linenum]) | |
4284 | if not check_macro: | |
4285 | return | |
4286 | ||
4287 | # Find end of the boolean expression by matching parentheses | |
4288 | (last_line, end_line, end_pos) = CloseExpression( | |
4289 | clean_lines, linenum, start_pos) | |
4290 | if end_pos < 0: | |
4291 | return | |
4292 | ||
4293 | # If the check macro is followed by something other than a | |
4294 | # semicolon, assume users will log their own custom error messages | |
4295 | # and don't suggest any replacements. | |
4296 | if not Match(r'\s*;', last_line[end_pos:]): | |
4297 | return | |
4298 | ||
4299 | if linenum == end_line: | |
4300 | expression = lines[linenum][start_pos + 1:end_pos - 1] | |
4301 | else: | |
4302 | expression = lines[linenum][start_pos + 1:] | |
4303 | for i in xrange(linenum + 1, end_line): | |
4304 | expression += lines[i] | |
4305 | expression += last_line[0:end_pos - 1] | |
4306 | ||
4307 | # Parse expression so that we can take parentheses into account. | |
4308 | # This avoids false positives for inputs like "CHECK((a < 4) == b)", | |
4309 | # which is not replaceable by CHECK_LE. | |
4310 | lhs = '' | |
4311 | rhs = '' | |
4312 | operator = None | |
4313 | while expression: | |
4314 | matched = Match(r'^\s*(<<|<<=|>>|>>=|->\*|->|&&|\|\||' | |
4315 | r'==|!=|>=|>|<=|<|\()(.*)$', expression) | |
4316 | if matched: | |
4317 | token = matched.group(1) | |
4318 | if token == '(': | |
4319 | # Parenthesized operand | |
4320 | expression = matched.group(2) | |
4321 | (end, _) = FindEndOfExpressionInLine(expression, 0, ['(']) | |
4322 | if end < 0: | |
4323 | return # Unmatched parenthesis | |
4324 | lhs += '(' + expression[0:end] | |
4325 | expression = expression[end:] | |
4326 | elif token in ('&&', '||'): | |
4327 | # Logical and/or operators. This means the expression | |
4328 | # contains more than one term, for example: | |
4329 | # CHECK(42 < a && a < b); | |
4330 | # | |
4331 | # These are not replaceable with CHECK_LE, so bail out early. | |
4332 | return | |
4333 | elif token in ('<<', '<<=', '>>', '>>=', '->*', '->'): | |
4334 | # Non-relational operator | |
4335 | lhs += token | |
4336 | expression = matched.group(2) | |
4337 | else: | |
4338 | # Relational operator | |
4339 | operator = token | |
4340 | rhs = matched.group(2) | |
4341 | break | |
4342 | else: | |
4343 | # Unparenthesized operand. Instead of appending to lhs one character | |
4344 | # at a time, we do another regular expression match to consume several | |
4345 | # characters at once if possible. Trivial benchmark shows that this | |
4346 | # is more efficient when the operands are longer than a single | |
4347 | # character, which is generally the case. | |
4348 | matched = Match(r'^([^-=!<>()&|]+)(.*)$', expression) | |
4349 | if not matched: | |
4350 | matched = Match(r'^(\s*\S)(.*)$', expression) | |
4351 | if not matched: | |
4352 | break | |
4353 | lhs += matched.group(1) | |
4354 | expression = matched.group(2) | |
4355 | ||
4356 | # Only apply checks if we got all parts of the boolean expression | |
4357 | if not (lhs and operator and rhs): | |
4358 | return | |
4359 | ||
4360 | # Check that rhs do not contain logical operators. We already know | |
4361 | # that lhs is fine since the loop above parses out && and ||. | |
4362 | if rhs.find('&&') > -1 or rhs.find('||') > -1: | |
4363 | return | |
4364 | ||
4365 | # At least one of the operands must be a constant literal. This is | |
4366 | # to avoid suggesting replacements for unprintable things like | |
4367 | # CHECK(variable != iterator) | |
4368 | # | |
4369 | # The following pattern matches decimal, hex integers, strings, and | |
4370 | # characters (in that order). | |
4371 | lhs = lhs.strip() | |
4372 | rhs = rhs.strip() | |
4373 | match_constant = r'^([-+]?(\d+|0[xX][0-9a-fA-F]+)[lLuU]{0,3}|".*"|\'.*\')$' | |
4374 | if Match(match_constant, lhs) or Match(match_constant, rhs): | |
4375 | # Note: since we know both lhs and rhs, we can provide a more | |
4376 | # descriptive error message like: | |
4377 | # Consider using CHECK_EQ(x, 42) instead of CHECK(x == 42) | |
4378 | # Instead of: | |
4379 | # Consider using CHECK_EQ instead of CHECK(a == b) | |
4380 | # | |
4381 | # We are still keeping the less descriptive message because if lhs | |
4382 | # or rhs gets long, the error message might become unreadable. | |
4383 | error(filename, linenum, 'readability/check', 2, | |
4384 | 'Consider using %s instead of %s(a %s b)' % ( | |
4385 | _CHECK_REPLACEMENT[check_macro][operator], | |
4386 | check_macro, operator)) | |
4387 | ||
4388 | ||
4389 | def CheckAltTokens(filename, clean_lines, linenum, error): | |
4390 | """Check alternative keywords being used in boolean expressions. | |
4391 | ||
4392 | Args: | |
4393 | filename: The name of the current file. | |
4394 | clean_lines: A CleansedLines instance containing the file. | |
4395 | linenum: The number of the line to check. | |
4396 | error: The function to call with any errors found. | |
4397 | """ | |
4398 | line = clean_lines.elided[linenum] | |
4399 | ||
4400 | # Avoid preprocessor lines | |
4401 | if Match(r'^\s*#', line): | |
4402 | return | |
4403 | ||
4404 | # Last ditch effort to avoid multi-line comments. This will not help | |
4405 | # if the comment started before the current line or ended after the | |
4406 | # current line, but it catches most of the false positives. At least, | |
4407 | # it provides a way to workaround this warning for people who use | |
4408 | # multi-line comments in preprocessor macros. | |
4409 | # | |
4410 | # TODO(unknown): remove this once cpplint has better support for | |
4411 | # multi-line comments. | |
4412 | if line.find('/*') >= 0 or line.find('*/') >= 0: | |
4413 | return | |
4414 | ||
4415 | for match in _ALT_TOKEN_REPLACEMENT_PATTERN.finditer(line): | |
4416 | error(filename, linenum, 'readability/alt_tokens', 2, | |
4417 | 'Use operator %s instead of %s' % ( | |
4418 | _ALT_TOKEN_REPLACEMENT[match.group(1)], match.group(1))) | |
4419 | ||
4420 | ||
4421 | def GetLineWidth(line): | |
4422 | """Determines the width of the line in column positions. | |
4423 | ||
4424 | Args: | |
4425 | line: A string, which may be a Unicode string. | |
4426 | ||
4427 | Returns: | |
4428 | The width of the line in column positions, accounting for Unicode | |
4429 | combining characters and wide characters. | |
4430 | """ | |
4431 | if isinstance(line, unicode): | |
4432 | width = 0 | |
4433 | for uc in unicodedata.normalize('NFC', line): | |
4434 | if unicodedata.east_asian_width(uc) in ('W', 'F'): | |
4435 | width += 2 | |
4436 | elif not unicodedata.combining(uc): | |
4437 | width += 1 | |
4438 | return width | |
4439 | else: | |
4440 | return len(line) | |
4441 | ||
4442 | ||
4443 | def CheckStyle(filename, clean_lines, linenum, file_extension, nesting_state, | |
4444 | error): | |
4445 | """Checks rules from the 'C++ style rules' section of cppguide.html. | |
4446 | ||
4447 | Most of these rules are hard to test (naming, comment style), but we | |
4448 | do what we can. In particular we check for 2-space indents, line lengths, | |
4449 | tab usage, spaces inside code, etc. | |
4450 | ||
4451 | Args: | |
4452 | filename: The name of the current file. | |
4453 | clean_lines: A CleansedLines instance containing the file. | |
4454 | linenum: The number of the line to check. | |
4455 | file_extension: The extension (without the dot) of the filename. | |
4456 | nesting_state: A NestingState instance which maintains information about | |
4457 | the current stack of nested blocks being parsed. | |
4458 | error: The function to call with any errors found. | |
4459 | """ | |
4460 | ||
4461 | # Don't use "elided" lines here, otherwise we can't check commented lines. | |
4462 | # Don't want to use "raw" either, because we don't want to check inside C++11 | |
4463 | # raw strings, | |
4464 | raw_lines = clean_lines.lines_without_raw_strings | |
4465 | line = raw_lines[linenum] | |
4466 | prev = raw_lines[linenum - 1] if linenum > 0 else '' | |
4467 | ||
4468 | if line.find('\t') != -1: | |
4469 | error(filename, linenum, 'whitespace/tab', 1, | |
4470 | 'Tab found; better to use spaces') | |
4471 | ||
4472 | # One or three blank spaces at the beginning of the line is weird; it's | |
4473 | # hard to reconcile that with 2-space indents. | |
4474 | # NOTE: here are the conditions rob pike used for his tests. Mine aren't | |
4475 | # as sophisticated, but it may be worth becoming so: RLENGTH==initial_spaces | |
4476 | # if(RLENGTH > 20) complain = 0; | |
4477 | # if(match($0, " +(error|private|public|protected):")) complain = 0; | |
4478 | # if(match(prev, "&& *$")) complain = 0; | |
4479 | # if(match(prev, "\\|\\| *$")) complain = 0; | |
4480 | # if(match(prev, "[\",=><] *$")) complain = 0; | |
4481 | # if(match($0, " <<")) complain = 0; | |
4482 | # if(match(prev, " +for \\(")) complain = 0; | |
4483 | # if(prevodd && match(prevprev, " +for \\(")) complain = 0; | |
4484 | scope_or_label_pattern = r'\s*\w+\s*:\s*\\?$' | |
4485 | classinfo = nesting_state.InnermostClass() | |
4486 | initial_spaces = 0 | |
4487 | cleansed_line = clean_lines.elided[linenum] | |
4488 | while initial_spaces < len(line) and line[initial_spaces] == ' ': | |
4489 | initial_spaces += 1 | |
4490 | # There are certain situations we allow one space, notably for | |
4491 | # section labels, and also lines containing multi-line raw strings. | |
4492 | # We also don't check for lines that look like continuation lines | |
4493 | # (of lines ending in double quotes, commas, equals, or angle brackets) | |
4494 | # because the rules for how to indent those are non-trivial. | |
4495 | if (not Search(r'[",=><] *$', prev) and | |
4496 | (initial_spaces == 1 or initial_spaces == 3) and | |
4497 | not Match(scope_or_label_pattern, cleansed_line) and | |
4498 | not (clean_lines.raw_lines[linenum] != line and | |
4499 | Match(r'^\s*""', line))): | |
4500 | error(filename, linenum, 'whitespace/indent', 3, | |
4501 | 'Weird number of spaces at line-start. ' | |
4502 | 'Are you using a 2-space indent?') | |
4503 | ||
4504 | if line and line[-1].isspace(): | |
4505 | error(filename, linenum, 'whitespace/end_of_line', 4, | |
4506 | 'Line ends in whitespace. Consider deleting these extra spaces.') | |
4507 | ||
4508 | # Check if the line is a header guard. | |
4509 | is_header_guard = False | |
4510 | if file_extension in GetHeaderExtensions(): | |
4511 | cppvar = GetHeaderGuardCPPVariable(filename) | |
4512 | if (line.startswith('#ifndef %s' % cppvar) or | |
4513 | line.startswith('#define %s' % cppvar) or | |
4514 | line.startswith('#endif // %s' % cppvar)): | |
4515 | is_header_guard = True | |
4516 | # #include lines and header guards can be long, since there's no clean way to | |
4517 | # split them. | |
4518 | # | |
4519 | # URLs can be long too. It's possible to split these, but it makes them | |
4520 | # harder to cut&paste. | |
4521 | # | |
4522 | # The "$Id:...$" comment may also get very long without it being the | |
4523 | # developers fault. | |
4524 | # | |
4525 | # Doxygen documentation copying can get pretty long when using an overloaded | |
4526 | # function declaration | |
4527 | if (not line.startswith('#include') and not is_header_guard and | |
4528 | not Match(r'^\s*//.*http(s?)://\S*$', line) and | |
4529 | not Match(r'^\s*//\s*[^\s]*$', line) and | |
4530 | not Match(r'^// \$Id:.*#[0-9]+ \$$', line) and | |
4531 | not Match(r'^\s*/// [@\\](copydoc|copydetails|copybrief) .*$', line)): | |
4532 | line_width = GetLineWidth(line) | |
4533 | if line_width > _line_length: | |
4534 | error(filename, linenum, 'whitespace/line_length', 2, | |
4535 | 'Lines should be <= %i characters long' % _line_length) | |
4536 | ||
4537 | if (cleansed_line.count(';') > 1 and | |
4538 | # allow simple single line lambdas | |
4539 | not Match(r'^[^{};]*\[[^\[\]]*\][^{}]*\{[^{}\n\r]*\}', | |
4540 | line) and | |
4541 | # for loops are allowed two ;'s (and may run over two lines). | |
4542 | cleansed_line.find('for') == -1 and | |
4543 | (GetPreviousNonBlankLine(clean_lines, linenum)[0].find('for') == -1 or | |
4544 | GetPreviousNonBlankLine(clean_lines, linenum)[0].find(';') != -1) and | |
4545 | # It's ok to have many commands in a switch case that fits in 1 line | |
4546 | not ((cleansed_line.find('case ') != -1 or | |
4547 | cleansed_line.find('default:') != -1) and | |
4548 | cleansed_line.find('break;') != -1)): | |
4549 | error(filename, linenum, 'whitespace/newline', 0, | |
4550 | 'More than one command on the same line') | |
4551 | ||
4552 | # Some more style checks | |
4553 | CheckBraces(filename, clean_lines, linenum, error) | |
4554 | CheckTrailingSemicolon(filename, clean_lines, linenum, error) | |
4555 | CheckEmptyBlockBody(filename, clean_lines, linenum, error) | |
4556 | CheckAccess(filename, clean_lines, linenum, nesting_state, error) | |
4557 | CheckSpacing(filename, clean_lines, linenum, nesting_state, error) | |
4558 | CheckOperatorSpacing(filename, clean_lines, linenum, error) | |
4559 | CheckParenthesisSpacing(filename, clean_lines, linenum, error) | |
4560 | CheckCommaSpacing(filename, clean_lines, linenum, error) | |
4561 | CheckBracesSpacing(filename, clean_lines, linenum, nesting_state, error) | |
4562 | CheckSpacingForFunctionCall(filename, clean_lines, linenum, error) | |
4563 | CheckCheck(filename, clean_lines, linenum, error) | |
4564 | CheckAltTokens(filename, clean_lines, linenum, error) | |
4565 | classinfo = nesting_state.InnermostClass() | |
4566 | if classinfo: | |
4567 | CheckSectionSpacing(filename, clean_lines, classinfo, linenum, error) | |
4568 | ||
4569 | ||
4570 | _RE_PATTERN_INCLUDE = re.compile(r'^\s*#\s*include\s*([<"])([^>"]*)[>"].*$') | |
4571 | # Matches the first component of a filename delimited by -s and _s. That is: | |
4572 | # _RE_FIRST_COMPONENT.match('foo').group(0) == 'foo' | |
4573 | # _RE_FIRST_COMPONENT.match('foo.cc').group(0) == 'foo' | |
4574 | # _RE_FIRST_COMPONENT.match('foo-bar_baz.cc').group(0) == 'foo' | |
4575 | # _RE_FIRST_COMPONENT.match('foo_bar-baz.cc').group(0) == 'foo' | |
4576 | _RE_FIRST_COMPONENT = re.compile(r'^[^-_.]+') | |
4577 | ||
4578 | ||
4579 | def _DropCommonSuffixes(filename): | |
4580 | """Drops common suffixes like _test.cc or -inl.h from filename. | |
4581 | ||
4582 | For example: | |
4583 | >>> _DropCommonSuffixes('foo/foo-inl.h') | |
4584 | 'foo/foo' | |
4585 | >>> _DropCommonSuffixes('foo/bar/foo.cc') | |
4586 | 'foo/bar/foo' | |
4587 | >>> _DropCommonSuffixes('foo/foo_internal.h') | |
4588 | 'foo/foo' | |
4589 | >>> _DropCommonSuffixes('foo/foo_unusualinternal.h') | |
4590 | 'foo/foo_unusualinternal' | |
4591 | ||
4592 | Args: | |
4593 | filename: The input filename. | |
4594 | ||
4595 | Returns: | |
4596 | The filename with the common suffix removed. | |
4597 | """ | |
4598 | for suffix in itertools.chain( | |
4599 | ('%s.%s' % (test_suffix.lstrip('_'), ext) | |
4600 | for test_suffix, ext in itertools.product(_test_suffixes, GetNonHeaderExtensions())), | |
4601 | ('%s.%s' % (suffix, ext) | |
4602 | for suffix, ext in itertools.product(['inl', 'imp', 'internal'], GetHeaderExtensions()))): | |
4603 | if (filename.endswith(suffix) and len(filename) > len(suffix) and | |
4604 | filename[-len(suffix) - 1] in ('-', '_')): | |
4605 | return filename[:-len(suffix) - 1] | |
4606 | return os.path.splitext(filename)[0] | |
4607 | ||
4608 | ||
4609 | def _ClassifyInclude(fileinfo, include, is_system): | |
4610 | """Figures out what kind of header 'include' is. | |
4611 | ||
4612 | Args: | |
4613 | fileinfo: The current file cpplint is running over. A FileInfo instance. | |
4614 | include: The path to a #included file. | |
4615 | is_system: True if the #include used <> rather than "". | |
4616 | ||
4617 | Returns: | |
4618 | One of the _XXX_HEADER constants. | |
4619 | ||
4620 | For example: | |
4621 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'stdio.h', True) | |
4622 | _C_SYS_HEADER | |
4623 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'string', True) | |
4624 | _CPP_SYS_HEADER | |
4625 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/foo.h', False) | |
4626 | _LIKELY_MY_HEADER | |
4627 | >>> _ClassifyInclude(FileInfo('foo/foo_unknown_extension.cc'), | |
4628 | ... 'bar/foo_other_ext.h', False) | |
4629 | _POSSIBLE_MY_HEADER | |
4630 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/bar.h', False) | |
4631 | _OTHER_HEADER | |
4632 | """ | |
4633 | # This is a list of all standard c++ header files, except | |
4634 | # those already checked for above. | |
4635 | is_cpp_h = include in _CPP_HEADERS | |
4636 | ||
4637 | # Headers with C++ extensions shouldn't be considered C system headers | |
4638 | if is_system and os.path.splitext(include)[1] in ['.hpp', '.hxx', '.h++']: | |
4639 | is_system = False | |
4640 | ||
4641 | if is_system: | |
4642 | if is_cpp_h: | |
4643 | return _CPP_SYS_HEADER | |
4644 | else: | |
4645 | return _C_SYS_HEADER | |
4646 | ||
4647 | # If the target file and the include we're checking share a | |
4648 | # basename when we drop common extensions, and the include | |
4649 | # lives in . , then it's likely to be owned by the target file. | |
4650 | target_dir, target_base = ( | |
4651 | os.path.split(_DropCommonSuffixes(fileinfo.RepositoryName()))) | |
4652 | include_dir, include_base = os.path.split(_DropCommonSuffixes(include)) | |
4653 | target_dir_pub = os.path.normpath(target_dir + '/../public') | |
4654 | target_dir_pub = target_dir_pub.replace('\\', '/') | |
4655 | if target_base == include_base and ( | |
4656 | include_dir == target_dir or | |
4657 | include_dir == target_dir_pub): | |
4658 | return _LIKELY_MY_HEADER | |
4659 | ||
4660 | # If the target and include share some initial basename | |
4661 | # component, it's possible the target is implementing the | |
4662 | # include, so it's allowed to be first, but we'll never | |
4663 | # complain if it's not there. | |
4664 | target_first_component = _RE_FIRST_COMPONENT.match(target_base) | |
4665 | include_first_component = _RE_FIRST_COMPONENT.match(include_base) | |
4666 | if (target_first_component and include_first_component and | |
4667 | target_first_component.group(0) == | |
4668 | include_first_component.group(0)): | |
4669 | return _POSSIBLE_MY_HEADER | |
4670 | ||
4671 | return _OTHER_HEADER | |
4672 | ||
4673 | ||
4674 | ||
4675 | def CheckIncludeLine(filename, clean_lines, linenum, include_state, error): | |
4676 | """Check rules that are applicable to #include lines. | |
4677 | ||
4678 | Strings on #include lines are NOT removed from elided line, to make | |
4679 | certain tasks easier. However, to prevent false positives, checks | |
4680 | applicable to #include lines in CheckLanguage must be put here. | |
4681 | ||
4682 | Args: | |
4683 | filename: The name of the current file. | |
4684 | clean_lines: A CleansedLines instance containing the file. | |
4685 | linenum: The number of the line to check. | |
4686 | include_state: An _IncludeState instance in which the headers are inserted. | |
4687 | error: The function to call with any errors found. | |
4688 | """ | |
4689 | fileinfo = FileInfo(filename) | |
4690 | line = clean_lines.lines[linenum] | |
4691 | ||
4692 | # "include" should use the new style "foo/bar.h" instead of just "bar.h" | |
4693 | # Only do this check if the included header follows google naming | |
4694 | # conventions. If not, assume that it's a 3rd party API that | |
4695 | # requires special include conventions. | |
4696 | # | |
4697 | # We also make an exception for Lua headers, which follow google | |
4698 | # naming convention but not the include convention. | |
4699 | match = Match(r'#include\s*"([^/]+\.h)"', line) | |
4700 | if match and not _THIRD_PARTY_HEADERS_PATTERN.match(match.group(1)): | |
4701 | error(filename, linenum, 'build/include_subdir', 4, | |
4702 | 'Include the directory when naming .h files') | |
4703 | ||
4704 | # we shouldn't include a file more than once. actually, there are a | |
4705 | # handful of instances where doing so is okay, but in general it's | |
4706 | # not. | |
4707 | match = _RE_PATTERN_INCLUDE.search(line) | |
4708 | if match: | |
4709 | include = match.group(2) | |
4710 | is_system = (match.group(1) == '<') | |
4711 | duplicate_line = include_state.FindHeader(include) | |
4712 | if duplicate_line >= 0: | |
4713 | error(filename, linenum, 'build/include', 4, | |
4714 | '"%s" already included at %s:%s' % | |
4715 | (include, filename, duplicate_line)) | |
4716 | return | |
4717 | ||
4718 | for extension in GetNonHeaderExtensions(): | |
4719 | if (include.endswith('.' + extension) and | |
4720 | os.path.dirname(fileinfo.RepositoryName()) != os.path.dirname(include)): | |
4721 | error(filename, linenum, 'build/include', 4, | |
4722 | 'Do not include .' + extension + ' files from other packages') | |
4723 | return | |
4724 | ||
4725 | if not _THIRD_PARTY_HEADERS_PATTERN.match(include): | |
4726 | include_state.include_list[-1].append((include, linenum)) | |
4727 | ||
4728 | # We want to ensure that headers appear in the right order: | |
4729 | # 1) for foo.cc, foo.h (preferred location) | |
4730 | # 2) c system files | |
4731 | # 3) cpp system files | |
4732 | # 4) for foo.cc, foo.h (deprecated location) | |
4733 | # 5) other google headers | |
4734 | # | |
4735 | # We classify each include statement as one of those 5 types | |
4736 | # using a number of techniques. The include_state object keeps | |
4737 | # track of the highest type seen, and complains if we see a | |
4738 | # lower type after that. | |
4739 | error_message = include_state.CheckNextIncludeOrder( | |
4740 | _ClassifyInclude(fileinfo, include, is_system)) | |
4741 | if error_message: | |
4742 | error(filename, linenum, 'build/include_order', 4, | |
4743 | '%s. Should be: %s.h, c system, c++ system, other.' % | |
4744 | (error_message, fileinfo.BaseName())) | |
4745 | canonical_include = include_state.CanonicalizeAlphabeticalOrder(include) | |
4746 | if not include_state.IsInAlphabeticalOrder( | |
4747 | clean_lines, linenum, canonical_include): | |
4748 | error(filename, linenum, 'build/include_alpha', 4, | |
4749 | 'Include "%s" not in alphabetical order' % include) | |
4750 | include_state.SetLastHeader(canonical_include) | |
4751 | ||
4752 | ||
4753 | ||
4754 | def _GetTextInside(text, start_pattern): | |
4755 | r"""Retrieves all the text between matching open and close parentheses. | |
4756 | ||
4757 | Given a string of lines and a regular expression string, retrieve all the text | |
4758 | following the expression and between opening punctuation symbols like | |
4759 | (, [, or {, and the matching close-punctuation symbol. This properly nested | |
4760 | occurrences of the punctuations, so for the text like | |
4761 | printf(a(), b(c())); | |
4762 | a call to _GetTextInside(text, r'printf\(') will return 'a(), b(c())'. | |
4763 | start_pattern must match string having an open punctuation symbol at the end. | |
4764 | ||
4765 | Args: | |
4766 | text: The lines to extract text. Its comments and strings must be elided. | |
4767 | It can be single line and can span multiple lines. | |
4768 | start_pattern: The regexp string indicating where to start extracting | |
4769 | the text. | |
4770 | Returns: | |
4771 | The extracted text. | |
4772 | None if either the opening string or ending punctuation could not be found. | |
4773 | """ | |
4774 | # TODO(unknown): Audit cpplint.py to see what places could be profitably | |
4775 | # rewritten to use _GetTextInside (and use inferior regexp matching today). | |
4776 | ||
4777 | # Give opening punctuations to get the matching close-punctuations. | |
4778 | matching_punctuation = {'(': ')', '{': '}', '[': ']'} | |
4779 | closing_punctuation = set(itervalues(matching_punctuation)) | |
4780 | ||
4781 | # Find the position to start extracting text. | |
4782 | match = re.search(start_pattern, text, re.M) | |
4783 | if not match: # start_pattern not found in text. | |
4784 | return None | |
4785 | start_position = match.end(0) | |
4786 | ||
4787 | assert start_position > 0, ( | |
4788 | 'start_pattern must ends with an opening punctuation.') | |
4789 | assert text[start_position - 1] in matching_punctuation, ( | |
4790 | 'start_pattern must ends with an opening punctuation.') | |
4791 | # Stack of closing punctuations we expect to have in text after position. | |
4792 | punctuation_stack = [matching_punctuation[text[start_position - 1]]] | |
4793 | position = start_position | |
4794 | while punctuation_stack and position < len(text): | |
4795 | if text[position] == punctuation_stack[-1]: | |
4796 | punctuation_stack.pop() | |
4797 | elif text[position] in closing_punctuation: | |
4798 | # A closing punctuation without matching opening punctuations. | |
4799 | return None | |
4800 | elif text[position] in matching_punctuation: | |
4801 | punctuation_stack.append(matching_punctuation[text[position]]) | |
4802 | position += 1 | |
4803 | if punctuation_stack: | |
4804 | # Opening punctuations left without matching close-punctuations. | |
4805 | return None | |
4806 | # punctuations match. | |
4807 | return text[start_position:position - 1] | |
4808 | ||
4809 | ||
4810 | # Patterns for matching call-by-reference parameters. | |
4811 | # | |
4812 | # Supports nested templates up to 2 levels deep using this messy pattern: | |
4813 | # < (?: < (?: < [^<>]* | |
4814 | # > | |
4815 | # | [^<>] )* | |
4816 | # > | |
4817 | # | [^<>] )* | |
4818 | # > | |
4819 | _RE_PATTERN_IDENT = r'[_a-zA-Z]\w*' # =~ [[:alpha:]][[:alnum:]]* | |
4820 | _RE_PATTERN_TYPE = ( | |
4821 | r'(?:const\s+)?(?:typename\s+|class\s+|struct\s+|union\s+|enum\s+)?' | |
4822 | r'(?:\w|' | |
4823 | r'\s*<(?:<(?:<[^<>]*>|[^<>])*>|[^<>])*>|' | |
4824 | r'::)+') | |
4825 | # A call-by-reference parameter ends with '& identifier'. | |
4826 | _RE_PATTERN_REF_PARAM = re.compile( | |
4827 | r'(' + _RE_PATTERN_TYPE + r'(?:\s*(?:\bconst\b|[*]))*\s*' | |
4828 | r'&\s*' + _RE_PATTERN_IDENT + r')\s*(?:=[^,()]+)?[,)]') | |
4829 | # A call-by-const-reference parameter either ends with 'const& identifier' | |
4830 | # or looks like 'const type& identifier' when 'type' is atomic. | |
4831 | _RE_PATTERN_CONST_REF_PARAM = ( | |
4832 | r'(?:.*\s*\bconst\s*&\s*' + _RE_PATTERN_IDENT + | |
4833 | r'|const\s+' + _RE_PATTERN_TYPE + r'\s*&\s*' + _RE_PATTERN_IDENT + r')') | |
4834 | # Stream types. | |
4835 | _RE_PATTERN_REF_STREAM_PARAM = ( | |
4836 | r'(?:.*stream\s*&\s*' + _RE_PATTERN_IDENT + r')') | |
4837 | ||
4838 | ||
4839 | def CheckLanguage(filename, clean_lines, linenum, file_extension, | |
4840 | include_state, nesting_state, error): | |
4841 | """Checks rules from the 'C++ language rules' section of cppguide.html. | |
4842 | ||
4843 | Some of these rules are hard to test (function overloading, using | |
4844 | uint32 inappropriately), but we do the best we can. | |
4845 | ||
4846 | Args: | |
4847 | filename: The name of the current file. | |
4848 | clean_lines: A CleansedLines instance containing the file. | |
4849 | linenum: The number of the line to check. | |
4850 | file_extension: The extension (without the dot) of the filename. | |
4851 | include_state: An _IncludeState instance in which the headers are inserted. | |
4852 | nesting_state: A NestingState instance which maintains information about | |
4853 | the current stack of nested blocks being parsed. | |
4854 | error: The function to call with any errors found. | |
4855 | """ | |
4856 | # If the line is empty or consists of entirely a comment, no need to | |
4857 | # check it. | |
4858 | line = clean_lines.elided[linenum] | |
4859 | if not line: | |
4860 | return | |
4861 | ||
4862 | match = _RE_PATTERN_INCLUDE.search(line) | |
4863 | if match: | |
4864 | CheckIncludeLine(filename, clean_lines, linenum, include_state, error) | |
4865 | return | |
4866 | ||
4867 | # Reset include state across preprocessor directives. This is meant | |
4868 | # to silence warnings for conditional includes. | |
4869 | match = Match(r'^\s*#\s*(if|ifdef|ifndef|elif|else|endif)\b', line) | |
4870 | if match: | |
4871 | include_state.ResetSection(match.group(1)) | |
4872 | ||
4873 | ||
4874 | # Perform other checks now that we are sure that this is not an include line | |
4875 | CheckCasts(filename, clean_lines, linenum, error) | |
4876 | CheckGlobalStatic(filename, clean_lines, linenum, error) | |
4877 | CheckPrintf(filename, clean_lines, linenum, error) | |
4878 | ||
4879 | if file_extension in GetHeaderExtensions(): | |
4880 | # TODO(unknown): check that 1-arg constructors are explicit. | |
4881 | # How to tell it's a constructor? | |
4882 | # (handled in CheckForNonStandardConstructs for now) | |
4883 | # TODO(unknown): check that classes declare or disable copy/assign | |
4884 | # (level 1 error) | |
4885 | pass | |
4886 | ||
4887 | # Check if people are using the verboten C basic types. The only exception | |
4888 | # we regularly allow is "unsigned short port" for port. | |
4889 | if Search(r'\bshort port\b', line): | |
4890 | if not Search(r'\bunsigned short port\b', line): | |
4891 | error(filename, linenum, 'runtime/int', 4, | |
4892 | 'Use "unsigned short" for ports, not "short"') | |
4893 | else: | |
4894 | match = Search(r'\b(short|long(?! +double)|long long)\b', line) | |
4895 | if match: | |
4896 | error(filename, linenum, 'runtime/int', 4, | |
4897 | 'Use int16/int64/etc, rather than the C type %s' % match.group(1)) | |
4898 | ||
4899 | # Check if some verboten operator overloading is going on | |
4900 | # TODO(unknown): catch out-of-line unary operator&: | |
4901 | # class X {}; | |
4902 | # int operator&(const X& x) { return 42; } // unary operator& | |
4903 | # The trick is it's hard to tell apart from binary operator&: | |
4904 | # class Y { int operator&(const Y& x) { return 23; } }; // binary operator& | |
4905 | if Search(r'\boperator\s*&\s*\(\s*\)', line): | |
4906 | error(filename, linenum, 'runtime/operator', 4, | |
4907 | 'Unary operator& is dangerous. Do not use it.') | |
4908 | ||
4909 | # Check for suspicious usage of "if" like | |
4910 | # } if (a == b) { | |
4911 | if Search(r'\}\s*if\s*\(', line): | |
4912 | error(filename, linenum, 'readability/braces', 4, | |
4913 | 'Did you mean "else if"? If not, start a new line for "if".') | |
4914 | ||
4915 | # Check for potential format string bugs like printf(foo). | |
4916 | # We constrain the pattern not to pick things like DocidForPrintf(foo). | |
4917 | # Not perfect but it can catch printf(foo.c_str()) and printf(foo->c_str()) | |
4918 | # TODO(unknown): Catch the following case. Need to change the calling | |
4919 | # convention of the whole function to process multiple line to handle it. | |
4920 | # printf( | |
4921 | # boy_this_is_a_really_long_variable_that_cannot_fit_on_the_prev_line); | |
4922 | printf_args = _GetTextInside(line, r'(?i)\b(string)?printf\s*\(') | |
4923 | if printf_args: | |
4924 | match = Match(r'([\w.\->()]+)$', printf_args) | |
4925 | if match and match.group(1) != '__VA_ARGS__': | |
4926 | function_name = re.search(r'\b((?:string)?printf)\s*\(', | |
4927 | line, re.I).group(1) | |
4928 | error(filename, linenum, 'runtime/printf', 4, | |
4929 | 'Potential format string bug. Do %s("%%s", %s) instead.' | |
4930 | % (function_name, match.group(1))) | |
4931 | ||
4932 | # Check for potential memset bugs like memset(buf, sizeof(buf), 0). | |
4933 | match = Search(r'memset\s*\(([^,]*),\s*([^,]*),\s*0\s*\)', line) | |
4934 | if match and not Match(r"^''|-?[0-9]+|0x[0-9A-Fa-f]$", match.group(2)): | |
4935 | error(filename, linenum, 'runtime/memset', 4, | |
4936 | 'Did you mean "memset(%s, 0, %s)"?' | |
4937 | % (match.group(1), match.group(2))) | |
4938 | ||
4939 | if Search(r'\busing namespace\b', line): | |
4940 | if Search(r'\bliterals\b', line): | |
4941 | error(filename, linenum, 'build/namespaces_literals', 5, | |
4942 | 'Do not use namespace using-directives. ' | |
4943 | 'Use using-declarations instead.') | |
4944 | else: | |
4945 | error(filename, linenum, 'build/namespaces', 5, | |
4946 | 'Do not use namespace using-directives. ' | |
4947 | 'Use using-declarations instead.') | |
4948 | ||
4949 | # Detect variable-length arrays. | |
4950 | match = Match(r'\s*(.+::)?(\w+) [a-z]\w*\[(.+)];', line) | |
4951 | if (match and match.group(2) != 'return' and match.group(2) != 'delete' and | |
4952 | match.group(3).find(']') == -1): | |
4953 | # Split the size using space and arithmetic operators as delimiters. | |
4954 | # If any of the resulting tokens are not compile time constants then | |
4955 | # report the error. | |
4956 | tokens = re.split(r'\s|\+|\-|\*|\/|<<|>>]', match.group(3)) | |
4957 | is_const = True | |
4958 | skip_next = False | |
4959 | for tok in tokens: | |
4960 | if skip_next: | |
4961 | skip_next = False | |
4962 | continue | |
4963 | ||
4964 | if Search(r'sizeof\(.+\)', tok): continue | |
4965 | if Search(r'arraysize\(\w+\)', tok): continue | |
4966 | ||
4967 | tok = tok.lstrip('(') | |
4968 | tok = tok.rstrip(')') | |
4969 | if not tok: continue | |
4970 | if Match(r'\d+', tok): continue | |
4971 | if Match(r'0[xX][0-9a-fA-F]+', tok): continue | |
4972 | if Match(r'k[A-Z0-9]\w*', tok): continue | |
4973 | if Match(r'(.+::)?k[A-Z0-9]\w*', tok): continue | |
4974 | if Match(r'(.+::)?[A-Z][A-Z0-9_]*', tok): continue | |
4975 | # A catch all for tricky sizeof cases, including 'sizeof expression', | |
4976 | # 'sizeof(*type)', 'sizeof(const type)', 'sizeof(struct StructName)' | |
4977 | # requires skipping the next token because we split on ' ' and '*'. | |
4978 | if tok.startswith('sizeof'): | |
4979 | skip_next = True | |
4980 | continue | |
4981 | is_const = False | |
4982 | break | |
4983 | if not is_const: | |
4984 | error(filename, linenum, 'runtime/arrays', 1, | |
4985 | 'Do not use variable-length arrays. Use an appropriately named ' | |
4986 | "('k' followed by CamelCase) compile-time constant for the size.") | |
4987 | ||
4988 | # Check for use of unnamed namespaces in header files. Registration | |
4989 | # macros are typically OK, so we allow use of "namespace {" on lines | |
4990 | # that end with backslashes. | |
4991 | if (file_extension in GetHeaderExtensions() | |
4992 | and Search(r'\bnamespace\s*{', line) | |
4993 | and line[-1] != '\\'): | |
4994 | error(filename, linenum, 'build/namespaces', 4, | |
4995 | 'Do not use unnamed namespaces in header files. See ' | |
4996 | 'https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Namespaces' | |
4997 | ' for more information.') | |
4998 | ||
4999 | ||
5000 | def CheckGlobalStatic(filename, clean_lines, linenum, error): | |
5001 | """Check for unsafe global or static objects. | |
5002 | ||
5003 | Args: | |
5004 | filename: The name of the current file. | |
5005 | clean_lines: A CleansedLines instance containing the file. | |
5006 | linenum: The number of the line to check. | |
5007 | error: The function to call with any errors found. | |
5008 | """ | |
5009 | line = clean_lines.elided[linenum] | |
5010 | ||
5011 | # Match two lines at a time to support multiline declarations | |
5012 | if linenum + 1 < clean_lines.NumLines() and not Search(r'[;({]', line): | |
5013 | line += clean_lines.elided[linenum + 1].strip() | |
5014 | ||
5015 | # Check for people declaring static/global STL strings at the top level. | |
5016 | # This is dangerous because the C++ language does not guarantee that | |
5017 | # globals with constructors are initialized before the first access, and | |
5018 | # also because globals can be destroyed when some threads are still running. | |
5019 | # TODO(unknown): Generalize this to also find static unique_ptr instances. | |
5020 | # TODO(unknown): File bugs for clang-tidy to find these. | |
5021 | match = Match( | |
5022 | r'((?:|static +)(?:|const +))(?::*std::)?string( +const)? +' | |
5023 | r'([a-zA-Z0-9_:]+)\b(.*)', | |
5024 | line) | |
5025 | ||
5026 | # Remove false positives: | |
5027 | # - String pointers (as opposed to values). | |
5028 | # string *pointer | |
5029 | # const string *pointer | |
5030 | # string const *pointer | |
5031 | # string *const pointer | |
5032 | # | |
5033 | # - Functions and template specializations. | |
5034 | # string Function<Type>(... | |
5035 | # string Class<Type>::Method(... | |
5036 | # | |
5037 | # - Operators. These are matched separately because operator names | |
5038 | # cross non-word boundaries, and trying to match both operators | |
5039 | # and functions at the same time would decrease accuracy of | |
5040 | # matching identifiers. | |
5041 | # string Class::operator*() | |
5042 | if (match and | |
5043 | not Search(r'\bstring\b(\s+const)?\s*[\*\&]\s*(const\s+)?\w', line) and | |
5044 | not Search(r'\boperator\W', line) and | |
5045 | not Match(r'\s*(<.*>)?(::[a-zA-Z0-9_]+)*\s*\(([^"]|$)', match.group(4))): | |
5046 | if Search(r'\bconst\b', line): | |
5047 | error(filename, linenum, 'runtime/string', 4, | |
5048 | 'For a static/global string constant, use a C style string ' | |
5049 | 'instead: "%schar%s %s[]".' % | |
5050 | (match.group(1), match.group(2) or '', match.group(3))) | |
5051 | else: | |
5052 | error(filename, linenum, 'runtime/string', 4, | |
5053 | 'Static/global string variables are not permitted.') | |
5054 | ||
5055 | if (Search(r'\b([A-Za-z0-9_]*_)\(\1\)', line) or | |
5056 | Search(r'\b([A-Za-z0-9_]*_)\(CHECK_NOTNULL\(\1\)\)', line)): | |
5057 | error(filename, linenum, 'runtime/init', 4, | |
5058 | 'You seem to be initializing a member variable with itself.') | |
5059 | ||
5060 | ||
5061 | def CheckPrintf(filename, clean_lines, linenum, error): | |
5062 | """Check for printf related issues. | |
5063 | ||
5064 | Args: | |
5065 | filename: The name of the current file. | |
5066 | clean_lines: A CleansedLines instance containing the file. | |
5067 | linenum: The number of the line to check. | |
5068 | error: The function to call with any errors found. | |
5069 | """ | |
5070 | line = clean_lines.elided[linenum] | |
5071 | ||
5072 | # When snprintf is used, the second argument shouldn't be a literal. | |
5073 | match = Search(r'snprintf\s*\(([^,]*),\s*([0-9]*)\s*,', line) | |
5074 | if match and match.group(2) != '0': | |
5075 | # If 2nd arg is zero, snprintf is used to calculate size. | |
5076 | error(filename, linenum, 'runtime/printf', 3, | |
5077 | 'If you can, use sizeof(%s) instead of %s as the 2nd arg ' | |
5078 | 'to snprintf.' % (match.group(1), match.group(2))) | |
5079 | ||
5080 | # Check if some verboten C functions are being used. | |
5081 | if Search(r'\bsprintf\s*\(', line): | |
5082 | error(filename, linenum, 'runtime/printf', 5, | |
5083 | 'Never use sprintf. Use snprintf instead.') | |
5084 | match = Search(r'\b(strcpy|strcat)\s*\(', line) | |
5085 | if match: | |
5086 | error(filename, linenum, 'runtime/printf', 4, | |
5087 | 'Almost always, snprintf is better than %s' % match.group(1)) | |
5088 | ||
5089 | ||
5090 | def IsDerivedFunction(clean_lines, linenum): | |
5091 | """Check if current line contains an inherited function. | |
5092 | ||
5093 | Args: | |
5094 | clean_lines: A CleansedLines instance containing the file. | |
5095 | linenum: The number of the line to check. | |
5096 | Returns: | |
5097 | True if current line contains a function with "override" | |
5098 | virt-specifier. | |
5099 | """ | |
5100 | # Scan back a few lines for start of current function | |
5101 | for i in xrange(linenum, max(-1, linenum - 10), -1): | |
5102 | match = Match(r'^([^()]*\w+)\(', clean_lines.elided[i]) | |
5103 | if match: | |
5104 | # Look for "override" after the matching closing parenthesis | |
5105 | line, _, closing_paren = CloseExpression( | |
5106 | clean_lines, i, len(match.group(1))) | |
5107 | return (closing_paren >= 0 and | |
5108 | Search(r'\boverride\b', line[closing_paren:])) | |
5109 | return False | |
5110 | ||
5111 | ||
5112 | def IsOutOfLineMethodDefinition(clean_lines, linenum): | |
5113 | """Check if current line contains an out-of-line method definition. | |
5114 | ||
5115 | Args: | |
5116 | clean_lines: A CleansedLines instance containing the file. | |
5117 | linenum: The number of the line to check. | |
5118 | Returns: | |
5119 | True if current line contains an out-of-line method definition. | |
5120 | """ | |
5121 | # Scan back a few lines for start of current function | |
5122 | for i in xrange(linenum, max(-1, linenum - 10), -1): | |
5123 | if Match(r'^([^()]*\w+)\(', clean_lines.elided[i]): | |
5124 | return Match(r'^[^()]*\w+::\w+\(', clean_lines.elided[i]) is not None | |
5125 | return False | |
5126 | ||
5127 | ||
5128 | def IsInitializerList(clean_lines, linenum): | |
5129 | """Check if current line is inside constructor initializer list. | |
5130 | ||
5131 | Args: | |
5132 | clean_lines: A CleansedLines instance containing the file. | |
5133 | linenum: The number of the line to check. | |
5134 | Returns: | |
5135 | True if current line appears to be inside constructor initializer | |
5136 | list, False otherwise. | |
5137 | """ | |
5138 | for i in xrange(linenum, 1, -1): | |
5139 | line = clean_lines.elided[i] | |
5140 | if i == linenum: | |
5141 | remove_function_body = Match(r'^(.*)\{\s*$', line) | |
5142 | if remove_function_body: | |
5143 | line = remove_function_body.group(1) | |
5144 | ||
5145 | if Search(r'\s:\s*\w+[({]', line): | |
5146 | # A lone colon tend to indicate the start of a constructor | |
5147 | # initializer list. It could also be a ternary operator, which | |
5148 | # also tend to appear in constructor initializer lists as | |
5149 | # opposed to parameter lists. | |
5150 | return True | |
5151 | if Search(r'\}\s*,\s*$', line): | |
5152 | # A closing brace followed by a comma is probably the end of a | |
5153 | # brace-initialized member in constructor initializer list. | |
5154 | return True | |
5155 | if Search(r'[{};]\s*$', line): | |
5156 | # Found one of the following: | |
5157 | # - A closing brace or semicolon, probably the end of the previous | |
5158 | # function. | |
5159 | # - An opening brace, probably the start of current class or namespace. | |
5160 | # | |
5161 | # Current line is probably not inside an initializer list since | |
5162 | # we saw one of those things without seeing the starting colon. | |
5163 | return False | |
5164 | ||
5165 | # Got to the beginning of the file without seeing the start of | |
5166 | # constructor initializer list. | |
5167 | return False | |
5168 | ||
5169 | ||
5170 | def CheckForNonConstReference(filename, clean_lines, linenum, | |
5171 | nesting_state, error): | |
5172 | """Check for non-const references. | |
5173 | ||
5174 | Separate from CheckLanguage since it scans backwards from current | |
5175 | line, instead of scanning forward. | |
5176 | ||
5177 | Args: | |
5178 | filename: The name of the current file. | |
5179 | clean_lines: A CleansedLines instance containing the file. | |
5180 | linenum: The number of the line to check. | |
5181 | nesting_state: A NestingState instance which maintains information about | |
5182 | the current stack of nested blocks being parsed. | |
5183 | error: The function to call with any errors found. | |
5184 | """ | |
5185 | # Do nothing if there is no '&' on current line. | |
5186 | line = clean_lines.elided[linenum] | |
5187 | if '&' not in line: | |
5188 | return | |
5189 | ||
5190 | # If a function is inherited, current function doesn't have much of | |
5191 | # a choice, so any non-const references should not be blamed on | |
5192 | # derived function. | |
5193 | if IsDerivedFunction(clean_lines, linenum): | |
5194 | return | |
5195 | ||
5196 | # Don't warn on out-of-line method definitions, as we would warn on the | |
5197 | # in-line declaration, if it isn't marked with 'override'. | |
5198 | if IsOutOfLineMethodDefinition(clean_lines, linenum): | |
5199 | return | |
5200 | ||
5201 | # Long type names may be broken across multiple lines, usually in one | |
5202 | # of these forms: | |
5203 | # LongType | |
5204 | # ::LongTypeContinued &identifier | |
5205 | # LongType:: | |
5206 | # LongTypeContinued &identifier | |
5207 | # LongType< | |
5208 | # ...>::LongTypeContinued &identifier | |
5209 | # | |
5210 | # If we detected a type split across two lines, join the previous | |
5211 | # line to current line so that we can match const references | |
5212 | # accordingly. | |
5213 | # | |
5214 | # Note that this only scans back one line, since scanning back | |
5215 | # arbitrary number of lines would be expensive. If you have a type | |
5216 | # that spans more than 2 lines, please use a typedef. | |
5217 | if linenum > 1: | |
5218 | previous = None | |
5219 | if Match(r'\s*::(?:[\w<>]|::)+\s*&\s*\S', line): | |
5220 | # previous_line\n + ::current_line | |
5221 | previous = Search(r'\b((?:const\s*)?(?:[\w<>]|::)+[\w<>])\s*$', | |
5222 | clean_lines.elided[linenum - 1]) | |
5223 | elif Match(r'\s*[a-zA-Z_]([\w<>]|::)+\s*&\s*\S', line): | |
5224 | # previous_line::\n + current_line | |
5225 | previous = Search(r'\b((?:const\s*)?(?:[\w<>]|::)+::)\s*$', | |
5226 | clean_lines.elided[linenum - 1]) | |
5227 | if previous: | |
5228 | line = previous.group(1) + line.lstrip() | |
5229 | else: | |
5230 | # Check for templated parameter that is split across multiple lines | |
5231 | endpos = line.rfind('>') | |
5232 | if endpos > -1: | |
5233 | (_, startline, startpos) = ReverseCloseExpression( | |
5234 | clean_lines, linenum, endpos) | |
5235 | if startpos > -1 and startline < linenum: | |
5236 | # Found the matching < on an earlier line, collect all | |
5237 | # pieces up to current line. | |
5238 | line = '' | |
5239 | for i in xrange(startline, linenum + 1): | |
5240 | line += clean_lines.elided[i].strip() | |
5241 | ||
5242 | # Check for non-const references in function parameters. A single '&' may | |
5243 | # found in the following places: | |
5244 | # inside expression: binary & for bitwise AND | |
5245 | # inside expression: unary & for taking the address of something | |
5246 | # inside declarators: reference parameter | |
5247 | # We will exclude the first two cases by checking that we are not inside a | |
5248 | # function body, including one that was just introduced by a trailing '{'. | |
5249 | # TODO(unknown): Doesn't account for 'catch(Exception& e)' [rare]. | |
5250 | if (nesting_state.previous_stack_top and | |
5251 | not (isinstance(nesting_state.previous_stack_top, _ClassInfo) or | |
5252 | isinstance(nesting_state.previous_stack_top, _NamespaceInfo))): | |
5253 | # Not at toplevel, not within a class, and not within a namespace | |
5254 | return | |
5255 | ||
5256 | # Avoid initializer lists. We only need to scan back from the | |
5257 | # current line for something that starts with ':'. | |
5258 | # | |
5259 | # We don't need to check the current line, since the '&' would | |
5260 | # appear inside the second set of parentheses on the current line as | |
5261 | # opposed to the first set. | |
5262 | if linenum > 0: | |
5263 | for i in xrange(linenum - 1, max(0, linenum - 10), -1): | |
5264 | previous_line = clean_lines.elided[i] | |
5265 | if not Search(r'[),]\s*$', previous_line): | |
5266 | break | |
5267 | if Match(r'^\s*:\s+\S', previous_line): | |
5268 | return | |
5269 | ||
5270 | # Avoid preprocessors | |
5271 | if Search(r'\\\s*$', line): | |
5272 | return | |
5273 | ||
5274 | # Avoid constructor initializer lists | |
5275 | if IsInitializerList(clean_lines, linenum): | |
5276 | return | |
5277 | ||
5278 | # We allow non-const references in a few standard places, like functions | |
5279 | # called "swap()" or iostream operators like "<<" or ">>". Do not check | |
5280 | # those function parameters. | |
5281 | # | |
5282 | # We also accept & in static_assert, which looks like a function but | |
5283 | # it's actually a declaration expression. | |
5284 | allowed_functions = (r'(?:[sS]wap(?:<\w:+>)?|' | |
5285 | r'operator\s*[<>][<>]|' | |
5286 | r'static_assert|COMPILE_ASSERT' | |
5287 | r')\s*\(') | |
5288 | if Search(allowed_functions, line): | |
5289 | return | |
5290 | elif not Search(r'\S+\([^)]*$', line): | |
5291 | # Don't see an allowed function on this line. Actually we | |
5292 | # didn't see any function name on this line, so this is likely a | |
5293 | # multi-line parameter list. Try a bit harder to catch this case. | |
5294 | for i in xrange(2): | |
5295 | if (linenum > i and | |
5296 | Search(allowed_functions, clean_lines.elided[linenum - i - 1])): | |
5297 | return | |
5298 | ||
5299 | decls = ReplaceAll(r'{[^}]*}', ' ', line) # exclude function body | |
5300 | for parameter in re.findall(_RE_PATTERN_REF_PARAM, decls): | |
5301 | if (not Match(_RE_PATTERN_CONST_REF_PARAM, parameter) and | |
5302 | not Match(_RE_PATTERN_REF_STREAM_PARAM, parameter)): | |
5303 | error(filename, linenum, 'runtime/references', 2, | |
5304 | 'Is this a non-const reference? ' | |
5305 | 'If so, make const or use a pointer: ' + | |
5306 | ReplaceAll(' *<', '<', parameter)) | |
5307 | ||
5308 | ||
5309 | def CheckCasts(filename, clean_lines, linenum, error): | |
5310 | """Various cast related checks. | |
5311 | ||
5312 | Args: | |
5313 | filename: The name of the current file. | |
5314 | clean_lines: A CleansedLines instance containing the file. | |
5315 | linenum: The number of the line to check. | |
5316 | error: The function to call with any errors found. | |
5317 | """ | |
5318 | line = clean_lines.elided[linenum] | |
5319 | ||
5320 | # Check to see if they're using an conversion function cast. | |
5321 | # I just try to capture the most common basic types, though there are more. | |
5322 | # Parameterless conversion functions, such as bool(), are allowed as they are | |
5323 | # probably a member operator declaration or default constructor. | |
5324 | match = Search( | |
5325 | r'(\bnew\s+(?:const\s+)?|\S<\s*(?:const\s+)?)?\b' | |
5326 | r'(int|float|double|bool|char|int32|uint32|int64|uint64)' | |
5327 | r'(\([^)].*)', line) | |
5328 | expecting_function = ExpectingFunctionArgs(clean_lines, linenum) | |
5329 | if match and not expecting_function: | |
5330 | matched_type = match.group(2) | |
5331 | ||
5332 | # matched_new_or_template is used to silence two false positives: | |
5333 | # - New operators | |
5334 | # - Template arguments with function types | |
5335 | # | |
5336 | # For template arguments, we match on types immediately following | |
5337 | # an opening bracket without any spaces. This is a fast way to | |
5338 | # silence the common case where the function type is the first | |
5339 | # template argument. False negative with less-than comparison is | |
5340 | # avoided because those operators are usually followed by a space. | |
5341 | # | |
5342 | # function<double(double)> // bracket + no space = false positive | |
5343 | # value < double(42) // bracket + space = true positive | |
5344 | matched_new_or_template = match.group(1) | |
5345 | ||
5346 | # Avoid arrays by looking for brackets that come after the closing | |
5347 | # parenthesis. | |
5348 | if Match(r'\([^()]+\)\s*\[', match.group(3)): | |
5349 | return | |
5350 | ||
5351 | # Other things to ignore: | |
5352 | # - Function pointers | |
5353 | # - Casts to pointer types | |
5354 | # - Placement new | |
5355 | # - Alias declarations | |
5356 | matched_funcptr = match.group(3) | |
5357 | if (matched_new_or_template is None and | |
5358 | not (matched_funcptr and | |
5359 | (Match(r'\((?:[^() ]+::\s*\*\s*)?[^() ]+\)\s*\(', | |
5360 | matched_funcptr) or | |
5361 | matched_funcptr.startswith('(*)'))) and | |
5362 | not Match(r'\s*using\s+\S+\s*=\s*' + matched_type, line) and | |
5363 | not Search(r'new\(\S+\)\s*' + matched_type, line)): | |
5364 | error(filename, linenum, 'readability/casting', 4, | |
5365 | 'Using deprecated casting style. ' | |
5366 | 'Use static_cast<%s>(...) instead' % | |
5367 | matched_type) | |
5368 | ||
5369 | if not expecting_function: | |
5370 | CheckCStyleCast(filename, clean_lines, linenum, 'static_cast', | |
5371 | r'\((int|float|double|bool|char|u?int(16|32|64))\)', error) | |
5372 | ||
5373 | # This doesn't catch all cases. Consider (const char * const)"hello". | |
5374 | # | |
5375 | # (char *) "foo" should always be a const_cast (reinterpret_cast won't | |
5376 | # compile). | |
5377 | if CheckCStyleCast(filename, clean_lines, linenum, 'const_cast', | |
5378 | r'\((char\s?\*+\s?)\)\s*"', error): | |
5379 | pass | |
5380 | else: | |
5381 | # Check pointer casts for other than string constants | |
5382 | CheckCStyleCast(filename, clean_lines, linenum, 'reinterpret_cast', | |
5383 | r'\((\w+\s?\*+\s?)\)', error) | |
5384 | ||
5385 | # In addition, we look for people taking the address of a cast. This | |
5386 | # is dangerous -- casts can assign to temporaries, so the pointer doesn't | |
5387 | # point where you think. | |
5388 | # | |
5389 | # Some non-identifier character is required before the '&' for the | |
5390 | # expression to be recognized as a cast. These are casts: | |
5391 | # expression = &static_cast<int*>(temporary()); | |
5392 | # function(&(int*)(temporary())); | |
5393 | # | |
5394 | # This is not a cast: | |
5395 | # reference_type&(int* function_param); | |
5396 | match = Search( | |
5397 | r'(?:[^\w]&\(([^)*][^)]*)\)[\w(])|' | |
5398 | r'(?:[^\w]&(static|dynamic|down|reinterpret)_cast\b)', line) | |
5399 | if match: | |
5400 | # Try a better error message when the & is bound to something | |
5401 | # dereferenced by the casted pointer, as opposed to the casted | |
5402 | # pointer itself. | |
5403 | parenthesis_error = False | |
5404 | match = Match(r'^(.*&(?:static|dynamic|down|reinterpret)_cast\b)<', line) | |
5405 | if match: | |
5406 | _, y1, x1 = CloseExpression(clean_lines, linenum, len(match.group(1))) | |
5407 | if x1 >= 0 and clean_lines.elided[y1][x1] == '(': | |
5408 | _, y2, x2 = CloseExpression(clean_lines, y1, x1) | |
5409 | if x2 >= 0: | |
5410 | extended_line = clean_lines.elided[y2][x2:] | |
5411 | if y2 < clean_lines.NumLines() - 1: | |
5412 | extended_line += clean_lines.elided[y2 + 1] | |
5413 | if Match(r'\s*(?:->|\[)', extended_line): | |
5414 | parenthesis_error = True | |
5415 | ||
5416 | if parenthesis_error: | |
5417 | error(filename, linenum, 'readability/casting', 4, | |
5418 | ('Are you taking an address of something dereferenced ' | |
5419 | 'from a cast? Wrapping the dereferenced expression in ' | |
5420 | 'parentheses will make the binding more obvious')) | |
5421 | else: | |
5422 | error(filename, linenum, 'runtime/casting', 4, | |
5423 | ('Are you taking an address of a cast? ' | |
5424 | 'This is dangerous: could be a temp var. ' | |
5425 | 'Take the address before doing the cast, rather than after')) | |
5426 | ||
5427 | ||
5428 | def CheckCStyleCast(filename, clean_lines, linenum, cast_type, pattern, error): | |
5429 | """Checks for a C-style cast by looking for the pattern. | |
5430 | ||
5431 | Args: | |
5432 | filename: The name of the current file. | |
5433 | clean_lines: A CleansedLines instance containing the file. | |
5434 | linenum: The number of the line to check. | |
5435 | cast_type: The string for the C++ cast to recommend. This is either | |
5436 | reinterpret_cast, static_cast, or const_cast, depending. | |
5437 | pattern: The regular expression used to find C-style casts. | |
5438 | error: The function to call with any errors found. | |
5439 | ||
5440 | Returns: | |
5441 | True if an error was emitted. | |
5442 | False otherwise. | |
5443 | """ | |
5444 | line = clean_lines.elided[linenum] | |
5445 | match = Search(pattern, line) | |
5446 | if not match: | |
5447 | return False | |
5448 | ||
5449 | # Exclude lines with keywords that tend to look like casts | |
5450 | context = line[0:match.start(1) - 1] | |
5451 | if Match(r'.*\b(?:sizeof|alignof|alignas|[_A-Z][_A-Z0-9]*)\s*$', context): | |
5452 | return False | |
5453 | ||
5454 | # Try expanding current context to see if we one level of | |
5455 | # parentheses inside a macro. | |
5456 | if linenum > 0: | |
5457 | for i in xrange(linenum - 1, max(0, linenum - 5), -1): | |
5458 | context = clean_lines.elided[i] + context | |
5459 | if Match(r'.*\b[_A-Z][_A-Z0-9]*\s*\((?:\([^()]*\)|[^()])*$', context): | |
5460 | return False | |
5461 | ||
5462 | # operator++(int) and operator--(int) | |
5463 | if context.endswith(' operator++') or context.endswith(' operator--'): | |
5464 | return False | |
5465 | ||
5466 | # A single unnamed argument for a function tends to look like old style cast. | |
5467 | # If we see those, don't issue warnings for deprecated casts. | |
5468 | remainder = line[match.end(0):] | |
5469 | if Match(r'^\s*(?:;|const\b|throw\b|final\b|override\b|[=>{),]|->)', | |
5470 | remainder): | |
5471 | return False | |
5472 | ||
5473 | # At this point, all that should be left is actual casts. | |
5474 | error(filename, linenum, 'readability/casting', 4, | |
5475 | 'Using C-style cast. Use %s<%s>(...) instead' % | |
5476 | (cast_type, match.group(1))) | |
5477 | ||
5478 | return True | |
5479 | ||
5480 | ||
5481 | def ExpectingFunctionArgs(clean_lines, linenum): | |
5482 | """Checks whether where function type arguments are expected. | |
5483 | ||
5484 | Args: | |
5485 | clean_lines: A CleansedLines instance containing the file. | |
5486 | linenum: The number of the line to check. | |
5487 | ||
5488 | Returns: | |
5489 | True if the line at 'linenum' is inside something that expects arguments | |
5490 | of function types. | |
5491 | """ | |
5492 | line = clean_lines.elided[linenum] | |
5493 | return (Match(r'^\s*MOCK_(CONST_)?METHOD\d+(_T)?\(', line) or | |
5494 | (linenum >= 2 and | |
5495 | (Match(r'^\s*MOCK_(?:CONST_)?METHOD\d+(?:_T)?\((?:\S+,)?\s*$', | |
5496 | clean_lines.elided[linenum - 1]) or | |
5497 | Match(r'^\s*MOCK_(?:CONST_)?METHOD\d+(?:_T)?\(\s*$', | |
5498 | clean_lines.elided[linenum - 2]) or | |
5499 | Search(r'\bstd::m?function\s*\<\s*$', | |
5500 | clean_lines.elided[linenum - 1])))) | |
5501 | ||
5502 | ||
5503 | _HEADERS_CONTAINING_TEMPLATES = ( | |
5504 | ('<deque>', ('deque',)), | |
5505 | ('<functional>', ('unary_function', 'binary_function', | |
5506 | 'plus', 'minus', 'multiplies', 'divides', 'modulus', | |
5507 | 'negate', | |
5508 | 'equal_to', 'not_equal_to', 'greater', 'less', | |
5509 | 'greater_equal', 'less_equal', | |
5510 | 'logical_and', 'logical_or', 'logical_not', | |
5511 | 'unary_negate', 'not1', 'binary_negate', 'not2', | |
5512 | 'bind1st', 'bind2nd', | |
5513 | 'pointer_to_unary_function', | |
5514 | 'pointer_to_binary_function', | |
5515 | 'ptr_fun', | |
5516 | 'mem_fun_t', 'mem_fun', 'mem_fun1_t', 'mem_fun1_ref_t', | |
5517 | 'mem_fun_ref_t', | |
5518 | 'const_mem_fun_t', 'const_mem_fun1_t', | |
5519 | 'const_mem_fun_ref_t', 'const_mem_fun1_ref_t', | |
5520 | 'mem_fun_ref', | |
5521 | )), | |
5522 | ('<limits>', ('numeric_limits',)), | |
5523 | ('<list>', ('list',)), | |
5524 | ('<map>', ('map', 'multimap',)), | |
5525 | ('<memory>', ('allocator', 'make_shared', 'make_unique', 'shared_ptr', | |
5526 | 'unique_ptr', 'weak_ptr')), | |
5527 | ('<queue>', ('queue', 'priority_queue',)), | |
5528 | ('<set>', ('set', 'multiset',)), | |
5529 | ('<stack>', ('stack',)), | |
5530 | ('<string>', ('char_traits', 'basic_string',)), | |
5531 | ('<tuple>', ('tuple',)), | |
5532 | ('<unordered_map>', ('unordered_map', 'unordered_multimap')), | |
5533 | ('<unordered_set>', ('unordered_set', 'unordered_multiset')), | |
5534 | ('<utility>', ('pair',)), | |
5535 | ('<vector>', ('vector',)), | |
5536 | ||
5537 | # gcc extensions. | |
5538 | # Note: std::hash is their hash, ::hash is our hash | |
5539 | ('<hash_map>', ('hash_map', 'hash_multimap',)), | |
5540 | ('<hash_set>', ('hash_set', 'hash_multiset',)), | |
5541 | ('<slist>', ('slist',)), | |
5542 | ) | |
5543 | ||
5544 | _HEADERS_MAYBE_TEMPLATES = ( | |
5545 | ('<algorithm>', ('copy', 'max', 'min', 'min_element', 'sort', | |
5546 | 'transform', | |
5547 | )), | |
5548 | ('<utility>', ('forward', 'make_pair', 'move', 'swap')), | |
5549 | ) | |
5550 | ||
5551 | _RE_PATTERN_STRING = re.compile(r'\bstring\b') | |
5552 | ||
5553 | _re_pattern_headers_maybe_templates = [] | |
5554 | for _header, _templates in _HEADERS_MAYBE_TEMPLATES: | |
5555 | for _template in _templates: | |
5556 | # Match max<type>(..., ...), max(..., ...), but not foo->max, foo.max or | |
5557 | # type::max(). | |
5558 | _re_pattern_headers_maybe_templates.append( | |
5559 | (re.compile(r'[^>.]\b' + _template + r'(<.*?>)?\([^\)]'), | |
5560 | _template, | |
5561 | _header)) | |
5562 | ||
5563 | # Other scripts may reach in and modify this pattern. | |
5564 | _re_pattern_templates = [] | |
5565 | for _header, _templates in _HEADERS_CONTAINING_TEMPLATES: | |
5566 | for _template in _templates: | |
5567 | _re_pattern_templates.append( | |
5568 | (re.compile(r'(\<|\b)' + _template + r'\s*\<'), | |
5569 | _template + '<>', | |
5570 | _header)) | |
5571 | ||
5572 | ||
5573 | def FilesBelongToSameModule(filename_cc, filename_h): | |
5574 | """Check if these two filenames belong to the same module. | |
5575 | ||
5576 | The concept of a 'module' here is a as follows: | |
5577 | foo.h, foo-inl.h, foo.cc, foo_test.cc and foo_unittest.cc belong to the | |
5578 | same 'module' if they are in the same directory. | |
5579 | some/path/public/xyzzy and some/path/internal/xyzzy are also considered | |
5580 | to belong to the same module here. | |
5581 | ||
5582 | If the filename_cc contains a longer path than the filename_h, for example, | |
5583 | '/absolute/path/to/base/sysinfo.cc', and this file would include | |
5584 | 'base/sysinfo.h', this function also produces the prefix needed to open the | |
5585 | header. This is used by the caller of this function to more robustly open the | |
5586 | header file. We don't have access to the real include paths in this context, | |
5587 | so we need this guesswork here. | |
5588 | ||
5589 | Known bugs: tools/base/bar.cc and base/bar.h belong to the same module | |
5590 | according to this implementation. Because of this, this function gives | |
5591 | some false positives. This should be sufficiently rare in practice. | |
5592 | ||
5593 | Args: | |
5594 | filename_cc: is the path for the source (e.g. .cc) file | |
5595 | filename_h: is the path for the header path | |
5596 | ||
5597 | Returns: | |
5598 | Tuple with a bool and a string: | |
5599 | bool: True if filename_cc and filename_h belong to the same module. | |
5600 | string: the additional prefix needed to open the header file. | |
5601 | """ | |
5602 | fileinfo_cc = FileInfo(filename_cc) | |
5603 | if not fileinfo_cc.Extension().lstrip('.') in GetNonHeaderExtensions(): | |
5604 | return (False, '') | |
5605 | ||
5606 | fileinfo_h = FileInfo(filename_h) | |
5607 | if not fileinfo_h.Extension().lstrip('.') in GetHeaderExtensions(): | |
5608 | return (False, '') | |
5609 | ||
5610 | filename_cc = filename_cc[:-(len(fileinfo_cc.Extension()))] | |
5611 | matched_test_suffix = Search(_TEST_FILE_SUFFIX, fileinfo_cc.BaseName()) | |
5612 | if matched_test_suffix: | |
5613 | filename_cc = filename_cc[:-len(matched_test_suffix.group(1))] | |
5614 | ||
5615 | filename_cc = filename_cc.replace('/public/', '/') | |
5616 | filename_cc = filename_cc.replace('/internal/', '/') | |
5617 | ||
5618 | filename_h = filename_h[:-(len(fileinfo_h.Extension()))] | |
5619 | if filename_h.endswith('-inl'): | |
5620 | filename_h = filename_h[:-len('-inl')] | |
5621 | filename_h = filename_h.replace('/public/', '/') | |
5622 | filename_h = filename_h.replace('/internal/', '/') | |
5623 | ||
5624 | files_belong_to_same_module = filename_cc.endswith(filename_h) | |
5625 | common_path = '' | |
5626 | if files_belong_to_same_module: | |
5627 | common_path = filename_cc[:-len(filename_h)] | |
5628 | return files_belong_to_same_module, common_path | |
5629 | ||
5630 | ||
5631 | def UpdateIncludeState(filename, include_dict, io=codecs): | |
5632 | """Fill up the include_dict with new includes found from the file. | |
5633 | ||
5634 | Args: | |
5635 | filename: the name of the header to read. | |
5636 | include_dict: a dictionary in which the headers are inserted. | |
5637 | io: The io factory to use to read the file. Provided for testability. | |
5638 | ||
5639 | Returns: | |
5640 | True if a header was successfully added. False otherwise. | |
5641 | """ | |
5642 | headerfile = None | |
5643 | try: | |
5644 | headerfile = io.open(filename, 'r', 'utf8', 'replace') | |
5645 | except IOError: | |
5646 | return False | |
5647 | linenum = 0 | |
5648 | for line in headerfile: | |
5649 | linenum += 1 | |
5650 | clean_line = CleanseComments(line) | |
5651 | match = _RE_PATTERN_INCLUDE.search(clean_line) | |
5652 | if match: | |
5653 | include = match.group(2) | |
5654 | include_dict.setdefault(include, linenum) | |
5655 | return True | |
5656 | ||
5657 | ||
5658 | def CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error, | |
5659 | io=codecs): | |
5660 | """Reports for missing stl includes. | |
5661 | ||
5662 | This function will output warnings to make sure you are including the headers | |
5663 | necessary for the stl containers and functions that you use. We only give one | |
5664 | reason to include a header. For example, if you use both equal_to<> and | |
5665 | less<> in a .h file, only one (the latter in the file) of these will be | |
5666 | reported as a reason to include the <functional>. | |
5667 | ||
5668 | Args: | |
5669 | filename: The name of the current file. | |
5670 | clean_lines: A CleansedLines instance containing the file. | |
5671 | include_state: An _IncludeState instance. | |
5672 | error: The function to call with any errors found. | |
5673 | io: The IO factory to use to read the header file. Provided for unittest | |
5674 | injection. | |
5675 | """ | |
5676 | required = {} # A map of header name to linenumber and the template entity. | |
5677 | # Example of required: { '<functional>': (1219, 'less<>') } | |
5678 | ||
5679 | for linenum in range(clean_lines.NumLines()): | |
5680 | line = clean_lines.elided[linenum] | |
5681 | if not line or line[0] == '#': | |
5682 | continue | |
5683 | ||
5684 | # String is special -- it is a non-templatized type in STL. | |
5685 | matched = _RE_PATTERN_STRING.search(line) | |
5686 | if matched: | |
5687 | # Don't warn about strings in non-STL namespaces: | |
5688 | # (We check only the first match per line; good enough.) | |
5689 | prefix = line[:matched.start()] | |
5690 | if prefix.endswith('std::') or not prefix.endswith('::'): | |
5691 | required['<string>'] = (linenum, 'string') | |
5692 | ||
5693 | for pattern, template, header in _re_pattern_headers_maybe_templates: | |
5694 | if pattern.search(line): | |
5695 | required[header] = (linenum, template) | |
5696 | ||
5697 | # The following function is just a speed up, no semantics are changed. | |
5698 | if not '<' in line: # Reduces the cpu time usage by skipping lines. | |
5699 | continue | |
5700 | ||
5701 | for pattern, template, header in _re_pattern_templates: | |
5702 | matched = pattern.search(line) | |
5703 | if matched: | |
5704 | # Don't warn about IWYU in non-STL namespaces: | |
5705 | # (We check only the first match per line; good enough.) | |
5706 | prefix = line[:matched.start()] | |
5707 | if prefix.endswith('std::') or not prefix.endswith('::'): | |
5708 | required[header] = (linenum, template) | |
5709 | ||
5710 | # The policy is that if you #include something in foo.h you don't need to | |
5711 | # include it again in foo.cc. Here, we will look at possible includes. | |
5712 | # Let's flatten the include_state include_list and copy it into a dictionary. | |
5713 | include_dict = dict([item for sublist in include_state.include_list | |
5714 | for item in sublist]) | |
5715 | ||
5716 | # Did we find the header for this file (if any) and successfully load it? | |
5717 | header_found = False | |
5718 | ||
5719 | # Use the absolute path so that matching works properly. | |
5720 | abs_filename = FileInfo(filename).FullName() | |
5721 | ||
5722 | # For Emacs's flymake. | |
5723 | # If cpplint is invoked from Emacs's flymake, a temporary file is generated | |
5724 | # by flymake and that file name might end with '_flymake.cc'. In that case, | |
5725 | # restore original file name here so that the corresponding header file can be | |
5726 | # found. | |
5727 | # e.g. If the file name is 'foo_flymake.cc', we should search for 'foo.h' | |
5728 | # instead of 'foo_flymake.h' | |
5729 | abs_filename = re.sub(r'_flymake\.cc$', '.cc', abs_filename) | |
5730 | ||
5731 | # include_dict is modified during iteration, so we iterate over a copy of | |
5732 | # the keys. | |
5733 | header_keys = list(include_dict.keys()) | |
5734 | for header in header_keys: | |
5735 | (same_module, common_path) = FilesBelongToSameModule(abs_filename, header) | |
5736 | fullpath = common_path + header | |
5737 | if same_module and UpdateIncludeState(fullpath, include_dict, io): | |
5738 | header_found = True | |
5739 | ||
5740 | # If we can't find the header file for a .cc, assume it's because we don't | |
5741 | # know where to look. In that case we'll give up as we're not sure they | |
5742 | # didn't include it in the .h file. | |
5743 | # TODO(unknown): Do a better job of finding .h files so we are confident that | |
5744 | # not having the .h file means there isn't one. | |
5745 | if not header_found: | |
5746 | for extension in GetNonHeaderExtensions(): | |
5747 | if filename.endswith('.' + extension): | |
5748 | return | |
5749 | ||
5750 | # All the lines have been processed, report the errors found. | |
5751 | for required_header_unstripped in sorted(required, key=required.__getitem__): | |
5752 | template = required[required_header_unstripped][1] | |
5753 | if required_header_unstripped.strip('<>"') not in include_dict: | |
5754 | error(filename, required[required_header_unstripped][0], | |
5755 | 'build/include_what_you_use', 4, | |
5756 | 'Add #include ' + required_header_unstripped + ' for ' + template) | |
5757 | ||
5758 | ||
5759 | _RE_PATTERN_EXPLICIT_MAKEPAIR = re.compile(r'\bmake_pair\s*<') | |
5760 | ||
5761 | ||
5762 | def CheckMakePairUsesDeduction(filename, clean_lines, linenum, error): | |
5763 | """Check that make_pair's template arguments are deduced. | |
5764 | ||
5765 | G++ 4.6 in C++11 mode fails badly if make_pair's template arguments are | |
5766 | specified explicitly, and such use isn't intended in any case. | |
5767 | ||
5768 | Args: | |
5769 | filename: The name of the current file. | |
5770 | clean_lines: A CleansedLines instance containing the file. | |
5771 | linenum: The number of the line to check. | |
5772 | error: The function to call with any errors found. | |
5773 | """ | |
5774 | line = clean_lines.elided[linenum] | |
5775 | match = _RE_PATTERN_EXPLICIT_MAKEPAIR.search(line) | |
5776 | if match: | |
5777 | error(filename, linenum, 'build/explicit_make_pair', | |
5778 | 4, # 4 = high confidence | |
5779 | 'For C++11-compatibility, omit template arguments from make_pair' | |
5780 | ' OR use pair directly OR if appropriate, construct a pair directly') | |
5781 | ||
5782 | ||
5783 | def CheckRedundantVirtual(filename, clean_lines, linenum, error): | |
5784 | """Check if line contains a redundant "virtual" function-specifier. | |
5785 | ||
5786 | Args: | |
5787 | filename: The name of the current file. | |
5788 | clean_lines: A CleansedLines instance containing the file. | |
5789 | linenum: The number of the line to check. | |
5790 | error: The function to call with any errors found. | |
5791 | """ | |
5792 | # Look for "virtual" on current line. | |
5793 | line = clean_lines.elided[linenum] | |
5794 | virtual = Match(r'^(.*)(\bvirtual\b)(.*)$', line) | |
5795 | if not virtual: return | |
5796 | ||
5797 | # Ignore "virtual" keywords that are near access-specifiers. These | |
5798 | # are only used in class base-specifier and do not apply to member | |
5799 | # functions. | |
5800 | if (Search(r'\b(public|protected|private)\s+$', virtual.group(1)) or | |
5801 | Match(r'^\s+(public|protected|private)\b', virtual.group(3))): | |
5802 | return | |
5803 | ||
5804 | # Ignore the "virtual" keyword from virtual base classes. Usually | |
5805 | # there is a column on the same line in these cases (virtual base | |
5806 | # classes are rare in google3 because multiple inheritance is rare). | |
5807 | if Match(r'^.*[^:]:[^:].*$', line): return | |
5808 | ||
5809 | # Look for the next opening parenthesis. This is the start of the | |
5810 | # parameter list (possibly on the next line shortly after virtual). | |
5811 | # TODO(unknown): doesn't work if there are virtual functions with | |
5812 | # decltype() or other things that use parentheses, but csearch suggests | |
5813 | # that this is rare. | |
5814 | end_col = -1 | |
5815 | end_line = -1 | |
5816 | start_col = len(virtual.group(2)) | |
5817 | for start_line in xrange(linenum, min(linenum + 3, clean_lines.NumLines())): | |
5818 | line = clean_lines.elided[start_line][start_col:] | |
5819 | parameter_list = Match(r'^([^(]*)\(', line) | |
5820 | if parameter_list: | |
5821 | # Match parentheses to find the end of the parameter list | |
5822 | (_, end_line, end_col) = CloseExpression( | |
5823 | clean_lines, start_line, start_col + len(parameter_list.group(1))) | |
5824 | break | |
5825 | start_col = 0 | |
5826 | ||
5827 | if end_col < 0: | |
5828 | return # Couldn't find end of parameter list, give up | |
5829 | ||
5830 | # Look for "override" or "final" after the parameter list | |
5831 | # (possibly on the next few lines). | |
5832 | for i in xrange(end_line, min(end_line + 3, clean_lines.NumLines())): | |
5833 | line = clean_lines.elided[i][end_col:] | |
5834 | match = Search(r'\b(override|final)\b', line) | |
5835 | if match: | |
5836 | error(filename, linenum, 'readability/inheritance', 4, | |
5837 | ('"virtual" is redundant since function is ' | |
5838 | 'already declared as "%s"' % match.group(1))) | |
5839 | ||
5840 | # Set end_col to check whole lines after we are done with the | |
5841 | # first line. | |
5842 | end_col = 0 | |
5843 | if Search(r'[^\w]\s*$', line): | |
5844 | break | |
5845 | ||
5846 | ||
5847 | def CheckRedundantOverrideOrFinal(filename, clean_lines, linenum, error): | |
5848 | """Check if line contains a redundant "override" or "final" virt-specifier. | |
5849 | ||
5850 | Args: | |
5851 | filename: The name of the current file. | |
5852 | clean_lines: A CleansedLines instance containing the file. | |
5853 | linenum: The number of the line to check. | |
5854 | error: The function to call with any errors found. | |
5855 | """ | |
5856 | # Look for closing parenthesis nearby. We need one to confirm where | |
5857 | # the declarator ends and where the virt-specifier starts to avoid | |
5858 | # false positives. | |
5859 | line = clean_lines.elided[linenum] | |
5860 | declarator_end = line.rfind(')') | |
5861 | if declarator_end >= 0: | |
5862 | fragment = line[declarator_end:] | |
5863 | else: | |
5864 | if linenum > 1 and clean_lines.elided[linenum - 1].rfind(')') >= 0: | |
5865 | fragment = line | |
5866 | else: | |
5867 | return | |
5868 | ||
5869 | # Check that at most one of "override" or "final" is present, not both | |
5870 | if Search(r'\boverride\b', fragment) and Search(r'\bfinal\b', fragment): | |
5871 | error(filename, linenum, 'readability/inheritance', 4, | |
5872 | ('"override" is redundant since function is ' | |
5873 | 'already declared as "final"')) | |
5874 | ||
5875 | ||
5876 | ||
5877 | ||
5878 | # Returns true if we are at a new block, and it is directly | |
5879 | # inside of a namespace. | |
5880 | def IsBlockInNameSpace(nesting_state, is_forward_declaration): | |
5881 | """Checks that the new block is directly in a namespace. | |
5882 | ||
5883 | Args: | |
5884 | nesting_state: The _NestingState object that contains info about our state. | |
5885 | is_forward_declaration: If the class is a forward declared class. | |
5886 | Returns: | |
5887 | Whether or not the new block is directly in a namespace. | |
5888 | """ | |
5889 | if is_forward_declaration: | |
5890 | return len(nesting_state.stack) >= 1 and ( | |
5891 | isinstance(nesting_state.stack[-1], _NamespaceInfo)) | |
5892 | ||
5893 | ||
5894 | return (len(nesting_state.stack) > 1 and | |
5895 | nesting_state.stack[-1].check_namespace_indentation and | |
5896 | isinstance(nesting_state.stack[-2], _NamespaceInfo)) | |
5897 | ||
5898 | ||
5899 | def ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item, | |
5900 | raw_lines_no_comments, linenum): | |
5901 | """This method determines if we should apply our namespace indentation check. | |
5902 | ||
5903 | Args: | |
5904 | nesting_state: The current nesting state. | |
5905 | is_namespace_indent_item: If we just put a new class on the stack, True. | |
5906 | If the top of the stack is not a class, or we did not recently | |
5907 | add the class, False. | |
5908 | raw_lines_no_comments: The lines without the comments. | |
5909 | linenum: The current line number we are processing. | |
5910 | ||
5911 | Returns: | |
5912 | True if we should apply our namespace indentation check. Currently, it | |
5913 | only works for classes and namespaces inside of a namespace. | |
5914 | """ | |
5915 | ||
5916 | is_forward_declaration = IsForwardClassDeclaration(raw_lines_no_comments, | |
5917 | linenum) | |
5918 | ||
5919 | if not (is_namespace_indent_item or is_forward_declaration): | |
5920 | return False | |
5921 | ||
5922 | # If we are in a macro, we do not want to check the namespace indentation. | |
5923 | if IsMacroDefinition(raw_lines_no_comments, linenum): | |
5924 | return False | |
5925 | ||
5926 | return IsBlockInNameSpace(nesting_state, is_forward_declaration) | |
5927 | ||
5928 | ||
5929 | # Call this method if the line is directly inside of a namespace. | |
5930 | # If the line above is blank (excluding comments) or the start of | |
5931 | # an inner namespace, it cannot be indented. | |
5932 | def CheckItemIndentationInNamespace(filename, raw_lines_no_comments, linenum, | |
5933 | error): | |
5934 | line = raw_lines_no_comments[linenum] | |
5935 | if Match(r'^\s+', line): | |
5936 | error(filename, linenum, 'runtime/indentation_namespace', 4, | |
5937 | 'Do not indent within a namespace') | |
5938 | ||
5939 | ||
5940 | def ProcessLine(filename, file_extension, clean_lines, line, | |
5941 | include_state, function_state, nesting_state, error, | |
5942 | extra_check_functions=None): | |
5943 | """Processes a single line in the file. | |
5944 | ||
5945 | Args: | |
5946 | filename: Filename of the file that is being processed. | |
5947 | file_extension: The extension (dot not included) of the file. | |
5948 | clean_lines: An array of strings, each representing a line of the file, | |
5949 | with comments stripped. | |
5950 | line: Number of line being processed. | |
5951 | include_state: An _IncludeState instance in which the headers are inserted. | |
5952 | function_state: A _FunctionState instance which counts function lines, etc. | |
5953 | nesting_state: A NestingState instance which maintains information about | |
5954 | the current stack of nested blocks being parsed. | |
5955 | error: A callable to which errors are reported, which takes 4 arguments: | |
5956 | filename, line number, error level, and message | |
5957 | extra_check_functions: An array of additional check functions that will be | |
5958 | run on each source line. Each function takes 4 | |
5959 | arguments: filename, clean_lines, line, error | |
5960 | """ | |
5961 | raw_lines = clean_lines.raw_lines | |
5962 | ParseNolintSuppressions(filename, raw_lines[line], line, error) | |
5963 | nesting_state.Update(filename, clean_lines, line, error) | |
5964 | CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line, | |
5965 | error) | |
5966 | if nesting_state.InAsmBlock(): return | |
5967 | CheckForFunctionLengths(filename, clean_lines, line, function_state, error) | |
5968 | CheckForMultilineCommentsAndStrings(filename, clean_lines, line, error) | |
5969 | CheckStyle(filename, clean_lines, line, file_extension, nesting_state, error) | |
5970 | CheckLanguage(filename, clean_lines, line, file_extension, include_state, | |
5971 | nesting_state, error) | |
5972 | CheckForNonConstReference(filename, clean_lines, line, nesting_state, error) | |
5973 | CheckForNonStandardConstructs(filename, clean_lines, line, | |
5974 | nesting_state, error) | |
5975 | CheckVlogArguments(filename, clean_lines, line, error) | |
5976 | CheckPosixThreading(filename, clean_lines, line, error) | |
5977 | CheckInvalidIncrement(filename, clean_lines, line, error) | |
5978 | CheckMakePairUsesDeduction(filename, clean_lines, line, error) | |
5979 | CheckRedundantVirtual(filename, clean_lines, line, error) | |
5980 | CheckRedundantOverrideOrFinal(filename, clean_lines, line, error) | |
5981 | if extra_check_functions: | |
5982 | for check_fn in extra_check_functions: | |
5983 | check_fn(filename, clean_lines, line, error) | |
5984 | ||
5985 | def FlagCxx11Features(filename, clean_lines, linenum, error): | |
5986 | """Flag those c++11 features that we only allow in certain places. | |
5987 | ||
5988 | Args: | |
5989 | filename: The name of the current file. | |
5990 | clean_lines: A CleansedLines instance containing the file. | |
5991 | linenum: The number of the line to check. | |
5992 | error: The function to call with any errors found. | |
5993 | """ | |
5994 | line = clean_lines.elided[linenum] | |
5995 | ||
5996 | include = Match(r'\s*#\s*include\s+[<"]([^<"]+)[">]', line) | |
5997 | ||
5998 | # Flag unapproved C++ TR1 headers. | |
5999 | if include and include.group(1).startswith('tr1/'): | |
6000 | error(filename, linenum, 'build/c++tr1', 5, | |
6001 | ('C++ TR1 headers such as <%s> are unapproved.') % include.group(1)) | |
6002 | ||
6003 | # Flag unapproved C++11 headers. | |
6004 | if include and include.group(1) in ('cfenv', | |
6005 | 'condition_variable', | |
6006 | 'fenv.h', | |
6007 | 'future', | |
6008 | 'mutex', | |
6009 | 'thread', | |
6010 | 'chrono', | |
6011 | 'ratio', | |
6012 | 'regex', | |
6013 | 'system_error', | |
6014 | ): | |
6015 | error(filename, linenum, 'build/c++11', 5, | |
6016 | ('<%s> is an unapproved C++11 header.') % include.group(1)) | |
6017 | ||
6018 | # The only place where we need to worry about C++11 keywords and library | |
6019 | # features in preprocessor directives is in macro definitions. | |
6020 | if Match(r'\s*#', line) and not Match(r'\s*#\s*define\b', line): return | |
6021 | ||
6022 | # These are classes and free functions. The classes are always | |
6023 | # mentioned as std::*, but we only catch the free functions if | |
6024 | # they're not found by ADL. They're alphabetical by header. | |
6025 | for top_name in ( | |
6026 | # type_traits | |
6027 | 'alignment_of', | |
6028 | 'aligned_union', | |
6029 | ): | |
6030 | if Search(r'\bstd::%s\b' % top_name, line): | |
6031 | error(filename, linenum, 'build/c++11', 5, | |
6032 | ('std::%s is an unapproved C++11 class or function. Send c-style ' | |
6033 | 'an example of where it would make your code more readable, and ' | |
6034 | 'they may let you use it.') % top_name) | |
6035 | ||
6036 | ||
6037 | def FlagCxx14Features(filename, clean_lines, linenum, error): | |
6038 | """Flag those C++14 features that we restrict. | |
6039 | ||
6040 | Args: | |
6041 | filename: The name of the current file. | |
6042 | clean_lines: A CleansedLines instance containing the file. | |
6043 | linenum: The number of the line to check. | |
6044 | error: The function to call with any errors found. | |
6045 | """ | |
6046 | line = clean_lines.elided[linenum] | |
6047 | ||
6048 | include = Match(r'\s*#\s*include\s+[<"]([^<"]+)[">]', line) | |
6049 | ||
6050 | # Flag unapproved C++14 headers. | |
6051 | if include and include.group(1) in ('scoped_allocator', 'shared_mutex'): | |
6052 | error(filename, linenum, 'build/c++14', 5, | |
6053 | ('<%s> is an unapproved C++14 header.') % include.group(1)) | |
6054 | ||
6055 | ||
6056 | def ProcessFileData(filename, file_extension, lines, error, | |
6057 | extra_check_functions=None): | |
6058 | """Performs lint checks and reports any errors to the given error function. | |
6059 | ||
6060 | Args: | |
6061 | filename: Filename of the file that is being processed. | |
6062 | file_extension: The extension (dot not included) of the file. | |
6063 | lines: An array of strings, each representing a line of the file, with the | |
6064 | last element being empty if the file is terminated with a newline. | |
6065 | error: A callable to which errors are reported, which takes 4 arguments: | |
6066 | filename, line number, error level, and message | |
6067 | extra_check_functions: An array of additional check functions that will be | |
6068 | run on each source line. Each function takes 4 | |
6069 | arguments: filename, clean_lines, line, error | |
6070 | """ | |
6071 | lines = (['// marker so line numbers and indices both start at 1'] + lines + | |
6072 | ['// marker so line numbers end in a known way']) | |
6073 | ||
6074 | include_state = _IncludeState() | |
6075 | function_state = _FunctionState() | |
6076 | nesting_state = NestingState() | |
6077 | ||
6078 | ResetNolintSuppressions() | |
6079 | ||
6080 | CheckForCopyright(filename, lines, error) | |
6081 | ProcessGlobalSuppresions(lines) | |
6082 | RemoveMultiLineComments(filename, lines, error) | |
6083 | clean_lines = CleansedLines(lines) | |
6084 | ||
6085 | if file_extension in GetHeaderExtensions(): | |
6086 | CheckForHeaderGuard(filename, clean_lines, error) | |
6087 | ||
6088 | for line in range(clean_lines.NumLines()): | |
6089 | ProcessLine(filename, file_extension, clean_lines, line, | |
6090 | include_state, function_state, nesting_state, error, | |
6091 | extra_check_functions) | |
6092 | FlagCxx11Features(filename, clean_lines, line, error) | |
6093 | nesting_state.CheckCompletedBlocks(filename, error) | |
6094 | ||
6095 | CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error) | |
6096 | ||
6097 | # Check that the .cc file has included its header if it exists. | |
6098 | if _IsSourceExtension(file_extension): | |
6099 | CheckHeaderFileIncluded(filename, include_state, error) | |
6100 | ||
6101 | # We check here rather than inside ProcessLine so that we see raw | |
6102 | # lines rather than "cleaned" lines. | |
6103 | CheckForBadCharacters(filename, lines, error) | |
6104 | ||
6105 | CheckForNewlineAtEOF(filename, lines, error) | |
6106 | ||
6107 | def ProcessConfigOverrides(filename): | |
6108 | """ Loads the configuration files and processes the config overrides. | |
6109 | ||
6110 | Args: | |
6111 | filename: The name of the file being processed by the linter. | |
6112 | ||
6113 | Returns: | |
6114 | False if the current |filename| should not be processed further. | |
6115 | """ | |
6116 | ||
6117 | abs_filename = os.path.abspath(filename) | |
6118 | cfg_filters = [] | |
6119 | keep_looking = True | |
6120 | while keep_looking: | |
6121 | abs_path, base_name = os.path.split(abs_filename) | |
6122 | if not base_name: | |
6123 | break # Reached the root directory. | |
6124 | ||
6125 | cfg_file = os.path.join(abs_path, "CPPLINT.cfg") | |
6126 | abs_filename = abs_path | |
6127 | if not os.path.isfile(cfg_file): | |
6128 | continue | |
6129 | ||
6130 | try: | |
6131 | with open(cfg_file) as file_handle: | |
6132 | for line in file_handle: | |
6133 | line, _, _ = line.partition('#') # Remove comments. | |
6134 | if not line.strip(): | |
6135 | continue | |
6136 | ||
6137 | name, _, val = line.partition('=') | |
6138 | name = name.strip() | |
6139 | val = val.strip() | |
6140 | if name == 'set noparent': | |
6141 | keep_looking = False | |
6142 | elif name == 'filter': | |
6143 | cfg_filters.append(val) | |
6144 | elif name == 'exclude_files': | |
6145 | # When matching exclude_files pattern, use the base_name of | |
6146 | # the current file name or the directory name we are processing. | |
6147 | # For example, if we are checking for lint errors in /foo/bar/baz.cc | |
6148 | # and we found the .cfg file at /foo/CPPLINT.cfg, then the config | |
6149 | # file's "exclude_files" filter is meant to be checked against "bar" | |
6150 | # and not "baz" nor "bar/baz.cc". | |
6151 | if base_name: | |
6152 | pattern = re.compile(val) | |
6153 | if pattern.match(base_name): | |
6154 | _cpplint_state.PrintInfo('Ignoring "%s": file excluded by ' | |
6155 | '"%s". File path component "%s" matches pattern "%s"\n' % | |
6156 | (filename, cfg_file, base_name, val)) | |
6157 | return False | |
6158 | elif name == 'linelength': | |
6159 | global _line_length | |
6160 | try: | |
6161 | _line_length = int(val) | |
6162 | except ValueError: | |
6163 | _cpplint_state.PrintError('Line length must be numeric.') | |
6164 | elif name == 'extensions': | |
6165 | global _valid_extensions | |
6166 | try: | |
6167 | extensions = [ext.strip() for ext in val.split(',')] | |
6168 | _valid_extensions = set(extensions) | |
6169 | except ValueError: | |
6170 | sys.stderr.write('Extensions should be a comma-separated list of values;' | |
6171 | 'for example: extensions=hpp,cpp\n' | |
6172 | 'This could not be parsed: "%s"' % (val,)) | |
6173 | elif name == 'headers': | |
6174 | global _header_extensions | |
6175 | try: | |
6176 | extensions = [ext.strip() for ext in val.split(',')] | |
6177 | _header_extensions = set(extensions) | |
6178 | except ValueError: | |
6179 | sys.stderr.write('Extensions should be a comma-separated list of values;' | |
6180 | 'for example: extensions=hpp,cpp\n' | |
6181 | 'This could not be parsed: "%s"' % (val,)) | |
6182 | elif name == 'root': | |
6183 | global _root | |
6184 | _root = val | |
6185 | else: | |
6186 | _cpplint_state.PrintError( | |
6187 | 'Invalid configuration option (%s) in file %s\n' % | |
6188 | (name, cfg_file)) | |
6189 | ||
6190 | except IOError: | |
6191 | _cpplint_state.PrintError( | |
6192 | "Skipping config file '%s': Can't open for reading\n" % cfg_file) | |
6193 | keep_looking = False | |
6194 | ||
6195 | # Apply all the accumulated filters in reverse order (top-level directory | |
6196 | # config options having the least priority). | |
6197 | for cfg_filter in reversed(cfg_filters): | |
6198 | _AddFilters(cfg_filter) | |
6199 | ||
6200 | return True | |
6201 | ||
6202 | ||
6203 | def ProcessFile(filename, vlevel, extra_check_functions=None): | |
6204 | """Does google-lint on a single file. | |
6205 | ||
6206 | Args: | |
6207 | filename: The name of the file to parse. | |
6208 | ||
6209 | vlevel: The level of errors to report. Every error of confidence | |
6210 | >= verbose_level will be reported. 0 is a good default. | |
6211 | ||
6212 | extra_check_functions: An array of additional check functions that will be | |
6213 | run on each source line. Each function takes 4 | |
6214 | arguments: filename, clean_lines, line, error | |
6215 | """ | |
6216 | ||
6217 | _SetVerboseLevel(vlevel) | |
6218 | _BackupFilters() | |
6219 | ||
6220 | if not ProcessConfigOverrides(filename): | |
6221 | _RestoreFilters() | |
6222 | return | |
6223 | ||
6224 | lf_lines = [] | |
6225 | crlf_lines = [] | |
6226 | try: | |
6227 | # Support the UNIX convention of using "-" for stdin. Note that | |
6228 | # we are not opening the file with universal newline support | |
6229 | # (which codecs doesn't support anyway), so the resulting lines do | |
6230 | # contain trailing '\r' characters if we are reading a file that | |
6231 | # has CRLF endings. | |
6232 | # If after the split a trailing '\r' is present, it is removed | |
6233 | # below. | |
6234 | if filename == '-': | |
6235 | lines = codecs.StreamReaderWriter(sys.stdin, | |
6236 | codecs.getreader('utf8'), | |
6237 | codecs.getwriter('utf8'), | |
6238 | 'replace').read().split('\n') | |
6239 | else: | |
6240 | lines = codecs.open(filename, 'r', 'utf8', 'replace').read().split('\n') | |
6241 | ||
6242 | # Remove trailing '\r'. | |
6243 | # The -1 accounts for the extra trailing blank line we get from split() | |
6244 | for linenum in range(len(lines) - 1): | |
6245 | if lines[linenum].endswith('\r'): | |
6246 | lines[linenum] = lines[linenum].rstrip('\r') | |
6247 | crlf_lines.append(linenum + 1) | |
6248 | else: | |
6249 | lf_lines.append(linenum + 1) | |
6250 | ||
6251 | except IOError: | |
6252 | _cpplint_state.PrintError( | |
6253 | "Skipping input '%s': Can't open for reading\n" % filename) | |
6254 | _RestoreFilters() | |
6255 | return | |
6256 | ||
6257 | # Note, if no dot is found, this will give the entire filename as the ext. | |
6258 | file_extension = filename[filename.rfind('.') + 1:] | |
6259 | ||
6260 | # When reading from stdin, the extension is unknown, so no cpplint tests | |
6261 | # should rely on the extension. | |
6262 | if filename != '-' and file_extension not in GetAllExtensions(): | |
6263 | _cpplint_state.PrintError('Ignoring %s; not a valid file name ' | |
6264 | '(%s)\n' % (filename, ', '.join(GetAllExtensions()))) | |
6265 | else: | |
6266 | ProcessFileData(filename, file_extension, lines, Error, | |
6267 | extra_check_functions) | |
6268 | ||
6269 | # If end-of-line sequences are a mix of LF and CR-LF, issue | |
6270 | # warnings on the lines with CR. | |
6271 | # | |
6272 | # Don't issue any warnings if all lines are uniformly LF or CR-LF, | |
6273 | # since critique can handle these just fine, and the style guide | |
6274 | # doesn't dictate a particular end of line sequence. | |
6275 | # | |
6276 | # We can't depend on os.linesep to determine what the desired | |
6277 | # end-of-line sequence should be, since that will return the | |
6278 | # server-side end-of-line sequence. | |
6279 | if lf_lines and crlf_lines: | |
6280 | # Warn on every line with CR. An alternative approach might be to | |
6281 | # check whether the file is mostly CRLF or just LF, and warn on the | |
6282 | # minority, we bias toward LF here since most tools prefer LF. | |
6283 | for linenum in crlf_lines: | |
6284 | Error(filename, linenum, 'whitespace/newline', 1, | |
6285 | 'Unexpected \\r (^M) found; better to use only \\n') | |
6286 | ||
6287 | _cpplint_state.PrintInfo('Done processing %s\n' % filename) | |
6288 | _RestoreFilters() | |
6289 | ||
6290 | ||
6291 | def PrintUsage(message): | |
6292 | """Prints a brief usage string and exits, optionally with an error message. | |
6293 | ||
6294 | Args: | |
6295 | message: The optional error message. | |
6296 | """ | |
6297 | sys.stderr.write(_USAGE) | |
6298 | ||
6299 | if message: | |
6300 | sys.exit('\nFATAL ERROR: ' + message) | |
6301 | else: | |
6302 | sys.exit(0) | |
6303 | ||
6304 | ||
6305 | def PrintCategories(): | |
6306 | """Prints a list of all the error-categories used by error messages. | |
6307 | ||
6308 | These are the categories used to filter messages via --filter. | |
6309 | """ | |
6310 | sys.stderr.write(''.join(' %s\n' % cat for cat in _ERROR_CATEGORIES)) | |
6311 | sys.exit(0) | |
6312 | ||
6313 | ||
6314 | def ParseArguments(args): | |
6315 | """Parses the command line arguments. | |
6316 | ||
6317 | This may set the output format and verbosity level as side-effects. | |
6318 | ||
6319 | Args: | |
6320 | args: The command line arguments: | |
6321 | ||
6322 | Returns: | |
6323 | The list of filenames to lint. | |
6324 | """ | |
6325 | try: | |
6326 | (opts, filenames) = getopt.getopt(args, '', ['help', 'output=', 'verbose=', | |
6327 | 'counting=', | |
6328 | 'filter=', | |
6329 | 'root=', | |
6330 | 'repository=', | |
6331 | 'linelength=', | |
6332 | 'extensions=', | |
6333 | 'exclude=', | |
6334 | 'headers=', | |
6335 | 'quiet', | |
6336 | 'recursive']) | |
6337 | except getopt.GetoptError: | |
6338 | PrintUsage('Invalid arguments.') | |
6339 | ||
6340 | verbosity = _VerboseLevel() | |
6341 | output_format = _OutputFormat() | |
6342 | filters = '' | |
6343 | counting_style = '' | |
6344 | recursive = False | |
6345 | ||
6346 | for (opt, val) in opts: | |
6347 | if opt == '--help': | |
6348 | PrintUsage(None) | |
6349 | elif opt == '--output': | |
6350 | if val not in ('emacs', 'vs7', 'eclipse', 'junit'): | |
6351 | PrintUsage('The only allowed output formats are emacs, vs7, eclipse ' | |
6352 | 'and junit.') | |
6353 | output_format = val | |
6354 | elif opt == '--verbose': | |
6355 | verbosity = int(val) | |
6356 | elif opt == '--filter': | |
6357 | filters = val | |
6358 | if not filters: | |
6359 | PrintCategories() | |
6360 | elif opt == '--counting': | |
6361 | if val not in ('total', 'toplevel', 'detailed'): | |
6362 | PrintUsage('Valid counting options are total, toplevel, and detailed') | |
6363 | counting_style = val | |
6364 | elif opt == '--root': | |
6365 | global _root | |
6366 | _root = val | |
6367 | elif opt == '--repository': | |
6368 | global _repository | |
6369 | _repository = val | |
6370 | elif opt == '--linelength': | |
6371 | global _line_length | |
6372 | try: | |
6373 | _line_length = int(val) | |
6374 | except ValueError: | |
6375 | PrintUsage('Line length must be digits.') | |
6376 | elif opt == '--exclude': | |
6377 | global _excludes | |
6378 | if not _excludes: | |
6379 | _excludes = set() | |
6380 | _excludes.update(glob.glob(val)) | |
6381 | elif opt == '--extensions': | |
6382 | global _valid_extensions | |
6383 | try: | |
6384 | _valid_extensions = set(val.split(',')) | |
6385 | except ValueError: | |
6386 | PrintUsage('Extensions must be comma separated list.') | |
6387 | elif opt == '--headers': | |
6388 | global _header_extensions | |
6389 | try: | |
6390 | _header_extensions = set(val.split(',')) | |
6391 | except ValueError: | |
6392 | PrintUsage('Extensions must be comma separated list.') | |
6393 | elif opt == '--recursive': | |
6394 | recursive = True | |
6395 | elif opt == '--quiet': | |
6396 | global _quiet | |
6397 | _quiet = True | |
6398 | ||
6399 | if not filenames: | |
6400 | PrintUsage('No files were specified.') | |
6401 | ||
6402 | if recursive: | |
6403 | filenames = _ExpandDirectories(filenames) | |
6404 | ||
6405 | if _excludes: | |
6406 | filenames = _FilterExcludedFiles(filenames) | |
6407 | ||
6408 | _SetOutputFormat(output_format) | |
6409 | _SetVerboseLevel(verbosity) | |
6410 | _SetFilters(filters) | |
6411 | _SetCountingStyle(counting_style) | |
6412 | ||
6413 | return filenames | |
6414 | ||
6415 | def _ExpandDirectories(filenames): | |
6416 | """Searches a list of filenames and replaces directories in the list with | |
6417 | all files descending from those directories. Files with extensions not in | |
6418 | the valid extensions list are excluded. | |
6419 | ||
6420 | Args: | |
6421 | filenames: A list of files or directories | |
6422 | ||
6423 | Returns: | |
6424 | A list of all files that are members of filenames or descended from a | |
6425 | directory in filenames | |
6426 | """ | |
6427 | expanded = set() | |
6428 | for filename in filenames: | |
6429 | if not os.path.isdir(filename): | |
6430 | expanded.add(filename) | |
6431 | continue | |
6432 | ||
6433 | for root, _, files in os.walk(filename): | |
6434 | for loopfile in files: | |
6435 | fullname = os.path.join(root, loopfile) | |
6436 | if fullname.startswith('.' + os.path.sep): | |
6437 | fullname = fullname[len('.' + os.path.sep):] | |
6438 | expanded.add(fullname) | |
6439 | ||
6440 | filtered = [] | |
6441 | for filename in expanded: | |
6442 | if os.path.splitext(filename)[1][1:] in GetAllExtensions(): | |
6443 | filtered.append(filename) | |
6444 | ||
6445 | return filtered | |
6446 | ||
6447 | def _FilterExcludedFiles(filenames): | |
6448 | """Filters out files listed in the --exclude command line switch. File paths | |
6449 | in the switch are evaluated relative to the current working directory | |
6450 | """ | |
6451 | exclude_paths = [os.path.abspath(f) for f in _excludes] | |
6452 | return [f for f in filenames if os.path.abspath(f) not in exclude_paths] | |
6453 | ||
6454 | def main(): | |
6455 | filenames = ParseArguments(sys.argv[1:]) | |
6456 | backup_err = sys.stderr | |
6457 | try: | |
6458 | # Change stderr to write with replacement characters so we don't die | |
6459 | # if we try to print something containing non-ASCII characters. | |
6460 | sys.stderr = codecs.StreamReader(sys.stderr, 'replace') | |
6461 | ||
6462 | _cpplint_state.ResetErrorCounts() | |
6463 | for filename in filenames: | |
6464 | ProcessFile(filename, _cpplint_state.verbose_level) | |
6465 | _cpplint_state.PrintErrorCounts() | |
6466 | ||
6467 | if _cpplint_state.output_format == 'junit': | |
6468 | sys.stderr.write(_cpplint_state.FormatJUnitXML()) | |
6469 | ||
6470 | finally: | |
6471 | sys.stderr = backup_err | |
6472 | ||
6473 | sys.exit(_cpplint_state.error_count > 0) | |
6474 | ||
6475 | ||
6476 | if __name__ == '__main__': | |
6477 | main() |