]>
Commit | Line | Data |
---|---|---|
1e59de90 TL |
1 | #!/usr/bin/env python |
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 sysconfig | |
56 | import unicodedata | |
57 | import xml.etree.ElementTree | |
58 | ||
59 | # if empty, use defaults | |
60 | _valid_extensions = set([]) | |
61 | ||
62 | __VERSION__ = '1.5.5' | |
63 | ||
64 | try: | |
65 | xrange # Python 2 | |
66 | except NameError: | |
67 | # -- pylint: disable=redefined-builtin | |
68 | xrange = range # Python 3 | |
69 | ||
70 | ||
71 | _USAGE = """ | |
72 | Syntax: cpplint.py [--verbose=#] [--output=emacs|eclipse|vs7|junit|sed|gsed] | |
73 | [--filter=-x,+y,...] | |
74 | [--counting=total|toplevel|detailed] [--root=subdir] | |
75 | [--repository=path] | |
76 | [--linelength=digits] [--headers=x,y,...] | |
77 | [--recursive] | |
78 | [--exclude=path] | |
79 | [--extensions=hpp,cpp,...] | |
80 | [--includeorder=default|standardcfirst] | |
81 | [--quiet] | |
82 | [--version] | |
83 | <file> [file] ... | |
84 | ||
85 | Style checker for C/C++ source files. | |
86 | This is a fork of the Google style checker with minor extensions. | |
87 | ||
88 | The style guidelines this tries to follow are those in | |
89 | https://google.github.io/styleguide/cppguide.html | |
90 | ||
91 | Every problem is given a confidence score from 1-5, with 5 meaning we are | |
92 | certain of the problem, and 1 meaning it could be a legitimate construct. | |
93 | This will miss some errors, and is not a substitute for a code review. | |
94 | ||
95 | To suppress false-positive errors of a certain category, add a | |
96 | 'NOLINT(category)' comment to the line. NOLINT or NOLINT(*) | |
97 | suppresses errors of all categories on that line. | |
98 | ||
99 | The files passed in will be linted; at least one file must be provided. | |
100 | Default linted extensions are %s. | |
101 | Other file types will be ignored. | |
102 | Change the extensions with the --extensions flag. | |
103 | ||
104 | Flags: | |
105 | ||
106 | output=emacs|eclipse|vs7|junit|sed|gsed | |
107 | By default, the output is formatted to ease emacs parsing. Visual Studio | |
108 | compatible output (vs7) may also be used. Further support exists for | |
109 | eclipse (eclipse), and JUnit (junit). XML parsers such as those used | |
110 | in Jenkins and Bamboo may also be used. | |
111 | The sed format outputs sed commands that should fix some of the errors. | |
112 | Note that this requires gnu sed. If that is installed as gsed on your | |
113 | system (common e.g. on macOS with homebrew) you can use the gsed output | |
114 | format. Sed commands are written to stdout, not stderr, so you should be | |
115 | able to pipe output straight to a shell to run the fixes. | |
116 | ||
117 | verbose=# | |
118 | Specify a number 0-5 to restrict errors to certain verbosity levels. | |
119 | Errors with lower verbosity levels have lower confidence and are more | |
120 | likely to be false positives. | |
121 | ||
122 | quiet | |
123 | Don't print anything if no errors are found. | |
124 | ||
125 | filter=-x,+y,... | |
126 | Specify a comma-separated list of category-filters to apply: only | |
127 | error messages whose category names pass the filters will be printed. | |
128 | (Category names are printed with the message and look like | |
129 | "[whitespace/indent]".) Filters are evaluated left to right. | |
130 | "-FOO" means "do not print categories that start with FOO". | |
131 | "+FOO" means "do print categories that start with FOO". | |
132 | ||
133 | Examples: --filter=-whitespace,+whitespace/braces | |
134 | --filter=-whitespace,-runtime/printf,+runtime/printf_format | |
135 | --filter=-,+build/include_what_you_use | |
136 | ||
137 | To see a list of all the categories used in cpplint, pass no arg: | |
138 | --filter= | |
139 | ||
140 | counting=total|toplevel|detailed | |
141 | The total number of errors found is always printed. If | |
142 | 'toplevel' is provided, then the count of errors in each of | |
143 | the top-level categories like 'build' and 'whitespace' will | |
144 | also be printed. If 'detailed' is provided, then a count | |
145 | is provided for each category like 'build/class'. | |
146 | ||
147 | repository=path | |
148 | The top level directory of the repository, used to derive the header | |
149 | guard CPP variable. By default, this is determined by searching for a | |
150 | path that contains .git, .hg, or .svn. When this flag is specified, the | |
151 | given path is used instead. This option allows the header guard CPP | |
152 | variable to remain consistent even if members of a team have different | |
153 | repository root directories (such as when checking out a subdirectory | |
154 | with SVN). In addition, users of non-mainstream version control systems | |
155 | can use this flag to ensure readable header guard CPP variables. | |
156 | ||
157 | Examples: | |
158 | Assuming that Alice checks out ProjectName and Bob checks out | |
159 | ProjectName/trunk and trunk contains src/chrome/ui/browser.h, then | |
160 | with no --repository flag, the header guard CPP variable will be: | |
161 | ||
162 | Alice => TRUNK_SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
163 | Bob => SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
164 | ||
165 | If Alice uses the --repository=trunk flag and Bob omits the flag or | |
166 | uses --repository=. then the header guard CPP variable will be: | |
167 | ||
168 | Alice => SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
169 | Bob => SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
170 | ||
171 | root=subdir | |
172 | The root directory used for deriving header guard CPP variable. | |
173 | This directory is relative to the top level directory of the repository | |
174 | which by default is determined by searching for a directory that contains | |
175 | .git, .hg, or .svn but can also be controlled with the --repository flag. | |
176 | If the specified directory does not exist, this flag is ignored. | |
177 | ||
178 | Examples: | |
179 | Assuming that src is the top level directory of the repository (and | |
180 | cwd=top/src), the header guard CPP variables for | |
181 | src/chrome/browser/ui/browser.h are: | |
182 | ||
183 | No flag => CHROME_BROWSER_UI_BROWSER_H_ | |
184 | --root=chrome => BROWSER_UI_BROWSER_H_ | |
185 | --root=chrome/browser => UI_BROWSER_H_ | |
186 | --root=.. => SRC_CHROME_BROWSER_UI_BROWSER_H_ | |
187 | ||
188 | linelength=digits | |
189 | This is the allowed line length for the project. The default value is | |
190 | 80 characters. | |
191 | ||
192 | Examples: | |
193 | --linelength=120 | |
194 | ||
195 | recursive | |
196 | Search for files to lint recursively. Each directory given in the list | |
197 | of files to be linted is replaced by all files that descend from that | |
198 | directory. Files with extensions not in the valid extensions list are | |
199 | excluded. | |
200 | ||
201 | exclude=path | |
202 | Exclude the given path from the list of files to be linted. Relative | |
203 | paths are evaluated relative to the current directory and shell globbing | |
204 | is performed. This flag can be provided multiple times to exclude | |
205 | multiple files. | |
206 | ||
207 | Examples: | |
208 | --exclude=one.cc | |
209 | --exclude=src/*.cc | |
210 | --exclude=src/*.cc --exclude=test/*.cc | |
211 | ||
212 | extensions=extension,extension,... | |
213 | The allowed file extensions that cpplint will check | |
214 | ||
215 | Examples: | |
216 | --extensions=%s | |
217 | ||
218 | includeorder=default|standardcfirst | |
219 | For the build/include_order rule, the default is to blindly assume angle | |
220 | bracket includes with file extension are c-system-headers (default), | |
221 | even knowing this will have false classifications. | |
222 | The default is established at google. | |
223 | standardcfirst means to instead use an allow-list of known c headers and | |
224 | treat all others as separate group of "other system headers". The C headers | |
225 | included are those of the C-standard lib and closely related ones. | |
226 | ||
227 | headers=x,y,... | |
228 | The header extensions that cpplint will treat as .h in checks. Values are | |
229 | automatically added to --extensions list. | |
230 | (by default, only files with extensions %s will be assumed to be headers) | |
231 | ||
232 | Examples: | |
233 | --headers=%s | |
234 | --headers=hpp,hxx | |
235 | --headers=hpp | |
236 | ||
237 | cpplint.py supports per-directory configurations specified in CPPLINT.cfg | |
238 | files. CPPLINT.cfg file can contain a number of key=value pairs. | |
239 | Currently the following options are supported: | |
240 | ||
241 | set noparent | |
242 | filter=+filter1,-filter2,... | |
243 | exclude_files=regex | |
244 | linelength=80 | |
245 | root=subdir | |
246 | headers=x,y,... | |
247 | ||
248 | "set noparent" option prevents cpplint from traversing directory tree | |
249 | upwards looking for more .cfg files in parent directories. This option | |
250 | is usually placed in the top-level project directory. | |
251 | ||
252 | The "filter" option is similar in function to --filter flag. It specifies | |
253 | message filters in addition to the |_DEFAULT_FILTERS| and those specified | |
254 | through --filter command-line flag. | |
255 | ||
256 | "exclude_files" allows to specify a regular expression to be matched against | |
257 | a file name. If the expression matches, the file is skipped and not run | |
258 | through the linter. | |
259 | ||
260 | "linelength" allows to specify the allowed line length for the project. | |
261 | ||
262 | The "root" option is similar in function to the --root flag (see example | |
263 | above). Paths are relative to the directory of the CPPLINT.cfg. | |
264 | ||
265 | The "headers" option is similar in function to the --headers flag | |
266 | (see example above). | |
267 | ||
268 | CPPLINT.cfg has an effect on files in the same directory and all | |
269 | sub-directories, unless overridden by a nested configuration file. | |
270 | ||
271 | Example file: | |
272 | filter=-build/include_order,+build/include_alpha | |
273 | exclude_files=.*\\.cc | |
274 | ||
275 | The above example disables build/include_order warning and enables | |
276 | build/include_alpha as well as excludes all .cc from being | |
277 | processed by linter, in the current directory (where the .cfg | |
278 | file is located) and all sub-directories. | |
279 | """ | |
280 | ||
281 | # We categorize each error message we print. Here are the categories. | |
282 | # We want an explicit list so we can list them all in cpplint --filter=. | |
283 | # If you add a new error message with a new category, add it to the list | |
284 | # here! cpplint_unittest.py should tell you if you forget to do this. | |
285 | _ERROR_CATEGORIES = [ | |
286 | 'build/class', | |
287 | 'build/c++11', | |
288 | 'build/c++14', | |
289 | 'build/c++tr1', | |
290 | 'build/deprecated', | |
291 | 'build/endif_comment', | |
292 | 'build/explicit_make_pair', | |
293 | 'build/forward_decl', | |
294 | 'build/header_guard', | |
295 | 'build/include', | |
296 | 'build/include_subdir', | |
297 | 'build/include_alpha', | |
298 | 'build/include_order', | |
299 | 'build/include_what_you_use', | |
300 | 'build/namespaces_headers', | |
301 | 'build/namespaces_literals', | |
302 | 'build/namespaces', | |
303 | 'build/printf_format', | |
304 | 'build/storage_class', | |
305 | 'legal/copyright', | |
306 | 'readability/alt_tokens', | |
307 | 'readability/braces', | |
308 | 'readability/casting', | |
309 | 'readability/check', | |
310 | 'readability/constructors', | |
311 | 'readability/fn_size', | |
312 | 'readability/inheritance', | |
313 | 'readability/multiline_comment', | |
314 | 'readability/multiline_string', | |
315 | 'readability/namespace', | |
316 | 'readability/nolint', | |
317 | 'readability/nul', | |
318 | 'readability/strings', | |
319 | 'readability/todo', | |
320 | 'readability/utf8', | |
321 | 'runtime/arrays', | |
322 | 'runtime/casting', | |
323 | 'runtime/explicit', | |
324 | 'runtime/int', | |
325 | 'runtime/init', | |
326 | 'runtime/invalid_increment', | |
327 | 'runtime/member_string_references', | |
328 | 'runtime/memset', | |
329 | 'runtime/indentation_namespace', | |
330 | 'runtime/operator', | |
331 | 'runtime/printf', | |
332 | 'runtime/printf_format', | |
333 | 'runtime/references', | |
334 | 'runtime/string', | |
335 | 'runtime/threadsafe_fn', | |
336 | 'runtime/vlog', | |
337 | 'whitespace/blank_line', | |
338 | 'whitespace/braces', | |
339 | 'whitespace/comma', | |
340 | 'whitespace/comments', | |
341 | 'whitespace/empty_conditional_body', | |
342 | 'whitespace/empty_if_body', | |
343 | 'whitespace/empty_loop_body', | |
344 | 'whitespace/end_of_line', | |
345 | 'whitespace/ending_newline', | |
346 | 'whitespace/forcolon', | |
347 | 'whitespace/indent', | |
348 | 'whitespace/line_length', | |
349 | 'whitespace/newline', | |
350 | 'whitespace/operators', | |
351 | 'whitespace/parens', | |
352 | 'whitespace/semicolon', | |
353 | 'whitespace/tab', | |
354 | 'whitespace/todo', | |
355 | ] | |
356 | ||
357 | # keywords to use with --outputs which generate stdout for machine processing | |
358 | _MACHINE_OUTPUTS = [ | |
359 | 'junit', | |
360 | 'sed', | |
361 | 'gsed' | |
362 | ] | |
363 | ||
364 | # These error categories are no longer enforced by cpplint, but for backwards- | |
365 | # compatibility they may still appear in NOLINT comments. | |
366 | _LEGACY_ERROR_CATEGORIES = [ | |
367 | 'readability/streams', | |
368 | 'readability/function', | |
369 | ] | |
370 | ||
371 | # The default state of the category filter. This is overridden by the --filter= | |
372 | # flag. By default all errors are on, so only add here categories that should be | |
373 | # off by default (i.e., categories that must be enabled by the --filter= flags). | |
374 | # All entries here should start with a '-' or '+', as in the --filter= flag. | |
375 | _DEFAULT_FILTERS = ['-build/include_alpha'] | |
376 | ||
377 | # The default list of categories suppressed for C (not C++) files. | |
378 | _DEFAULT_C_SUPPRESSED_CATEGORIES = [ | |
379 | 'readability/casting', | |
380 | ] | |
381 | ||
382 | # The default list of categories suppressed for Linux Kernel files. | |
383 | _DEFAULT_KERNEL_SUPPRESSED_CATEGORIES = [ | |
384 | 'whitespace/tab', | |
385 | ] | |
386 | ||
387 | # We used to check for high-bit characters, but after much discussion we | |
388 | # decided those were OK, as long as they were in UTF-8 and didn't represent | |
389 | # hard-coded international strings, which belong in a separate i18n file. | |
390 | ||
391 | # C++ headers | |
392 | _CPP_HEADERS = frozenset([ | |
393 | # Legacy | |
394 | 'algobase.h', | |
395 | 'algo.h', | |
396 | 'alloc.h', | |
397 | 'builtinbuf.h', | |
398 | 'bvector.h', | |
399 | 'complex.h', | |
400 | 'defalloc.h', | |
401 | 'deque.h', | |
402 | 'editbuf.h', | |
403 | 'fstream.h', | |
404 | 'function.h', | |
405 | 'hash_map', | |
406 | 'hash_map.h', | |
407 | 'hash_set', | |
408 | 'hash_set.h', | |
409 | 'hashtable.h', | |
410 | 'heap.h', | |
411 | 'indstream.h', | |
412 | 'iomanip.h', | |
413 | 'iostream.h', | |
414 | 'istream.h', | |
415 | 'iterator.h', | |
416 | 'list.h', | |
417 | 'map.h', | |
418 | 'multimap.h', | |
419 | 'multiset.h', | |
420 | 'ostream.h', | |
421 | 'pair.h', | |
422 | 'parsestream.h', | |
423 | 'pfstream.h', | |
424 | 'procbuf.h', | |
425 | 'pthread_alloc', | |
426 | 'pthread_alloc.h', | |
427 | 'rope', | |
428 | 'rope.h', | |
429 | 'ropeimpl.h', | |
430 | 'set.h', | |
431 | 'slist', | |
432 | 'slist.h', | |
433 | 'stack.h', | |
434 | 'stdiostream.h', | |
435 | 'stl_alloc.h', | |
436 | 'stl_relops.h', | |
437 | 'streambuf.h', | |
438 | 'stream.h', | |
439 | 'strfile.h', | |
440 | 'strstream.h', | |
441 | 'tempbuf.h', | |
442 | 'tree.h', | |
443 | 'type_traits.h', | |
444 | 'vector.h', | |
445 | # 17.6.1.2 C++ library headers | |
446 | 'algorithm', | |
447 | 'array', | |
448 | 'atomic', | |
449 | 'bitset', | |
450 | 'chrono', | |
451 | 'codecvt', | |
452 | 'complex', | |
453 | 'condition_variable', | |
454 | 'deque', | |
455 | 'exception', | |
456 | 'forward_list', | |
457 | 'fstream', | |
458 | 'functional', | |
459 | 'future', | |
460 | 'initializer_list', | |
461 | 'iomanip', | |
462 | 'ios', | |
463 | 'iosfwd', | |
464 | 'iostream', | |
465 | 'istream', | |
466 | 'iterator', | |
467 | 'limits', | |
468 | 'list', | |
469 | 'locale', | |
470 | 'map', | |
471 | 'memory', | |
472 | 'mutex', | |
473 | 'new', | |
474 | 'numeric', | |
475 | 'ostream', | |
476 | 'queue', | |
477 | 'random', | |
478 | 'ratio', | |
479 | 'regex', | |
480 | 'scoped_allocator', | |
481 | 'set', | |
482 | 'sstream', | |
483 | 'stack', | |
484 | 'stdexcept', | |
485 | 'streambuf', | |
486 | 'string', | |
487 | 'strstream', | |
488 | 'system_error', | |
489 | 'thread', | |
490 | 'tuple', | |
491 | 'typeindex', | |
492 | 'typeinfo', | |
493 | 'type_traits', | |
494 | 'unordered_map', | |
495 | 'unordered_set', | |
496 | 'utility', | |
497 | 'valarray', | |
498 | 'vector', | |
499 | # 17.6.1.2 C++14 headers | |
500 | 'shared_mutex', | |
501 | # 17.6.1.2 C++17 headers | |
502 | 'any', | |
503 | 'charconv', | |
504 | 'codecvt', | |
505 | 'execution', | |
506 | 'filesystem', | |
507 | 'memory_resource', | |
508 | 'optional', | |
509 | 'string_view', | |
510 | 'variant', | |
511 | # 17.6.1.2 C++ headers for C library facilities | |
512 | 'cassert', | |
513 | 'ccomplex', | |
514 | 'cctype', | |
515 | 'cerrno', | |
516 | 'cfenv', | |
517 | 'cfloat', | |
518 | 'cinttypes', | |
519 | 'ciso646', | |
520 | 'climits', | |
521 | 'clocale', | |
522 | 'cmath', | |
523 | 'csetjmp', | |
524 | 'csignal', | |
525 | 'cstdalign', | |
526 | 'cstdarg', | |
527 | 'cstdbool', | |
528 | 'cstddef', | |
529 | 'cstdint', | |
530 | 'cstdio', | |
531 | 'cstdlib', | |
532 | 'cstring', | |
533 | 'ctgmath', | |
534 | 'ctime', | |
535 | 'cuchar', | |
536 | 'cwchar', | |
537 | 'cwctype', | |
538 | ]) | |
539 | ||
540 | # C headers | |
541 | _C_HEADERS = frozenset([ | |
542 | # System C headers | |
543 | 'assert.h', | |
544 | 'complex.h', | |
545 | 'ctype.h', | |
546 | 'errno.h', | |
547 | 'fenv.h', | |
548 | 'float.h', | |
549 | 'inttypes.h', | |
550 | 'iso646.h', | |
551 | 'limits.h', | |
552 | 'locale.h', | |
553 | 'math.h', | |
554 | 'setjmp.h', | |
555 | 'signal.h', | |
556 | 'stdalign.h', | |
557 | 'stdarg.h', | |
558 | 'stdatomic.h', | |
559 | 'stdbool.h', | |
560 | 'stddef.h', | |
561 | 'stdint.h', | |
562 | 'stdio.h', | |
563 | 'stdlib.h', | |
564 | 'stdnoreturn.h', | |
565 | 'string.h', | |
566 | 'tgmath.h', | |
567 | 'threads.h', | |
568 | 'time.h', | |
569 | 'uchar.h', | |
570 | 'wchar.h', | |
571 | 'wctype.h', | |
572 | # additional POSIX C headers | |
573 | 'aio.h', | |
574 | 'arpa/inet.h', | |
575 | 'cpio.h', | |
576 | 'dirent.h', | |
577 | 'dlfcn.h', | |
578 | 'fcntl.h', | |
579 | 'fmtmsg.h', | |
580 | 'fnmatch.h', | |
581 | 'ftw.h', | |
582 | 'glob.h', | |
583 | 'grp.h', | |
584 | 'iconv.h', | |
585 | 'langinfo.h', | |
586 | 'libgen.h', | |
587 | 'monetary.h', | |
588 | 'mqueue.h', | |
589 | 'ndbm.h', | |
590 | 'net/if.h', | |
591 | 'netdb.h', | |
592 | 'netinet/in.h', | |
593 | 'netinet/tcp.h', | |
594 | 'nl_types.h', | |
595 | 'poll.h', | |
596 | 'pthread.h', | |
597 | 'pwd.h', | |
598 | 'regex.h', | |
599 | 'sched.h', | |
600 | 'search.h', | |
601 | 'semaphore.h', | |
602 | 'setjmp.h', | |
603 | 'signal.h', | |
604 | 'spawn.h', | |
605 | 'strings.h', | |
606 | 'stropts.h', | |
607 | 'syslog.h', | |
608 | 'tar.h', | |
609 | 'termios.h', | |
610 | 'trace.h', | |
611 | 'ulimit.h', | |
612 | 'unistd.h', | |
613 | 'utime.h', | |
614 | 'utmpx.h', | |
615 | 'wordexp.h', | |
616 | # additional GNUlib headers | |
617 | 'a.out.h', | |
618 | 'aliases.h', | |
619 | 'alloca.h', | |
620 | 'ar.h', | |
621 | 'argp.h', | |
622 | 'argz.h', | |
623 | 'byteswap.h', | |
624 | 'crypt.h', | |
625 | 'endian.h', | |
626 | 'envz.h', | |
627 | 'err.h', | |
628 | 'error.h', | |
629 | 'execinfo.h', | |
630 | 'fpu_control.h', | |
631 | 'fstab.h', | |
632 | 'fts.h', | |
633 | 'getopt.h', | |
634 | 'gshadow.h', | |
635 | 'ieee754.h', | |
636 | 'ifaddrs.h', | |
637 | 'libintl.h', | |
638 | 'mcheck.h', | |
639 | 'mntent.h', | |
640 | 'obstack.h', | |
641 | 'paths.h', | |
642 | 'printf.h', | |
643 | 'pty.h', | |
644 | 'resolv.h', | |
645 | 'shadow.h', | |
646 | 'sysexits.h', | |
647 | 'ttyent.h', | |
648 | # Additional linux glibc headers | |
649 | 'dlfcn.h', | |
650 | 'elf.h', | |
651 | 'features.h', | |
652 | 'gconv.h', | |
653 | 'gnu-versions.h', | |
654 | 'lastlog.h', | |
655 | 'libio.h', | |
656 | 'link.h', | |
657 | 'malloc.h', | |
658 | 'memory.h', | |
659 | 'netash/ash.h', | |
660 | 'netatalk/at.h', | |
661 | 'netax25/ax25.h', | |
662 | 'neteconet/ec.h', | |
663 | 'netipx/ipx.h', | |
664 | 'netiucv/iucv.h', | |
665 | 'netpacket/packet.h', | |
666 | 'netrom/netrom.h', | |
667 | 'netrose/rose.h', | |
668 | 'nfs/nfs.h', | |
669 | 'nl_types.h', | |
670 | 'nss.h', | |
671 | 're_comp.h', | |
672 | 'regexp.h', | |
673 | 'sched.h', | |
674 | 'sgtty.h', | |
675 | 'stab.h', | |
676 | 'stdc-predef.h', | |
677 | 'stdio_ext.h', | |
678 | 'syscall.h', | |
679 | 'termio.h', | |
680 | 'thread_db.h', | |
681 | 'ucontext.h', | |
682 | 'ustat.h', | |
683 | 'utmp.h', | |
684 | 'values.h', | |
685 | 'wait.h', | |
686 | 'xlocale.h', | |
687 | # Hardware specific headers | |
688 | 'arm_neon.h', | |
689 | 'emmintrin.h', | |
690 | 'xmmintin.h', | |
691 | ]) | |
692 | ||
693 | # Folders of C libraries so commonly used in C++, | |
694 | # that they have parity with standard C libraries. | |
695 | C_STANDARD_HEADER_FOLDERS = frozenset([ | |
696 | # standard C library | |
697 | "sys", | |
698 | # glibc for linux | |
699 | "arpa", | |
700 | "asm-generic", | |
701 | "bits", | |
702 | "gnu", | |
703 | "net", | |
704 | "netinet", | |
705 | "protocols", | |
706 | "rpc", | |
707 | "rpcsvc", | |
708 | "scsi", | |
709 | # linux kernel header | |
710 | "drm", | |
711 | "linux", | |
712 | "misc", | |
713 | "mtd", | |
714 | "rdma", | |
715 | "sound", | |
716 | "video", | |
717 | "xen", | |
718 | ]) | |
719 | ||
720 | # Type names | |
721 | _TYPES = re.compile( | |
722 | r'^(?:' | |
723 | # [dcl.type.simple] | |
724 | r'(char(16_t|32_t)?)|wchar_t|' | |
725 | r'bool|short|int|long|signed|unsigned|float|double|' | |
726 | # [support.types] | |
727 | r'(ptrdiff_t|size_t|max_align_t|nullptr_t)|' | |
728 | # [cstdint.syn] | |
729 | r'(u?int(_fast|_least)?(8|16|32|64)_t)|' | |
730 | r'(u?int(max|ptr)_t)|' | |
731 | r')$') | |
732 | ||
733 | ||
734 | # These headers are excluded from [build/include] and [build/include_order] | |
735 | # checks: | |
736 | # - Anything not following google file name conventions (containing an | |
737 | # uppercase character, such as Python.h or nsStringAPI.h, for example). | |
738 | # - Lua headers. | |
739 | _THIRD_PARTY_HEADERS_PATTERN = re.compile( | |
740 | r'^(?:[^/]*[A-Z][^/]*\.h|lua\.h|lauxlib\.h|lualib\.h)$') | |
741 | ||
742 | # Pattern for matching FileInfo.BaseName() against test file name | |
743 | _test_suffixes = ['_test', '_regtest', '_unittest'] | |
744 | _TEST_FILE_SUFFIX = '(' + '|'.join(_test_suffixes) + r')$' | |
745 | ||
746 | # Pattern that matches only complete whitespace, possibly across multiple lines. | |
747 | _EMPTY_CONDITIONAL_BODY_PATTERN = re.compile(r'^\s*$', re.DOTALL) | |
748 | ||
749 | # Assertion macros. These are defined in base/logging.h and | |
750 | # testing/base/public/gunit.h. | |
751 | _CHECK_MACROS = [ | |
752 | 'DCHECK', 'CHECK', | |
753 | 'EXPECT_TRUE', 'ASSERT_TRUE', | |
754 | 'EXPECT_FALSE', 'ASSERT_FALSE', | |
755 | ] | |
756 | ||
757 | # Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE | |
758 | _CHECK_REPLACEMENT = dict([(macro_var, {}) for macro_var in _CHECK_MACROS]) | |
759 | ||
760 | for op, replacement in [('==', 'EQ'), ('!=', 'NE'), | |
761 | ('>=', 'GE'), ('>', 'GT'), | |
762 | ('<=', 'LE'), ('<', 'LT')]: | |
763 | _CHECK_REPLACEMENT['DCHECK'][op] = 'DCHECK_%s' % replacement | |
764 | _CHECK_REPLACEMENT['CHECK'][op] = 'CHECK_%s' % replacement | |
765 | _CHECK_REPLACEMENT['EXPECT_TRUE'][op] = 'EXPECT_%s' % replacement | |
766 | _CHECK_REPLACEMENT['ASSERT_TRUE'][op] = 'ASSERT_%s' % replacement | |
767 | ||
768 | for op, inv_replacement in [('==', 'NE'), ('!=', 'EQ'), | |
769 | ('>=', 'LT'), ('>', 'LE'), | |
770 | ('<=', 'GT'), ('<', 'GE')]: | |
771 | _CHECK_REPLACEMENT['EXPECT_FALSE'][op] = 'EXPECT_%s' % inv_replacement | |
772 | _CHECK_REPLACEMENT['ASSERT_FALSE'][op] = 'ASSERT_%s' % inv_replacement | |
773 | ||
774 | # Alternative tokens and their replacements. For full list, see section 2.5 | |
775 | # Alternative tokens [lex.digraph] in the C++ standard. | |
776 | # | |
777 | # Digraphs (such as '%:') are not included here since it's a mess to | |
778 | # match those on a word boundary. | |
779 | _ALT_TOKEN_REPLACEMENT = { | |
780 | 'and': '&&', | |
781 | 'bitor': '|', | |
782 | 'or': '||', | |
783 | 'xor': '^', | |
784 | 'compl': '~', | |
785 | 'bitand': '&', | |
786 | 'and_eq': '&=', | |
787 | 'or_eq': '|=', | |
788 | 'xor_eq': '^=', | |
789 | 'not': '!', | |
790 | 'not_eq': '!=' | |
791 | } | |
792 | ||
793 | # Compile regular expression that matches all the above keywords. The "[ =()]" | |
794 | # bit is meant to avoid matching these keywords outside of boolean expressions. | |
795 | # | |
796 | # False positives include C-style multi-line comments and multi-line strings | |
797 | # but those have always been troublesome for cpplint. | |
798 | _ALT_TOKEN_REPLACEMENT_PATTERN = re.compile( | |
799 | r'[ =()](' + ('|'.join(_ALT_TOKEN_REPLACEMENT.keys())) + r')(?=[ (]|$)') | |
800 | ||
801 | ||
802 | # These constants define types of headers for use with | |
803 | # _IncludeState.CheckNextIncludeOrder(). | |
804 | _C_SYS_HEADER = 1 | |
805 | _CPP_SYS_HEADER = 2 | |
806 | _OTHER_SYS_HEADER = 3 | |
807 | _LIKELY_MY_HEADER = 4 | |
808 | _POSSIBLE_MY_HEADER = 5 | |
809 | _OTHER_HEADER = 6 | |
810 | ||
811 | # These constants define the current inline assembly state | |
812 | _NO_ASM = 0 # Outside of inline assembly block | |
813 | _INSIDE_ASM = 1 # Inside inline assembly block | |
814 | _END_ASM = 2 # Last line of inline assembly block | |
815 | _BLOCK_ASM = 3 # The whole block is an inline assembly block | |
816 | ||
817 | # Match start of assembly blocks | |
818 | _MATCH_ASM = re.compile(r'^\s*(?:asm|_asm|__asm|__asm__)' | |
819 | r'(?:\s+(volatile|__volatile__))?' | |
820 | r'\s*[{(]') | |
821 | ||
822 | # Match strings that indicate we're working on a C (not C++) file. | |
823 | _SEARCH_C_FILE = re.compile(r'\b(?:LINT_C_FILE|' | |
824 | r'vim?:\s*.*(\s*|:)filetype=c(\s*|:|$))') | |
825 | ||
826 | # Match string that indicates we're working on a Linux Kernel file. | |
827 | _SEARCH_KERNEL_FILE = re.compile(r'\b(?:LINT_KERNEL_FILE)') | |
828 | ||
829 | # Commands for sed to fix the problem | |
830 | _SED_FIXUPS = { | |
831 | 'Remove spaces around =': r's/ = /=/', | |
832 | 'Remove spaces around !=': r's/ != /!=/', | |
833 | 'Remove space before ( in if (': r's/if (/if(/', | |
834 | 'Remove space before ( in for (': r's/for (/for(/', | |
835 | 'Remove space before ( in while (': r's/while (/while(/', | |
836 | 'Remove space before ( in switch (': r's/switch (/switch(/', | |
837 | 'Should have a space between // and comment': r's/\/\//\/\/ /', | |
838 | 'Missing space before {': r's/\([^ ]\){/\1 {/', | |
839 | 'Tab found, replace by spaces': r's/\t/ /g', | |
840 | 'Line ends in whitespace. Consider deleting these extra spaces.': r's/\s*$//', | |
841 | 'You don\'t need a ; after a }': r's/};/}/', | |
842 | 'Missing space after ,': r's/,\([^ ]\)/, \1/g', | |
843 | } | |
844 | ||
845 | _regexp_compile_cache = {} | |
846 | ||
847 | # {str, set(int)}: a map from error categories to sets of linenumbers | |
848 | # on which those errors are expected and should be suppressed. | |
849 | _error_suppressions = {} | |
850 | ||
851 | # The root directory used for deriving header guard CPP variable. | |
852 | # This is set by --root flag. | |
853 | _root = None | |
854 | _root_debug = False | |
855 | ||
856 | # The top level repository directory. If set, _root is calculated relative to | |
857 | # this directory instead of the directory containing version control artifacts. | |
858 | # This is set by the --repository flag. | |
859 | _repository = None | |
860 | ||
861 | # Files to exclude from linting. This is set by the --exclude flag. | |
862 | _excludes = None | |
863 | ||
864 | # Whether to supress all PrintInfo messages, UNRELATED to --quiet flag | |
865 | _quiet = False | |
866 | ||
867 | # The allowed line length of files. | |
868 | # This is set by --linelength flag. | |
869 | _line_length = 80 | |
870 | ||
871 | # This allows to use different include order rule than default | |
872 | _include_order = "default" | |
873 | ||
874 | try: | |
875 | unicode | |
876 | except NameError: | |
877 | # -- pylint: disable=redefined-builtin | |
878 | basestring = unicode = str | |
879 | ||
880 | try: | |
881 | long | |
882 | except NameError: | |
883 | # -- pylint: disable=redefined-builtin | |
884 | long = int | |
885 | ||
886 | if sys.version_info < (3,): | |
887 | # -- pylint: disable=no-member | |
888 | # BINARY_TYPE = str | |
889 | itervalues = dict.itervalues | |
890 | iteritems = dict.iteritems | |
891 | else: | |
892 | # BINARY_TYPE = bytes | |
893 | itervalues = dict.values | |
894 | iteritems = dict.items | |
895 | ||
896 | def unicode_escape_decode(x): | |
897 | if sys.version_info < (3,): | |
898 | return codecs.unicode_escape_decode(x)[0] | |
899 | else: | |
900 | return x | |
901 | ||
902 | # Treat all headers starting with 'h' equally: .h, .hpp, .hxx etc. | |
903 | # This is set by --headers flag. | |
904 | _hpp_headers = set([]) | |
905 | ||
906 | # {str, bool}: a map from error categories to booleans which indicate if the | |
907 | # category should be suppressed for every line. | |
908 | _global_error_suppressions = {} | |
909 | ||
910 | def ProcessHppHeadersOption(val): | |
911 | global _hpp_headers | |
912 | try: | |
913 | _hpp_headers = {ext.strip() for ext in val.split(',')} | |
914 | except ValueError: | |
915 | PrintUsage('Header extensions must be comma separated list.') | |
916 | ||
917 | def ProcessIncludeOrderOption(val): | |
918 | if val is None or val == "default": | |
919 | pass | |
920 | elif val == "standardcfirst": | |
921 | global _include_order | |
922 | _include_order = val | |
923 | else: | |
924 | PrintUsage('Invalid includeorder value %s. Expected default|standardcfirst') | |
925 | ||
926 | def IsHeaderExtension(file_extension): | |
927 | return file_extension in GetHeaderExtensions() | |
928 | ||
929 | def GetHeaderExtensions(): | |
930 | if _hpp_headers: | |
931 | return _hpp_headers | |
932 | if _valid_extensions: | |
933 | return {h for h in _valid_extensions if 'h' in h} | |
934 | return set(['h', 'hh', 'hpp', 'hxx', 'h++', 'cuh']) | |
935 | ||
936 | # The allowed extensions for file names | |
937 | # This is set by --extensions flag | |
938 | def GetAllExtensions(): | |
939 | return GetHeaderExtensions().union(_valid_extensions or set( | |
940 | ['c', 'cc', 'cpp', 'cxx', 'c++', 'cu'])) | |
941 | ||
942 | def ProcessExtensionsOption(val): | |
943 | global _valid_extensions | |
944 | try: | |
945 | extensions = [ext.strip() for ext in val.split(',')] | |
946 | _valid_extensions = set(extensions) | |
947 | except ValueError: | |
948 | PrintUsage('Extensions should be a comma-separated list of values;' | |
949 | 'for example: extensions=hpp,cpp\n' | |
950 | 'This could not be parsed: "%s"' % (val,)) | |
951 | ||
952 | def GetNonHeaderExtensions(): | |
953 | return GetAllExtensions().difference(GetHeaderExtensions()) | |
954 | ||
955 | def ParseNolintSuppressions(filename, raw_line, linenum, error): | |
956 | """Updates the global list of line error-suppressions. | |
957 | ||
958 | Parses any NOLINT comments on the current line, updating the global | |
959 | error_suppressions store. Reports an error if the NOLINT comment | |
960 | was malformed. | |
961 | ||
962 | Args: | |
963 | filename: str, the name of the input file. | |
964 | raw_line: str, the line of input text, with comments. | |
965 | linenum: int, the number of the current line. | |
966 | error: function, an error handler. | |
967 | """ | |
968 | matched = Search(r'\bNOLINT(NEXTLINE)?\b(\([^)]+\))?', raw_line) | |
969 | if matched: | |
970 | if matched.group(1): | |
971 | suppressed_line = linenum + 1 | |
972 | else: | |
973 | suppressed_line = linenum | |
974 | category = matched.group(2) | |
975 | if category in (None, '(*)'): # => "suppress all" | |
976 | _error_suppressions.setdefault(None, set()).add(suppressed_line) | |
977 | else: | |
978 | if category.startswith('(') and category.endswith(')'): | |
979 | category = category[1:-1] | |
980 | if category in _ERROR_CATEGORIES: | |
981 | _error_suppressions.setdefault(category, set()).add(suppressed_line) | |
982 | elif category not in _LEGACY_ERROR_CATEGORIES: | |
983 | error(filename, linenum, 'readability/nolint', 5, | |
984 | 'Unknown NOLINT error category: %s' % category) | |
985 | ||
986 | ||
987 | def ProcessGlobalSuppresions(lines): | |
988 | """Updates the list of global error suppressions. | |
989 | ||
990 | Parses any lint directives in the file that have global effect. | |
991 | ||
992 | Args: | |
993 | lines: An array of strings, each representing a line of the file, with the | |
994 | last element being empty if the file is terminated with a newline. | |
995 | """ | |
996 | for line in lines: | |
997 | if _SEARCH_C_FILE.search(line): | |
998 | for category in _DEFAULT_C_SUPPRESSED_CATEGORIES: | |
999 | _global_error_suppressions[category] = True | |
1000 | if _SEARCH_KERNEL_FILE.search(line): | |
1001 | for category in _DEFAULT_KERNEL_SUPPRESSED_CATEGORIES: | |
1002 | _global_error_suppressions[category] = True | |
1003 | ||
1004 | ||
1005 | def ResetNolintSuppressions(): | |
1006 | """Resets the set of NOLINT suppressions to empty.""" | |
1007 | _error_suppressions.clear() | |
1008 | _global_error_suppressions.clear() | |
1009 | ||
1010 | ||
1011 | def IsErrorSuppressedByNolint(category, linenum): | |
1012 | """Returns true if the specified error category is suppressed on this line. | |
1013 | ||
1014 | Consults the global error_suppressions map populated by | |
1015 | ParseNolintSuppressions/ProcessGlobalSuppresions/ResetNolintSuppressions. | |
1016 | ||
1017 | Args: | |
1018 | category: str, the category of the error. | |
1019 | linenum: int, the current line number. | |
1020 | Returns: | |
1021 | bool, True iff the error should be suppressed due to a NOLINT comment or | |
1022 | global suppression. | |
1023 | """ | |
1024 | return (_global_error_suppressions.get(category, False) or | |
1025 | linenum in _error_suppressions.get(category, set()) or | |
1026 | linenum in _error_suppressions.get(None, set())) | |
1027 | ||
1028 | ||
1029 | def Match(pattern, s): | |
1030 | """Matches the string with the pattern, caching the compiled regexp.""" | |
1031 | # The regexp compilation caching is inlined in both Match and Search for | |
1032 | # performance reasons; factoring it out into a separate function turns out | |
1033 | # to be noticeably expensive. | |
1034 | if pattern not in _regexp_compile_cache: | |
1035 | _regexp_compile_cache[pattern] = sre_compile.compile(pattern) | |
1036 | return _regexp_compile_cache[pattern].match(s) | |
1037 | ||
1038 | ||
1039 | def ReplaceAll(pattern, rep, s): | |
1040 | """Replaces instances of pattern in a string with a replacement. | |
1041 | ||
1042 | The compiled regex is kept in a cache shared by Match and Search. | |
1043 | ||
1044 | Args: | |
1045 | pattern: regex pattern | |
1046 | rep: replacement text | |
1047 | s: search string | |
1048 | ||
1049 | Returns: | |
1050 | string with replacements made (or original string if no replacements) | |
1051 | """ | |
1052 | if pattern not in _regexp_compile_cache: | |
1053 | _regexp_compile_cache[pattern] = sre_compile.compile(pattern) | |
1054 | return _regexp_compile_cache[pattern].sub(rep, s) | |
1055 | ||
1056 | ||
1057 | def Search(pattern, s): | |
1058 | """Searches the string for the pattern, caching the compiled regexp.""" | |
1059 | if pattern not in _regexp_compile_cache: | |
1060 | _regexp_compile_cache[pattern] = sre_compile.compile(pattern) | |
1061 | return _regexp_compile_cache[pattern].search(s) | |
1062 | ||
1063 | ||
1064 | def _IsSourceExtension(s): | |
1065 | """File extension (excluding dot) matches a source file extension.""" | |
1066 | return s in GetNonHeaderExtensions() | |
1067 | ||
1068 | ||
1069 | class _IncludeState(object): | |
1070 | """Tracks line numbers for includes, and the order in which includes appear. | |
1071 | ||
1072 | include_list contains list of lists of (header, line number) pairs. | |
1073 | It's a lists of lists rather than just one flat list to make it | |
1074 | easier to update across preprocessor boundaries. | |
1075 | ||
1076 | Call CheckNextIncludeOrder() once for each header in the file, passing | |
1077 | in the type constants defined above. Calls in an illegal order will | |
1078 | raise an _IncludeError with an appropriate error message. | |
1079 | ||
1080 | """ | |
1081 | # self._section will move monotonically through this set. If it ever | |
1082 | # needs to move backwards, CheckNextIncludeOrder will raise an error. | |
1083 | _INITIAL_SECTION = 0 | |
1084 | _MY_H_SECTION = 1 | |
1085 | _C_SECTION = 2 | |
1086 | _CPP_SECTION = 3 | |
1087 | _OTHER_SYS_SECTION = 4 | |
1088 | _OTHER_H_SECTION = 5 | |
1089 | ||
1090 | _TYPE_NAMES = { | |
1091 | _C_SYS_HEADER: 'C system header', | |
1092 | _CPP_SYS_HEADER: 'C++ system header', | |
1093 | _OTHER_SYS_HEADER: 'other system header', | |
1094 | _LIKELY_MY_HEADER: 'header this file implements', | |
1095 | _POSSIBLE_MY_HEADER: 'header this file may implement', | |
1096 | _OTHER_HEADER: 'other header', | |
1097 | } | |
1098 | _SECTION_NAMES = { | |
1099 | _INITIAL_SECTION: "... nothing. (This can't be an error.)", | |
1100 | _MY_H_SECTION: 'a header this file implements', | |
1101 | _C_SECTION: 'C system header', | |
1102 | _CPP_SECTION: 'C++ system header', | |
1103 | _OTHER_SYS_SECTION: 'other system header', | |
1104 | _OTHER_H_SECTION: 'other header', | |
1105 | } | |
1106 | ||
1107 | def __init__(self): | |
1108 | self.include_list = [[]] | |
1109 | self._section = None | |
1110 | self._last_header = None | |
1111 | self.ResetSection('') | |
1112 | ||
1113 | def FindHeader(self, header): | |
1114 | """Check if a header has already been included. | |
1115 | ||
1116 | Args: | |
1117 | header: header to check. | |
1118 | Returns: | |
1119 | Line number of previous occurrence, or -1 if the header has not | |
1120 | been seen before. | |
1121 | """ | |
1122 | for section_list in self.include_list: | |
1123 | for f in section_list: | |
1124 | if f[0] == header: | |
1125 | return f[1] | |
1126 | return -1 | |
1127 | ||
1128 | def ResetSection(self, directive): | |
1129 | """Reset section checking for preprocessor directive. | |
1130 | ||
1131 | Args: | |
1132 | directive: preprocessor directive (e.g. "if", "else"). | |
1133 | """ | |
1134 | # The name of the current section. | |
1135 | self._section = self._INITIAL_SECTION | |
1136 | # The path of last found header. | |
1137 | self._last_header = '' | |
1138 | ||
1139 | # Update list of includes. Note that we never pop from the | |
1140 | # include list. | |
1141 | if directive in ('if', 'ifdef', 'ifndef'): | |
1142 | self.include_list.append([]) | |
1143 | elif directive in ('else', 'elif'): | |
1144 | self.include_list[-1] = [] | |
1145 | ||
1146 | def SetLastHeader(self, header_path): | |
1147 | self._last_header = header_path | |
1148 | ||
1149 | def CanonicalizeAlphabeticalOrder(self, header_path): | |
1150 | """Returns a path canonicalized for alphabetical comparison. | |
1151 | ||
1152 | - replaces "-" with "_" so they both cmp the same. | |
1153 | - removes '-inl' since we don't require them to be after the main header. | |
1154 | - lowercase everything, just in case. | |
1155 | ||
1156 | Args: | |
1157 | header_path: Path to be canonicalized. | |
1158 | ||
1159 | Returns: | |
1160 | Canonicalized path. | |
1161 | """ | |
1162 | return header_path.replace('-inl.h', '.h').replace('-', '_').lower() | |
1163 | ||
1164 | def IsInAlphabeticalOrder(self, clean_lines, linenum, header_path): | |
1165 | """Check if a header is in alphabetical order with the previous header. | |
1166 | ||
1167 | Args: | |
1168 | clean_lines: A CleansedLines instance containing the file. | |
1169 | linenum: The number of the line to check. | |
1170 | header_path: Canonicalized header to be checked. | |
1171 | ||
1172 | Returns: | |
1173 | Returns true if the header is in alphabetical order. | |
1174 | """ | |
1175 | # If previous section is different from current section, _last_header will | |
1176 | # be reset to empty string, so it's always less than current header. | |
1177 | # | |
1178 | # If previous line was a blank line, assume that the headers are | |
1179 | # intentionally sorted the way they are. | |
1180 | if (self._last_header > header_path and | |
1181 | Match(r'^\s*#\s*include\b', clean_lines.elided[linenum - 1])): | |
1182 | return False | |
1183 | return True | |
1184 | ||
1185 | def CheckNextIncludeOrder(self, header_type): | |
1186 | """Returns a non-empty error message if the next header is out of order. | |
1187 | ||
1188 | This function also updates the internal state to be ready to check | |
1189 | the next include. | |
1190 | ||
1191 | Args: | |
1192 | header_type: One of the _XXX_HEADER constants defined above. | |
1193 | ||
1194 | Returns: | |
1195 | The empty string if the header is in the right order, or an | |
1196 | error message describing what's wrong. | |
1197 | ||
1198 | """ | |
1199 | error_message = ('Found %s after %s' % | |
1200 | (self._TYPE_NAMES[header_type], | |
1201 | self._SECTION_NAMES[self._section])) | |
1202 | ||
1203 | last_section = self._section | |
1204 | ||
1205 | if header_type == _C_SYS_HEADER: | |
1206 | if self._section <= self._C_SECTION: | |
1207 | self._section = self._C_SECTION | |
1208 | else: | |
1209 | self._last_header = '' | |
1210 | return error_message | |
1211 | elif header_type == _CPP_SYS_HEADER: | |
1212 | if self._section <= self._CPP_SECTION: | |
1213 | self._section = self._CPP_SECTION | |
1214 | else: | |
1215 | self._last_header = '' | |
1216 | return error_message | |
1217 | elif header_type == _OTHER_SYS_HEADER: | |
1218 | if self._section <= self._OTHER_SYS_SECTION: | |
1219 | self._section = self._OTHER_SYS_SECTION | |
1220 | else: | |
1221 | self._last_header = '' | |
1222 | return error_message | |
1223 | elif header_type == _LIKELY_MY_HEADER: | |
1224 | if self._section <= self._MY_H_SECTION: | |
1225 | self._section = self._MY_H_SECTION | |
1226 | else: | |
1227 | self._section = self._OTHER_H_SECTION | |
1228 | elif header_type == _POSSIBLE_MY_HEADER: | |
1229 | if self._section <= self._MY_H_SECTION: | |
1230 | self._section = self._MY_H_SECTION | |
1231 | else: | |
1232 | # This will always be the fallback because we're not sure | |
1233 | # enough that the header is associated with this file. | |
1234 | self._section = self._OTHER_H_SECTION | |
1235 | else: | |
1236 | assert header_type == _OTHER_HEADER | |
1237 | self._section = self._OTHER_H_SECTION | |
1238 | ||
1239 | if last_section != self._section: | |
1240 | self._last_header = '' | |
1241 | ||
1242 | return '' | |
1243 | ||
1244 | ||
1245 | class _CppLintState(object): | |
1246 | """Maintains module-wide state..""" | |
1247 | ||
1248 | def __init__(self): | |
1249 | self.verbose_level = 1 # global setting. | |
1250 | self.error_count = 0 # global count of reported errors | |
1251 | # filters to apply when emitting error messages | |
1252 | self.filters = _DEFAULT_FILTERS[:] | |
1253 | # backup of filter list. Used to restore the state after each file. | |
1254 | self._filters_backup = self.filters[:] | |
1255 | self.counting = 'total' # In what way are we counting errors? | |
1256 | self.errors_by_category = {} # string to int dict storing error counts | |
1257 | self.quiet = False # Suppress non-error messagess? | |
1258 | ||
1259 | # output format: | |
1260 | # "emacs" - format that emacs can parse (default) | |
1261 | # "eclipse" - format that eclipse can parse | |
1262 | # "vs7" - format that Microsoft Visual Studio 7 can parse | |
1263 | # "junit" - format that Jenkins, Bamboo, etc can parse | |
1264 | # "sed" - returns a gnu sed command to fix the problem | |
1265 | # "gsed" - like sed, but names the command gsed, e.g. for macOS homebrew users | |
1266 | self.output_format = 'emacs' | |
1267 | ||
1268 | # For JUnit output, save errors and failures until the end so that they | |
1269 | # can be written into the XML | |
1270 | self._junit_errors = [] | |
1271 | self._junit_failures = [] | |
1272 | ||
1273 | def SetOutputFormat(self, output_format): | |
1274 | """Sets the output format for errors.""" | |
1275 | self.output_format = output_format | |
1276 | ||
1277 | def SetQuiet(self, quiet): | |
1278 | """Sets the module's quiet settings, and returns the previous setting.""" | |
1279 | last_quiet = self.quiet | |
1280 | self.quiet = quiet | |
1281 | return last_quiet | |
1282 | ||
1283 | def SetVerboseLevel(self, level): | |
1284 | """Sets the module's verbosity, and returns the previous setting.""" | |
1285 | last_verbose_level = self.verbose_level | |
1286 | self.verbose_level = level | |
1287 | return last_verbose_level | |
1288 | ||
1289 | def SetCountingStyle(self, counting_style): | |
1290 | """Sets the module's counting options.""" | |
1291 | self.counting = counting_style | |
1292 | ||
1293 | def SetFilters(self, filters): | |
1294 | """Sets the error-message filters. | |
1295 | ||
1296 | These filters are applied when deciding whether to emit a given | |
1297 | error message. | |
1298 | ||
1299 | Args: | |
1300 | filters: A string of comma-separated filters (eg "+whitespace/indent"). | |
1301 | Each filter should start with + or -; else we die. | |
1302 | ||
1303 | Raises: | |
1304 | ValueError: The comma-separated filters did not all start with '+' or '-'. | |
1305 | E.g. "-,+whitespace,-whitespace/indent,whitespace/badfilter" | |
1306 | """ | |
1307 | # Default filters always have less priority than the flag ones. | |
1308 | self.filters = _DEFAULT_FILTERS[:] | |
1309 | self.AddFilters(filters) | |
1310 | ||
1311 | def AddFilters(self, filters): | |
1312 | """ Adds more filters to the existing list of error-message filters. """ | |
1313 | for filt in filters.split(','): | |
1314 | clean_filt = filt.strip() | |
1315 | if clean_filt: | |
1316 | self.filters.append(clean_filt) | |
1317 | for filt in self.filters: | |
1318 | if not (filt.startswith('+') or filt.startswith('-')): | |
1319 | raise ValueError('Every filter in --filters must start with + or -' | |
1320 | ' (%s does not)' % filt) | |
1321 | ||
1322 | def BackupFilters(self): | |
1323 | """ Saves the current filter list to backup storage.""" | |
1324 | self._filters_backup = self.filters[:] | |
1325 | ||
1326 | def RestoreFilters(self): | |
1327 | """ Restores filters previously backed up.""" | |
1328 | self.filters = self._filters_backup[:] | |
1329 | ||
1330 | def ResetErrorCounts(self): | |
1331 | """Sets the module's error statistic back to zero.""" | |
1332 | self.error_count = 0 | |
1333 | self.errors_by_category = {} | |
1334 | ||
1335 | def IncrementErrorCount(self, category): | |
1336 | """Bumps the module's error statistic.""" | |
1337 | self.error_count += 1 | |
1338 | if self.counting in ('toplevel', 'detailed'): | |
1339 | if self.counting != 'detailed': | |
1340 | category = category.split('/')[0] | |
1341 | if category not in self.errors_by_category: | |
1342 | self.errors_by_category[category] = 0 | |
1343 | self.errors_by_category[category] += 1 | |
1344 | ||
1345 | def PrintErrorCounts(self): | |
1346 | """Print a summary of errors by category, and the total.""" | |
1347 | for category, count in sorted(iteritems(self.errors_by_category)): | |
1348 | self.PrintInfo('Category \'%s\' errors found: %d\n' % | |
1349 | (category, count)) | |
1350 | if self.error_count > 0: | |
1351 | self.PrintInfo('Total errors found: %d\n' % self.error_count) | |
1352 | ||
1353 | def PrintInfo(self, message): | |
1354 | # _quiet does not represent --quiet flag. | |
1355 | # Hide infos from stdout to keep stdout pure for machine consumption | |
1356 | if not _quiet and self.output_format not in _MACHINE_OUTPUTS: | |
1357 | sys.stdout.write(message) | |
1358 | ||
1359 | def PrintError(self, message): | |
1360 | if self.output_format == 'junit': | |
1361 | self._junit_errors.append(message) | |
1362 | else: | |
1363 | sys.stderr.write(message) | |
1364 | ||
1365 | def AddJUnitFailure(self, filename, linenum, message, category, confidence): | |
1366 | self._junit_failures.append((filename, linenum, message, category, | |
1367 | confidence)) | |
1368 | ||
1369 | def FormatJUnitXML(self): | |
1370 | num_errors = len(self._junit_errors) | |
1371 | num_failures = len(self._junit_failures) | |
1372 | ||
1373 | testsuite = xml.etree.ElementTree.Element('testsuite') | |
1374 | testsuite.attrib['errors'] = str(num_errors) | |
1375 | testsuite.attrib['failures'] = str(num_failures) | |
1376 | testsuite.attrib['name'] = 'cpplint' | |
1377 | ||
1378 | if num_errors == 0 and num_failures == 0: | |
1379 | testsuite.attrib['tests'] = str(1) | |
1380 | xml.etree.ElementTree.SubElement(testsuite, 'testcase', name='passed') | |
1381 | ||
1382 | else: | |
1383 | testsuite.attrib['tests'] = str(num_errors + num_failures) | |
1384 | if num_errors > 0: | |
1385 | testcase = xml.etree.ElementTree.SubElement(testsuite, 'testcase') | |
1386 | testcase.attrib['name'] = 'errors' | |
1387 | error = xml.etree.ElementTree.SubElement(testcase, 'error') | |
1388 | error.text = '\n'.join(self._junit_errors) | |
1389 | if num_failures > 0: | |
1390 | # Group failures by file | |
1391 | failed_file_order = [] | |
1392 | failures_by_file = {} | |
1393 | for failure in self._junit_failures: | |
1394 | failed_file = failure[0] | |
1395 | if failed_file not in failed_file_order: | |
1396 | failed_file_order.append(failed_file) | |
1397 | failures_by_file[failed_file] = [] | |
1398 | failures_by_file[failed_file].append(failure) | |
1399 | # Create a testcase for each file | |
1400 | for failed_file in failed_file_order: | |
1401 | failures = failures_by_file[failed_file] | |
1402 | testcase = xml.etree.ElementTree.SubElement(testsuite, 'testcase') | |
1403 | testcase.attrib['name'] = failed_file | |
1404 | failure = xml.etree.ElementTree.SubElement(testcase, 'failure') | |
1405 | template = '{0}: {1} [{2}] [{3}]' | |
1406 | texts = [template.format(f[1], f[2], f[3], f[4]) for f in failures] | |
1407 | failure.text = '\n'.join(texts) | |
1408 | ||
1409 | xml_decl = '<?xml version="1.0" encoding="UTF-8" ?>\n' | |
1410 | return xml_decl + xml.etree.ElementTree.tostring(testsuite, 'utf-8').decode('utf-8') | |
1411 | ||
1412 | ||
1413 | _cpplint_state = _CppLintState() | |
1414 | ||
1415 | ||
1416 | def _OutputFormat(): | |
1417 | """Gets the module's output format.""" | |
1418 | return _cpplint_state.output_format | |
1419 | ||
1420 | ||
1421 | def _SetOutputFormat(output_format): | |
1422 | """Sets the module's output format.""" | |
1423 | _cpplint_state.SetOutputFormat(output_format) | |
1424 | ||
1425 | def _Quiet(): | |
1426 | """Return's the module's quiet setting.""" | |
1427 | return _cpplint_state.quiet | |
1428 | ||
1429 | def _SetQuiet(quiet): | |
1430 | """Set the module's quiet status, and return previous setting.""" | |
1431 | return _cpplint_state.SetQuiet(quiet) | |
1432 | ||
1433 | ||
1434 | def _VerboseLevel(): | |
1435 | """Returns the module's verbosity setting.""" | |
1436 | return _cpplint_state.verbose_level | |
1437 | ||
1438 | ||
1439 | def _SetVerboseLevel(level): | |
1440 | """Sets the module's verbosity, and returns the previous setting.""" | |
1441 | return _cpplint_state.SetVerboseLevel(level) | |
1442 | ||
1443 | ||
1444 | def _SetCountingStyle(level): | |
1445 | """Sets the module's counting options.""" | |
1446 | _cpplint_state.SetCountingStyle(level) | |
1447 | ||
1448 | ||
1449 | def _Filters(): | |
1450 | """Returns the module's list of output filters, as a list.""" | |
1451 | return _cpplint_state.filters | |
1452 | ||
1453 | ||
1454 | def _SetFilters(filters): | |
1455 | """Sets the module's error-message filters. | |
1456 | ||
1457 | These filters are applied when deciding whether to emit a given | |
1458 | error message. | |
1459 | ||
1460 | Args: | |
1461 | filters: A string of comma-separated filters (eg "whitespace/indent"). | |
1462 | Each filter should start with + or -; else we die. | |
1463 | """ | |
1464 | _cpplint_state.SetFilters(filters) | |
1465 | ||
1466 | def _AddFilters(filters): | |
1467 | """Adds more filter overrides. | |
1468 | ||
1469 | Unlike _SetFilters, this function does not reset the current list of filters | |
1470 | available. | |
1471 | ||
1472 | Args: | |
1473 | filters: A string of comma-separated filters (eg "whitespace/indent"). | |
1474 | Each filter should start with + or -; else we die. | |
1475 | """ | |
1476 | _cpplint_state.AddFilters(filters) | |
1477 | ||
1478 | def _BackupFilters(): | |
1479 | """ Saves the current filter list to backup storage.""" | |
1480 | _cpplint_state.BackupFilters() | |
1481 | ||
1482 | def _RestoreFilters(): | |
1483 | """ Restores filters previously backed up.""" | |
1484 | _cpplint_state.RestoreFilters() | |
1485 | ||
1486 | class _FunctionState(object): | |
1487 | """Tracks current function name and the number of lines in its body.""" | |
1488 | ||
1489 | _NORMAL_TRIGGER = 250 # for --v=0, 500 for --v=1, etc. | |
1490 | _TEST_TRIGGER = 400 # about 50% more than _NORMAL_TRIGGER. | |
1491 | ||
1492 | def __init__(self): | |
1493 | self.in_a_function = False | |
1494 | self.lines_in_function = 0 | |
1495 | self.current_function = '' | |
1496 | ||
1497 | def Begin(self, function_name): | |
1498 | """Start analyzing function body. | |
1499 | ||
1500 | Args: | |
1501 | function_name: The name of the function being tracked. | |
1502 | """ | |
1503 | self.in_a_function = True | |
1504 | self.lines_in_function = 0 | |
1505 | self.current_function = function_name | |
1506 | ||
1507 | def Count(self): | |
1508 | """Count line in current function body.""" | |
1509 | if self.in_a_function: | |
1510 | self.lines_in_function += 1 | |
1511 | ||
1512 | def Check(self, error, filename, linenum): | |
1513 | """Report if too many lines in function body. | |
1514 | ||
1515 | Args: | |
1516 | error: The function to call with any errors found. | |
1517 | filename: The name of the current file. | |
1518 | linenum: The number of the line to check. | |
1519 | """ | |
1520 | if not self.in_a_function: | |
1521 | return | |
1522 | ||
1523 | if Match(r'T(EST|est)', self.current_function): | |
1524 | base_trigger = self._TEST_TRIGGER | |
1525 | else: | |
1526 | base_trigger = self._NORMAL_TRIGGER | |
1527 | trigger = base_trigger * 2**_VerboseLevel() | |
1528 | ||
1529 | if self.lines_in_function > trigger: | |
1530 | error_level = int(math.log(self.lines_in_function / base_trigger, 2)) | |
1531 | # 50 => 0, 100 => 1, 200 => 2, 400 => 3, 800 => 4, 1600 => 5, ... | |
1532 | if error_level > 5: | |
1533 | error_level = 5 | |
1534 | error(filename, linenum, 'readability/fn_size', error_level, | |
1535 | 'Small and focused functions are preferred:' | |
1536 | ' %s has %d non-comment lines' | |
1537 | ' (error triggered by exceeding %d lines).' % ( | |
1538 | self.current_function, self.lines_in_function, trigger)) | |
1539 | ||
1540 | def End(self): | |
1541 | """Stop analyzing function body.""" | |
1542 | self.in_a_function = False | |
1543 | ||
1544 | ||
1545 | class _IncludeError(Exception): | |
1546 | """Indicates a problem with the include order in a file.""" | |
1547 | pass | |
1548 | ||
1549 | ||
1550 | class FileInfo(object): | |
1551 | """Provides utility functions for filenames. | |
1552 | ||
1553 | FileInfo provides easy access to the components of a file's path | |
1554 | relative to the project root. | |
1555 | """ | |
1556 | ||
1557 | def __init__(self, filename): | |
1558 | self._filename = filename | |
1559 | ||
1560 | def FullName(self): | |
1561 | """Make Windows paths like Unix.""" | |
1562 | return os.path.abspath(self._filename).replace('\\', '/') | |
1563 | ||
1564 | def RepositoryName(self): | |
1565 | r"""FullName after removing the local path to the repository. | |
1566 | ||
1567 | If we have a real absolute path name here we can try to do something smart: | |
1568 | detecting the root of the checkout and truncating /path/to/checkout from | |
1569 | the name so that we get header guards that don't include things like | |
1570 | "C:\\Documents and Settings\\..." or "/home/username/..." in them and thus | |
1571 | people on different computers who have checked the source out to different | |
1572 | locations won't see bogus errors. | |
1573 | """ | |
1574 | fullname = self.FullName() | |
1575 | ||
1576 | if os.path.exists(fullname): | |
1577 | project_dir = os.path.dirname(fullname) | |
1578 | ||
1579 | # If the user specified a repository path, it exists, and the file is | |
1580 | # contained in it, use the specified repository path | |
1581 | if _repository: | |
1582 | repo = FileInfo(_repository).FullName() | |
1583 | root_dir = project_dir | |
1584 | while os.path.exists(root_dir): | |
1585 | # allow case insensitive compare on Windows | |
1586 | if os.path.normcase(root_dir) == os.path.normcase(repo): | |
1587 | return os.path.relpath(fullname, root_dir).replace('\\', '/') | |
1588 | one_up_dir = os.path.dirname(root_dir) | |
1589 | if one_up_dir == root_dir: | |
1590 | break | |
1591 | root_dir = one_up_dir | |
1592 | ||
1593 | if os.path.exists(os.path.join(project_dir, ".svn")): | |
1594 | # If there's a .svn file in the current directory, we recursively look | |
1595 | # up the directory tree for the top of the SVN checkout | |
1596 | root_dir = project_dir | |
1597 | one_up_dir = os.path.dirname(root_dir) | |
1598 | while os.path.exists(os.path.join(one_up_dir, ".svn")): | |
1599 | root_dir = os.path.dirname(root_dir) | |
1600 | one_up_dir = os.path.dirname(one_up_dir) | |
1601 | ||
1602 | prefix = os.path.commonprefix([root_dir, project_dir]) | |
1603 | return fullname[len(prefix) + 1:] | |
1604 | ||
1605 | # Not SVN <= 1.6? Try to find a git, hg, or svn top level directory by | |
1606 | # searching up from the current path. | |
1607 | root_dir = current_dir = os.path.dirname(fullname) | |
1608 | while current_dir != os.path.dirname(current_dir): | |
1609 | if (os.path.exists(os.path.join(current_dir, ".git")) or | |
1610 | os.path.exists(os.path.join(current_dir, ".hg")) or | |
1611 | os.path.exists(os.path.join(current_dir, ".svn"))): | |
1612 | root_dir = current_dir | |
1613 | current_dir = os.path.dirname(current_dir) | |
1614 | ||
1615 | if (os.path.exists(os.path.join(root_dir, ".git")) or | |
1616 | os.path.exists(os.path.join(root_dir, ".hg")) or | |
1617 | os.path.exists(os.path.join(root_dir, ".svn"))): | |
1618 | prefix = os.path.commonprefix([root_dir, project_dir]) | |
1619 | return fullname[len(prefix) + 1:] | |
1620 | ||
1621 | # Don't know what to do; header guard warnings may be wrong... | |
1622 | return fullname | |
1623 | ||
1624 | def Split(self): | |
1625 | """Splits the file into the directory, basename, and extension. | |
1626 | ||
1627 | For 'chrome/browser/browser.cc', Split() would | |
1628 | return ('chrome/browser', 'browser', '.cc') | |
1629 | ||
1630 | Returns: | |
1631 | A tuple of (directory, basename, extension). | |
1632 | """ | |
1633 | ||
1634 | googlename = self.RepositoryName() | |
1635 | project, rest = os.path.split(googlename) | |
1636 | return (project,) + os.path.splitext(rest) | |
1637 | ||
1638 | def BaseName(self): | |
1639 | """File base name - text after the final slash, before the final period.""" | |
1640 | return self.Split()[1] | |
1641 | ||
1642 | def Extension(self): | |
1643 | """File extension - text following the final period, includes that period.""" | |
1644 | return self.Split()[2] | |
1645 | ||
1646 | def NoExtension(self): | |
1647 | """File has no source file extension.""" | |
1648 | return '/'.join(self.Split()[0:2]) | |
1649 | ||
1650 | def IsSource(self): | |
1651 | """File has a source file extension.""" | |
1652 | return _IsSourceExtension(self.Extension()[1:]) | |
1653 | ||
1654 | ||
1655 | def _ShouldPrintError(category, confidence, linenum): | |
1656 | """If confidence >= verbose, category passes filter and is not suppressed.""" | |
1657 | ||
1658 | # There are three ways we might decide not to print an error message: | |
1659 | # a "NOLINT(category)" comment appears in the source, | |
1660 | # the verbosity level isn't high enough, or the filters filter it out. | |
1661 | if IsErrorSuppressedByNolint(category, linenum): | |
1662 | return False | |
1663 | ||
1664 | if confidence < _cpplint_state.verbose_level: | |
1665 | return False | |
1666 | ||
1667 | is_filtered = False | |
1668 | for one_filter in _Filters(): | |
1669 | if one_filter.startswith('-'): | |
1670 | if category.startswith(one_filter[1:]): | |
1671 | is_filtered = True | |
1672 | elif one_filter.startswith('+'): | |
1673 | if category.startswith(one_filter[1:]): | |
1674 | is_filtered = False | |
1675 | else: | |
1676 | assert False # should have been checked for in SetFilter. | |
1677 | if is_filtered: | |
1678 | return False | |
1679 | ||
1680 | return True | |
1681 | ||
1682 | ||
1683 | def Error(filename, linenum, category, confidence, message): | |
1684 | """Logs the fact we've found a lint error. | |
1685 | ||
1686 | We log where the error was found, and also our confidence in the error, | |
1687 | that is, how certain we are this is a legitimate style regression, and | |
1688 | not a misidentification or a use that's sometimes justified. | |
1689 | ||
1690 | False positives can be suppressed by the use of | |
1691 | "cpplint(category)" comments on the offending line. These are | |
1692 | parsed into _error_suppressions. | |
1693 | ||
1694 | Args: | |
1695 | filename: The name of the file containing the error. | |
1696 | linenum: The number of the line containing the error. | |
1697 | category: A string used to describe the "category" this bug | |
1698 | falls under: "whitespace", say, or "runtime". Categories | |
1699 | may have a hierarchy separated by slashes: "whitespace/indent". | |
1700 | confidence: A number from 1-5 representing a confidence score for | |
1701 | the error, with 5 meaning that we are certain of the problem, | |
1702 | and 1 meaning that it could be a legitimate construct. | |
1703 | message: The error message. | |
1704 | """ | |
1705 | if _ShouldPrintError(category, confidence, linenum): | |
1706 | _cpplint_state.IncrementErrorCount(category) | |
1707 | if _cpplint_state.output_format == 'vs7': | |
1708 | _cpplint_state.PrintError('%s(%s): error cpplint: [%s] %s [%d]\n' % ( | |
1709 | filename, linenum, category, message, confidence)) | |
1710 | elif _cpplint_state.output_format == 'eclipse': | |
1711 | sys.stderr.write('%s:%s: warning: %s [%s] [%d]\n' % ( | |
1712 | filename, linenum, message, category, confidence)) | |
1713 | elif _cpplint_state.output_format == 'junit': | |
1714 | _cpplint_state.AddJUnitFailure(filename, linenum, message, category, | |
1715 | confidence) | |
1716 | elif _cpplint_state.output_format in ['sed', 'gsed']: | |
1717 | if message in _SED_FIXUPS: | |
1718 | sys.stdout.write(_cpplint_state.output_format + " -i '%s%s' %s # %s [%s] [%d]\n" % ( | |
1719 | linenum, _SED_FIXUPS[message], filename, message, category, confidence)) | |
1720 | else: | |
1721 | sys.stderr.write('# %s:%s: "%s" [%s] [%d]\n' % ( | |
1722 | filename, linenum, message, category, confidence)) | |
1723 | else: | |
1724 | final_message = '%s:%s: %s [%s] [%d]\n' % ( | |
1725 | filename, linenum, message, category, confidence) | |
1726 | sys.stderr.write(final_message) | |
1727 | ||
1728 | # Matches standard C++ escape sequences per 2.13.2.3 of the C++ standard. | |
1729 | _RE_PATTERN_CLEANSE_LINE_ESCAPES = re.compile( | |
1730 | r'\\([abfnrtv?"\\\']|\d+|x[0-9a-fA-F]+)') | |
1731 | # Match a single C style comment on the same line. | |
1732 | _RE_PATTERN_C_COMMENTS = r'/\*(?:[^*]|\*(?!/))*\*/' | |
1733 | # Matches multi-line C style comments. | |
1734 | # This RE is a little bit more complicated than one might expect, because we | |
1735 | # have to take care of space removals tools so we can handle comments inside | |
1736 | # statements better. | |
1737 | # The current rule is: We only clear spaces from both sides when we're at the | |
1738 | # end of the line. Otherwise, we try to remove spaces from the right side, | |
1739 | # if this doesn't work we try on left side but only if there's a non-character | |
1740 | # on the right. | |
1741 | _RE_PATTERN_CLEANSE_LINE_C_COMMENTS = re.compile( | |
1742 | r'(\s*' + _RE_PATTERN_C_COMMENTS + r'\s*$|' + | |
1743 | _RE_PATTERN_C_COMMENTS + r'\s+|' + | |
1744 | r'\s+' + _RE_PATTERN_C_COMMENTS + r'(?=\W)|' + | |
1745 | _RE_PATTERN_C_COMMENTS + r')') | |
1746 | ||
1747 | ||
1748 | def IsCppString(line): | |
1749 | """Does line terminate so, that the next symbol is in string constant. | |
1750 | ||
1751 | This function does not consider single-line nor multi-line comments. | |
1752 | ||
1753 | Args: | |
1754 | line: is a partial line of code starting from the 0..n. | |
1755 | ||
1756 | Returns: | |
1757 | True, if next character appended to 'line' is inside a | |
1758 | string constant. | |
1759 | """ | |
1760 | ||
1761 | line = line.replace(r'\\', 'XX') # after this, \\" does not match to \" | |
1762 | return ((line.count('"') - line.count(r'\"') - line.count("'\"'")) & 1) == 1 | |
1763 | ||
1764 | ||
1765 | def CleanseRawStrings(raw_lines): | |
1766 | """Removes C++11 raw strings from lines. | |
1767 | ||
1768 | Before: | |
1769 | static const char kData[] = R"( | |
1770 | multi-line string | |
1771 | )"; | |
1772 | ||
1773 | After: | |
1774 | static const char kData[] = "" | |
1775 | (replaced by blank line) | |
1776 | ""; | |
1777 | ||
1778 | Args: | |
1779 | raw_lines: list of raw lines. | |
1780 | ||
1781 | Returns: | |
1782 | list of lines with C++11 raw strings replaced by empty strings. | |
1783 | """ | |
1784 | ||
1785 | delimiter = None | |
1786 | lines_without_raw_strings = [] | |
1787 | for line in raw_lines: | |
1788 | if delimiter: | |
1789 | # Inside a raw string, look for the end | |
1790 | end = line.find(delimiter) | |
1791 | if end >= 0: | |
1792 | # Found the end of the string, match leading space for this | |
1793 | # line and resume copying the original lines, and also insert | |
1794 | # a "" on the last line. | |
1795 | leading_space = Match(r'^(\s*)\S', line) | |
1796 | line = leading_space.group(1) + '""' + line[end + len(delimiter):] | |
1797 | delimiter = None | |
1798 | else: | |
1799 | # Haven't found the end yet, append a blank line. | |
1800 | line = '""' | |
1801 | ||
1802 | # Look for beginning of a raw string, and replace them with | |
1803 | # empty strings. This is done in a loop to handle multiple raw | |
1804 | # strings on the same line. | |
1805 | while delimiter is None: | |
1806 | # Look for beginning of a raw string. | |
1807 | # See 2.14.15 [lex.string] for syntax. | |
1808 | # | |
1809 | # Once we have matched a raw string, we check the prefix of the | |
1810 | # line to make sure that the line is not part of a single line | |
1811 | # comment. It's done this way because we remove raw strings | |
1812 | # before removing comments as opposed to removing comments | |
1813 | # before removing raw strings. This is because there are some | |
1814 | # cpplint checks that requires the comments to be preserved, but | |
1815 | # we don't want to check comments that are inside raw strings. | |
1816 | matched = Match(r'^(.*?)\b(?:R|u8R|uR|UR|LR)"([^\s\\()]*)\((.*)$', line) | |
1817 | if (matched and | |
1818 | not Match(r'^([^\'"]|\'(\\.|[^\'])*\'|"(\\.|[^"])*")*//', | |
1819 | matched.group(1))): | |
1820 | delimiter = ')' + matched.group(2) + '"' | |
1821 | ||
1822 | end = matched.group(3).find(delimiter) | |
1823 | if end >= 0: | |
1824 | # Raw string ended on same line | |
1825 | line = (matched.group(1) + '""' + | |
1826 | matched.group(3)[end + len(delimiter):]) | |
1827 | delimiter = None | |
1828 | else: | |
1829 | # Start of a multi-line raw string | |
1830 | line = matched.group(1) + '""' | |
1831 | else: | |
1832 | break | |
1833 | ||
1834 | lines_without_raw_strings.append(line) | |
1835 | ||
1836 | # TODO(unknown): if delimiter is not None here, we might want to | |
1837 | # emit a warning for unterminated string. | |
1838 | return lines_without_raw_strings | |
1839 | ||
1840 | ||
1841 | def FindNextMultiLineCommentStart(lines, lineix): | |
1842 | """Find the beginning marker for a multiline comment.""" | |
1843 | while lineix < len(lines): | |
1844 | if lines[lineix].strip().startswith('/*'): | |
1845 | # Only return this marker if the comment goes beyond this line | |
1846 | if lines[lineix].strip().find('*/', 2) < 0: | |
1847 | return lineix | |
1848 | lineix += 1 | |
1849 | return len(lines) | |
1850 | ||
1851 | ||
1852 | def FindNextMultiLineCommentEnd(lines, lineix): | |
1853 | """We are inside a comment, find the end marker.""" | |
1854 | while lineix < len(lines): | |
1855 | if lines[lineix].strip().endswith('*/'): | |
1856 | return lineix | |
1857 | lineix += 1 | |
1858 | return len(lines) | |
1859 | ||
1860 | ||
1861 | def RemoveMultiLineCommentsFromRange(lines, begin, end): | |
1862 | """Clears a range of lines for multi-line comments.""" | |
1863 | # Having // <empty> comments makes the lines non-empty, so we will not get | |
1864 | # unnecessary blank line warnings later in the code. | |
1865 | for i in range(begin, end): | |
1866 | lines[i] = '/**/' | |
1867 | ||
1868 | ||
1869 | def RemoveMultiLineComments(filename, lines, error): | |
1870 | """Removes multiline (c-style) comments from lines.""" | |
1871 | lineix = 0 | |
1872 | while lineix < len(lines): | |
1873 | lineix_begin = FindNextMultiLineCommentStart(lines, lineix) | |
1874 | if lineix_begin >= len(lines): | |
1875 | return | |
1876 | lineix_end = FindNextMultiLineCommentEnd(lines, lineix_begin) | |
1877 | if lineix_end >= len(lines): | |
1878 | error(filename, lineix_begin + 1, 'readability/multiline_comment', 5, | |
1879 | 'Could not find end of multi-line comment') | |
1880 | return | |
1881 | RemoveMultiLineCommentsFromRange(lines, lineix_begin, lineix_end + 1) | |
1882 | lineix = lineix_end + 1 | |
1883 | ||
1884 | ||
1885 | def CleanseComments(line): | |
1886 | """Removes //-comments and single-line C-style /* */ comments. | |
1887 | ||
1888 | Args: | |
1889 | line: A line of C++ source. | |
1890 | ||
1891 | Returns: | |
1892 | The line with single-line comments removed. | |
1893 | """ | |
1894 | commentpos = line.find('//') | |
1895 | if commentpos != -1 and not IsCppString(line[:commentpos]): | |
1896 | line = line[:commentpos].rstrip() | |
1897 | # get rid of /* ... */ | |
1898 | return _RE_PATTERN_CLEANSE_LINE_C_COMMENTS.sub('', line) | |
1899 | ||
1900 | ||
1901 | class CleansedLines(object): | |
1902 | """Holds 4 copies of all lines with different preprocessing applied to them. | |
1903 | ||
1904 | 1) elided member contains lines without strings and comments. | |
1905 | 2) lines member contains lines without comments. | |
1906 | 3) raw_lines member contains all the lines without processing. | |
1907 | 4) lines_without_raw_strings member is same as raw_lines, but with C++11 raw | |
1908 | strings removed. | |
1909 | All these members are of <type 'list'>, and of the same length. | |
1910 | """ | |
1911 | ||
1912 | def __init__(self, lines): | |
1913 | self.elided = [] | |
1914 | self.lines = [] | |
1915 | self.raw_lines = lines | |
1916 | self.num_lines = len(lines) | |
1917 | self.lines_without_raw_strings = CleanseRawStrings(lines) | |
1918 | for linenum in range(len(self.lines_without_raw_strings)): | |
1919 | self.lines.append(CleanseComments( | |
1920 | self.lines_without_raw_strings[linenum])) | |
1921 | elided = self._CollapseStrings(self.lines_without_raw_strings[linenum]) | |
1922 | self.elided.append(CleanseComments(elided)) | |
1923 | ||
1924 | def NumLines(self): | |
1925 | """Returns the number of lines represented.""" | |
1926 | return self.num_lines | |
1927 | ||
1928 | @staticmethod | |
1929 | def _CollapseStrings(elided): | |
1930 | """Collapses strings and chars on a line to simple "" or '' blocks. | |
1931 | ||
1932 | We nix strings first so we're not fooled by text like '"http://"' | |
1933 | ||
1934 | Args: | |
1935 | elided: The line being processed. | |
1936 | ||
1937 | Returns: | |
1938 | The line with collapsed strings. | |
1939 | """ | |
1940 | if _RE_PATTERN_INCLUDE.match(elided): | |
1941 | return elided | |
1942 | ||
1943 | # Remove escaped characters first to make quote/single quote collapsing | |
1944 | # basic. Things that look like escaped characters shouldn't occur | |
1945 | # outside of strings and chars. | |
1946 | elided = _RE_PATTERN_CLEANSE_LINE_ESCAPES.sub('', elided) | |
1947 | ||
1948 | # Replace quoted strings and digit separators. Both single quotes | |
1949 | # and double quotes are processed in the same loop, otherwise | |
1950 | # nested quotes wouldn't work. | |
1951 | collapsed = '' | |
1952 | while True: | |
1953 | # Find the first quote character | |
1954 | match = Match(r'^([^\'"]*)([\'"])(.*)$', elided) | |
1955 | if not match: | |
1956 | collapsed += elided | |
1957 | break | |
1958 | head, quote, tail = match.groups() | |
1959 | ||
1960 | if quote == '"': | |
1961 | # Collapse double quoted strings | |
1962 | second_quote = tail.find('"') | |
1963 | if second_quote >= 0: | |
1964 | collapsed += head + '""' | |
1965 | elided = tail[second_quote + 1:] | |
1966 | else: | |
1967 | # Unmatched double quote, don't bother processing the rest | |
1968 | # of the line since this is probably a multiline string. | |
1969 | collapsed += elided | |
1970 | break | |
1971 | else: | |
1972 | # Found single quote, check nearby text to eliminate digit separators. | |
1973 | # | |
1974 | # There is no special handling for floating point here, because | |
1975 | # the integer/fractional/exponent parts would all be parsed | |
1976 | # correctly as long as there are digits on both sides of the | |
1977 | # separator. So we are fine as long as we don't see something | |
1978 | # like "0.'3" (gcc 4.9.0 will not allow this literal). | |
1979 | if Search(r'\b(?:0[bBxX]?|[1-9])[0-9a-fA-F]*$', head): | |
1980 | match_literal = Match(r'^((?:\'?[0-9a-zA-Z_])*)(.*)$', "'" + tail) | |
1981 | collapsed += head + match_literal.group(1).replace("'", '') | |
1982 | elided = match_literal.group(2) | |
1983 | else: | |
1984 | second_quote = tail.find('\'') | |
1985 | if second_quote >= 0: | |
1986 | collapsed += head + "''" | |
1987 | elided = tail[second_quote + 1:] | |
1988 | else: | |
1989 | # Unmatched single quote | |
1990 | collapsed += elided | |
1991 | break | |
1992 | ||
1993 | return collapsed | |
1994 | ||
1995 | ||
1996 | def FindEndOfExpressionInLine(line, startpos, stack): | |
1997 | """Find the position just after the end of current parenthesized expression. | |
1998 | ||
1999 | Args: | |
2000 | line: a CleansedLines line. | |
2001 | startpos: start searching at this position. | |
2002 | stack: nesting stack at startpos. | |
2003 | ||
2004 | Returns: | |
2005 | On finding matching end: (index just after matching end, None) | |
2006 | On finding an unclosed expression: (-1, None) | |
2007 | Otherwise: (-1, new stack at end of this line) | |
2008 | """ | |
2009 | for i in xrange(startpos, len(line)): | |
2010 | char = line[i] | |
2011 | if char in '([{': | |
2012 | # Found start of parenthesized expression, push to expression stack | |
2013 | stack.append(char) | |
2014 | elif char == '<': | |
2015 | # Found potential start of template argument list | |
2016 | if i > 0 and line[i - 1] == '<': | |
2017 | # Left shift operator | |
2018 | if stack and stack[-1] == '<': | |
2019 | stack.pop() | |
2020 | if not stack: | |
2021 | return (-1, None) | |
2022 | elif i > 0 and Search(r'\boperator\s*$', line[0:i]): | |
2023 | # operator<, don't add to stack | |
2024 | continue | |
2025 | else: | |
2026 | # Tentative start of template argument list | |
2027 | stack.append('<') | |
2028 | elif char in ')]}': | |
2029 | # Found end of parenthesized expression. | |
2030 | # | |
2031 | # If we are currently expecting a matching '>', the pending '<' | |
2032 | # must have been an operator. Remove them from expression stack. | |
2033 | while stack and stack[-1] == '<': | |
2034 | stack.pop() | |
2035 | if not stack: | |
2036 | return (-1, None) | |
2037 | if ((stack[-1] == '(' and char == ')') or | |
2038 | (stack[-1] == '[' and char == ']') or | |
2039 | (stack[-1] == '{' and char == '}')): | |
2040 | stack.pop() | |
2041 | if not stack: | |
2042 | return (i + 1, None) | |
2043 | else: | |
2044 | # Mismatched parentheses | |
2045 | return (-1, None) | |
2046 | elif char == '>': | |
2047 | # Found potential end of template argument list. | |
2048 | ||
2049 | # Ignore "->" and operator functions | |
2050 | if (i > 0 and | |
2051 | (line[i - 1] == '-' or Search(r'\boperator\s*$', line[0:i - 1]))): | |
2052 | continue | |
2053 | ||
2054 | # Pop the stack if there is a matching '<'. Otherwise, ignore | |
2055 | # this '>' since it must be an operator. | |
2056 | if stack: | |
2057 | if stack[-1] == '<': | |
2058 | stack.pop() | |
2059 | if not stack: | |
2060 | return (i + 1, None) | |
2061 | elif char == ';': | |
2062 | # Found something that look like end of statements. If we are currently | |
2063 | # expecting a '>', the matching '<' must have been an operator, since | |
2064 | # template argument list should not contain statements. | |
2065 | while stack and stack[-1] == '<': | |
2066 | stack.pop() | |
2067 | if not stack: | |
2068 | return (-1, None) | |
2069 | ||
2070 | # Did not find end of expression or unbalanced parentheses on this line | |
2071 | return (-1, stack) | |
2072 | ||
2073 | ||
2074 | def CloseExpression(clean_lines, linenum, pos): | |
2075 | """If input points to ( or { or [ or <, finds the position that closes it. | |
2076 | ||
2077 | If lines[linenum][pos] points to a '(' or '{' or '[' or '<', finds the | |
2078 | linenum/pos that correspond to the closing of the expression. | |
2079 | ||
2080 | TODO(unknown): cpplint spends a fair bit of time matching parentheses. | |
2081 | Ideally we would want to index all opening and closing parentheses once | |
2082 | and have CloseExpression be just a simple lookup, but due to preprocessor | |
2083 | tricks, this is not so easy. | |
2084 | ||
2085 | Args: | |
2086 | clean_lines: A CleansedLines instance containing the file. | |
2087 | linenum: The number of the line to check. | |
2088 | pos: A position on the line. | |
2089 | ||
2090 | Returns: | |
2091 | A tuple (line, linenum, pos) pointer *past* the closing brace, or | |
2092 | (line, len(lines), -1) if we never find a close. Note we ignore | |
2093 | strings and comments when matching; and the line we return is the | |
2094 | 'cleansed' line at linenum. | |
2095 | """ | |
2096 | ||
2097 | line = clean_lines.elided[linenum] | |
2098 | if (line[pos] not in '({[<') or Match(r'<[<=]', line[pos:]): | |
2099 | return (line, clean_lines.NumLines(), -1) | |
2100 | ||
2101 | # Check first line | |
2102 | (end_pos, stack) = FindEndOfExpressionInLine(line, pos, []) | |
2103 | if end_pos > -1: | |
2104 | return (line, linenum, end_pos) | |
2105 | ||
2106 | # Continue scanning forward | |
2107 | while stack and linenum < clean_lines.NumLines() - 1: | |
2108 | linenum += 1 | |
2109 | line = clean_lines.elided[linenum] | |
2110 | (end_pos, stack) = FindEndOfExpressionInLine(line, 0, stack) | |
2111 | if end_pos > -1: | |
2112 | return (line, linenum, end_pos) | |
2113 | ||
2114 | # Did not find end of expression before end of file, give up | |
2115 | return (line, clean_lines.NumLines(), -1) | |
2116 | ||
2117 | ||
2118 | def FindStartOfExpressionInLine(line, endpos, stack): | |
2119 | """Find position at the matching start of current expression. | |
2120 | ||
2121 | This is almost the reverse of FindEndOfExpressionInLine, but note | |
2122 | that the input position and returned position differs by 1. | |
2123 | ||
2124 | Args: | |
2125 | line: a CleansedLines line. | |
2126 | endpos: start searching at this position. | |
2127 | stack: nesting stack at endpos. | |
2128 | ||
2129 | Returns: | |
2130 | On finding matching start: (index at matching start, None) | |
2131 | On finding an unclosed expression: (-1, None) | |
2132 | Otherwise: (-1, new stack at beginning of this line) | |
2133 | """ | |
2134 | i = endpos | |
2135 | while i >= 0: | |
2136 | char = line[i] | |
2137 | if char in ')]}': | |
2138 | # Found end of expression, push to expression stack | |
2139 | stack.append(char) | |
2140 | elif char == '>': | |
2141 | # Found potential end of template argument list. | |
2142 | # | |
2143 | # Ignore it if it's a "->" or ">=" or "operator>" | |
2144 | if (i > 0 and | |
2145 | (line[i - 1] == '-' or | |
2146 | Match(r'\s>=\s', line[i - 1:]) or | |
2147 | Search(r'\boperator\s*$', line[0:i]))): | |
2148 | i -= 1 | |
2149 | else: | |
2150 | stack.append('>') | |
2151 | elif char == '<': | |
2152 | # Found potential start of template argument list | |
2153 | if i > 0 and line[i - 1] == '<': | |
2154 | # Left shift operator | |
2155 | i -= 1 | |
2156 | else: | |
2157 | # If there is a matching '>', we can pop the expression stack. | |
2158 | # Otherwise, ignore this '<' since it must be an operator. | |
2159 | if stack and stack[-1] == '>': | |
2160 | stack.pop() | |
2161 | if not stack: | |
2162 | return (i, None) | |
2163 | elif char in '([{': | |
2164 | # Found start of expression. | |
2165 | # | |
2166 | # If there are any unmatched '>' on the stack, they must be | |
2167 | # operators. Remove those. | |
2168 | while stack and stack[-1] == '>': | |
2169 | stack.pop() | |
2170 | if not stack: | |
2171 | return (-1, None) | |
2172 | if ((char == '(' and stack[-1] == ')') or | |
2173 | (char == '[' and stack[-1] == ']') or | |
2174 | (char == '{' and stack[-1] == '}')): | |
2175 | stack.pop() | |
2176 | if not stack: | |
2177 | return (i, None) | |
2178 | else: | |
2179 | # Mismatched parentheses | |
2180 | return (-1, None) | |
2181 | elif char == ';': | |
2182 | # Found something that look like end of statements. If we are currently | |
2183 | # expecting a '<', the matching '>' must have been an operator, since | |
2184 | # template argument list should not contain statements. | |
2185 | while stack and stack[-1] == '>': | |
2186 | stack.pop() | |
2187 | if not stack: | |
2188 | return (-1, None) | |
2189 | ||
2190 | i -= 1 | |
2191 | ||
2192 | return (-1, stack) | |
2193 | ||
2194 | ||
2195 | def ReverseCloseExpression(clean_lines, linenum, pos): | |
2196 | """If input points to ) or } or ] or >, finds the position that opens it. | |
2197 | ||
2198 | If lines[linenum][pos] points to a ')' or '}' or ']' or '>', finds the | |
2199 | linenum/pos that correspond to the opening of the expression. | |
2200 | ||
2201 | Args: | |
2202 | clean_lines: A CleansedLines instance containing the file. | |
2203 | linenum: The number of the line to check. | |
2204 | pos: A position on the line. | |
2205 | ||
2206 | Returns: | |
2207 | A tuple (line, linenum, pos) pointer *at* the opening brace, or | |
2208 | (line, 0, -1) if we never find the matching opening brace. Note | |
2209 | we ignore strings and comments when matching; and the line we | |
2210 | return is the 'cleansed' line at linenum. | |
2211 | """ | |
2212 | line = clean_lines.elided[linenum] | |
2213 | if line[pos] not in ')}]>': | |
2214 | return (line, 0, -1) | |
2215 | ||
2216 | # Check last line | |
2217 | (start_pos, stack) = FindStartOfExpressionInLine(line, pos, []) | |
2218 | if start_pos > -1: | |
2219 | return (line, linenum, start_pos) | |
2220 | ||
2221 | # Continue scanning backward | |
2222 | while stack and linenum > 0: | |
2223 | linenum -= 1 | |
2224 | line = clean_lines.elided[linenum] | |
2225 | (start_pos, stack) = FindStartOfExpressionInLine(line, len(line) - 1, stack) | |
2226 | if start_pos > -1: | |
2227 | return (line, linenum, start_pos) | |
2228 | ||
2229 | # Did not find start of expression before beginning of file, give up | |
2230 | return (line, 0, -1) | |
2231 | ||
2232 | ||
2233 | def CheckForCopyright(filename, lines, error): | |
2234 | """Logs an error if no Copyright message appears at the top of the file.""" | |
2235 | ||
2236 | # We'll say it should occur by line 10. Don't forget there's a | |
2237 | # placeholder line at the front. | |
2238 | for line in xrange(1, min(len(lines), 11)): | |
2239 | if re.search(r'Copyright', lines[line], re.I): break | |
2240 | else: # means no copyright line was found | |
2241 | error(filename, 0, 'legal/copyright', 5, | |
2242 | 'No copyright message found. ' | |
2243 | 'You should have a line: "Copyright [year] <Copyright Owner>"') | |
2244 | ||
2245 | ||
2246 | def GetIndentLevel(line): | |
2247 | """Return the number of leading spaces in line. | |
2248 | ||
2249 | Args: | |
2250 | line: A string to check. | |
2251 | ||
2252 | Returns: | |
2253 | An integer count of leading spaces, possibly zero. | |
2254 | """ | |
2255 | indent = Match(r'^( *)\S', line) | |
2256 | if indent: | |
2257 | return len(indent.group(1)) | |
2258 | else: | |
2259 | return 0 | |
2260 | ||
2261 | def PathSplitToList(path): | |
2262 | """Returns the path split into a list by the separator. | |
2263 | ||
2264 | Args: | |
2265 | path: An absolute or relative path (e.g. '/a/b/c/' or '../a') | |
2266 | ||
2267 | Returns: | |
2268 | A list of path components (e.g. ['a', 'b', 'c]). | |
2269 | """ | |
2270 | lst = [] | |
2271 | while True: | |
2272 | (head, tail) = os.path.split(path) | |
2273 | if head == path: # absolute paths end | |
2274 | lst.append(head) | |
2275 | break | |
2276 | if tail == path: # relative paths end | |
2277 | lst.append(tail) | |
2278 | break | |
2279 | ||
2280 | path = head | |
2281 | lst.append(tail) | |
2282 | ||
2283 | lst.reverse() | |
2284 | return lst | |
2285 | ||
2286 | def GetHeaderGuardCPPVariable(filename): | |
2287 | """Returns the CPP variable that should be used as a header guard. | |
2288 | ||
2289 | Args: | |
2290 | filename: The name of a C++ header file. | |
2291 | ||
2292 | Returns: | |
2293 | The CPP variable that should be used as a header guard in the | |
2294 | named file. | |
2295 | ||
2296 | """ | |
2297 | ||
2298 | # Restores original filename in case that cpplint is invoked from Emacs's | |
2299 | # flymake. | |
2300 | filename = re.sub(r'_flymake\.h$', '.h', filename) | |
2301 | filename = re.sub(r'/\.flymake/([^/]*)$', r'/\1', filename) | |
2302 | # Replace 'c++' with 'cpp'. | |
2303 | filename = filename.replace('C++', 'cpp').replace('c++', 'cpp') | |
2304 | ||
2305 | fileinfo = FileInfo(filename) | |
2306 | file_path_from_root = fileinfo.RepositoryName() | |
2307 | ||
2308 | def FixupPathFromRoot(): | |
2309 | if _root_debug: | |
2310 | sys.stderr.write("\n_root fixup, _root = '%s', repository name = '%s'\n" | |
2311 | % (_root, fileinfo.RepositoryName())) | |
2312 | ||
2313 | # Process the file path with the --root flag if it was set. | |
2314 | if not _root: | |
2315 | if _root_debug: | |
2316 | sys.stderr.write("_root unspecified\n") | |
2317 | return file_path_from_root | |
2318 | ||
2319 | def StripListPrefix(lst, prefix): | |
2320 | # f(['x', 'y'], ['w, z']) -> None (not a valid prefix) | |
2321 | if lst[:len(prefix)] != prefix: | |
2322 | return None | |
2323 | # f(['a, 'b', 'c', 'd'], ['a', 'b']) -> ['c', 'd'] | |
2324 | return lst[(len(prefix)):] | |
2325 | ||
2326 | # root behavior: | |
2327 | # --root=subdir , lstrips subdir from the header guard | |
2328 | maybe_path = StripListPrefix(PathSplitToList(file_path_from_root), | |
2329 | PathSplitToList(_root)) | |
2330 | ||
2331 | if _root_debug: | |
2332 | sys.stderr.write(("_root lstrip (maybe_path=%s, file_path_from_root=%s," + | |
2333 | " _root=%s)\n") % (maybe_path, file_path_from_root, _root)) | |
2334 | ||
2335 | if maybe_path: | |
2336 | return os.path.join(*maybe_path) | |
2337 | ||
2338 | # --root=.. , will prepend the outer directory to the header guard | |
2339 | full_path = fileinfo.FullName() | |
2340 | # adapt slashes for windows | |
2341 | root_abspath = os.path.abspath(_root).replace('\\', '/') | |
2342 | ||
2343 | maybe_path = StripListPrefix(PathSplitToList(full_path), | |
2344 | PathSplitToList(root_abspath)) | |
2345 | ||
2346 | if _root_debug: | |
2347 | sys.stderr.write(("_root prepend (maybe_path=%s, full_path=%s, " + | |
2348 | "root_abspath=%s)\n") % (maybe_path, full_path, root_abspath)) | |
2349 | ||
2350 | if maybe_path: | |
2351 | return os.path.join(*maybe_path) | |
2352 | ||
2353 | if _root_debug: | |
2354 | sys.stderr.write("_root ignore, returning %s\n" % (file_path_from_root)) | |
2355 | ||
2356 | # --root=FAKE_DIR is ignored | |
2357 | return file_path_from_root | |
2358 | ||
2359 | file_path_from_root = FixupPathFromRoot() | |
2360 | return re.sub(r'[^a-zA-Z0-9]', '_', file_path_from_root).upper() + '_' | |
2361 | ||
2362 | ||
2363 | def CheckForHeaderGuard(filename, clean_lines, error): | |
2364 | """Checks that the file contains a header guard. | |
2365 | ||
2366 | Logs an error if no #ifndef header guard is present. For other | |
2367 | headers, checks that the full pathname is used. | |
2368 | ||
2369 | Args: | |
2370 | filename: The name of the C++ header file. | |
2371 | clean_lines: A CleansedLines instance containing the file. | |
2372 | error: The function to call with any errors found. | |
2373 | """ | |
2374 | ||
2375 | # Don't check for header guards if there are error suppression | |
2376 | # comments somewhere in this file. | |
2377 | # | |
2378 | # Because this is silencing a warning for a nonexistent line, we | |
2379 | # only support the very specific NOLINT(build/header_guard) syntax, | |
2380 | # and not the general NOLINT or NOLINT(*) syntax. | |
2381 | raw_lines = clean_lines.lines_without_raw_strings | |
2382 | for i in raw_lines: | |
2383 | if Search(r'//\s*NOLINT\(build/header_guard\)', i): | |
2384 | return | |
2385 | ||
2386 | # Allow pragma once instead of header guards | |
2387 | for i in raw_lines: | |
2388 | if Search(r'^\s*#pragma\s+once', i): | |
2389 | return | |
2390 | ||
2391 | cppvar = GetHeaderGuardCPPVariable(filename) | |
2392 | ||
2393 | ifndef = '' | |
2394 | ifndef_linenum = 0 | |
2395 | define = '' | |
2396 | endif = '' | |
2397 | endif_linenum = 0 | |
2398 | for linenum, line in enumerate(raw_lines): | |
2399 | linesplit = line.split() | |
2400 | if len(linesplit) >= 2: | |
2401 | # find the first occurrence of #ifndef and #define, save arg | |
2402 | if not ifndef and linesplit[0] == '#ifndef': | |
2403 | # set ifndef to the header guard presented on the #ifndef line. | |
2404 | ifndef = linesplit[1] | |
2405 | ifndef_linenum = linenum | |
2406 | if not define and linesplit[0] == '#define': | |
2407 | define = linesplit[1] | |
2408 | # find the last occurrence of #endif, save entire line | |
2409 | if line.startswith('#endif'): | |
2410 | endif = line | |
2411 | endif_linenum = linenum | |
2412 | ||
2413 | if not ifndef or not define or ifndef != define: | |
2414 | error(filename, 0, 'build/header_guard', 5, | |
2415 | 'No #ifndef header guard found, suggested CPP variable is: %s' % | |
2416 | cppvar) | |
2417 | return | |
2418 | ||
2419 | # The guard should be PATH_FILE_H_, but we also allow PATH_FILE_H__ | |
2420 | # for backward compatibility. | |
2421 | if ifndef != cppvar: | |
2422 | error_level = 0 | |
2423 | if ifndef != cppvar + '_': | |
2424 | error_level = 5 | |
2425 | ||
2426 | ParseNolintSuppressions(filename, raw_lines[ifndef_linenum], ifndef_linenum, | |
2427 | error) | |
2428 | error(filename, ifndef_linenum, 'build/header_guard', error_level, | |
2429 | '#ifndef header guard has wrong style, please use: %s' % cppvar) | |
2430 | ||
2431 | # Check for "//" comments on endif line. | |
2432 | ParseNolintSuppressions(filename, raw_lines[endif_linenum], endif_linenum, | |
2433 | error) | |
2434 | match = Match(r'#endif\s*//\s*' + cppvar + r'(_)?\b', endif) | |
2435 | if match: | |
2436 | if match.group(1) == '_': | |
2437 | # Issue low severity warning for deprecated double trailing underscore | |
2438 | error(filename, endif_linenum, 'build/header_guard', 0, | |
2439 | '#endif line should be "#endif // %s"' % cppvar) | |
2440 | return | |
2441 | ||
2442 | # Didn't find the corresponding "//" comment. If this file does not | |
2443 | # contain any "//" comments at all, it could be that the compiler | |
2444 | # only wants "/**/" comments, look for those instead. | |
2445 | no_single_line_comments = True | |
2446 | for i in xrange(1, len(raw_lines) - 1): | |
2447 | line = raw_lines[i] | |
2448 | if Match(r'^(?:(?:\'(?:\.|[^\'])*\')|(?:"(?:\.|[^"])*")|[^\'"])*//', line): | |
2449 | no_single_line_comments = False | |
2450 | break | |
2451 | ||
2452 | if no_single_line_comments: | |
2453 | match = Match(r'#endif\s*/\*\s*' + cppvar + r'(_)?\s*\*/', endif) | |
2454 | if match: | |
2455 | if match.group(1) == '_': | |
2456 | # Low severity warning for double trailing underscore | |
2457 | error(filename, endif_linenum, 'build/header_guard', 0, | |
2458 | '#endif line should be "#endif /* %s */"' % cppvar) | |
2459 | return | |
2460 | ||
2461 | # Didn't find anything | |
2462 | error(filename, endif_linenum, 'build/header_guard', 5, | |
2463 | '#endif line should be "#endif // %s"' % cppvar) | |
2464 | ||
2465 | ||
2466 | def CheckHeaderFileIncluded(filename, include_state, error): | |
2467 | """Logs an error if a source file does not include its header.""" | |
2468 | ||
2469 | # Do not check test files | |
2470 | fileinfo = FileInfo(filename) | |
2471 | if Search(_TEST_FILE_SUFFIX, fileinfo.BaseName()): | |
2472 | return | |
2473 | ||
2474 | for ext in GetHeaderExtensions(): | |
2475 | basefilename = filename[0:len(filename) - len(fileinfo.Extension())] | |
2476 | headerfile = basefilename + '.' + ext | |
2477 | if not os.path.exists(headerfile): | |
2478 | continue | |
2479 | headername = FileInfo(headerfile).RepositoryName() | |
2480 | first_include = None | |
2481 | include_uses_unix_dir_aliases = False | |
2482 | for section_list in include_state.include_list: | |
2483 | for f in section_list: | |
2484 | include_text = f[0] | |
2485 | if "./" in include_text: | |
2486 | include_uses_unix_dir_aliases = True | |
2487 | if headername in include_text or include_text in headername: | |
2488 | return | |
2489 | if not first_include: | |
2490 | first_include = f[1] | |
2491 | ||
2492 | message = '%s should include its header file %s' % (fileinfo.RepositoryName(), headername) | |
2493 | if include_uses_unix_dir_aliases: | |
2494 | message += ". Relative paths like . and .. are not allowed." | |
2495 | ||
2496 | error(filename, first_include, 'build/include', 5, message) | |
2497 | ||
2498 | ||
2499 | def CheckForBadCharacters(filename, lines, error): | |
2500 | """Logs an error for each line containing bad characters. | |
2501 | ||
2502 | Two kinds of bad characters: | |
2503 | ||
2504 | 1. Unicode replacement characters: These indicate that either the file | |
2505 | contained invalid UTF-8 (likely) or Unicode replacement characters (which | |
2506 | it shouldn't). Note that it's possible for this to throw off line | |
2507 | numbering if the invalid UTF-8 occurred adjacent to a newline. | |
2508 | ||
2509 | 2. NUL bytes. These are problematic for some tools. | |
2510 | ||
2511 | Args: | |
2512 | filename: The name of the current file. | |
2513 | lines: An array of strings, each representing a line of the file. | |
2514 | error: The function to call with any errors found. | |
2515 | """ | |
2516 | for linenum, line in enumerate(lines): | |
2517 | if unicode_escape_decode('\ufffd') in line: | |
2518 | error(filename, linenum, 'readability/utf8', 5, | |
2519 | 'Line contains invalid UTF-8 (or Unicode replacement character).') | |
2520 | if '\0' in line: | |
2521 | error(filename, linenum, 'readability/nul', 5, 'Line contains NUL byte.') | |
2522 | ||
2523 | ||
2524 | def CheckForNewlineAtEOF(filename, lines, error): | |
2525 | """Logs an error if there is no newline char at the end of the file. | |
2526 | ||
2527 | Args: | |
2528 | filename: The name of the current file. | |
2529 | lines: An array of strings, each representing a line of the file. | |
2530 | error: The function to call with any errors found. | |
2531 | """ | |
2532 | ||
2533 | # The array lines() was created by adding two newlines to the | |
2534 | # original file (go figure), then splitting on \n. | |
2535 | # To verify that the file ends in \n, we just have to make sure the | |
2536 | # last-but-two element of lines() exists and is empty. | |
2537 | if len(lines) < 3 or lines[-2]: | |
2538 | error(filename, len(lines) - 2, 'whitespace/ending_newline', 5, | |
2539 | 'Could not find a newline character at the end of the file.') | |
2540 | ||
2541 | ||
2542 | def CheckForMultilineCommentsAndStrings(filename, clean_lines, linenum, error): | |
2543 | """Logs an error if we see /* ... */ or "..." that extend past one line. | |
2544 | ||
2545 | /* ... */ comments are legit inside macros, for one line. | |
2546 | Otherwise, we prefer // comments, so it's ok to warn about the | |
2547 | other. Likewise, it's ok for strings to extend across multiple | |
2548 | lines, as long as a line continuation character (backslash) | |
2549 | terminates each line. Although not currently prohibited by the C++ | |
2550 | style guide, it's ugly and unnecessary. We don't do well with either | |
2551 | in this lint program, so we warn about both. | |
2552 | ||
2553 | Args: | |
2554 | filename: The name of the current file. | |
2555 | clean_lines: A CleansedLines instance containing the file. | |
2556 | linenum: The number of the line to check. | |
2557 | error: The function to call with any errors found. | |
2558 | """ | |
2559 | line = clean_lines.elided[linenum] | |
2560 | ||
2561 | # Remove all \\ (escaped backslashes) from the line. They are OK, and the | |
2562 | # second (escaped) slash may trigger later \" detection erroneously. | |
2563 | line = line.replace('\\\\', '') | |
2564 | ||
2565 | if line.count('/*') > line.count('*/'): | |
2566 | error(filename, linenum, 'readability/multiline_comment', 5, | |
2567 | 'Complex multi-line /*...*/-style comment found. ' | |
2568 | 'Lint may give bogus warnings. ' | |
2569 | 'Consider replacing these with //-style comments, ' | |
2570 | 'with #if 0...#endif, ' | |
2571 | 'or with more clearly structured multi-line comments.') | |
2572 | ||
2573 | if (line.count('"') - line.count('\\"')) % 2: | |
2574 | error(filename, linenum, 'readability/multiline_string', 5, | |
2575 | 'Multi-line string ("...") found. This lint script doesn\'t ' | |
2576 | 'do well with such strings, and may give bogus warnings. ' | |
2577 | 'Use C++11 raw strings or concatenation instead.') | |
2578 | ||
2579 | ||
2580 | # (non-threadsafe name, thread-safe alternative, validation pattern) | |
2581 | # | |
2582 | # The validation pattern is used to eliminate false positives such as: | |
2583 | # _rand(); // false positive due to substring match. | |
2584 | # ->rand(); // some member function rand(). | |
2585 | # ACMRandom rand(seed); // some variable named rand. | |
2586 | # ISAACRandom rand(); // another variable named rand. | |
2587 | # | |
2588 | # Basically we require the return value of these functions to be used | |
2589 | # in some expression context on the same line by matching on some | |
2590 | # operator before the function name. This eliminates constructors and | |
2591 | # member function calls. | |
2592 | _UNSAFE_FUNC_PREFIX = r'(?:[-+*/=%^&|(<]\s*|>\s+)' | |
2593 | _THREADING_LIST = ( | |
2594 | ('asctime(', 'asctime_r(', _UNSAFE_FUNC_PREFIX + r'asctime\([^)]+\)'), | |
2595 | ('ctime(', 'ctime_r(', _UNSAFE_FUNC_PREFIX + r'ctime\([^)]+\)'), | |
2596 | ('getgrgid(', 'getgrgid_r(', _UNSAFE_FUNC_PREFIX + r'getgrgid\([^)]+\)'), | |
2597 | ('getgrnam(', 'getgrnam_r(', _UNSAFE_FUNC_PREFIX + r'getgrnam\([^)]+\)'), | |
2598 | ('getlogin(', 'getlogin_r(', _UNSAFE_FUNC_PREFIX + r'getlogin\(\)'), | |
2599 | ('getpwnam(', 'getpwnam_r(', _UNSAFE_FUNC_PREFIX + r'getpwnam\([^)]+\)'), | |
2600 | ('getpwuid(', 'getpwuid_r(', _UNSAFE_FUNC_PREFIX + r'getpwuid\([^)]+\)'), | |
2601 | ('gmtime(', 'gmtime_r(', _UNSAFE_FUNC_PREFIX + r'gmtime\([^)]+\)'), | |
2602 | ('localtime(', 'localtime_r(', _UNSAFE_FUNC_PREFIX + r'localtime\([^)]+\)'), | |
2603 | ('rand(', 'rand_r(', _UNSAFE_FUNC_PREFIX + r'rand\(\)'), | |
2604 | ('strtok(', 'strtok_r(', | |
2605 | _UNSAFE_FUNC_PREFIX + r'strtok\([^)]+\)'), | |
2606 | ('ttyname(', 'ttyname_r(', _UNSAFE_FUNC_PREFIX + r'ttyname\([^)]+\)'), | |
2607 | ) | |
2608 | ||
2609 | ||
2610 | def CheckPosixThreading(filename, clean_lines, linenum, error): | |
2611 | """Checks for calls to thread-unsafe functions. | |
2612 | ||
2613 | Much code has been originally written without consideration of | |
2614 | multi-threading. Also, engineers are relying on their old experience; | |
2615 | they have learned posix before threading extensions were added. These | |
2616 | tests guide the engineers to use thread-safe functions (when using | |
2617 | posix directly). | |
2618 | ||
2619 | Args: | |
2620 | filename: The name of the current file. | |
2621 | clean_lines: A CleansedLines instance containing the file. | |
2622 | linenum: The number of the line to check. | |
2623 | error: The function to call with any errors found. | |
2624 | """ | |
2625 | line = clean_lines.elided[linenum] | |
2626 | for single_thread_func, multithread_safe_func, pattern in _THREADING_LIST: | |
2627 | # Additional pattern matching check to confirm that this is the | |
2628 | # function we are looking for | |
2629 | if Search(pattern, line): | |
2630 | error(filename, linenum, 'runtime/threadsafe_fn', 2, | |
2631 | 'Consider using ' + multithread_safe_func + | |
2632 | '...) instead of ' + single_thread_func + | |
2633 | '...) for improved thread safety.') | |
2634 | ||
2635 | ||
2636 | def CheckVlogArguments(filename, clean_lines, linenum, error): | |
2637 | """Checks that VLOG() is only used for defining a logging level. | |
2638 | ||
2639 | For example, VLOG(2) is correct. VLOG(INFO), VLOG(WARNING), VLOG(ERROR), and | |
2640 | VLOG(FATAL) are not. | |
2641 | ||
2642 | Args: | |
2643 | filename: The name of the current file. | |
2644 | clean_lines: A CleansedLines instance containing the file. | |
2645 | linenum: The number of the line to check. | |
2646 | error: The function to call with any errors found. | |
2647 | """ | |
2648 | line = clean_lines.elided[linenum] | |
2649 | if Search(r'\bVLOG\((INFO|ERROR|WARNING|DFATAL|FATAL)\)', line): | |
2650 | error(filename, linenum, 'runtime/vlog', 5, | |
2651 | 'VLOG() should be used with numeric verbosity level. ' | |
2652 | 'Use LOG() if you want symbolic severity levels.') | |
2653 | ||
2654 | # Matches invalid increment: *count++, which moves pointer instead of | |
2655 | # incrementing a value. | |
2656 | _RE_PATTERN_INVALID_INCREMENT = re.compile( | |
2657 | r'^\s*\*\w+(\+\+|--);') | |
2658 | ||
2659 | ||
2660 | def CheckInvalidIncrement(filename, clean_lines, linenum, error): | |
2661 | """Checks for invalid increment *count++. | |
2662 | ||
2663 | For example following function: | |
2664 | void increment_counter(int* count) { | |
2665 | *count++; | |
2666 | } | |
2667 | is invalid, because it effectively does count++, moving pointer, and should | |
2668 | be replaced with ++*count, (*count)++ or *count += 1. | |
2669 | ||
2670 | Args: | |
2671 | filename: The name of the current file. | |
2672 | clean_lines: A CleansedLines instance containing the file. | |
2673 | linenum: The number of the line to check. | |
2674 | error: The function to call with any errors found. | |
2675 | """ | |
2676 | line = clean_lines.elided[linenum] | |
2677 | if _RE_PATTERN_INVALID_INCREMENT.match(line): | |
2678 | error(filename, linenum, 'runtime/invalid_increment', 5, | |
2679 | 'Changing pointer instead of value (or unused value of operator*).') | |
2680 | ||
2681 | ||
2682 | def IsMacroDefinition(clean_lines, linenum): | |
2683 | if Search(r'^#define', clean_lines[linenum]): | |
2684 | return True | |
2685 | ||
2686 | if linenum > 0 and Search(r'\\$', clean_lines[linenum - 1]): | |
2687 | return True | |
2688 | ||
2689 | return False | |
2690 | ||
2691 | ||
2692 | def IsForwardClassDeclaration(clean_lines, linenum): | |
2693 | return Match(r'^\s*(\btemplate\b)*.*class\s+\w+;\s*$', clean_lines[linenum]) | |
2694 | ||
2695 | ||
2696 | class _BlockInfo(object): | |
2697 | """Stores information about a generic block of code.""" | |
2698 | ||
2699 | def __init__(self, linenum, seen_open_brace): | |
2700 | self.starting_linenum = linenum | |
2701 | self.seen_open_brace = seen_open_brace | |
2702 | self.open_parentheses = 0 | |
2703 | self.inline_asm = _NO_ASM | |
2704 | self.check_namespace_indentation = False | |
2705 | ||
2706 | def CheckBegin(self, filename, clean_lines, linenum, error): | |
2707 | """Run checks that applies to text up to the opening brace. | |
2708 | ||
2709 | This is mostly for checking the text after the class identifier | |
2710 | and the "{", usually where the base class is specified. For other | |
2711 | blocks, there isn't much to check, so we always pass. | |
2712 | ||
2713 | Args: | |
2714 | filename: The name of the current file. | |
2715 | clean_lines: A CleansedLines instance containing the file. | |
2716 | linenum: The number of the line to check. | |
2717 | error: The function to call with any errors found. | |
2718 | """ | |
2719 | pass | |
2720 | ||
2721 | def CheckEnd(self, filename, clean_lines, linenum, error): | |
2722 | """Run checks that applies to text after the closing brace. | |
2723 | ||
2724 | This is mostly used for checking end of namespace comments. | |
2725 | ||
2726 | Args: | |
2727 | filename: The name of the current file. | |
2728 | clean_lines: A CleansedLines instance containing the file. | |
2729 | linenum: The number of the line to check. | |
2730 | error: The function to call with any errors found. | |
2731 | """ | |
2732 | pass | |
2733 | ||
2734 | def IsBlockInfo(self): | |
2735 | """Returns true if this block is a _BlockInfo. | |
2736 | ||
2737 | This is convenient for verifying that an object is an instance of | |
2738 | a _BlockInfo, but not an instance of any of the derived classes. | |
2739 | ||
2740 | Returns: | |
2741 | True for this class, False for derived classes. | |
2742 | """ | |
2743 | return self.__class__ == _BlockInfo | |
2744 | ||
2745 | ||
2746 | class _ExternCInfo(_BlockInfo): | |
2747 | """Stores information about an 'extern "C"' block.""" | |
2748 | ||
2749 | def __init__(self, linenum): | |
2750 | _BlockInfo.__init__(self, linenum, True) | |
2751 | ||
2752 | ||
2753 | class _ClassInfo(_BlockInfo): | |
2754 | """Stores information about a class.""" | |
2755 | ||
2756 | def __init__(self, name, class_or_struct, clean_lines, linenum): | |
2757 | _BlockInfo.__init__(self, linenum, False) | |
2758 | self.name = name | |
2759 | self.is_derived = False | |
2760 | self.check_namespace_indentation = True | |
2761 | if class_or_struct == 'struct': | |
2762 | self.access = 'public' | |
2763 | self.is_struct = True | |
2764 | else: | |
2765 | self.access = 'private' | |
2766 | self.is_struct = False | |
2767 | ||
2768 | # Remember initial indentation level for this class. Using raw_lines here | |
2769 | # instead of elided to account for leading comments. | |
2770 | self.class_indent = GetIndentLevel(clean_lines.raw_lines[linenum]) | |
2771 | ||
2772 | # Try to find the end of the class. This will be confused by things like: | |
2773 | # class A { | |
2774 | # } *x = { ... | |
2775 | # | |
2776 | # But it's still good enough for CheckSectionSpacing. | |
2777 | self.last_line = 0 | |
2778 | depth = 0 | |
2779 | for i in range(linenum, clean_lines.NumLines()): | |
2780 | line = clean_lines.elided[i] | |
2781 | depth += line.count('{') - line.count('}') | |
2782 | if not depth: | |
2783 | self.last_line = i | |
2784 | break | |
2785 | ||
2786 | def CheckBegin(self, filename, clean_lines, linenum, error): | |
2787 | # Look for a bare ':' | |
2788 | if Search('(^|[^:]):($|[^:])', clean_lines.elided[linenum]): | |
2789 | self.is_derived = True | |
2790 | ||
2791 | def CheckEnd(self, filename, clean_lines, linenum, error): | |
2792 | # If there is a DISALLOW macro, it should appear near the end of | |
2793 | # the class. | |
2794 | seen_last_thing_in_class = False | |
2795 | for i in xrange(linenum - 1, self.starting_linenum, -1): | |
2796 | match = Search( | |
2797 | r'\b(DISALLOW_COPY_AND_ASSIGN|DISALLOW_IMPLICIT_CONSTRUCTORS)\(' + | |
2798 | self.name + r'\)', | |
2799 | clean_lines.elided[i]) | |
2800 | if match: | |
2801 | if seen_last_thing_in_class: | |
2802 | error(filename, i, 'readability/constructors', 3, | |
2803 | match.group(1) + ' should be the last thing in the class') | |
2804 | break | |
2805 | ||
2806 | if not Match(r'^\s*$', clean_lines.elided[i]): | |
2807 | seen_last_thing_in_class = True | |
2808 | ||
2809 | # Check that closing brace is aligned with beginning of the class. | |
2810 | # Only do this if the closing brace is indented by only whitespaces. | |
2811 | # This means we will not check single-line class definitions. | |
2812 | indent = Match(r'^( *)\}', clean_lines.elided[linenum]) | |
2813 | if indent and len(indent.group(1)) != self.class_indent: | |
2814 | if self.is_struct: | |
2815 | parent = 'struct ' + self.name | |
2816 | else: | |
2817 | parent = 'class ' + self.name | |
2818 | error(filename, linenum, 'whitespace/indent', 3, | |
2819 | 'Closing brace should be aligned with beginning of %s' % parent) | |
2820 | ||
2821 | ||
2822 | class _NamespaceInfo(_BlockInfo): | |
2823 | """Stores information about a namespace.""" | |
2824 | ||
2825 | def __init__(self, name, linenum): | |
2826 | _BlockInfo.__init__(self, linenum, False) | |
2827 | self.name = name or '' | |
2828 | self.check_namespace_indentation = True | |
2829 | ||
2830 | def CheckEnd(self, filename, clean_lines, linenum, error): | |
2831 | """Check end of namespace comments.""" | |
2832 | line = clean_lines.raw_lines[linenum] | |
2833 | ||
2834 | # Check how many lines is enclosed in this namespace. Don't issue | |
2835 | # warning for missing namespace comments if there aren't enough | |
2836 | # lines. However, do apply checks if there is already an end of | |
2837 | # namespace comment and it's incorrect. | |
2838 | # | |
2839 | # TODO(unknown): We always want to check end of namespace comments | |
2840 | # if a namespace is large, but sometimes we also want to apply the | |
2841 | # check if a short namespace contained nontrivial things (something | |
2842 | # other than forward declarations). There is currently no logic on | |
2843 | # deciding what these nontrivial things are, so this check is | |
2844 | # triggered by namespace size only, which works most of the time. | |
2845 | if (linenum - self.starting_linenum < 10 | |
2846 | and not Match(r'^\s*};*\s*(//|/\*).*\bnamespace\b', line)): | |
2847 | return | |
2848 | ||
2849 | # Look for matching comment at end of namespace. | |
2850 | # | |
2851 | # Note that we accept C style "/* */" comments for terminating | |
2852 | # namespaces, so that code that terminate namespaces inside | |
2853 | # preprocessor macros can be cpplint clean. | |
2854 | # | |
2855 | # We also accept stuff like "// end of namespace <name>." with the | |
2856 | # period at the end. | |
2857 | # | |
2858 | # Besides these, we don't accept anything else, otherwise we might | |
2859 | # get false negatives when existing comment is a substring of the | |
2860 | # expected namespace. | |
2861 | if self.name: | |
2862 | # Named namespace | |
2863 | if not Match((r'^\s*};*\s*(//|/\*).*\bnamespace\s+' + | |
2864 | re.escape(self.name) + r'[\*/\.\\\s]*$'), | |
2865 | line): | |
2866 | error(filename, linenum, 'readability/namespace', 5, | |
2867 | 'Namespace should be terminated with "// namespace %s"' % | |
2868 | self.name) | |
2869 | else: | |
2870 | # Anonymous namespace | |
2871 | if not Match(r'^\s*};*\s*(//|/\*).*\bnamespace[\*/\.\\\s]*$', line): | |
2872 | # If "// namespace anonymous" or "// anonymous namespace (more text)", | |
2873 | # mention "// anonymous namespace" as an acceptable form | |
2874 | if Match(r'^\s*}.*\b(namespace anonymous|anonymous namespace)\b', line): | |
2875 | error(filename, linenum, 'readability/namespace', 5, | |
2876 | 'Anonymous namespace should be terminated with "// namespace"' | |
2877 | ' or "// anonymous namespace"') | |
2878 | else: | |
2879 | error(filename, linenum, 'readability/namespace', 5, | |
2880 | 'Anonymous namespace should be terminated with "// namespace"') | |
2881 | ||
2882 | ||
2883 | class _PreprocessorInfo(object): | |
2884 | """Stores checkpoints of nesting stacks when #if/#else is seen.""" | |
2885 | ||
2886 | def __init__(self, stack_before_if): | |
2887 | # The entire nesting stack before #if | |
2888 | self.stack_before_if = stack_before_if | |
2889 | ||
2890 | # The entire nesting stack up to #else | |
2891 | self.stack_before_else = [] | |
2892 | ||
2893 | # Whether we have already seen #else or #elif | |
2894 | self.seen_else = False | |
2895 | ||
2896 | ||
2897 | class NestingState(object): | |
2898 | """Holds states related to parsing braces.""" | |
2899 | ||
2900 | def __init__(self): | |
2901 | # Stack for tracking all braces. An object is pushed whenever we | |
2902 | # see a "{", and popped when we see a "}". Only 3 types of | |
2903 | # objects are possible: | |
2904 | # - _ClassInfo: a class or struct. | |
2905 | # - _NamespaceInfo: a namespace. | |
2906 | # - _BlockInfo: some other type of block. | |
2907 | self.stack = [] | |
2908 | ||
2909 | # Top of the previous stack before each Update(). | |
2910 | # | |
2911 | # Because the nesting_stack is updated at the end of each line, we | |
2912 | # had to do some convoluted checks to find out what is the current | |
2913 | # scope at the beginning of the line. This check is simplified by | |
2914 | # saving the previous top of nesting stack. | |
2915 | # | |
2916 | # We could save the full stack, but we only need the top. Copying | |
2917 | # the full nesting stack would slow down cpplint by ~10%. | |
2918 | self.previous_stack_top = [] | |
2919 | ||
2920 | # Stack of _PreprocessorInfo objects. | |
2921 | self.pp_stack = [] | |
2922 | ||
2923 | def SeenOpenBrace(self): | |
2924 | """Check if we have seen the opening brace for the innermost block. | |
2925 | ||
2926 | Returns: | |
2927 | True if we have seen the opening brace, False if the innermost | |
2928 | block is still expecting an opening brace. | |
2929 | """ | |
2930 | return (not self.stack) or self.stack[-1].seen_open_brace | |
2931 | ||
2932 | def InNamespaceBody(self): | |
2933 | """Check if we are currently one level inside a namespace body. | |
2934 | ||
2935 | Returns: | |
2936 | True if top of the stack is a namespace block, False otherwise. | |
2937 | """ | |
2938 | return self.stack and isinstance(self.stack[-1], _NamespaceInfo) | |
2939 | ||
2940 | def InExternC(self): | |
2941 | """Check if we are currently one level inside an 'extern "C"' block. | |
2942 | ||
2943 | Returns: | |
2944 | True if top of the stack is an extern block, False otherwise. | |
2945 | """ | |
2946 | return self.stack and isinstance(self.stack[-1], _ExternCInfo) | |
2947 | ||
2948 | def InClassDeclaration(self): | |
2949 | """Check if we are currently one level inside a class or struct declaration. | |
2950 | ||
2951 | Returns: | |
2952 | True if top of the stack is a class/struct, False otherwise. | |
2953 | """ | |
2954 | return self.stack and isinstance(self.stack[-1], _ClassInfo) | |
2955 | ||
2956 | def InAsmBlock(self): | |
2957 | """Check if we are currently one level inside an inline ASM block. | |
2958 | ||
2959 | Returns: | |
2960 | True if the top of the stack is a block containing inline ASM. | |
2961 | """ | |
2962 | return self.stack and self.stack[-1].inline_asm != _NO_ASM | |
2963 | ||
2964 | def InTemplateArgumentList(self, clean_lines, linenum, pos): | |
2965 | """Check if current position is inside template argument list. | |
2966 | ||
2967 | Args: | |
2968 | clean_lines: A CleansedLines instance containing the file. | |
2969 | linenum: The number of the line to check. | |
2970 | pos: position just after the suspected template argument. | |
2971 | Returns: | |
2972 | True if (linenum, pos) is inside template arguments. | |
2973 | """ | |
2974 | while linenum < clean_lines.NumLines(): | |
2975 | # Find the earliest character that might indicate a template argument | |
2976 | line = clean_lines.elided[linenum] | |
2977 | match = Match(r'^[^{};=\[\]\.<>]*(.)', line[pos:]) | |
2978 | if not match: | |
2979 | linenum += 1 | |
2980 | pos = 0 | |
2981 | continue | |
2982 | token = match.group(1) | |
2983 | pos += len(match.group(0)) | |
2984 | ||
2985 | # These things do not look like template argument list: | |
2986 | # class Suspect { | |
2987 | # class Suspect x; } | |
2988 | if token in ('{', '}', ';'): return False | |
2989 | ||
2990 | # These things look like template argument list: | |
2991 | # template <class Suspect> | |
2992 | # template <class Suspect = default_value> | |
2993 | # template <class Suspect[]> | |
2994 | # template <class Suspect...> | |
2995 | if token in ('>', '=', '[', ']', '.'): return True | |
2996 | ||
2997 | # Check if token is an unmatched '<'. | |
2998 | # If not, move on to the next character. | |
2999 | if token != '<': | |
3000 | pos += 1 | |
3001 | if pos >= len(line): | |
3002 | linenum += 1 | |
3003 | pos = 0 | |
3004 | continue | |
3005 | ||
3006 | # We can't be sure if we just find a single '<', and need to | |
3007 | # find the matching '>'. | |
3008 | (_, end_line, end_pos) = CloseExpression(clean_lines, linenum, pos - 1) | |
3009 | if end_pos < 0: | |
3010 | # Not sure if template argument list or syntax error in file | |
3011 | return False | |
3012 | linenum = end_line | |
3013 | pos = end_pos | |
3014 | return False | |
3015 | ||
3016 | def UpdatePreprocessor(self, line): | |
3017 | """Update preprocessor stack. | |
3018 | ||
3019 | We need to handle preprocessors due to classes like this: | |
3020 | #ifdef SWIG | |
3021 | struct ResultDetailsPageElementExtensionPoint { | |
3022 | #else | |
3023 | struct ResultDetailsPageElementExtensionPoint : public Extension { | |
3024 | #endif | |
3025 | ||
3026 | We make the following assumptions (good enough for most files): | |
3027 | - Preprocessor condition evaluates to true from #if up to first | |
3028 | #else/#elif/#endif. | |
3029 | ||
3030 | - Preprocessor condition evaluates to false from #else/#elif up | |
3031 | to #endif. We still perform lint checks on these lines, but | |
3032 | these do not affect nesting stack. | |
3033 | ||
3034 | Args: | |
3035 | line: current line to check. | |
3036 | """ | |
3037 | if Match(r'^\s*#\s*(if|ifdef|ifndef)\b', line): | |
3038 | # Beginning of #if block, save the nesting stack here. The saved | |
3039 | # stack will allow us to restore the parsing state in the #else case. | |
3040 | self.pp_stack.append(_PreprocessorInfo(copy.deepcopy(self.stack))) | |
3041 | elif Match(r'^\s*#\s*(else|elif)\b', line): | |
3042 | # Beginning of #else block | |
3043 | if self.pp_stack: | |
3044 | if not self.pp_stack[-1].seen_else: | |
3045 | # This is the first #else or #elif block. Remember the | |
3046 | # whole nesting stack up to this point. This is what we | |
3047 | # keep after the #endif. | |
3048 | self.pp_stack[-1].seen_else = True | |
3049 | self.pp_stack[-1].stack_before_else = copy.deepcopy(self.stack) | |
3050 | ||
3051 | # Restore the stack to how it was before the #if | |
3052 | self.stack = copy.deepcopy(self.pp_stack[-1].stack_before_if) | |
3053 | else: | |
3054 | # TODO(unknown): unexpected #else, issue warning? | |
3055 | pass | |
3056 | elif Match(r'^\s*#\s*endif\b', line): | |
3057 | # End of #if or #else blocks. | |
3058 | if self.pp_stack: | |
3059 | # If we saw an #else, we will need to restore the nesting | |
3060 | # stack to its former state before the #else, otherwise we | |
3061 | # will just continue from where we left off. | |
3062 | if self.pp_stack[-1].seen_else: | |
3063 | # Here we can just use a shallow copy since we are the last | |
3064 | # reference to it. | |
3065 | self.stack = self.pp_stack[-1].stack_before_else | |
3066 | # Drop the corresponding #if | |
3067 | self.pp_stack.pop() | |
3068 | else: | |
3069 | # TODO(unknown): unexpected #endif, issue warning? | |
3070 | pass | |
3071 | ||
3072 | # TODO(unknown): Update() is too long, but we will refactor later. | |
3073 | def Update(self, filename, clean_lines, linenum, error): | |
3074 | """Update nesting state with current line. | |
3075 | ||
3076 | Args: | |
3077 | filename: The name of the current file. | |
3078 | clean_lines: A CleansedLines instance containing the file. | |
3079 | linenum: The number of the line to check. | |
3080 | error: The function to call with any errors found. | |
3081 | """ | |
3082 | line = clean_lines.elided[linenum] | |
3083 | ||
3084 | # Remember top of the previous nesting stack. | |
3085 | # | |
3086 | # The stack is always pushed/popped and not modified in place, so | |
3087 | # we can just do a shallow copy instead of copy.deepcopy. Using | |
3088 | # deepcopy would slow down cpplint by ~28%. | |
3089 | if self.stack: | |
3090 | self.previous_stack_top = self.stack[-1] | |
3091 | else: | |
3092 | self.previous_stack_top = None | |
3093 | ||
3094 | # Update pp_stack | |
3095 | self.UpdatePreprocessor(line) | |
3096 | ||
3097 | # Count parentheses. This is to avoid adding struct arguments to | |
3098 | # the nesting stack. | |
3099 | if self.stack: | |
3100 | inner_block = self.stack[-1] | |
3101 | depth_change = line.count('(') - line.count(')') | |
3102 | inner_block.open_parentheses += depth_change | |
3103 | ||
3104 | # Also check if we are starting or ending an inline assembly block. | |
3105 | if inner_block.inline_asm in (_NO_ASM, _END_ASM): | |
3106 | if (depth_change != 0 and | |
3107 | inner_block.open_parentheses == 1 and | |
3108 | _MATCH_ASM.match(line)): | |
3109 | # Enter assembly block | |
3110 | inner_block.inline_asm = _INSIDE_ASM | |
3111 | else: | |
3112 | # Not entering assembly block. If previous line was _END_ASM, | |
3113 | # we will now shift to _NO_ASM state. | |
3114 | inner_block.inline_asm = _NO_ASM | |
3115 | elif (inner_block.inline_asm == _INSIDE_ASM and | |
3116 | inner_block.open_parentheses == 0): | |
3117 | # Exit assembly block | |
3118 | inner_block.inline_asm = _END_ASM | |
3119 | ||
3120 | # Consume namespace declaration at the beginning of the line. Do | |
3121 | # this in a loop so that we catch same line declarations like this: | |
3122 | # namespace proto2 { namespace bridge { class MessageSet; } } | |
3123 | while True: | |
3124 | # Match start of namespace. The "\b\s*" below catches namespace | |
3125 | # declarations even if it weren't followed by a whitespace, this | |
3126 | # is so that we don't confuse our namespace checker. The | |
3127 | # missing spaces will be flagged by CheckSpacing. | |
3128 | namespace_decl_match = Match(r'^\s*namespace\b\s*([:\w]+)?(.*)$', line) | |
3129 | if not namespace_decl_match: | |
3130 | break | |
3131 | ||
3132 | new_namespace = _NamespaceInfo(namespace_decl_match.group(1), linenum) | |
3133 | self.stack.append(new_namespace) | |
3134 | ||
3135 | line = namespace_decl_match.group(2) | |
3136 | if line.find('{') != -1: | |
3137 | new_namespace.seen_open_brace = True | |
3138 | line = line[line.find('{') + 1:] | |
3139 | ||
3140 | # Look for a class declaration in whatever is left of the line | |
3141 | # after parsing namespaces. The regexp accounts for decorated classes | |
3142 | # such as in: | |
3143 | # class LOCKABLE API Object { | |
3144 | # }; | |
3145 | class_decl_match = Match( | |
3146 | r'^(\s*(?:template\s*<[\w\s<>,:=]*>\s*)?' | |
3147 | r'(class|struct)\s+(?:[a-zA-Z0-9_]+\s+)*(\w+(?:::\w+)*))' | |
3148 | r'(.*)$', line) | |
3149 | if (class_decl_match and | |
3150 | (not self.stack or self.stack[-1].open_parentheses == 0)): | |
3151 | # We do not want to accept classes that are actually template arguments: | |
3152 | # template <class Ignore1, | |
3153 | # class Ignore2 = Default<Args>, | |
3154 | # template <Args> class Ignore3> | |
3155 | # void Function() {}; | |
3156 | # | |
3157 | # To avoid template argument cases, we scan forward and look for | |
3158 | # an unmatched '>'. If we see one, assume we are inside a | |
3159 | # template argument list. | |
3160 | end_declaration = len(class_decl_match.group(1)) | |
3161 | if not self.InTemplateArgumentList(clean_lines, linenum, end_declaration): | |
3162 | self.stack.append(_ClassInfo( | |
3163 | class_decl_match.group(3), class_decl_match.group(2), | |
3164 | clean_lines, linenum)) | |
3165 | line = class_decl_match.group(4) | |
3166 | ||
3167 | # If we have not yet seen the opening brace for the innermost block, | |
3168 | # run checks here. | |
3169 | if not self.SeenOpenBrace(): | |
3170 | self.stack[-1].CheckBegin(filename, clean_lines, linenum, error) | |
3171 | ||
3172 | # Update access control if we are inside a class/struct | |
3173 | if self.stack and isinstance(self.stack[-1], _ClassInfo): | |
3174 | classinfo = self.stack[-1] | |
3175 | access_match = Match( | |
3176 | r'^(.*)\b(public|private|protected|signals)(\s+(?:slots\s*)?)?' | |
3177 | r':(?:[^:]|$)', | |
3178 | line) | |
3179 | if access_match: | |
3180 | classinfo.access = access_match.group(2) | |
3181 | ||
3182 | # Check that access keywords are indented +1 space. Skip this | |
3183 | # check if the keywords are not preceded by whitespaces. | |
3184 | indent = access_match.group(1) | |
3185 | if (len(indent) != classinfo.class_indent + 1 and | |
3186 | Match(r'^\s*$', indent)): | |
3187 | if classinfo.is_struct: | |
3188 | parent = 'struct ' + classinfo.name | |
3189 | else: | |
3190 | parent = 'class ' + classinfo.name | |
3191 | slots = '' | |
3192 | if access_match.group(3): | |
3193 | slots = access_match.group(3) | |
3194 | error(filename, linenum, 'whitespace/indent', 3, | |
3195 | '%s%s: should be indented +1 space inside %s' % ( | |
3196 | access_match.group(2), slots, parent)) | |
3197 | ||
3198 | # Consume braces or semicolons from what's left of the line | |
3199 | while True: | |
3200 | # Match first brace, semicolon, or closed parenthesis. | |
3201 | matched = Match(r'^[^{;)}]*([{;)}])(.*)$', line) | |
3202 | if not matched: | |
3203 | break | |
3204 | ||
3205 | token = matched.group(1) | |
3206 | if token == '{': | |
3207 | # If namespace or class hasn't seen a opening brace yet, mark | |
3208 | # namespace/class head as complete. Push a new block onto the | |
3209 | # stack otherwise. | |
3210 | if not self.SeenOpenBrace(): | |
3211 | self.stack[-1].seen_open_brace = True | |
3212 | elif Match(r'^extern\s*"[^"]*"\s*\{', line): | |
3213 | self.stack.append(_ExternCInfo(linenum)) | |
3214 | else: | |
3215 | self.stack.append(_BlockInfo(linenum, True)) | |
3216 | if _MATCH_ASM.match(line): | |
3217 | self.stack[-1].inline_asm = _BLOCK_ASM | |
3218 | ||
3219 | elif token == ';' or token == ')': | |
3220 | # If we haven't seen an opening brace yet, but we already saw | |
3221 | # a semicolon, this is probably a forward declaration. Pop | |
3222 | # the stack for these. | |
3223 | # | |
3224 | # Similarly, if we haven't seen an opening brace yet, but we | |
3225 | # already saw a closing parenthesis, then these are probably | |
3226 | # function arguments with extra "class" or "struct" keywords. | |
3227 | # Also pop these stack for these. | |
3228 | if not self.SeenOpenBrace(): | |
3229 | self.stack.pop() | |
3230 | else: # token == '}' | |
3231 | # Perform end of block checks and pop the stack. | |
3232 | if self.stack: | |
3233 | self.stack[-1].CheckEnd(filename, clean_lines, linenum, error) | |
3234 | self.stack.pop() | |
3235 | line = matched.group(2) | |
3236 | ||
3237 | def InnermostClass(self): | |
3238 | """Get class info on the top of the stack. | |
3239 | ||
3240 | Returns: | |
3241 | A _ClassInfo object if we are inside a class, or None otherwise. | |
3242 | """ | |
3243 | for i in range(len(self.stack), 0, -1): | |
3244 | classinfo = self.stack[i - 1] | |
3245 | if isinstance(classinfo, _ClassInfo): | |
3246 | return classinfo | |
3247 | return None | |
3248 | ||
3249 | def CheckCompletedBlocks(self, filename, error): | |
3250 | """Checks that all classes and namespaces have been completely parsed. | |
3251 | ||
3252 | Call this when all lines in a file have been processed. | |
3253 | Args: | |
3254 | filename: The name of the current file. | |
3255 | error: The function to call with any errors found. | |
3256 | """ | |
3257 | # Note: This test can result in false positives if #ifdef constructs | |
3258 | # get in the way of brace matching. See the testBuildClass test in | |
3259 | # cpplint_unittest.py for an example of this. | |
3260 | for obj in self.stack: | |
3261 | if isinstance(obj, _ClassInfo): | |
3262 | error(filename, obj.starting_linenum, 'build/class', 5, | |
3263 | 'Failed to find complete declaration of class %s' % | |
3264 | obj.name) | |
3265 | elif isinstance(obj, _NamespaceInfo): | |
3266 | error(filename, obj.starting_linenum, 'build/namespaces', 5, | |
3267 | 'Failed to find complete declaration of namespace %s' % | |
3268 | obj.name) | |
3269 | ||
3270 | ||
3271 | def CheckForNonStandardConstructs(filename, clean_lines, linenum, | |
3272 | nesting_state, error): | |
3273 | r"""Logs an error if we see certain non-ANSI constructs ignored by gcc-2. | |
3274 | ||
3275 | Complain about several constructs which gcc-2 accepts, but which are | |
3276 | not standard C++. Warning about these in lint is one way to ease the | |
3277 | transition to new compilers. | |
3278 | - put storage class first (e.g. "static const" instead of "const static"). | |
3279 | - "%lld" instead of %qd" in printf-type functions. | |
3280 | - "%1$d" is non-standard in printf-type functions. | |
3281 | - "\%" is an undefined character escape sequence. | |
3282 | - text after #endif is not allowed. | |
3283 | - invalid inner-style forward declaration. | |
3284 | - >? and <? operators, and their >?= and <?= cousins. | |
3285 | ||
3286 | Additionally, check for constructor/destructor style violations and reference | |
3287 | members, as it is very convenient to do so while checking for | |
3288 | gcc-2 compliance. | |
3289 | ||
3290 | Args: | |
3291 | filename: The name of the current file. | |
3292 | clean_lines: A CleansedLines instance containing the file. | |
3293 | linenum: The number of the line to check. | |
3294 | nesting_state: A NestingState instance which maintains information about | |
3295 | the current stack of nested blocks being parsed. | |
3296 | error: A callable to which errors are reported, which takes 4 arguments: | |
3297 | filename, line number, error level, and message | |
3298 | """ | |
3299 | ||
3300 | # Remove comments from the line, but leave in strings for now. | |
3301 | line = clean_lines.lines[linenum] | |
3302 | ||
3303 | if Search(r'printf\s*\(.*".*%[-+ ]?\d*q', line): | |
3304 | error(filename, linenum, 'runtime/printf_format', 3, | |
3305 | '%q in format strings is deprecated. Use %ll instead.') | |
3306 | ||
3307 | if Search(r'printf\s*\(.*".*%\d+\$', line): | |
3308 | error(filename, linenum, 'runtime/printf_format', 2, | |
3309 | '%N$ formats are unconventional. Try rewriting to avoid them.') | |
3310 | ||
3311 | # Remove escaped backslashes before looking for undefined escapes. | |
3312 | line = line.replace('\\\\', '') | |
3313 | ||
3314 | if Search(r'("|\').*\\(%|\[|\(|{)', line): | |
3315 | error(filename, linenum, 'build/printf_format', 3, | |
3316 | '%, [, (, and { are undefined character escapes. Unescape them.') | |
3317 | ||
3318 | # For the rest, work with both comments and strings removed. | |
3319 | line = clean_lines.elided[linenum] | |
3320 | ||
3321 | if Search(r'\b(const|volatile|void|char|short|int|long' | |
3322 | r'|float|double|signed|unsigned' | |
3323 | r'|schar|u?int8|u?int16|u?int32|u?int64)' | |
3324 | r'\s+(register|static|extern|typedef)\b', | |
3325 | line): | |
3326 | error(filename, linenum, 'build/storage_class', 5, | |
3327 | 'Storage-class specifier (static, extern, typedef, etc) should be ' | |
3328 | 'at the beginning of the declaration.') | |
3329 | ||
3330 | if Match(r'\s*#\s*endif\s*[^/\s]+', line): | |
3331 | error(filename, linenum, 'build/endif_comment', 5, | |
3332 | 'Uncommented text after #endif is non-standard. Use a comment.') | |
3333 | ||
3334 | if Match(r'\s*class\s+(\w+\s*::\s*)+\w+\s*;', line): | |
3335 | error(filename, linenum, 'build/forward_decl', 5, | |
3336 | 'Inner-style forward declarations are invalid. Remove this line.') | |
3337 | ||
3338 | if Search(r'(\w+|[+-]?\d+(\.\d*)?)\s*(<|>)\?=?\s*(\w+|[+-]?\d+)(\.\d*)?', | |
3339 | line): | |
3340 | error(filename, linenum, 'build/deprecated', 3, | |
3341 | '>? and <? (max and min) operators are non-standard and deprecated.') | |
3342 | ||
3343 | if Search(r'^\s*const\s*string\s*&\s*\w+\s*;', line): | |
3344 | # TODO(unknown): Could it be expanded safely to arbitrary references, | |
3345 | # without triggering too many false positives? The first | |
3346 | # attempt triggered 5 warnings for mostly benign code in the regtest, hence | |
3347 | # the restriction. | |
3348 | # Here's the original regexp, for the reference: | |
3349 | # type_name = r'\w+((\s*::\s*\w+)|(\s*<\s*\w+?\s*>))?' | |
3350 | # r'\s*const\s*' + type_name + '\s*&\s*\w+\s*;' | |
3351 | error(filename, linenum, 'runtime/member_string_references', 2, | |
3352 | 'const string& members are dangerous. It is much better to use ' | |
3353 | 'alternatives, such as pointers or simple constants.') | |
3354 | ||
3355 | # Everything else in this function operates on class declarations. | |
3356 | # Return early if the top of the nesting stack is not a class, or if | |
3357 | # the class head is not completed yet. | |
3358 | classinfo = nesting_state.InnermostClass() | |
3359 | if not classinfo or not classinfo.seen_open_brace: | |
3360 | return | |
3361 | ||
3362 | # The class may have been declared with namespace or classname qualifiers. | |
3363 | # The constructor and destructor will not have those qualifiers. | |
3364 | base_classname = classinfo.name.split('::')[-1] | |
3365 | ||
3366 | # Look for single-argument constructors that aren't marked explicit. | |
3367 | # Technically a valid construct, but against style. | |
3368 | explicit_constructor_match = Match( | |
3369 | r'\s+(?:(?:inline|constexpr)\s+)*(explicit\s+)?' | |
3370 | r'(?:(?:inline|constexpr)\s+)*%s\s*' | |
3371 | r'\(((?:[^()]|\([^()]*\))*)\)' | |
3372 | % re.escape(base_classname), | |
3373 | line) | |
3374 | ||
3375 | if explicit_constructor_match: | |
3376 | is_marked_explicit = explicit_constructor_match.group(1) | |
3377 | ||
3378 | if not explicit_constructor_match.group(2): | |
3379 | constructor_args = [] | |
3380 | else: | |
3381 | constructor_args = explicit_constructor_match.group(2).split(',') | |
3382 | ||
3383 | # collapse arguments so that commas in template parameter lists and function | |
3384 | # argument parameter lists don't split arguments in two | |
3385 | i = 0 | |
3386 | while i < len(constructor_args): | |
3387 | constructor_arg = constructor_args[i] | |
3388 | while (constructor_arg.count('<') > constructor_arg.count('>') or | |
3389 | constructor_arg.count('(') > constructor_arg.count(')')): | |
3390 | constructor_arg += ',' + constructor_args[i + 1] | |
3391 | del constructor_args[i + 1] | |
3392 | constructor_args[i] = constructor_arg | |
3393 | i += 1 | |
3394 | ||
3395 | variadic_args = [arg for arg in constructor_args if '&&...' in arg] | |
3396 | defaulted_args = [arg for arg in constructor_args if '=' in arg] | |
3397 | noarg_constructor = (not constructor_args or # empty arg list | |
3398 | # 'void' arg specifier | |
3399 | (len(constructor_args) == 1 and | |
3400 | constructor_args[0].strip() == 'void')) | |
3401 | onearg_constructor = ((len(constructor_args) == 1 and # exactly one arg | |
3402 | not noarg_constructor) or | |
3403 | # all but at most one arg defaulted | |
3404 | (len(constructor_args) >= 1 and | |
3405 | not noarg_constructor and | |
3406 | len(defaulted_args) >= len(constructor_args) - 1) or | |
3407 | # variadic arguments with zero or one argument | |
3408 | (len(constructor_args) <= 2 and | |
3409 | len(variadic_args) >= 1)) | |
3410 | initializer_list_constructor = bool( | |
3411 | onearg_constructor and | |
3412 | Search(r'\bstd\s*::\s*initializer_list\b', constructor_args[0])) | |
3413 | copy_constructor = bool( | |
3414 | onearg_constructor and | |
3415 | Match(r'((const\s+(volatile\s+)?)?|(volatile\s+(const\s+)?))?' | |
3416 | r'%s(\s*<[^>]*>)?(\s+const)?\s*(?:<\w+>\s*)?&' | |
3417 | % re.escape(base_classname), constructor_args[0].strip())) | |
3418 | ||
3419 | if (not is_marked_explicit and | |
3420 | onearg_constructor and | |
3421 | not initializer_list_constructor and | |
3422 | not copy_constructor): | |
3423 | if defaulted_args or variadic_args: | |
3424 | error(filename, linenum, 'runtime/explicit', 5, | |
3425 | 'Constructors callable with one argument ' | |
3426 | 'should be marked explicit.') | |
3427 | else: | |
3428 | error(filename, linenum, 'runtime/explicit', 5, | |
3429 | 'Single-parameter constructors should be marked explicit.') | |
3430 | elif is_marked_explicit and not onearg_constructor: | |
3431 | if noarg_constructor: | |
3432 | error(filename, linenum, 'runtime/explicit', 5, | |
3433 | 'Zero-parameter constructors should not be marked explicit.') | |
3434 | ||
3435 | ||
3436 | def CheckSpacingForFunctionCall(filename, clean_lines, linenum, error): | |
3437 | """Checks for the correctness of various spacing around function calls. | |
3438 | ||
3439 | Args: | |
3440 | filename: The name of the current file. | |
3441 | clean_lines: A CleansedLines instance containing the file. | |
3442 | linenum: The number of the line to check. | |
3443 | error: The function to call with any errors found. | |
3444 | """ | |
3445 | line = clean_lines.elided[linenum] | |
3446 | ||
3447 | # Since function calls often occur inside if/for/while/switch | |
3448 | # expressions - which have their own, more liberal conventions - we | |
3449 | # first see if we should be looking inside such an expression for a | |
3450 | # function call, to which we can apply more strict standards. | |
3451 | fncall = line # if there's no control flow construct, look at whole line | |
3452 | for pattern in (r'\bif\s*\((.*)\)\s*{', | |
3453 | r'\bfor\s*\((.*)\)\s*{', | |
3454 | r'\bwhile\s*\((.*)\)\s*[{;]', | |
3455 | r'\bswitch\s*\((.*)\)\s*{'): | |
3456 | match = Search(pattern, line) | |
3457 | if match: | |
3458 | fncall = match.group(1) # look inside the parens for function calls | |
3459 | break | |
3460 | ||
3461 | # Except in if/for/while/switch, there should never be space | |
3462 | # immediately inside parens (eg "f( 3, 4 )"). We make an exception | |
3463 | # for nested parens ( (a+b) + c ). Likewise, there should never be | |
3464 | # a space before a ( when it's a function argument. I assume it's a | |
3465 | # function argument when the char before the whitespace is legal in | |
3466 | # a function name (alnum + _) and we're not starting a macro. Also ignore | |
3467 | # pointers and references to arrays and functions coz they're too tricky: | |
3468 | # we use a very simple way to recognize these: | |
3469 | # " (something)(maybe-something)" or | |
3470 | # " (something)(maybe-something," or | |
3471 | # " (something)[something]" | |
3472 | # Note that we assume the contents of [] to be short enough that | |
3473 | # they'll never need to wrap. | |
3474 | if ( # Ignore control structures. | |
3475 | not Search(r'\b(if|elif|for|while|switch|return|new|delete|catch|sizeof)\b', | |
3476 | fncall) and | |
3477 | # Ignore pointers/references to functions. | |
3478 | not Search(r' \([^)]+\)\([^)]*(\)|,$)', fncall) and | |
3479 | # Ignore pointers/references to arrays. | |
3480 | not Search(r' \([^)]+\)\[[^\]]+\]', fncall)): | |
3481 | if Search(r'\w\s*\(\s(?!\s*\\$)', fncall): # a ( used for a fn call | |
3482 | error(filename, linenum, 'whitespace/parens', 4, | |
3483 | 'Extra space after ( in function call') | |
3484 | elif Search(r'\(\s+(?!(\s*\\)|\()', fncall): | |
3485 | error(filename, linenum, 'whitespace/parens', 2, | |
3486 | 'Extra space after (') | |
3487 | if (Search(r'\w\s+\(', fncall) and | |
3488 | not Search(r'_{0,2}asm_{0,2}\s+_{0,2}volatile_{0,2}\s+\(', fncall) and | |
3489 | not Search(r'#\s*define|typedef|using\s+\w+\s*=', fncall) and | |
3490 | not Search(r'\w\s+\((\w+::)*\*\w+\)\(', fncall) and | |
3491 | not Search(r'\bcase\s+\(', fncall)): | |
3492 | # TODO(unknown): Space after an operator function seem to be a common | |
3493 | # error, silence those for now by restricting them to highest verbosity. | |
3494 | if Search(r'\boperator_*\b', line): | |
3495 | error(filename, linenum, 'whitespace/parens', 0, | |
3496 | 'Extra space before ( in function call') | |
3497 | else: | |
3498 | error(filename, linenum, 'whitespace/parens', 4, | |
3499 | 'Extra space before ( in function call') | |
3500 | # If the ) is followed only by a newline or a { + newline, assume it's | |
3501 | # part of a control statement (if/while/etc), and don't complain | |
3502 | if Search(r'[^)]\s+\)\s*[^{\s]', fncall): | |
3503 | # If the closing parenthesis is preceded by only whitespaces, | |
3504 | # try to give a more descriptive error message. | |
3505 | if Search(r'^\s+\)', fncall): | |
3506 | error(filename, linenum, 'whitespace/parens', 2, | |
3507 | 'Closing ) should be moved to the previous line') | |
3508 | else: | |
3509 | error(filename, linenum, 'whitespace/parens', 2, | |
3510 | 'Extra space before )') | |
3511 | ||
3512 | ||
3513 | def IsBlankLine(line): | |
3514 | """Returns true if the given line is blank. | |
3515 | ||
3516 | We consider a line to be blank if the line is empty or consists of | |
3517 | only white spaces. | |
3518 | ||
3519 | Args: | |
3520 | line: A line of a string. | |
3521 | ||
3522 | Returns: | |
3523 | True, if the given line is blank. | |
3524 | """ | |
3525 | return not line or line.isspace() | |
3526 | ||
3527 | ||
3528 | def CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line, | |
3529 | error): | |
3530 | is_namespace_indent_item = ( | |
3531 | len(nesting_state.stack) > 1 and | |
3532 | nesting_state.stack[-1].check_namespace_indentation and | |
3533 | isinstance(nesting_state.previous_stack_top, _NamespaceInfo) and | |
3534 | nesting_state.previous_stack_top == nesting_state.stack[-2]) | |
3535 | ||
3536 | if ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item, | |
3537 | clean_lines.elided, line): | |
3538 | CheckItemIndentationInNamespace(filename, clean_lines.elided, | |
3539 | line, error) | |
3540 | ||
3541 | ||
3542 | def CheckForFunctionLengths(filename, clean_lines, linenum, | |
3543 | function_state, error): | |
3544 | """Reports for long function bodies. | |
3545 | ||
3546 | For an overview why this is done, see: | |
3547 | https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions | |
3548 | ||
3549 | Uses a simplistic algorithm assuming other style guidelines | |
3550 | (especially spacing) are followed. | |
3551 | Only checks unindented functions, so class members are unchecked. | |
3552 | Trivial bodies are unchecked, so constructors with huge initializer lists | |
3553 | may be missed. | |
3554 | Blank/comment lines are not counted so as to avoid encouraging the removal | |
3555 | of vertical space and comments just to get through a lint check. | |
3556 | NOLINT *on the last line of a function* disables this check. | |
3557 | ||
3558 | Args: | |
3559 | filename: The name of the current file. | |
3560 | clean_lines: A CleansedLines instance containing the file. | |
3561 | linenum: The number of the line to check. | |
3562 | function_state: Current function name and lines in body so far. | |
3563 | error: The function to call with any errors found. | |
3564 | """ | |
3565 | lines = clean_lines.lines | |
3566 | line = lines[linenum] | |
3567 | joined_line = '' | |
3568 | ||
3569 | starting_func = False | |
3570 | regexp = r'(\w(\w|::|\*|\&|\s)*)\(' # decls * & space::name( ... | |
3571 | match_result = Match(regexp, line) | |
3572 | if match_result: | |
3573 | # If the name is all caps and underscores, figure it's a macro and | |
3574 | # ignore it, unless it's TEST or TEST_F. | |
3575 | function_name = match_result.group(1).split()[-1] | |
3576 | if function_name == 'TEST' or function_name == 'TEST_F' or ( | |
3577 | not Match(r'[A-Z_]+$', function_name)): | |
3578 | starting_func = True | |
3579 | ||
3580 | if starting_func: | |
3581 | body_found = False | |
3582 | for start_linenum in xrange(linenum, clean_lines.NumLines()): | |
3583 | start_line = lines[start_linenum] | |
3584 | joined_line += ' ' + start_line.lstrip() | |
3585 | if Search(r'(;|})', start_line): # Declarations and trivial functions | |
3586 | body_found = True | |
3587 | break # ... ignore | |
3588 | if Search(r'{', start_line): | |
3589 | body_found = True | |
3590 | function = Search(r'((\w|:)*)\(', line).group(1) | |
3591 | if Match(r'TEST', function): # Handle TEST... macros | |
3592 | parameter_regexp = Search(r'(\(.*\))', joined_line) | |
3593 | if parameter_regexp: # Ignore bad syntax | |
3594 | function += parameter_regexp.group(1) | |
3595 | else: | |
3596 | function += '()' | |
3597 | function_state.Begin(function) | |
3598 | break | |
3599 | if not body_found: | |
3600 | # No body for the function (or evidence of a non-function) was found. | |
3601 | error(filename, linenum, 'readability/fn_size', 5, | |
3602 | 'Lint failed to find start of function body.') | |
3603 | elif Match(r'^\}\s*$', line): # function end | |
3604 | function_state.Check(error, filename, linenum) | |
3605 | function_state.End() | |
3606 | elif not Match(r'^\s*$', line): | |
3607 | function_state.Count() # Count non-blank/non-comment lines. | |
3608 | ||
3609 | ||
3610 | _RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?') | |
3611 | ||
3612 | ||
3613 | def CheckComment(line, filename, linenum, next_line_start, error): | |
3614 | """Checks for common mistakes in comments. | |
3615 | ||
3616 | Args: | |
3617 | line: The line in question. | |
3618 | filename: The name of the current file. | |
3619 | linenum: The number of the line to check. | |
3620 | next_line_start: The first non-whitespace column of the next line. | |
3621 | error: The function to call with any errors found. | |
3622 | """ | |
3623 | commentpos = line.find('//') | |
3624 | if commentpos != -1: | |
3625 | # Check if the // may be in quotes. If so, ignore it | |
3626 | if re.sub(r'\\.', '', line[0:commentpos]).count('"') % 2 == 0: | |
3627 | # Allow one space for new scopes, two spaces otherwise: | |
3628 | if (not (Match(r'^.*{ *//', line) and next_line_start == commentpos) and | |
3629 | ((commentpos >= 1 and | |
3630 | line[commentpos-1] not in string.whitespace) or | |
3631 | (commentpos >= 2 and | |
3632 | line[commentpos-2] not in string.whitespace))): | |
3633 | error(filename, linenum, 'whitespace/comments', 2, | |
3634 | 'At least two spaces is best between code and comments') | |
3635 | ||
3636 | # Checks for common mistakes in TODO comments. | |
3637 | comment = line[commentpos:] | |
3638 | match = _RE_PATTERN_TODO.match(comment) | |
3639 | if match: | |
3640 | # One whitespace is correct; zero whitespace is handled elsewhere. | |
3641 | leading_whitespace = match.group(1) | |
3642 | if len(leading_whitespace) > 1: | |
3643 | error(filename, linenum, 'whitespace/todo', 2, | |
3644 | 'Too many spaces before TODO') | |
3645 | ||
3646 | username = match.group(2) | |
3647 | if not username: | |
3648 | error(filename, linenum, 'readability/todo', 2, | |
3649 | 'Missing username in TODO; it should look like ' | |
3650 | '"// TODO(my_username): Stuff."') | |
3651 | ||
3652 | middle_whitespace = match.group(3) | |
3653 | # Comparisons made explicit for correctness -- pylint: disable=g-explicit-bool-comparison | |
3654 | if middle_whitespace != ' ' and middle_whitespace != '': | |
3655 | error(filename, linenum, 'whitespace/todo', 2, | |
3656 | 'TODO(my_username) should be followed by a space') | |
3657 | ||
3658 | # If the comment contains an alphanumeric character, there | |
3659 | # should be a space somewhere between it and the // unless | |
3660 | # it's a /// or //! Doxygen comment. | |
3661 | if (Match(r'//[^ ]*\w', comment) and | |
3662 | not Match(r'(///|//\!)(\s+|$)', comment)): | |
3663 | error(filename, linenum, 'whitespace/comments', 4, | |
3664 | 'Should have a space between // and comment') | |
3665 | ||
3666 | ||
3667 | def CheckSpacing(filename, clean_lines, linenum, nesting_state, error): | |
3668 | """Checks for the correctness of various spacing issues in the code. | |
3669 | ||
3670 | Things we check for: spaces around operators, spaces after | |
3671 | if/for/while/switch, no spaces around parens in function calls, two | |
3672 | spaces between code and comment, don't start a block with a blank | |
3673 | line, don't end a function with a blank line, don't add a blank line | |
3674 | after public/protected/private, don't have too many blank lines in a row. | |
3675 | ||
3676 | Args: | |
3677 | filename: The name of the current file. | |
3678 | clean_lines: A CleansedLines instance containing the file. | |
3679 | linenum: The number of the line to check. | |
3680 | nesting_state: A NestingState instance which maintains information about | |
3681 | the current stack of nested blocks being parsed. | |
3682 | error: The function to call with any errors found. | |
3683 | """ | |
3684 | ||
3685 | # Don't use "elided" lines here, otherwise we can't check commented lines. | |
3686 | # Don't want to use "raw" either, because we don't want to check inside C++11 | |
3687 | # raw strings, | |
3688 | raw = clean_lines.lines_without_raw_strings | |
3689 | line = raw[linenum] | |
3690 | ||
3691 | # Before nixing comments, check if the line is blank for no good | |
3692 | # reason. This includes the first line after a block is opened, and | |
3693 | # blank lines at the end of a function (ie, right before a line like '}' | |
3694 | # | |
3695 | # Skip all the blank line checks if we are immediately inside a | |
3696 | # namespace body. In other words, don't issue blank line warnings | |
3697 | # for this block: | |
3698 | # namespace { | |
3699 | # | |
3700 | # } | |
3701 | # | |
3702 | # A warning about missing end of namespace comments will be issued instead. | |
3703 | # | |
3704 | # Also skip blank line checks for 'extern "C"' blocks, which are formatted | |
3705 | # like namespaces. | |
3706 | if (IsBlankLine(line) and | |
3707 | not nesting_state.InNamespaceBody() and | |
3708 | not nesting_state.InExternC()): | |
3709 | elided = clean_lines.elided | |
3710 | prev_line = elided[linenum - 1] | |
3711 | prevbrace = prev_line.rfind('{') | |
3712 | # TODO(unknown): Don't complain if line before blank line, and line after, | |
3713 | # both start with alnums and are indented the same amount. | |
3714 | # This ignores whitespace at the start of a namespace block | |
3715 | # because those are not usually indented. | |
3716 | if prevbrace != -1 and prev_line[prevbrace:].find('}') == -1: | |
3717 | # OK, we have a blank line at the start of a code block. Before we | |
3718 | # complain, we check if it is an exception to the rule: The previous | |
3719 | # non-empty line has the parameters of a function header that are indented | |
3720 | # 4 spaces (because they did not fit in a 80 column line when placed on | |
3721 | # the same line as the function name). We also check for the case where | |
3722 | # the previous line is indented 6 spaces, which may happen when the | |
3723 | # initializers of a constructor do not fit into a 80 column line. | |
3724 | exception = False | |
3725 | if Match(r' {6}\w', prev_line): # Initializer list? | |
3726 | # We are looking for the opening column of initializer list, which | |
3727 | # should be indented 4 spaces to cause 6 space indentation afterwards. | |
3728 | search_position = linenum-2 | |
3729 | while (search_position >= 0 | |
3730 | and Match(r' {6}\w', elided[search_position])): | |
3731 | search_position -= 1 | |
3732 | exception = (search_position >= 0 | |
3733 | and elided[search_position][:5] == ' :') | |
3734 | else: | |
3735 | # Search for the function arguments or an initializer list. We use a | |
3736 | # simple heuristic here: If the line is indented 4 spaces; and we have a | |
3737 | # closing paren, without the opening paren, followed by an opening brace | |
3738 | # or colon (for initializer lists) we assume that it is the last line of | |
3739 | # a function header. If we have a colon indented 4 spaces, it is an | |
3740 | # initializer list. | |
3741 | exception = (Match(r' {4}\w[^\(]*\)\s*(const\s*)?(\{\s*$|:)', | |
3742 | prev_line) | |
3743 | or Match(r' {4}:', prev_line)) | |
3744 | ||
3745 | if not exception: | |
3746 | error(filename, linenum, 'whitespace/blank_line', 2, | |
3747 | 'Redundant blank line at the start of a code block ' | |
3748 | 'should be deleted.') | |
3749 | # Ignore blank lines at the end of a block in a long if-else | |
3750 | # chain, like this: | |
3751 | # if (condition1) { | |
3752 | # // Something followed by a blank line | |
3753 | # | |
3754 | # } else if (condition2) { | |
3755 | # // Something else | |
3756 | # } | |
3757 | if linenum + 1 < clean_lines.NumLines(): | |
3758 | next_line = raw[linenum + 1] | |
3759 | if (next_line | |
3760 | and Match(r'\s*}', next_line) | |
3761 | and next_line.find('} else ') == -1): | |
3762 | error(filename, linenum, 'whitespace/blank_line', 3, | |
3763 | 'Redundant blank line at the end of a code block ' | |
3764 | 'should be deleted.') | |
3765 | ||
3766 | matched = Match(r'\s*(public|protected|private):', prev_line) | |
3767 | if matched: | |
3768 | error(filename, linenum, 'whitespace/blank_line', 3, | |
3769 | 'Do not leave a blank line after "%s:"' % matched.group(1)) | |
3770 | ||
3771 | # Next, check comments | |
3772 | next_line_start = 0 | |
3773 | if linenum + 1 < clean_lines.NumLines(): | |
3774 | next_line = raw[linenum + 1] | |
3775 | next_line_start = len(next_line) - len(next_line.lstrip()) | |
3776 | CheckComment(line, filename, linenum, next_line_start, error) | |
3777 | ||
3778 | # get rid of comments and strings | |
3779 | line = clean_lines.elided[linenum] | |
3780 | ||
3781 | # You shouldn't have spaces before your brackets, except for C++11 attributes | |
3782 | # or maybe after 'delete []', 'return []() {};', or 'auto [abc, ...] = ...;'. | |
3783 | if (Search(r'\w\s+\[(?!\[)', line) and | |
3784 | not Search(r'(?:auto&?|delete|return)\s+\[', line)): | |
3785 | error(filename, linenum, 'whitespace/braces', 5, | |
3786 | 'Extra space before [') | |
3787 | ||
3788 | # In range-based for, we wanted spaces before and after the colon, but | |
3789 | # not around "::" tokens that might appear. | |
3790 | if (Search(r'for *\(.*[^:]:[^: ]', line) or | |
3791 | Search(r'for *\(.*[^: ]:[^:]', line)): | |
3792 | error(filename, linenum, 'whitespace/forcolon', 2, | |
3793 | 'Missing space around colon in range-based for loop') | |
3794 | ||
3795 | ||
3796 | def CheckOperatorSpacing(filename, clean_lines, linenum, error): | |
3797 | """Checks for horizontal spacing around operators. | |
3798 | ||
3799 | Args: | |
3800 | filename: The name of the current file. | |
3801 | clean_lines: A CleansedLines instance containing the file. | |
3802 | linenum: The number of the line to check. | |
3803 | error: The function to call with any errors found. | |
3804 | """ | |
3805 | line = clean_lines.elided[linenum] | |
3806 | ||
3807 | # Don't try to do spacing checks for operator methods. Do this by | |
3808 | # replacing the troublesome characters with something else, | |
3809 | # preserving column position for all other characters. | |
3810 | # | |
3811 | # The replacement is done repeatedly to avoid false positives from | |
3812 | # operators that call operators. | |
3813 | while True: | |
3814 | match = Match(r'^(.*\boperator\b)(\S+)(\s*\(.*)$', line) | |
3815 | if match: | |
3816 | line = match.group(1) + ('_' * len(match.group(2))) + match.group(3) | |
3817 | else: | |
3818 | break | |
3819 | ||
3820 | # We allow no-spaces around = within an if: "if ( (a=Foo()) == 0 )". | |
3821 | # Otherwise not. Note we only check for non-spaces on *both* sides; | |
3822 | # sometimes people put non-spaces on one side when aligning ='s among | |
3823 | # many lines (not that this is behavior that I approve of...) | |
3824 | if ((Search(r'[\w.]=', line) or | |
3825 | Search(r'=[\w.]', line)) | |
3826 | and not Search(r'\b(if|while|for) ', line) | |
3827 | # Operators taken from [lex.operators] in C++11 standard. | |
3828 | and not Search(r'(>=|<=|==|!=|&=|\^=|\|=|\+=|\*=|\/=|\%=)', line) | |
3829 | and not Search(r'operator=', line)): | |
3830 | error(filename, linenum, 'whitespace/operators', 4, | |
3831 | 'Missing spaces around =') | |
3832 | ||
3833 | # It's ok not to have spaces around binary operators like + - * /, but if | |
3834 | # there's too little whitespace, we get concerned. It's hard to tell, | |
3835 | # though, so we punt on this one for now. TODO. | |
3836 | ||
3837 | # You should always have whitespace around binary operators. | |
3838 | # | |
3839 | # Check <= and >= first to avoid false positives with < and >, then | |
3840 | # check non-include lines for spacing around < and >. | |
3841 | # | |
3842 | # If the operator is followed by a comma, assume it's be used in a | |
3843 | # macro context and don't do any checks. This avoids false | |
3844 | # positives. | |
3845 | # | |
3846 | # Note that && is not included here. This is because there are too | |
3847 | # many false positives due to RValue references. | |
3848 | match = Search(r'[^<>=!\s](==|!=|<=|>=|\|\|)[^<>=!\s,;\)]', line) | |
3849 | if match: | |
3850 | error(filename, linenum, 'whitespace/operators', 3, | |
3851 | 'Missing spaces around %s' % match.group(1)) | |
3852 | elif not Match(r'#.*include', line): | |
3853 | # Look for < that is not surrounded by spaces. This is only | |
3854 | # triggered if both sides are missing spaces, even though | |
3855 | # technically should should flag if at least one side is missing a | |
3856 | # space. This is done to avoid some false positives with shifts. | |
3857 | match = Match(r'^(.*[^\s<])<[^\s=<,]', line) | |
3858 | if match: | |
3859 | (_, _, end_pos) = CloseExpression( | |
3860 | clean_lines, linenum, len(match.group(1))) | |
3861 | if end_pos <= -1: | |
3862 | error(filename, linenum, 'whitespace/operators', 3, | |
3863 | 'Missing spaces around <') | |
3864 | ||
3865 | # Look for > that is not surrounded by spaces. Similar to the | |
3866 | # above, we only trigger if both sides are missing spaces to avoid | |
3867 | # false positives with shifts. | |
3868 | match = Match(r'^(.*[^-\s>])>[^\s=>,]', line) | |
3869 | if match: | |
3870 | (_, _, start_pos) = ReverseCloseExpression( | |
3871 | clean_lines, linenum, len(match.group(1))) | |
3872 | if start_pos <= -1: | |
3873 | error(filename, linenum, 'whitespace/operators', 3, | |
3874 | 'Missing spaces around >') | |
3875 | ||
3876 | # We allow no-spaces around << when used like this: 10<<20, but | |
3877 | # not otherwise (particularly, not when used as streams) | |
3878 | # | |
3879 | # We also allow operators following an opening parenthesis, since | |
3880 | # those tend to be macros that deal with operators. | |
3881 | match = Search(r'(operator|[^\s(<])(?:L|UL|LL|ULL|l|ul|ll|ull)?<<([^\s,=<])', line) | |
3882 | if (match and not (match.group(1).isdigit() and match.group(2).isdigit()) and | |
3883 | not (match.group(1) == 'operator' and match.group(2) == ';')): | |
3884 | error(filename, linenum, 'whitespace/operators', 3, | |
3885 | 'Missing spaces around <<') | |
3886 | ||
3887 | # We allow no-spaces around >> for almost anything. This is because | |
3888 | # C++11 allows ">>" to close nested templates, which accounts for | |
3889 | # most cases when ">>" is not followed by a space. | |
3890 | # | |
3891 | # We still warn on ">>" followed by alpha character, because that is | |
3892 | # likely due to ">>" being used for right shifts, e.g.: | |
3893 | # value >> alpha | |
3894 | # | |
3895 | # When ">>" is used to close templates, the alphanumeric letter that | |
3896 | # follows would be part of an identifier, and there should still be | |
3897 | # a space separating the template type and the identifier. | |
3898 | # type<type<type>> alpha | |
3899 | match = Search(r'>>[a-zA-Z_]', line) | |
3900 | if match: | |
3901 | error(filename, linenum, 'whitespace/operators', 3, | |
3902 | 'Missing spaces around >>') | |
3903 | ||
3904 | # There shouldn't be space around unary operators | |
3905 | match = Search(r'(!\s|~\s|[\s]--[\s;]|[\s]\+\+[\s;])', line) | |
3906 | if match: | |
3907 | error(filename, linenum, 'whitespace/operators', 4, | |
3908 | 'Extra space for operator %s' % match.group(1)) | |
3909 | ||
3910 | ||
3911 | def CheckParenthesisSpacing(filename, clean_lines, linenum, error): | |
3912 | """Checks for horizontal spacing around parentheses. | |
3913 | ||
3914 | Args: | |
3915 | filename: The name of the current file. | |
3916 | clean_lines: A CleansedLines instance containing the file. | |
3917 | linenum: The number of the line to check. | |
3918 | error: The function to call with any errors found. | |
3919 | """ | |
3920 | line = clean_lines.elided[linenum] | |
3921 | ||
3922 | # No spaces after an if, while, switch, or for | |
3923 | match = Search(r' (if\(|for\(|while\(|switch\()', line) | |
3924 | if match: | |
3925 | error(filename, linenum, 'whitespace/parens', 5, | |
3926 | 'Missing space before ( in %s' % match.group(1)) | |
3927 | ||
3928 | # For if/for/while/switch, the left and right parens should be | |
3929 | # consistent about how many spaces are inside the parens, and | |
3930 | # there should either be zero or one spaces inside the parens. | |
3931 | # We don't want: "if ( foo)" or "if ( foo )". | |
3932 | # Exception: "for ( ; foo; bar)" and "for (foo; bar; )" are allowed. | |
3933 | match = Search(r'\b(if|for|while|switch)\s*' | |
3934 | r'\(([ ]*)(.).*[^ ]+([ ]*)\)\s*{\s*$', | |
3935 | line) | |
3936 | if match: | |
3937 | if len(match.group(2)) != len(match.group(4)): | |
3938 | if not (match.group(3) == ';' and | |
3939 | len(match.group(2)) == 1 + len(match.group(4)) or | |
3940 | not match.group(2) and Search(r'\bfor\s*\(.*; \)', line)): | |
3941 | error(filename, linenum, 'whitespace/parens', 5, | |
3942 | 'Mismatching spaces inside () in %s' % match.group(1)) | |
3943 | if len(match.group(2)) not in [0, 1]: | |
3944 | error(filename, linenum, 'whitespace/parens', 5, | |
3945 | 'Should have zero or one spaces inside ( and ) in %s' % | |
3946 | match.group(1)) | |
3947 | ||
3948 | ||
3949 | def CheckCommaSpacing(filename, clean_lines, linenum, error): | |
3950 | """Checks for horizontal spacing near commas and semicolons. | |
3951 | ||
3952 | Args: | |
3953 | filename: The name of the current file. | |
3954 | clean_lines: A CleansedLines instance containing the file. | |
3955 | linenum: The number of the line to check. | |
3956 | error: The function to call with any errors found. | |
3957 | """ | |
3958 | raw = clean_lines.lines_without_raw_strings | |
3959 | line = clean_lines.elided[linenum] | |
3960 | ||
3961 | # You should always have a space after a comma (either as fn arg or operator) | |
3962 | # | |
3963 | # This does not apply when the non-space character following the | |
3964 | # comma is another comma, since the only time when that happens is | |
3965 | # for empty macro arguments. | |
3966 | # | |
3967 | # We run this check in two passes: first pass on elided lines to | |
3968 | # verify that lines contain missing whitespaces, second pass on raw | |
3969 | # lines to confirm that those missing whitespaces are not due to | |
3970 | # elided comments. | |
3971 | if (Search(r',[^,\s]', ReplaceAll(r'\boperator\s*,\s*\(', 'F(', line)) and | |
3972 | Search(r',[^,\s]', raw[linenum])): | |
3973 | error(filename, linenum, 'whitespace/comma', 3, | |
3974 | 'Missing space after ,') | |
3975 | ||
3976 | # You should always have a space after a semicolon | |
3977 | # except for few corner cases | |
3978 | # TODO(unknown): clarify if 'if (1) { return 1;}' is requires one more | |
3979 | # space after ; | |
3980 | if Search(r';[^\s};\\)/]', line): | |
3981 | error(filename, linenum, 'whitespace/semicolon', 3, | |
3982 | 'Missing space after ;') | |
3983 | ||
3984 | ||
3985 | def _IsType(clean_lines, nesting_state, expr): | |
3986 | """Check if expression looks like a type name, returns true if so. | |
3987 | ||
3988 | Args: | |
3989 | clean_lines: A CleansedLines instance containing the file. | |
3990 | nesting_state: A NestingState instance which maintains information about | |
3991 | the current stack of nested blocks being parsed. | |
3992 | expr: The expression to check. | |
3993 | Returns: | |
3994 | True, if token looks like a type. | |
3995 | """ | |
3996 | # Keep only the last token in the expression | |
3997 | last_word = Match(r'^.*(\b\S+)$', expr) | |
3998 | if last_word: | |
3999 | token = last_word.group(1) | |
4000 | else: | |
4001 | token = expr | |
4002 | ||
4003 | # Match native types and stdint types | |
4004 | if _TYPES.match(token): | |
4005 | return True | |
4006 | ||
4007 | # Try a bit harder to match templated types. Walk up the nesting | |
4008 | # stack until we find something that resembles a typename | |
4009 | # declaration for what we are looking for. | |
4010 | typename_pattern = (r'\b(?:typename|class|struct)\s+' + re.escape(token) + | |
4011 | r'\b') | |
4012 | block_index = len(nesting_state.stack) - 1 | |
4013 | while block_index >= 0: | |
4014 | if isinstance(nesting_state.stack[block_index], _NamespaceInfo): | |
4015 | return False | |
4016 | ||
4017 | # Found where the opening brace is. We want to scan from this | |
4018 | # line up to the beginning of the function, minus a few lines. | |
4019 | # template <typename Type1, // stop scanning here | |
4020 | # ...> | |
4021 | # class C | |
4022 | # : public ... { // start scanning here | |
4023 | last_line = nesting_state.stack[block_index].starting_linenum | |
4024 | ||
4025 | next_block_start = 0 | |
4026 | if block_index > 0: | |
4027 | next_block_start = nesting_state.stack[block_index - 1].starting_linenum | |
4028 | first_line = last_line | |
4029 | while first_line >= next_block_start: | |
4030 | if clean_lines.elided[first_line].find('template') >= 0: | |
4031 | break | |
4032 | first_line -= 1 | |
4033 | if first_line < next_block_start: | |
4034 | # Didn't find any "template" keyword before reaching the next block, | |
4035 | # there are probably no template things to check for this block | |
4036 | block_index -= 1 | |
4037 | continue | |
4038 | ||
4039 | # Look for typename in the specified range | |
4040 | for i in xrange(first_line, last_line + 1, 1): | |
4041 | if Search(typename_pattern, clean_lines.elided[i]): | |
4042 | return True | |
4043 | block_index -= 1 | |
4044 | ||
4045 | return False | |
4046 | ||
4047 | ||
4048 | def CheckBracesSpacing(filename, clean_lines, linenum, nesting_state, error): | |
4049 | """Checks for horizontal spacing near commas. | |
4050 | ||
4051 | Args: | |
4052 | filename: The name of the current file. | |
4053 | clean_lines: A CleansedLines instance containing the file. | |
4054 | linenum: The number of the line to check. | |
4055 | nesting_state: A NestingState instance which maintains information about | |
4056 | the current stack of nested blocks being parsed. | |
4057 | error: The function to call with any errors found. | |
4058 | """ | |
4059 | line = clean_lines.elided[linenum] | |
4060 | ||
4061 | # Except after an opening paren, or after another opening brace (in case of | |
4062 | # an initializer list, for instance), you should have spaces before your | |
4063 | # braces when they are delimiting blocks, classes, namespaces etc. | |
4064 | # And since you should never have braces at the beginning of a line, | |
4065 | # this is an easy test. Except that braces used for initialization don't | |
4066 | # follow the same rule; we often don't want spaces before those. | |
4067 | match = Match(r'^(.*[^ ({>]){', line) | |
4068 | ||
4069 | if match: | |
4070 | # Try a bit harder to check for brace initialization. This | |
4071 | # happens in one of the following forms: | |
4072 | # Constructor() : initializer_list_{} { ... } | |
4073 | # Constructor{}.MemberFunction() | |
4074 | # Type variable{}; | |
4075 | # FunctionCall(type{}, ...); | |
4076 | # LastArgument(..., type{}); | |
4077 | # LOG(INFO) << type{} << " ..."; | |
4078 | # map_of_type[{...}] = ...; | |
4079 | # ternary = expr ? new type{} : nullptr; | |
4080 | # OuterTemplate<InnerTemplateConstructor<Type>{}> | |
4081 | # | |
4082 | # We check for the character following the closing brace, and | |
4083 | # silence the warning if it's one of those listed above, i.e. | |
4084 | # "{.;,)<>]:". | |
4085 | # | |
4086 | # To account for nested initializer list, we allow any number of | |
4087 | # closing braces up to "{;,)<". We can't simply silence the | |
4088 | # warning on first sight of closing brace, because that would | |
4089 | # cause false negatives for things that are not initializer lists. | |
4090 | # Silence this: But not this: | |
4091 | # Outer{ if (...) { | |
4092 | # Inner{...} if (...){ // Missing space before { | |
4093 | # }; } | |
4094 | # | |
4095 | # There is a false negative with this approach if people inserted | |
4096 | # spurious semicolons, e.g. "if (cond){};", but we will catch the | |
4097 | # spurious semicolon with a separate check. | |
4098 | leading_text = match.group(1) | |
4099 | (endline, endlinenum, endpos) = CloseExpression( | |
4100 | clean_lines, linenum, len(match.group(1))) | |
4101 | trailing_text = '' | |
4102 | if endpos > -1: | |
4103 | trailing_text = endline[endpos:] | |
4104 | for offset in xrange(endlinenum + 1, | |
4105 | min(endlinenum + 3, clean_lines.NumLines() - 1)): | |
4106 | trailing_text += clean_lines.elided[offset] | |
4107 | # We also suppress warnings for `uint64_t{expression}` etc., as the style | |
4108 | # guide recommends brace initialization for integral types to avoid | |
4109 | # overflow/truncation. | |
4110 | if (not Match(r'^[\s}]*[{.;,)<>\]:]', trailing_text) | |
4111 | and not _IsType(clean_lines, nesting_state, leading_text)): | |
4112 | error(filename, linenum, 'whitespace/braces', 5, | |
4113 | 'Missing space before {') | |
4114 | ||
4115 | # Make sure '} else {' has spaces. | |
4116 | if Search(r'}else', line): | |
4117 | error(filename, linenum, 'whitespace/braces', 5, | |
4118 | 'Missing space before else') | |
4119 | ||
4120 | # You shouldn't have a space before a semicolon at the end of the line. | |
4121 | # There's a special case for "for" since the style guide allows space before | |
4122 | # the semicolon there. | |
4123 | if Search(r':\s*;\s*$', line): | |
4124 | error(filename, linenum, 'whitespace/semicolon', 5, | |
4125 | 'Semicolon defining empty statement. Use {} instead.') | |
4126 | elif Search(r'^\s*;\s*$', line): | |
4127 | error(filename, linenum, 'whitespace/semicolon', 5, | |
4128 | 'Line contains only semicolon. If this should be an empty statement, ' | |
4129 | 'use {} instead.') | |
4130 | elif (Search(r'\s+;\s*$', line) and | |
4131 | not Search(r'\bfor\b', line)): | |
4132 | error(filename, linenum, 'whitespace/semicolon', 5, | |
4133 | 'Extra space before last semicolon. If this should be an empty ' | |
4134 | 'statement, use {} instead.') | |
4135 | ||
4136 | ||
4137 | def IsDecltype(clean_lines, linenum, column): | |
4138 | """Check if the token ending on (linenum, column) is decltype(). | |
4139 | ||
4140 | Args: | |
4141 | clean_lines: A CleansedLines instance containing the file. | |
4142 | linenum: the number of the line to check. | |
4143 | column: end column of the token to check. | |
4144 | Returns: | |
4145 | True if this token is decltype() expression, False otherwise. | |
4146 | """ | |
4147 | (text, _, start_col) = ReverseCloseExpression(clean_lines, linenum, column) | |
4148 | if start_col < 0: | |
4149 | return False | |
4150 | if Search(r'\bdecltype\s*$', text[0:start_col]): | |
4151 | return True | |
4152 | return False | |
4153 | ||
4154 | def CheckSectionSpacing(filename, clean_lines, class_info, linenum, error): | |
4155 | """Checks for additional blank line issues related to sections. | |
4156 | ||
4157 | Currently the only thing checked here is blank line before protected/private. | |
4158 | ||
4159 | Args: | |
4160 | filename: The name of the current file. | |
4161 | clean_lines: A CleansedLines instance containing the file. | |
4162 | class_info: A _ClassInfo objects. | |
4163 | linenum: The number of the line to check. | |
4164 | error: The function to call with any errors found. | |
4165 | """ | |
4166 | # Skip checks if the class is small, where small means 25 lines or less. | |
4167 | # 25 lines seems like a good cutoff since that's the usual height of | |
4168 | # terminals, and any class that can't fit in one screen can't really | |
4169 | # be considered "small". | |
4170 | # | |
4171 | # Also skip checks if we are on the first line. This accounts for | |
4172 | # classes that look like | |
4173 | # class Foo { public: ... }; | |
4174 | # | |
4175 | # If we didn't find the end of the class, last_line would be zero, | |
4176 | # and the check will be skipped by the first condition. | |
4177 | if (class_info.last_line - class_info.starting_linenum <= 24 or | |
4178 | linenum <= class_info.starting_linenum): | |
4179 | return | |
4180 | ||
4181 | matched = Match(r'\s*(public|protected|private):', clean_lines.lines[linenum]) | |
4182 | if matched: | |
4183 | # Issue warning if the line before public/protected/private was | |
4184 | # not a blank line, but don't do this if the previous line contains | |
4185 | # "class" or "struct". This can happen two ways: | |
4186 | # - We are at the beginning of the class. | |
4187 | # - We are forward-declaring an inner class that is semantically | |
4188 | # private, but needed to be public for implementation reasons. | |
4189 | # Also ignores cases where the previous line ends with a backslash as can be | |
4190 | # common when defining classes in C macros. | |
4191 | prev_line = clean_lines.lines[linenum - 1] | |
4192 | if (not IsBlankLine(prev_line) and | |
4193 | not Search(r'\b(class|struct)\b', prev_line) and | |
4194 | not Search(r'\\$', prev_line)): | |
4195 | # Try a bit harder to find the beginning of the class. This is to | |
4196 | # account for multi-line base-specifier lists, e.g.: | |
4197 | # class Derived | |
4198 | # : public Base { | |
4199 | end_class_head = class_info.starting_linenum | |
4200 | for i in range(class_info.starting_linenum, linenum): | |
4201 | if Search(r'\{\s*$', clean_lines.lines[i]): | |
4202 | end_class_head = i | |
4203 | break | |
4204 | if end_class_head < linenum - 1: | |
4205 | error(filename, linenum, 'whitespace/blank_line', 3, | |
4206 | '"%s:" should be preceded by a blank line' % matched.group(1)) | |
4207 | ||
4208 | ||
4209 | def GetPreviousNonBlankLine(clean_lines, linenum): | |
4210 | """Return the most recent non-blank line and its line number. | |
4211 | ||
4212 | Args: | |
4213 | clean_lines: A CleansedLines instance containing the file contents. | |
4214 | linenum: The number of the line to check. | |
4215 | ||
4216 | Returns: | |
4217 | A tuple with two elements. The first element is the contents of the last | |
4218 | non-blank line before the current line, or the empty string if this is the | |
4219 | first non-blank line. The second is the line number of that line, or -1 | |
4220 | if this is the first non-blank line. | |
4221 | """ | |
4222 | ||
4223 | prevlinenum = linenum - 1 | |
4224 | while prevlinenum >= 0: | |
4225 | prevline = clean_lines.elided[prevlinenum] | |
4226 | if not IsBlankLine(prevline): # if not a blank line... | |
4227 | return (prevline, prevlinenum) | |
4228 | prevlinenum -= 1 | |
4229 | return ('', -1) | |
4230 | ||
4231 | ||
4232 | def CheckBraces(filename, clean_lines, linenum, error): | |
4233 | """Looks for misplaced braces (e.g. at the end of line). | |
4234 | ||
4235 | Args: | |
4236 | filename: The name of the current file. | |
4237 | clean_lines: A CleansedLines instance containing the file. | |
4238 | linenum: The number of the line to check. | |
4239 | error: The function to call with any errors found. | |
4240 | """ | |
4241 | ||
4242 | line = clean_lines.elided[linenum] # get rid of comments and strings | |
4243 | ||
4244 | if Match(r'\s*{\s*$', line): | |
4245 | # We allow an open brace to start a line in the case where someone is using | |
4246 | # braces in a block to explicitly create a new scope, which is commonly used | |
4247 | # to control the lifetime of stack-allocated variables. Braces are also | |
4248 | # used for brace initializers inside function calls. We don't detect this | |
4249 | # perfectly: we just don't complain if the last non-whitespace character on | |
4250 | # the previous non-blank line is ',', ';', ':', '(', '{', or '}', or if the | |
4251 | # previous line starts a preprocessor block. We also allow a brace on the | |
4252 | # following line if it is part of an array initialization and would not fit | |
4253 | # within the 80 character limit of the preceding line. | |
4254 | prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] | |
4255 | if (not Search(r'[,;:}{(]\s*$', prevline) and | |
4256 | not Match(r'\s*#', prevline) and | |
4257 | not (GetLineWidth(prevline) > _line_length - 2 and '[]' in prevline)): | |
4258 | error(filename, linenum, 'whitespace/braces', 4, | |
4259 | '{ should almost always be at the end of the previous line') | |
4260 | ||
4261 | # An else clause should be on the same line as the preceding closing brace. | |
4262 | if Match(r'\s*else\b\s*(?:if\b|\{|$)', line): | |
4263 | prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] | |
4264 | if Match(r'\s*}\s*$', prevline): | |
4265 | error(filename, linenum, 'whitespace/newline', 4, | |
4266 | 'An else should appear on the same line as the preceding }') | |
4267 | ||
4268 | # If braces come on one side of an else, they should be on both. | |
4269 | # However, we have to worry about "else if" that spans multiple lines! | |
4270 | if Search(r'else if\s*\(', line): # could be multi-line if | |
4271 | brace_on_left = bool(Search(r'}\s*else if\s*\(', line)) | |
4272 | # find the ( after the if | |
4273 | pos = line.find('else if') | |
4274 | pos = line.find('(', pos) | |
4275 | if pos > 0: | |
4276 | (endline, _, endpos) = CloseExpression(clean_lines, linenum, pos) | |
4277 | brace_on_right = endline[endpos:].find('{') != -1 | |
4278 | if brace_on_left != brace_on_right: # must be brace after if | |
4279 | error(filename, linenum, 'readability/braces', 5, | |
4280 | 'If an else has a brace on one side, it should have it on both') | |
4281 | elif Search(r'}\s*else[^{]*$', line) or Match(r'[^}]*else\s*{', line): | |
4282 | error(filename, linenum, 'readability/braces', 5, | |
4283 | 'If an else has a brace on one side, it should have it on both') | |
4284 | ||
4285 | # Likewise, an else should never have the else clause on the same line | |
4286 | if Search(r'\belse [^\s{]', line) and not Search(r'\belse if\b', line): | |
4287 | error(filename, linenum, 'whitespace/newline', 4, | |
4288 | 'Else clause should never be on same line as else (use 2 lines)') | |
4289 | ||
4290 | # In the same way, a do/while should never be on one line | |
4291 | if Match(r'\s*do [^\s{]', line): | |
4292 | error(filename, linenum, 'whitespace/newline', 4, | |
4293 | 'do/while clauses should not be on a single line') | |
4294 | ||
4295 | # Check single-line if/else bodies. The style guide says 'curly braces are not | |
4296 | # required for single-line statements'. We additionally allow multi-line, | |
4297 | # single statements, but we reject anything with more than one semicolon in | |
4298 | # it. This means that the first semicolon after the if should be at the end of | |
4299 | # its line, and the line after that should have an indent level equal to or | |
4300 | # lower than the if. We also check for ambiguous if/else nesting without | |
4301 | # braces. | |
4302 | if_else_match = Search(r'\b(if\s*(|constexpr)\s*\(|else\b)', line) | |
4303 | if if_else_match and not Match(r'\s*#', line): | |
4304 | if_indent = GetIndentLevel(line) | |
4305 | endline, endlinenum, endpos = line, linenum, if_else_match.end() | |
4306 | if_match = Search(r'\bif\s*(|constexpr)\s*\(', line) | |
4307 | if if_match: | |
4308 | # This could be a multiline if condition, so find the end first. | |
4309 | pos = if_match.end() - 1 | |
4310 | (endline, endlinenum, endpos) = CloseExpression(clean_lines, linenum, pos) | |
4311 | # Check for an opening brace, either directly after the if or on the next | |
4312 | # line. If found, this isn't a single-statement conditional. | |
4313 | if (not Match(r'\s*{', endline[endpos:]) | |
4314 | and not (Match(r'\s*$', endline[endpos:]) | |
4315 | and endlinenum < (len(clean_lines.elided) - 1) | |
4316 | and Match(r'\s*{', clean_lines.elided[endlinenum + 1]))): | |
4317 | while (endlinenum < len(clean_lines.elided) | |
4318 | and ';' not in clean_lines.elided[endlinenum][endpos:]): | |
4319 | endlinenum += 1 | |
4320 | endpos = 0 | |
4321 | if endlinenum < len(clean_lines.elided): | |
4322 | endline = clean_lines.elided[endlinenum] | |
4323 | # We allow a mix of whitespace and closing braces (e.g. for one-liner | |
4324 | # methods) and a single \ after the semicolon (for macros) | |
4325 | endpos = endline.find(';') | |
4326 | if not Match(r';[\s}]*(\\?)$', endline[endpos:]): | |
4327 | # Semicolon isn't the last character, there's something trailing. | |
4328 | # Output a warning if the semicolon is not contained inside | |
4329 | # a lambda expression. | |
4330 | if not Match(r'^[^{};]*\[[^\[\]]*\][^{}]*\{[^{}]*\}\s*\)*[;,]\s*$', | |
4331 | endline): | |
4332 | error(filename, linenum, 'readability/braces', 4, | |
4333 | 'If/else bodies with multiple statements require braces') | |
4334 | elif endlinenum < len(clean_lines.elided) - 1: | |
4335 | # Make sure the next line is dedented | |
4336 | next_line = clean_lines.elided[endlinenum + 1] | |
4337 | next_indent = GetIndentLevel(next_line) | |
4338 | # With ambiguous nested if statements, this will error out on the | |
4339 | # if that *doesn't* match the else, regardless of whether it's the | |
4340 | # inner one or outer one. | |
4341 | if (if_match and Match(r'\s*else\b', next_line) | |
4342 | and next_indent != if_indent): | |
4343 | error(filename, linenum, 'readability/braces', 4, | |
4344 | 'Else clause should be indented at the same level as if. ' | |
4345 | 'Ambiguous nested if/else chains require braces.') | |
4346 | elif next_indent > if_indent: | |
4347 | error(filename, linenum, 'readability/braces', 4, | |
4348 | 'If/else bodies with multiple statements require braces') | |
4349 | ||
4350 | ||
4351 | def CheckTrailingSemicolon(filename, clean_lines, linenum, error): | |
4352 | """Looks for redundant trailing semicolon. | |
4353 | ||
4354 | Args: | |
4355 | filename: The name of the current file. | |
4356 | clean_lines: A CleansedLines instance containing the file. | |
4357 | linenum: The number of the line to check. | |
4358 | error: The function to call with any errors found. | |
4359 | """ | |
4360 | ||
4361 | line = clean_lines.elided[linenum] | |
4362 | ||
4363 | # Block bodies should not be followed by a semicolon. Due to C++11 | |
4364 | # brace initialization, there are more places where semicolons are | |
4365 | # required than not, so we explicitly list the allowed rules rather | |
4366 | # than listing the disallowed ones. These are the places where "};" | |
4367 | # should be replaced by just "}": | |
4368 | # 1. Some flavor of block following closing parenthesis: | |
4369 | # for (;;) {}; | |
4370 | # while (...) {}; | |
4371 | # switch (...) {}; | |
4372 | # Function(...) {}; | |
4373 | # if (...) {}; | |
4374 | # if (...) else if (...) {}; | |
4375 | # | |
4376 | # 2. else block: | |
4377 | # if (...) else {}; | |
4378 | # | |
4379 | # 3. const member function: | |
4380 | # Function(...) const {}; | |
4381 | # | |
4382 | # 4. Block following some statement: | |
4383 | # x = 42; | |
4384 | # {}; | |
4385 | # | |
4386 | # 5. Block at the beginning of a function: | |
4387 | # Function(...) { | |
4388 | # {}; | |
4389 | # } | |
4390 | # | |
4391 | # Note that naively checking for the preceding "{" will also match | |
4392 | # braces inside multi-dimensional arrays, but this is fine since | |
4393 | # that expression will not contain semicolons. | |
4394 | # | |
4395 | # 6. Block following another block: | |
4396 | # while (true) {} | |
4397 | # {}; | |
4398 | # | |
4399 | # 7. End of namespaces: | |
4400 | # namespace {}; | |
4401 | # | |
4402 | # These semicolons seems far more common than other kinds of | |
4403 | # redundant semicolons, possibly due to people converting classes | |
4404 | # to namespaces. For now we do not warn for this case. | |
4405 | # | |
4406 | # Try matching case 1 first. | |
4407 | match = Match(r'^(.*\)\s*)\{', line) | |
4408 | if match: | |
4409 | # Matched closing parenthesis (case 1). Check the token before the | |
4410 | # matching opening parenthesis, and don't warn if it looks like a | |
4411 | # macro. This avoids these false positives: | |
4412 | # - macro that defines a base class | |
4413 | # - multi-line macro that defines a base class | |
4414 | # - macro that defines the whole class-head | |
4415 | # | |
4416 | # But we still issue warnings for macros that we know are safe to | |
4417 | # warn, specifically: | |
4418 | # - TEST, TEST_F, TEST_P, MATCHER, MATCHER_P | |
4419 | # - TYPED_TEST | |
4420 | # - INTERFACE_DEF | |
4421 | # - EXCLUSIVE_LOCKS_REQUIRED, SHARED_LOCKS_REQUIRED, LOCKS_EXCLUDED: | |
4422 | # | |
4423 | # We implement a list of safe macros instead of a list of | |
4424 | # unsafe macros, even though the latter appears less frequently in | |
4425 | # google code and would have been easier to implement. This is because | |
4426 | # the downside for getting the allowed checks wrong means some extra | |
4427 | # semicolons, while the downside for getting disallowed checks wrong | |
4428 | # would result in compile errors. | |
4429 | # | |
4430 | # In addition to macros, we also don't want to warn on | |
4431 | # - Compound literals | |
4432 | # - Lambdas | |
4433 | # - alignas specifier with anonymous structs | |
4434 | # - decltype | |
4435 | closing_brace_pos = match.group(1).rfind(')') | |
4436 | opening_parenthesis = ReverseCloseExpression( | |
4437 | clean_lines, linenum, closing_brace_pos) | |
4438 | if opening_parenthesis[2] > -1: | |
4439 | line_prefix = opening_parenthesis[0][0:opening_parenthesis[2]] | |
4440 | macro = Search(r'\b([A-Z_][A-Z0-9_]*)\s*$', line_prefix) | |
4441 | func = Match(r'^(.*\])\s*$', line_prefix) | |
4442 | if ((macro and | |
4443 | macro.group(1) not in ( | |
4444 | 'TEST', 'TEST_F', 'MATCHER', 'MATCHER_P', 'TYPED_TEST', | |
4445 | 'EXCLUSIVE_LOCKS_REQUIRED', 'SHARED_LOCKS_REQUIRED', | |
4446 | 'LOCKS_EXCLUDED', 'INTERFACE_DEF')) or | |
4447 | (func and not Search(r'\boperator\s*\[\s*\]', func.group(1))) or | |
4448 | Search(r'\b(?:struct|union)\s+alignas\s*$', line_prefix) or | |
4449 | Search(r'\bdecltype$', line_prefix) or | |
4450 | Search(r'\s+=\s*$', line_prefix)): | |
4451 | match = None | |
4452 | if (match and | |
4453 | opening_parenthesis[1] > 1 and | |
4454 | Search(r'\]\s*$', clean_lines.elided[opening_parenthesis[1] - 1])): | |
4455 | # Multi-line lambda-expression | |
4456 | match = None | |
4457 | ||
4458 | else: | |
4459 | # Try matching cases 2-3. | |
4460 | match = Match(r'^(.*(?:else|\)\s*const)\s*)\{', line) | |
4461 | if not match: | |
4462 | # Try matching cases 4-6. These are always matched on separate lines. | |
4463 | # | |
4464 | # Note that we can't simply concatenate the previous line to the | |
4465 | # current line and do a single match, otherwise we may output | |
4466 | # duplicate warnings for the blank line case: | |
4467 | # if (cond) { | |
4468 | # // blank line | |
4469 | # } | |
4470 | prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] | |
4471 | if prevline and Search(r'[;{}]\s*$', prevline): | |
4472 | match = Match(r'^(\s*)\{', line) | |
4473 | ||
4474 | # Check matching closing brace | |
4475 | if match: | |
4476 | (endline, endlinenum, endpos) = CloseExpression( | |
4477 | clean_lines, linenum, len(match.group(1))) | |
4478 | if endpos > -1 and Match(r'^\s*;', endline[endpos:]): | |
4479 | # Current {} pair is eligible for semicolon check, and we have found | |
4480 | # the redundant semicolon, output warning here. | |
4481 | # | |
4482 | # Note: because we are scanning forward for opening braces, and | |
4483 | # outputting warnings for the matching closing brace, if there are | |
4484 | # nested blocks with trailing semicolons, we will get the error | |
4485 | # messages in reversed order. | |
4486 | ||
4487 | # We need to check the line forward for NOLINT | |
4488 | raw_lines = clean_lines.raw_lines | |
4489 | ParseNolintSuppressions(filename, raw_lines[endlinenum-1], endlinenum-1, | |
4490 | error) | |
4491 | ParseNolintSuppressions(filename, raw_lines[endlinenum], endlinenum, | |
4492 | error) | |
4493 | ||
4494 | error(filename, endlinenum, 'readability/braces', 4, | |
4495 | "You don't need a ; after a }") | |
4496 | ||
4497 | ||
4498 | def CheckEmptyBlockBody(filename, clean_lines, linenum, error): | |
4499 | """Look for empty loop/conditional body with only a single semicolon. | |
4500 | ||
4501 | Args: | |
4502 | filename: The name of the current file. | |
4503 | clean_lines: A CleansedLines instance containing the file. | |
4504 | linenum: The number of the line to check. | |
4505 | error: The function to call with any errors found. | |
4506 | """ | |
4507 | ||
4508 | # Search for loop keywords at the beginning of the line. Because only | |
4509 | # whitespaces are allowed before the keywords, this will also ignore most | |
4510 | # do-while-loops, since those lines should start with closing brace. | |
4511 | # | |
4512 | # We also check "if" blocks here, since an empty conditional block | |
4513 | # is likely an error. | |
4514 | line = clean_lines.elided[linenum] | |
4515 | matched = Match(r'\s*(for|while|if)\s*\(', line) | |
4516 | if matched: | |
4517 | # Find the end of the conditional expression. | |
4518 | (end_line, end_linenum, end_pos) = CloseExpression( | |
4519 | clean_lines, linenum, line.find('(')) | |
4520 | ||
4521 | # Output warning if what follows the condition expression is a semicolon. | |
4522 | # No warning for all other cases, including whitespace or newline, since we | |
4523 | # have a separate check for semicolons preceded by whitespace. | |
4524 | if end_pos >= 0 and Match(r';', end_line[end_pos:]): | |
4525 | if matched.group(1) == 'if': | |
4526 | error(filename, end_linenum, 'whitespace/empty_conditional_body', 5, | |
4527 | 'Empty conditional bodies should use {}') | |
4528 | else: | |
4529 | error(filename, end_linenum, 'whitespace/empty_loop_body', 5, | |
4530 | 'Empty loop bodies should use {} or continue') | |
4531 | ||
4532 | # Check for if statements that have completely empty bodies (no comments) | |
4533 | # and no else clauses. | |
4534 | if end_pos >= 0 and matched.group(1) == 'if': | |
4535 | # Find the position of the opening { for the if statement. | |
4536 | # Return without logging an error if it has no brackets. | |
4537 | opening_linenum = end_linenum | |
4538 | opening_line_fragment = end_line[end_pos:] | |
4539 | # Loop until EOF or find anything that's not whitespace or opening {. | |
4540 | while not Search(r'^\s*\{', opening_line_fragment): | |
4541 | if Search(r'^(?!\s*$)', opening_line_fragment): | |
4542 | # Conditional has no brackets. | |
4543 | return | |
4544 | opening_linenum += 1 | |
4545 | if opening_linenum == len(clean_lines.elided): | |
4546 | # Couldn't find conditional's opening { or any code before EOF. | |
4547 | return | |
4548 | opening_line_fragment = clean_lines.elided[opening_linenum] | |
4549 | # Set opening_line (opening_line_fragment may not be entire opening line). | |
4550 | opening_line = clean_lines.elided[opening_linenum] | |
4551 | ||
4552 | # Find the position of the closing }. | |
4553 | opening_pos = opening_line_fragment.find('{') | |
4554 | if opening_linenum == end_linenum: | |
4555 | # We need to make opening_pos relative to the start of the entire line. | |
4556 | opening_pos += end_pos | |
4557 | (closing_line, closing_linenum, closing_pos) = CloseExpression( | |
4558 | clean_lines, opening_linenum, opening_pos) | |
4559 | if closing_pos < 0: | |
4560 | return | |
4561 | ||
4562 | # Now construct the body of the conditional. This consists of the portion | |
4563 | # of the opening line after the {, all lines until the closing line, | |
4564 | # and the portion of the closing line before the }. | |
4565 | if (clean_lines.raw_lines[opening_linenum] != | |
4566 | CleanseComments(clean_lines.raw_lines[opening_linenum])): | |
4567 | # Opening line ends with a comment, so conditional isn't empty. | |
4568 | return | |
4569 | if closing_linenum > opening_linenum: | |
4570 | # Opening line after the {. Ignore comments here since we checked above. | |
4571 | bodylist = list(opening_line[opening_pos+1:]) | |
4572 | # All lines until closing line, excluding closing line, with comments. | |
4573 | bodylist.extend(clean_lines.raw_lines[opening_linenum+1:closing_linenum]) | |
4574 | # Closing line before the }. Won't (and can't) have comments. | |
4575 | bodylist.append(clean_lines.elided[closing_linenum][:closing_pos-1]) | |
4576 | body = '\n'.join(bodylist) | |
4577 | else: | |
4578 | # If statement has brackets and fits on a single line. | |
4579 | body = opening_line[opening_pos+1:closing_pos-1] | |
4580 | ||
4581 | # Check if the body is empty | |
4582 | if not _EMPTY_CONDITIONAL_BODY_PATTERN.search(body): | |
4583 | return | |
4584 | # The body is empty. Now make sure there's not an else clause. | |
4585 | current_linenum = closing_linenum | |
4586 | current_line_fragment = closing_line[closing_pos:] | |
4587 | # Loop until EOF or find anything that's not whitespace or else clause. | |
4588 | while Search(r'^\s*$|^(?=\s*else)', current_line_fragment): | |
4589 | if Search(r'^(?=\s*else)', current_line_fragment): | |
4590 | # Found an else clause, so don't log an error. | |
4591 | return | |
4592 | current_linenum += 1 | |
4593 | if current_linenum == len(clean_lines.elided): | |
4594 | break | |
4595 | current_line_fragment = clean_lines.elided[current_linenum] | |
4596 | ||
4597 | # The body is empty and there's no else clause until EOF or other code. | |
4598 | error(filename, end_linenum, 'whitespace/empty_if_body', 4, | |
4599 | ('If statement had no body and no else clause')) | |
4600 | ||
4601 | ||
4602 | def FindCheckMacro(line): | |
4603 | """Find a replaceable CHECK-like macro. | |
4604 | ||
4605 | Args: | |
4606 | line: line to search on. | |
4607 | Returns: | |
4608 | (macro name, start position), or (None, -1) if no replaceable | |
4609 | macro is found. | |
4610 | """ | |
4611 | for macro in _CHECK_MACROS: | |
4612 | i = line.find(macro) | |
4613 | if i >= 0: | |
4614 | # Find opening parenthesis. Do a regular expression match here | |
4615 | # to make sure that we are matching the expected CHECK macro, as | |
4616 | # opposed to some other macro that happens to contain the CHECK | |
4617 | # substring. | |
4618 | matched = Match(r'^(.*\b' + macro + r'\s*)\(', line) | |
4619 | if not matched: | |
4620 | continue | |
4621 | return (macro, len(matched.group(1))) | |
4622 | return (None, -1) | |
4623 | ||
4624 | ||
4625 | def CheckCheck(filename, clean_lines, linenum, error): | |
4626 | """Checks the use of CHECK and EXPECT macros. | |
4627 | ||
4628 | Args: | |
4629 | filename: The name of the current file. | |
4630 | clean_lines: A CleansedLines instance containing the file. | |
4631 | linenum: The number of the line to check. | |
4632 | error: The function to call with any errors found. | |
4633 | """ | |
4634 | ||
4635 | # Decide the set of replacement macros that should be suggested | |
4636 | lines = clean_lines.elided | |
4637 | (check_macro, start_pos) = FindCheckMacro(lines[linenum]) | |
4638 | if not check_macro: | |
4639 | return | |
4640 | ||
4641 | # Find end of the boolean expression by matching parentheses | |
4642 | (last_line, end_line, end_pos) = CloseExpression( | |
4643 | clean_lines, linenum, start_pos) | |
4644 | if end_pos < 0: | |
4645 | return | |
4646 | ||
4647 | # If the check macro is followed by something other than a | |
4648 | # semicolon, assume users will log their own custom error messages | |
4649 | # and don't suggest any replacements. | |
4650 | if not Match(r'\s*;', last_line[end_pos:]): | |
4651 | return | |
4652 | ||
4653 | if linenum == end_line: | |
4654 | expression = lines[linenum][start_pos + 1:end_pos - 1] | |
4655 | else: | |
4656 | expression = lines[linenum][start_pos + 1:] | |
4657 | for i in xrange(linenum + 1, end_line): | |
4658 | expression += lines[i] | |
4659 | expression += last_line[0:end_pos - 1] | |
4660 | ||
4661 | # Parse expression so that we can take parentheses into account. | |
4662 | # This avoids false positives for inputs like "CHECK((a < 4) == b)", | |
4663 | # which is not replaceable by CHECK_LE. | |
4664 | lhs = '' | |
4665 | rhs = '' | |
4666 | operator = None | |
4667 | while expression: | |
4668 | matched = Match(r'^\s*(<<|<<=|>>|>>=|->\*|->|&&|\|\||' | |
4669 | r'==|!=|>=|>|<=|<|\()(.*)$', expression) | |
4670 | if matched: | |
4671 | token = matched.group(1) | |
4672 | if token == '(': | |
4673 | # Parenthesized operand | |
4674 | expression = matched.group(2) | |
4675 | (end, _) = FindEndOfExpressionInLine(expression, 0, ['(']) | |
4676 | if end < 0: | |
4677 | return # Unmatched parenthesis | |
4678 | lhs += '(' + expression[0:end] | |
4679 | expression = expression[end:] | |
4680 | elif token in ('&&', '||'): | |
4681 | # Logical and/or operators. This means the expression | |
4682 | # contains more than one term, for example: | |
4683 | # CHECK(42 < a && a < b); | |
4684 | # | |
4685 | # These are not replaceable with CHECK_LE, so bail out early. | |
4686 | return | |
4687 | elif token in ('<<', '<<=', '>>', '>>=', '->*', '->'): | |
4688 | # Non-relational operator | |
4689 | lhs += token | |
4690 | expression = matched.group(2) | |
4691 | else: | |
4692 | # Relational operator | |
4693 | operator = token | |
4694 | rhs = matched.group(2) | |
4695 | break | |
4696 | else: | |
4697 | # Unparenthesized operand. Instead of appending to lhs one character | |
4698 | # at a time, we do another regular expression match to consume several | |
4699 | # characters at once if possible. Trivial benchmark shows that this | |
4700 | # is more efficient when the operands are longer than a single | |
4701 | # character, which is generally the case. | |
4702 | matched = Match(r'^([^-=!<>()&|]+)(.*)$', expression) | |
4703 | if not matched: | |
4704 | matched = Match(r'^(\s*\S)(.*)$', expression) | |
4705 | if not matched: | |
4706 | break | |
4707 | lhs += matched.group(1) | |
4708 | expression = matched.group(2) | |
4709 | ||
4710 | # Only apply checks if we got all parts of the boolean expression | |
4711 | if not (lhs and operator and rhs): | |
4712 | return | |
4713 | ||
4714 | # Check that rhs do not contain logical operators. We already know | |
4715 | # that lhs is fine since the loop above parses out && and ||. | |
4716 | if rhs.find('&&') > -1 or rhs.find('||') > -1: | |
4717 | return | |
4718 | ||
4719 | # At least one of the operands must be a constant literal. This is | |
4720 | # to avoid suggesting replacements for unprintable things like | |
4721 | # CHECK(variable != iterator) | |
4722 | # | |
4723 | # The following pattern matches decimal, hex integers, strings, and | |
4724 | # characters (in that order). | |
4725 | lhs = lhs.strip() | |
4726 | rhs = rhs.strip() | |
4727 | match_constant = r'^([-+]?(\d+|0[xX][0-9a-fA-F]+)[lLuU]{0,3}|".*"|\'.*\')$' | |
4728 | if Match(match_constant, lhs) or Match(match_constant, rhs): | |
4729 | # Note: since we know both lhs and rhs, we can provide a more | |
4730 | # descriptive error message like: | |
4731 | # Consider using CHECK_EQ(x, 42) instead of CHECK(x == 42) | |
4732 | # Instead of: | |
4733 | # Consider using CHECK_EQ instead of CHECK(a == b) | |
4734 | # | |
4735 | # We are still keeping the less descriptive message because if lhs | |
4736 | # or rhs gets long, the error message might become unreadable. | |
4737 | error(filename, linenum, 'readability/check', 2, | |
4738 | 'Consider using %s instead of %s(a %s b)' % ( | |
4739 | _CHECK_REPLACEMENT[check_macro][operator], | |
4740 | check_macro, operator)) | |
4741 | ||
4742 | ||
4743 | def CheckAltTokens(filename, clean_lines, linenum, error): | |
4744 | """Check alternative keywords being used in boolean expressions. | |
4745 | ||
4746 | Args: | |
4747 | filename: The name of the current file. | |
4748 | clean_lines: A CleansedLines instance containing the file. | |
4749 | linenum: The number of the line to check. | |
4750 | error: The function to call with any errors found. | |
4751 | """ | |
4752 | line = clean_lines.elided[linenum] | |
4753 | ||
4754 | # Avoid preprocessor lines | |
4755 | if Match(r'^\s*#', line): | |
4756 | return | |
4757 | ||
4758 | # Last ditch effort to avoid multi-line comments. This will not help | |
4759 | # if the comment started before the current line or ended after the | |
4760 | # current line, but it catches most of the false positives. At least, | |
4761 | # it provides a way to workaround this warning for people who use | |
4762 | # multi-line comments in preprocessor macros. | |
4763 | # | |
4764 | # TODO(unknown): remove this once cpplint has better support for | |
4765 | # multi-line comments. | |
4766 | if line.find('/*') >= 0 or line.find('*/') >= 0: | |
4767 | return | |
4768 | ||
4769 | for match in _ALT_TOKEN_REPLACEMENT_PATTERN.finditer(line): | |
4770 | error(filename, linenum, 'readability/alt_tokens', 2, | |
4771 | 'Use operator %s instead of %s' % ( | |
4772 | _ALT_TOKEN_REPLACEMENT[match.group(1)], match.group(1))) | |
4773 | ||
4774 | ||
4775 | def GetLineWidth(line): | |
4776 | """Determines the width of the line in column positions. | |
4777 | ||
4778 | Args: | |
4779 | line: A string, which may be a Unicode string. | |
4780 | ||
4781 | Returns: | |
4782 | The width of the line in column positions, accounting for Unicode | |
4783 | combining characters and wide characters. | |
4784 | """ | |
4785 | if isinstance(line, unicode): | |
4786 | width = 0 | |
4787 | for uc in unicodedata.normalize('NFC', line): | |
4788 | if unicodedata.east_asian_width(uc) in ('W', 'F'): | |
4789 | width += 2 | |
4790 | elif not unicodedata.combining(uc): | |
4791 | # Issue 337 | |
4792 | # https://mail.python.org/pipermail/python-list/2012-August/628809.html | |
4793 | if (sys.version_info.major, sys.version_info.minor) <= (3, 2): | |
4794 | # https://github.com/python/cpython/blob/2.7/Include/unicodeobject.h#L81 | |
4795 | is_wide_build = sysconfig.get_config_var("Py_UNICODE_SIZE") >= 4 | |
4796 | # https://github.com/python/cpython/blob/2.7/Objects/unicodeobject.c#L564 | |
4797 | is_low_surrogate = 0xDC00 <= ord(uc) <= 0xDFFF | |
4798 | if not is_wide_build and is_low_surrogate: | |
4799 | width -= 1 | |
4800 | ||
4801 | width += 1 | |
4802 | return width | |
4803 | else: | |
4804 | return len(line) | |
4805 | ||
4806 | ||
4807 | def CheckStyle(filename, clean_lines, linenum, file_extension, nesting_state, | |
4808 | error): | |
4809 | """Checks rules from the 'C++ style rules' section of cppguide.html. | |
4810 | ||
4811 | Most of these rules are hard to test (naming, comment style), but we | |
4812 | do what we can. In particular we check for 2-space indents, line lengths, | |
4813 | tab usage, spaces inside code, etc. | |
4814 | ||
4815 | Args: | |
4816 | filename: The name of the current file. | |
4817 | clean_lines: A CleansedLines instance containing the file. | |
4818 | linenum: The number of the line to check. | |
4819 | file_extension: The extension (without the dot) of the filename. | |
4820 | nesting_state: A NestingState instance which maintains information about | |
4821 | the current stack of nested blocks being parsed. | |
4822 | error: The function to call with any errors found. | |
4823 | """ | |
4824 | ||
4825 | # Don't use "elided" lines here, otherwise we can't check commented lines. | |
4826 | # Don't want to use "raw" either, because we don't want to check inside C++11 | |
4827 | # raw strings, | |
4828 | raw_lines = clean_lines.lines_without_raw_strings | |
4829 | line = raw_lines[linenum] | |
4830 | prev = raw_lines[linenum - 1] if linenum > 0 else '' | |
4831 | ||
4832 | if line.find('\t') != -1: | |
4833 | error(filename, linenum, 'whitespace/tab', 1, | |
4834 | 'Tab found; better to use spaces') | |
4835 | ||
4836 | # One or three blank spaces at the beginning of the line is weird; it's | |
4837 | # hard to reconcile that with 2-space indents. | |
4838 | # NOTE: here are the conditions rob pike used for his tests. Mine aren't | |
4839 | # as sophisticated, but it may be worth becoming so: RLENGTH==initial_spaces | |
4840 | # if(RLENGTH > 20) complain = 0; | |
4841 | # if(match($0, " +(error|private|public|protected):")) complain = 0; | |
4842 | # if(match(prev, "&& *$")) complain = 0; | |
4843 | # if(match(prev, "\\|\\| *$")) complain = 0; | |
4844 | # if(match(prev, "[\",=><] *$")) complain = 0; | |
4845 | # if(match($0, " <<")) complain = 0; | |
4846 | # if(match(prev, " +for \\(")) complain = 0; | |
4847 | # if(prevodd && match(prevprev, " +for \\(")) complain = 0; | |
4848 | scope_or_label_pattern = r'\s*(?:public|private|protected|signals)(?:\s+(?:slots\s*)?)?:\s*\\?$' | |
4849 | classinfo = nesting_state.InnermostClass() | |
4850 | initial_spaces = 0 | |
4851 | cleansed_line = clean_lines.elided[linenum] | |
4852 | while initial_spaces < len(line) and line[initial_spaces] == ' ': | |
4853 | initial_spaces += 1 | |
4854 | # There are certain situations we allow one space, notably for | |
4855 | # section labels, and also lines containing multi-line raw strings. | |
4856 | # We also don't check for lines that look like continuation lines | |
4857 | # (of lines ending in double quotes, commas, equals, or angle brackets) | |
4858 | # because the rules for how to indent those are non-trivial. | |
4859 | if (not Search(r'[",=><] *$', prev) and | |
4860 | (initial_spaces == 1 or initial_spaces == 3) and | |
4861 | not Match(scope_or_label_pattern, cleansed_line) and | |
4862 | not (clean_lines.raw_lines[linenum] != line and | |
4863 | Match(r'^\s*""', line))): | |
4864 | error(filename, linenum, 'whitespace/indent', 3, | |
4865 | 'Weird number of spaces at line-start. ' | |
4866 | 'Are you using a 2-space indent?') | |
4867 | ||
4868 | if line and line[-1].isspace(): | |
4869 | error(filename, linenum, 'whitespace/end_of_line', 4, | |
4870 | 'Line ends in whitespace. Consider deleting these extra spaces.') | |
4871 | ||
4872 | # Check if the line is a header guard. | |
4873 | is_header_guard = False | |
4874 | if IsHeaderExtension(file_extension): | |
4875 | cppvar = GetHeaderGuardCPPVariable(filename) | |
4876 | if (line.startswith('#ifndef %s' % cppvar) or | |
4877 | line.startswith('#define %s' % cppvar) or | |
4878 | line.startswith('#endif // %s' % cppvar)): | |
4879 | is_header_guard = True | |
4880 | # #include lines and header guards can be long, since there's no clean way to | |
4881 | # split them. | |
4882 | # | |
4883 | # URLs can be long too. It's possible to split these, but it makes them | |
4884 | # harder to cut&paste. | |
4885 | # | |
4886 | # The "$Id:...$" comment may also get very long without it being the | |
4887 | # developers fault. | |
4888 | # | |
4889 | # Doxygen documentation copying can get pretty long when using an overloaded | |
4890 | # function declaration | |
4891 | if (not line.startswith('#include') and not is_header_guard and | |
4892 | not Match(r'^\s*//.*http(s?)://\S*$', line) and | |
4893 | not Match(r'^\s*//\s*[^\s]*$', line) and | |
4894 | not Match(r'^// \$Id:.*#[0-9]+ \$$', line) and | |
4895 | not Match(r'^\s*/// [@\\](copydoc|copydetails|copybrief) .*$', line)): | |
4896 | line_width = GetLineWidth(line) | |
4897 | if line_width > _line_length: | |
4898 | error(filename, linenum, 'whitespace/line_length', 2, | |
4899 | 'Lines should be <= %i characters long' % _line_length) | |
4900 | ||
4901 | if (cleansed_line.count(';') > 1 and | |
4902 | # allow simple single line lambdas | |
4903 | not Match(r'^[^{};]*\[[^\[\]]*\][^{}]*\{[^{}\n\r]*\}', | |
4904 | line) and | |
4905 | # for loops are allowed two ;'s (and may run over two lines). | |
4906 | cleansed_line.find('for') == -1 and | |
4907 | (GetPreviousNonBlankLine(clean_lines, linenum)[0].find('for') == -1 or | |
4908 | GetPreviousNonBlankLine(clean_lines, linenum)[0].find(';') != -1) and | |
4909 | # It's ok to have many commands in a switch case that fits in 1 line | |
4910 | not ((cleansed_line.find('case ') != -1 or | |
4911 | cleansed_line.find('default:') != -1) and | |
4912 | cleansed_line.find('break;') != -1)): | |
4913 | error(filename, linenum, 'whitespace/newline', 0, | |
4914 | 'More than one command on the same line') | |
4915 | ||
4916 | # Some more style checks | |
4917 | CheckBraces(filename, clean_lines, linenum, error) | |
4918 | CheckTrailingSemicolon(filename, clean_lines, linenum, error) | |
4919 | CheckEmptyBlockBody(filename, clean_lines, linenum, error) | |
4920 | CheckSpacing(filename, clean_lines, linenum, nesting_state, error) | |
4921 | CheckOperatorSpacing(filename, clean_lines, linenum, error) | |
4922 | CheckParenthesisSpacing(filename, clean_lines, linenum, error) | |
4923 | CheckCommaSpacing(filename, clean_lines, linenum, error) | |
4924 | CheckBracesSpacing(filename, clean_lines, linenum, nesting_state, error) | |
4925 | CheckSpacingForFunctionCall(filename, clean_lines, linenum, error) | |
4926 | CheckCheck(filename, clean_lines, linenum, error) | |
4927 | CheckAltTokens(filename, clean_lines, linenum, error) | |
4928 | classinfo = nesting_state.InnermostClass() | |
4929 | if classinfo: | |
4930 | CheckSectionSpacing(filename, clean_lines, classinfo, linenum, error) | |
4931 | ||
4932 | ||
4933 | _RE_PATTERN_INCLUDE = re.compile(r'^\s*#\s*include\s*([<"])([^>"]*)[>"].*$') | |
4934 | # Matches the first component of a filename delimited by -s and _s. That is: | |
4935 | # _RE_FIRST_COMPONENT.match('foo').group(0) == 'foo' | |
4936 | # _RE_FIRST_COMPONENT.match('foo.cc').group(0) == 'foo' | |
4937 | # _RE_FIRST_COMPONENT.match('foo-bar_baz.cc').group(0) == 'foo' | |
4938 | # _RE_FIRST_COMPONENT.match('foo_bar-baz.cc').group(0) == 'foo' | |
4939 | _RE_FIRST_COMPONENT = re.compile(r'^[^-_.]+') | |
4940 | ||
4941 | ||
4942 | def _DropCommonSuffixes(filename): | |
4943 | """Drops common suffixes like _test.cc or -inl.h from filename. | |
4944 | ||
4945 | For example: | |
4946 | >>> _DropCommonSuffixes('foo/foo-inl.h') | |
4947 | 'foo/foo' | |
4948 | >>> _DropCommonSuffixes('foo/bar/foo.cc') | |
4949 | 'foo/bar/foo' | |
4950 | >>> _DropCommonSuffixes('foo/foo_internal.h') | |
4951 | 'foo/foo' | |
4952 | >>> _DropCommonSuffixes('foo/foo_unusualinternal.h') | |
4953 | 'foo/foo_unusualinternal' | |
4954 | ||
4955 | Args: | |
4956 | filename: The input filename. | |
4957 | ||
4958 | Returns: | |
4959 | The filename with the common suffix removed. | |
4960 | """ | |
4961 | for suffix in itertools.chain( | |
4962 | ('%s.%s' % (test_suffix.lstrip('_'), ext) | |
4963 | for test_suffix, ext in itertools.product(_test_suffixes, GetNonHeaderExtensions())), | |
4964 | ('%s.%s' % (suffix, ext) | |
4965 | for suffix, ext in itertools.product(['inl', 'imp', 'internal'], GetHeaderExtensions()))): | |
4966 | if (filename.endswith(suffix) and len(filename) > len(suffix) and | |
4967 | filename[-len(suffix) - 1] in ('-', '_')): | |
4968 | return filename[:-len(suffix) - 1] | |
4969 | return os.path.splitext(filename)[0] | |
4970 | ||
4971 | ||
4972 | def _ClassifyInclude(fileinfo, include, used_angle_brackets, include_order="default"): | |
4973 | """Figures out what kind of header 'include' is. | |
4974 | ||
4975 | Args: | |
4976 | fileinfo: The current file cpplint is running over. A FileInfo instance. | |
4977 | include: The path to a #included file. | |
4978 | used_angle_brackets: True if the #include used <> rather than "". | |
4979 | include_order: "default" or other value allowed in program arguments | |
4980 | ||
4981 | Returns: | |
4982 | One of the _XXX_HEADER constants. | |
4983 | ||
4984 | For example: | |
4985 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'stdio.h', True) | |
4986 | _C_SYS_HEADER | |
4987 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'string', True) | |
4988 | _CPP_SYS_HEADER | |
4989 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/foo.h', True, "standardcfirst") | |
4990 | _OTHER_SYS_HEADER | |
4991 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/foo.h', False) | |
4992 | _LIKELY_MY_HEADER | |
4993 | >>> _ClassifyInclude(FileInfo('foo/foo_unknown_extension.cc'), | |
4994 | ... 'bar/foo_other_ext.h', False) | |
4995 | _POSSIBLE_MY_HEADER | |
4996 | >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/bar.h', False) | |
4997 | _OTHER_HEADER | |
4998 | """ | |
4999 | # This is a list of all standard c++ header files, except | |
5000 | # those already checked for above. | |
5001 | is_cpp_header = include in _CPP_HEADERS | |
5002 | ||
5003 | # Mark include as C header if in list or in a known folder for standard-ish C headers. | |
5004 | is_std_c_header = (include_order == "default") or (include in _C_HEADERS | |
5005 | # additional linux glibc header folders | |
5006 | or Search(r'(?:%s)\/.*\.h' % "|".join(C_STANDARD_HEADER_FOLDERS), include)) | |
5007 | ||
5008 | # Headers with C++ extensions shouldn't be considered C system headers | |
5009 | is_system = used_angle_brackets and not os.path.splitext(include)[1] in ['.hpp', '.hxx', '.h++'] | |
5010 | ||
5011 | if is_system: | |
5012 | if is_cpp_header: | |
5013 | return _CPP_SYS_HEADER | |
5014 | if is_std_c_header: | |
5015 | return _C_SYS_HEADER | |
5016 | else: | |
5017 | return _OTHER_SYS_HEADER | |
5018 | ||
5019 | # If the target file and the include we're checking share a | |
5020 | # basename when we drop common extensions, and the include | |
5021 | # lives in . , then it's likely to be owned by the target file. | |
5022 | target_dir, target_base = ( | |
5023 | os.path.split(_DropCommonSuffixes(fileinfo.RepositoryName()))) | |
5024 | include_dir, include_base = os.path.split(_DropCommonSuffixes(include)) | |
5025 | target_dir_pub = os.path.normpath(target_dir + '/../public') | |
5026 | target_dir_pub = target_dir_pub.replace('\\', '/') | |
5027 | if target_base == include_base and ( | |
5028 | include_dir == target_dir or | |
5029 | include_dir == target_dir_pub): | |
5030 | return _LIKELY_MY_HEADER | |
5031 | ||
5032 | # If the target and include share some initial basename | |
5033 | # component, it's possible the target is implementing the | |
5034 | # include, so it's allowed to be first, but we'll never | |
5035 | # complain if it's not there. | |
5036 | target_first_component = _RE_FIRST_COMPONENT.match(target_base) | |
5037 | include_first_component = _RE_FIRST_COMPONENT.match(include_base) | |
5038 | if (target_first_component and include_first_component and | |
5039 | target_first_component.group(0) == | |
5040 | include_first_component.group(0)): | |
5041 | return _POSSIBLE_MY_HEADER | |
5042 | ||
5043 | return _OTHER_HEADER | |
5044 | ||
5045 | ||
5046 | ||
5047 | def CheckIncludeLine(filename, clean_lines, linenum, include_state, error): | |
5048 | """Check rules that are applicable to #include lines. | |
5049 | ||
5050 | Strings on #include lines are NOT removed from elided line, to make | |
5051 | certain tasks easier. However, to prevent false positives, checks | |
5052 | applicable to #include lines in CheckLanguage must be put here. | |
5053 | ||
5054 | Args: | |
5055 | filename: The name of the current file. | |
5056 | clean_lines: A CleansedLines instance containing the file. | |
5057 | linenum: The number of the line to check. | |
5058 | include_state: An _IncludeState instance in which the headers are inserted. | |
5059 | error: The function to call with any errors found. | |
5060 | """ | |
5061 | fileinfo = FileInfo(filename) | |
5062 | line = clean_lines.lines[linenum] | |
5063 | ||
5064 | # "include" should use the new style "foo/bar.h" instead of just "bar.h" | |
5065 | # Only do this check if the included header follows google naming | |
5066 | # conventions. If not, assume that it's a 3rd party API that | |
5067 | # requires special include conventions. | |
5068 | # | |
5069 | # We also make an exception for Lua headers, which follow google | |
5070 | # naming convention but not the include convention. | |
5071 | match = Match(r'#include\s*"([^/]+\.h)"', line) | |
5072 | if match and not _THIRD_PARTY_HEADERS_PATTERN.match(match.group(1)): | |
5073 | error(filename, linenum, 'build/include_subdir', 4, | |
5074 | 'Include the directory when naming .h files') | |
5075 | ||
5076 | # we shouldn't include a file more than once. actually, there are a | |
5077 | # handful of instances where doing so is okay, but in general it's | |
5078 | # not. | |
5079 | match = _RE_PATTERN_INCLUDE.search(line) | |
5080 | if match: | |
5081 | include = match.group(2) | |
5082 | used_angle_brackets = (match.group(1) == '<') | |
5083 | duplicate_line = include_state.FindHeader(include) | |
5084 | if duplicate_line >= 0: | |
5085 | error(filename, linenum, 'build/include', 4, | |
5086 | '"%s" already included at %s:%s' % | |
5087 | (include, filename, duplicate_line)) | |
5088 | return | |
5089 | ||
5090 | for extension in GetNonHeaderExtensions(): | |
5091 | if (include.endswith('.' + extension) and | |
5092 | os.path.dirname(fileinfo.RepositoryName()) != os.path.dirname(include)): | |
5093 | error(filename, linenum, 'build/include', 4, | |
5094 | 'Do not include .' + extension + ' files from other packages') | |
5095 | return | |
5096 | ||
5097 | # We DO want to include a 3rd party looking header if it matches the | |
5098 | # filename. Otherwise we get an erroneous error "...should include its | |
5099 | # header" error later. | |
5100 | third_src_header = False | |
5101 | for ext in GetHeaderExtensions(): | |
5102 | basefilename = filename[0:len(filename) - len(fileinfo.Extension())] | |
5103 | headerfile = basefilename + '.' + ext | |
5104 | headername = FileInfo(headerfile).RepositoryName() | |
5105 | if headername in include or include in headername: | |
5106 | third_src_header = True | |
5107 | break | |
5108 | ||
5109 | if third_src_header or not _THIRD_PARTY_HEADERS_PATTERN.match(include): | |
5110 | include_state.include_list[-1].append((include, linenum)) | |
5111 | ||
5112 | # We want to ensure that headers appear in the right order: | |
5113 | # 1) for foo.cc, foo.h (preferred location) | |
5114 | # 2) c system files | |
5115 | # 3) cpp system files | |
5116 | # 4) for foo.cc, foo.h (deprecated location) | |
5117 | # 5) other google headers | |
5118 | # | |
5119 | # We classify each include statement as one of those 5 types | |
5120 | # using a number of techniques. The include_state object keeps | |
5121 | # track of the highest type seen, and complains if we see a | |
5122 | # lower type after that. | |
5123 | error_message = include_state.CheckNextIncludeOrder( | |
5124 | _ClassifyInclude(fileinfo, include, used_angle_brackets, _include_order)) | |
5125 | if error_message: | |
5126 | error(filename, linenum, 'build/include_order', 4, | |
5127 | '%s. Should be: %s.h, c system, c++ system, other.' % | |
5128 | (error_message, fileinfo.BaseName())) | |
5129 | canonical_include = include_state.CanonicalizeAlphabeticalOrder(include) | |
5130 | if not include_state.IsInAlphabeticalOrder( | |
5131 | clean_lines, linenum, canonical_include): | |
5132 | error(filename, linenum, 'build/include_alpha', 4, | |
5133 | 'Include "%s" not in alphabetical order' % include) | |
5134 | include_state.SetLastHeader(canonical_include) | |
5135 | ||
5136 | ||
5137 | ||
5138 | def _GetTextInside(text, start_pattern): | |
5139 | r"""Retrieves all the text between matching open and close parentheses. | |
5140 | ||
5141 | Given a string of lines and a regular expression string, retrieve all the text | |
5142 | following the expression and between opening punctuation symbols like | |
5143 | (, [, or {, and the matching close-punctuation symbol. This properly nested | |
5144 | occurrences of the punctuations, so for the text like | |
5145 | printf(a(), b(c())); | |
5146 | a call to _GetTextInside(text, r'printf\(') will return 'a(), b(c())'. | |
5147 | start_pattern must match string having an open punctuation symbol at the end. | |
5148 | ||
5149 | Args: | |
5150 | text: The lines to extract text. Its comments and strings must be elided. | |
5151 | It can be single line and can span multiple lines. | |
5152 | start_pattern: The regexp string indicating where to start extracting | |
5153 | the text. | |
5154 | Returns: | |
5155 | The extracted text. | |
5156 | None if either the opening string or ending punctuation could not be found. | |
5157 | """ | |
5158 | # TODO(unknown): Audit cpplint.py to see what places could be profitably | |
5159 | # rewritten to use _GetTextInside (and use inferior regexp matching today). | |
5160 | ||
5161 | # Give opening punctuations to get the matching close-punctuations. | |
5162 | matching_punctuation = {'(': ')', '{': '}', '[': ']'} | |
5163 | closing_punctuation = set(itervalues(matching_punctuation)) | |
5164 | ||
5165 | # Find the position to start extracting text. | |
5166 | match = re.search(start_pattern, text, re.M) | |
5167 | if not match: # start_pattern not found in text. | |
5168 | return None | |
5169 | start_position = match.end(0) | |
5170 | ||
5171 | assert start_position > 0, ( | |
5172 | 'start_pattern must ends with an opening punctuation.') | |
5173 | assert text[start_position - 1] in matching_punctuation, ( | |
5174 | 'start_pattern must ends with an opening punctuation.') | |
5175 | # Stack of closing punctuations we expect to have in text after position. | |
5176 | punctuation_stack = [matching_punctuation[text[start_position - 1]]] | |
5177 | position = start_position | |
5178 | while punctuation_stack and position < len(text): | |
5179 | if text[position] == punctuation_stack[-1]: | |
5180 | punctuation_stack.pop() | |
5181 | elif text[position] in closing_punctuation: | |
5182 | # A closing punctuation without matching opening punctuations. | |
5183 | return None | |
5184 | elif text[position] in matching_punctuation: | |
5185 | punctuation_stack.append(matching_punctuation[text[position]]) | |
5186 | position += 1 | |
5187 | if punctuation_stack: | |
5188 | # Opening punctuations left without matching close-punctuations. | |
5189 | return None | |
5190 | # punctuations match. | |
5191 | return text[start_position:position - 1] | |
5192 | ||
5193 | ||
5194 | # Patterns for matching call-by-reference parameters. | |
5195 | # | |
5196 | # Supports nested templates up to 2 levels deep using this messy pattern: | |
5197 | # < (?: < (?: < [^<>]* | |
5198 | # > | |
5199 | # | [^<>] )* | |
5200 | # > | |
5201 | # | [^<>] )* | |
5202 | # > | |
5203 | _RE_PATTERN_IDENT = r'[_a-zA-Z]\w*' # =~ [[:alpha:]][[:alnum:]]* | |
5204 | _RE_PATTERN_TYPE = ( | |
5205 | r'(?:const\s+)?(?:typename\s+|class\s+|struct\s+|union\s+|enum\s+)?' | |
5206 | r'(?:\w|' | |
5207 | r'\s*<(?:<(?:<[^<>]*>|[^<>])*>|[^<>])*>|' | |
5208 | r'::)+') | |
5209 | # A call-by-reference parameter ends with '& identifier'. | |
5210 | _RE_PATTERN_REF_PARAM = re.compile( | |
5211 | r'(' + _RE_PATTERN_TYPE + r'(?:\s*(?:\bconst\b|[*]))*\s*' | |
5212 | r'&\s*' + _RE_PATTERN_IDENT + r')\s*(?:=[^,()]+)?[,)]') | |
5213 | # A call-by-const-reference parameter either ends with 'const& identifier' | |
5214 | # or looks like 'const type& identifier' when 'type' is atomic. | |
5215 | _RE_PATTERN_CONST_REF_PARAM = ( | |
5216 | r'(?:.*\s*\bconst\s*&\s*' + _RE_PATTERN_IDENT + | |
5217 | r'|const\s+' + _RE_PATTERN_TYPE + r'\s*&\s*' + _RE_PATTERN_IDENT + r')') | |
5218 | # Stream types. | |
5219 | _RE_PATTERN_REF_STREAM_PARAM = ( | |
5220 | r'(?:.*stream\s*&\s*' + _RE_PATTERN_IDENT + r')') | |
5221 | ||
5222 | ||
5223 | def CheckLanguage(filename, clean_lines, linenum, file_extension, | |
5224 | include_state, nesting_state, error): | |
5225 | """Checks rules from the 'C++ language rules' section of cppguide.html. | |
5226 | ||
5227 | Some of these rules are hard to test (function overloading, using | |
5228 | uint32 inappropriately), but we do the best we can. | |
5229 | ||
5230 | Args: | |
5231 | filename: The name of the current file. | |
5232 | clean_lines: A CleansedLines instance containing the file. | |
5233 | linenum: The number of the line to check. | |
5234 | file_extension: The extension (without the dot) of the filename. | |
5235 | include_state: An _IncludeState instance in which the headers are inserted. | |
5236 | nesting_state: A NestingState instance which maintains information about | |
5237 | the current stack of nested blocks being parsed. | |
5238 | error: The function to call with any errors found. | |
5239 | """ | |
5240 | # If the line is empty or consists of entirely a comment, no need to | |
5241 | # check it. | |
5242 | line = clean_lines.elided[linenum] | |
5243 | if not line: | |
5244 | return | |
5245 | ||
5246 | match = _RE_PATTERN_INCLUDE.search(line) | |
5247 | if match: | |
5248 | CheckIncludeLine(filename, clean_lines, linenum, include_state, error) | |
5249 | return | |
5250 | ||
5251 | # Reset include state across preprocessor directives. This is meant | |
5252 | # to silence warnings for conditional includes. | |
5253 | match = Match(r'^\s*#\s*(if|ifdef|ifndef|elif|else|endif)\b', line) | |
5254 | if match: | |
5255 | include_state.ResetSection(match.group(1)) | |
5256 | ||
5257 | ||
5258 | # Perform other checks now that we are sure that this is not an include line | |
5259 | CheckCasts(filename, clean_lines, linenum, error) | |
5260 | CheckGlobalStatic(filename, clean_lines, linenum, error) | |
5261 | CheckPrintf(filename, clean_lines, linenum, error) | |
5262 | ||
5263 | if IsHeaderExtension(file_extension): | |
5264 | # TODO(unknown): check that 1-arg constructors are explicit. | |
5265 | # How to tell it's a constructor? | |
5266 | # (handled in CheckForNonStandardConstructs for now) | |
5267 | # TODO(unknown): check that classes declare or disable copy/assign | |
5268 | # (level 1 error) | |
5269 | pass | |
5270 | ||
5271 | # Check if people are using the verboten C basic types. The only exception | |
5272 | # we regularly allow is "unsigned short port" for port. | |
5273 | if Search(r'\bshort port\b', line): | |
5274 | if not Search(r'\bunsigned short port\b', line): | |
5275 | error(filename, linenum, 'runtime/int', 4, | |
5276 | 'Use "unsigned short" for ports, not "short"') | |
5277 | else: | |
5278 | match = Search(r'\b(short|long(?! +double)|long long)\b', line) | |
5279 | if match: | |
5280 | error(filename, linenum, 'runtime/int', 4, | |
5281 | 'Use int16/int64/etc, rather than the C type %s' % match.group(1)) | |
5282 | ||
5283 | # Check if some verboten operator overloading is going on | |
5284 | # TODO(unknown): catch out-of-line unary operator&: | |
5285 | # class X {}; | |
5286 | # int operator&(const X& x) { return 42; } // unary operator& | |
5287 | # The trick is it's hard to tell apart from binary operator&: | |
5288 | # class Y { int operator&(const Y& x) { return 23; } }; // binary operator& | |
5289 | if Search(r'\boperator\s*&\s*\(\s*\)', line): | |
5290 | error(filename, linenum, 'runtime/operator', 4, | |
5291 | 'Unary operator& is dangerous. Do not use it.') | |
5292 | ||
5293 | # Check for suspicious usage of "if" like | |
5294 | # } if (a == b) { | |
5295 | if Search(r'\}\s*if\s*\(', line): | |
5296 | error(filename, linenum, 'readability/braces', 4, | |
5297 | 'Did you mean "else if"? If not, start a new line for "if".') | |
5298 | ||
5299 | # Check for potential format string bugs like printf(foo). | |
5300 | # We constrain the pattern not to pick things like DocidForPrintf(foo). | |
5301 | # Not perfect but it can catch printf(foo.c_str()) and printf(foo->c_str()) | |
5302 | # TODO(unknown): Catch the following case. Need to change the calling | |
5303 | # convention of the whole function to process multiple line to handle it. | |
5304 | # printf( | |
5305 | # boy_this_is_a_really_long_variable_that_cannot_fit_on_the_prev_line); | |
5306 | printf_args = _GetTextInside(line, r'(?i)\b(string)?printf\s*\(') | |
5307 | if printf_args: | |
5308 | match = Match(r'([\w.\->()]+)$', printf_args) | |
5309 | if match and match.group(1) != '__VA_ARGS__': | |
5310 | function_name = re.search(r'\b((?:string)?printf)\s*\(', | |
5311 | line, re.I).group(1) | |
5312 | error(filename, linenum, 'runtime/printf', 4, | |
5313 | 'Potential format string bug. Do %s("%%s", %s) instead.' | |
5314 | % (function_name, match.group(1))) | |
5315 | ||
5316 | # Check for potential memset bugs like memset(buf, sizeof(buf), 0). | |
5317 | match = Search(r'memset\s*\(([^,]*),\s*([^,]*),\s*0\s*\)', line) | |
5318 | if match and not Match(r"^''|-?[0-9]+|0x[0-9A-Fa-f]$", match.group(2)): | |
5319 | error(filename, linenum, 'runtime/memset', 4, | |
5320 | 'Did you mean "memset(%s, 0, %s)"?' | |
5321 | % (match.group(1), match.group(2))) | |
5322 | ||
5323 | if Search(r'\busing namespace\b', line): | |
5324 | if Search(r'\bliterals\b', line): | |
5325 | error(filename, linenum, 'build/namespaces_literals', 5, | |
5326 | 'Do not use namespace using-directives. ' | |
5327 | 'Use using-declarations instead.') | |
5328 | else: | |
5329 | error(filename, linenum, 'build/namespaces', 5, | |
5330 | 'Do not use namespace using-directives. ' | |
5331 | 'Use using-declarations instead.') | |
5332 | ||
5333 | # Detect variable-length arrays. | |
5334 | match = Match(r'\s*(.+::)?(\w+) [a-z]\w*\[(.+)];', line) | |
5335 | if (match and match.group(2) != 'return' and match.group(2) != 'delete' and | |
5336 | match.group(3).find(']') == -1): | |
5337 | # Split the size using space and arithmetic operators as delimiters. | |
5338 | # If any of the resulting tokens are not compile time constants then | |
5339 | # report the error. | |
5340 | tokens = re.split(r'\s|\+|\-|\*|\/|<<|>>]', match.group(3)) | |
5341 | is_const = True | |
5342 | skip_next = False | |
5343 | for tok in tokens: | |
5344 | if skip_next: | |
5345 | skip_next = False | |
5346 | continue | |
5347 | ||
5348 | if Search(r'sizeof\(.+\)', tok): continue | |
5349 | if Search(r'arraysize\(\w+\)', tok): continue | |
5350 | ||
5351 | tok = tok.lstrip('(') | |
5352 | tok = tok.rstrip(')') | |
5353 | if not tok: continue | |
5354 | if Match(r'\d+', tok): continue | |
5355 | if Match(r'0[xX][0-9a-fA-F]+', tok): continue | |
5356 | if Match(r'k[A-Z0-9]\w*', tok): continue | |
5357 | if Match(r'(.+::)?k[A-Z0-9]\w*', tok): continue | |
5358 | if Match(r'(.+::)?[A-Z][A-Z0-9_]*', tok): continue | |
5359 | # A catch all for tricky sizeof cases, including 'sizeof expression', | |
5360 | # 'sizeof(*type)', 'sizeof(const type)', 'sizeof(struct StructName)' | |
5361 | # requires skipping the next token because we split on ' ' and '*'. | |
5362 | if tok.startswith('sizeof'): | |
5363 | skip_next = True | |
5364 | continue | |
5365 | is_const = False | |
5366 | break | |
5367 | if not is_const: | |
5368 | error(filename, linenum, 'runtime/arrays', 1, | |
5369 | 'Do not use variable-length arrays. Use an appropriately named ' | |
5370 | "('k' followed by CamelCase) compile-time constant for the size.") | |
5371 | ||
5372 | # Check for use of unnamed namespaces in header files. Registration | |
5373 | # macros are typically OK, so we allow use of "namespace {" on lines | |
5374 | # that end with backslashes. | |
5375 | if (IsHeaderExtension(file_extension) | |
5376 | and Search(r'\bnamespace\s*{', line) | |
5377 | and line[-1] != '\\'): | |
5378 | error(filename, linenum, 'build/namespaces_headers', 4, | |
5379 | 'Do not use unnamed namespaces in header files. See ' | |
5380 | 'https://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Namespaces' | |
5381 | ' for more information.') | |
5382 | ||
5383 | ||
5384 | def CheckGlobalStatic(filename, clean_lines, linenum, error): | |
5385 | """Check for unsafe global or static objects. | |
5386 | ||
5387 | Args: | |
5388 | filename: The name of the current file. | |
5389 | clean_lines: A CleansedLines instance containing the file. | |
5390 | linenum: The number of the line to check. | |
5391 | error: The function to call with any errors found. | |
5392 | """ | |
5393 | line = clean_lines.elided[linenum] | |
5394 | ||
5395 | # Match two lines at a time to support multiline declarations | |
5396 | if linenum + 1 < clean_lines.NumLines() and not Search(r'[;({]', line): | |
5397 | line += clean_lines.elided[linenum + 1].strip() | |
5398 | ||
5399 | # Check for people declaring static/global STL strings at the top level. | |
5400 | # This is dangerous because the C++ language does not guarantee that | |
5401 | # globals with constructors are initialized before the first access, and | |
5402 | # also because globals can be destroyed when some threads are still running. | |
5403 | # TODO(unknown): Generalize this to also find static unique_ptr instances. | |
5404 | # TODO(unknown): File bugs for clang-tidy to find these. | |
5405 | match = Match( | |
5406 | r'((?:|static +)(?:|const +))(?::*std::)?string( +const)? +' | |
5407 | r'([a-zA-Z0-9_:]+)\b(.*)', | |
5408 | line) | |
5409 | ||
5410 | # Remove false positives: | |
5411 | # - String pointers (as opposed to values). | |
5412 | # string *pointer | |
5413 | # const string *pointer | |
5414 | # string const *pointer | |
5415 | # string *const pointer | |
5416 | # | |
5417 | # - Functions and template specializations. | |
5418 | # string Function<Type>(... | |
5419 | # string Class<Type>::Method(... | |
5420 | # | |
5421 | # - Operators. These are matched separately because operator names | |
5422 | # cross non-word boundaries, and trying to match both operators | |
5423 | # and functions at the same time would decrease accuracy of | |
5424 | # matching identifiers. | |
5425 | # string Class::operator*() | |
5426 | if (match and | |
5427 | not Search(r'\bstring\b(\s+const)?\s*[\*\&]\s*(const\s+)?\w', line) and | |
5428 | not Search(r'\boperator\W', line) and | |
5429 | not Match(r'\s*(<.*>)?(::[a-zA-Z0-9_]+)*\s*\(([^"]|$)', match.group(4))): | |
5430 | if Search(r'\bconst\b', line): | |
5431 | error(filename, linenum, 'runtime/string', 4, | |
5432 | 'For a static/global string constant, use a C style string ' | |
5433 | 'instead: "%schar%s %s[]".' % | |
5434 | (match.group(1), match.group(2) or '', match.group(3))) | |
5435 | else: | |
5436 | error(filename, linenum, 'runtime/string', 4, | |
5437 | 'Static/global string variables are not permitted.') | |
5438 | ||
5439 | if (Search(r'\b([A-Za-z0-9_]*_)\(\1\)', line) or | |
5440 | Search(r'\b([A-Za-z0-9_]*_)\(CHECK_NOTNULL\(\1\)\)', line)): | |
5441 | error(filename, linenum, 'runtime/init', 4, | |
5442 | 'You seem to be initializing a member variable with itself.') | |
5443 | ||
5444 | ||
5445 | def CheckPrintf(filename, clean_lines, linenum, error): | |
5446 | """Check for printf related issues. | |
5447 | ||
5448 | Args: | |
5449 | filename: The name of the current file. | |
5450 | clean_lines: A CleansedLines instance containing the file. | |
5451 | linenum: The number of the line to check. | |
5452 | error: The function to call with any errors found. | |
5453 | """ | |
5454 | line = clean_lines.elided[linenum] | |
5455 | ||
5456 | # When snprintf is used, the second argument shouldn't be a literal. | |
5457 | match = Search(r'snprintf\s*\(([^,]*),\s*([0-9]*)\s*,', line) | |
5458 | if match and match.group(2) != '0': | |
5459 | # If 2nd arg is zero, snprintf is used to calculate size. | |
5460 | error(filename, linenum, 'runtime/printf', 3, | |
5461 | 'If you can, use sizeof(%s) instead of %s as the 2nd arg ' | |
5462 | 'to snprintf.' % (match.group(1), match.group(2))) | |
5463 | ||
5464 | # Check if some verboten C functions are being used. | |
5465 | if Search(r'\bsprintf\s*\(', line): | |
5466 | error(filename, linenum, 'runtime/printf', 5, | |
5467 | 'Never use sprintf. Use snprintf instead.') | |
5468 | match = Search(r'\b(strcpy|strcat)\s*\(', line) | |
5469 | if match: | |
5470 | error(filename, linenum, 'runtime/printf', 4, | |
5471 | 'Almost always, snprintf is better than %s' % match.group(1)) | |
5472 | ||
5473 | ||
5474 | def IsDerivedFunction(clean_lines, linenum): | |
5475 | """Check if current line contains an inherited function. | |
5476 | ||
5477 | Args: | |
5478 | clean_lines: A CleansedLines instance containing the file. | |
5479 | linenum: The number of the line to check. | |
5480 | Returns: | |
5481 | True if current line contains a function with "override" | |
5482 | virt-specifier. | |
5483 | """ | |
5484 | # Scan back a few lines for start of current function | |
5485 | for i in xrange(linenum, max(-1, linenum - 10), -1): | |
5486 | match = Match(r'^([^()]*\w+)\(', clean_lines.elided[i]) | |
5487 | if match: | |
5488 | # Look for "override" after the matching closing parenthesis | |
5489 | line, _, closing_paren = CloseExpression( | |
5490 | clean_lines, i, len(match.group(1))) | |
5491 | return (closing_paren >= 0 and | |
5492 | Search(r'\boverride\b', line[closing_paren:])) | |
5493 | return False | |
5494 | ||
5495 | ||
5496 | def IsOutOfLineMethodDefinition(clean_lines, linenum): | |
5497 | """Check if current line contains an out-of-line method definition. | |
5498 | ||
5499 | Args: | |
5500 | clean_lines: A CleansedLines instance containing the file. | |
5501 | linenum: The number of the line to check. | |
5502 | Returns: | |
5503 | True if current line contains an out-of-line method definition. | |
5504 | """ | |
5505 | # Scan back a few lines for start of current function | |
5506 | for i in xrange(linenum, max(-1, linenum - 10), -1): | |
5507 | if Match(r'^([^()]*\w+)\(', clean_lines.elided[i]): | |
5508 | return Match(r'^[^()]*\w+::\w+\(', clean_lines.elided[i]) is not None | |
5509 | return False | |
5510 | ||
5511 | ||
5512 | def IsInitializerList(clean_lines, linenum): | |
5513 | """Check if current line is inside constructor initializer list. | |
5514 | ||
5515 | Args: | |
5516 | clean_lines: A CleansedLines instance containing the file. | |
5517 | linenum: The number of the line to check. | |
5518 | Returns: | |
5519 | True if current line appears to be inside constructor initializer | |
5520 | list, False otherwise. | |
5521 | """ | |
5522 | for i in xrange(linenum, 1, -1): | |
5523 | line = clean_lines.elided[i] | |
5524 | if i == linenum: | |
5525 | remove_function_body = Match(r'^(.*)\{\s*$', line) | |
5526 | if remove_function_body: | |
5527 | line = remove_function_body.group(1) | |
5528 | ||
5529 | if Search(r'\s:\s*\w+[({]', line): | |
5530 | # A lone colon tend to indicate the start of a constructor | |
5531 | # initializer list. It could also be a ternary operator, which | |
5532 | # also tend to appear in constructor initializer lists as | |
5533 | # opposed to parameter lists. | |
5534 | return True | |
5535 | if Search(r'\}\s*,\s*$', line): | |
5536 | # A closing brace followed by a comma is probably the end of a | |
5537 | # brace-initialized member in constructor initializer list. | |
5538 | return True | |
5539 | if Search(r'[{};]\s*$', line): | |
5540 | # Found one of the following: | |
5541 | # - A closing brace or semicolon, probably the end of the previous | |
5542 | # function. | |
5543 | # - An opening brace, probably the start of current class or namespace. | |
5544 | # | |
5545 | # Current line is probably not inside an initializer list since | |
5546 | # we saw one of those things without seeing the starting colon. | |
5547 | return False | |
5548 | ||
5549 | # Got to the beginning of the file without seeing the start of | |
5550 | # constructor initializer list. | |
5551 | return False | |
5552 | ||
5553 | ||
5554 | def CheckForNonConstReference(filename, clean_lines, linenum, | |
5555 | nesting_state, error): | |
5556 | """Check for non-const references. | |
5557 | ||
5558 | Separate from CheckLanguage since it scans backwards from current | |
5559 | line, instead of scanning forward. | |
5560 | ||
5561 | Args: | |
5562 | filename: The name of the current file. | |
5563 | clean_lines: A CleansedLines instance containing the file. | |
5564 | linenum: The number of the line to check. | |
5565 | nesting_state: A NestingState instance which maintains information about | |
5566 | the current stack of nested blocks being parsed. | |
5567 | error: The function to call with any errors found. | |
5568 | """ | |
5569 | # Do nothing if there is no '&' on current line. | |
5570 | line = clean_lines.elided[linenum] | |
5571 | if '&' not in line: | |
5572 | return | |
5573 | ||
5574 | # If a function is inherited, current function doesn't have much of | |
5575 | # a choice, so any non-const references should not be blamed on | |
5576 | # derived function. | |
5577 | if IsDerivedFunction(clean_lines, linenum): | |
5578 | return | |
5579 | ||
5580 | # Don't warn on out-of-line method definitions, as we would warn on the | |
5581 | # in-line declaration, if it isn't marked with 'override'. | |
5582 | if IsOutOfLineMethodDefinition(clean_lines, linenum): | |
5583 | return | |
5584 | ||
5585 | # Long type names may be broken across multiple lines, usually in one | |
5586 | # of these forms: | |
5587 | # LongType | |
5588 | # ::LongTypeContinued &identifier | |
5589 | # LongType:: | |
5590 | # LongTypeContinued &identifier | |
5591 | # LongType< | |
5592 | # ...>::LongTypeContinued &identifier | |
5593 | # | |
5594 | # If we detected a type split across two lines, join the previous | |
5595 | # line to current line so that we can match const references | |
5596 | # accordingly. | |
5597 | # | |
5598 | # Note that this only scans back one line, since scanning back | |
5599 | # arbitrary number of lines would be expensive. If you have a type | |
5600 | # that spans more than 2 lines, please use a typedef. | |
5601 | if linenum > 1: | |
5602 | previous = None | |
5603 | if Match(r'\s*::(?:[\w<>]|::)+\s*&\s*\S', line): | |
5604 | # previous_line\n + ::current_line | |
5605 | previous = Search(r'\b((?:const\s*)?(?:[\w<>]|::)+[\w<>])\s*$', | |
5606 | clean_lines.elided[linenum - 1]) | |
5607 | elif Match(r'\s*[a-zA-Z_]([\w<>]|::)+\s*&\s*\S', line): | |
5608 | # previous_line::\n + current_line | |
5609 | previous = Search(r'\b((?:const\s*)?(?:[\w<>]|::)+::)\s*$', | |
5610 | clean_lines.elided[linenum - 1]) | |
5611 | if previous: | |
5612 | line = previous.group(1) + line.lstrip() | |
5613 | else: | |
5614 | # Check for templated parameter that is split across multiple lines | |
5615 | endpos = line.rfind('>') | |
5616 | if endpos > -1: | |
5617 | (_, startline, startpos) = ReverseCloseExpression( | |
5618 | clean_lines, linenum, endpos) | |
5619 | if startpos > -1 and startline < linenum: | |
5620 | # Found the matching < on an earlier line, collect all | |
5621 | # pieces up to current line. | |
5622 | line = '' | |
5623 | for i in xrange(startline, linenum + 1): | |
5624 | line += clean_lines.elided[i].strip() | |
5625 | ||
5626 | # Check for non-const references in function parameters. A single '&' may | |
5627 | # found in the following places: | |
5628 | # inside expression: binary & for bitwise AND | |
5629 | # inside expression: unary & for taking the address of something | |
5630 | # inside declarators: reference parameter | |
5631 | # We will exclude the first two cases by checking that we are not inside a | |
5632 | # function body, including one that was just introduced by a trailing '{'. | |
5633 | # TODO(unknown): Doesn't account for 'catch(Exception& e)' [rare]. | |
5634 | if (nesting_state.previous_stack_top and | |
5635 | not (isinstance(nesting_state.previous_stack_top, _ClassInfo) or | |
5636 | isinstance(nesting_state.previous_stack_top, _NamespaceInfo))): | |
5637 | # Not at toplevel, not within a class, and not within a namespace | |
5638 | return | |
5639 | ||
5640 | # Avoid initializer lists. We only need to scan back from the | |
5641 | # current line for something that starts with ':'. | |
5642 | # | |
5643 | # We don't need to check the current line, since the '&' would | |
5644 | # appear inside the second set of parentheses on the current line as | |
5645 | # opposed to the first set. | |
5646 | if linenum > 0: | |
5647 | for i in xrange(linenum - 1, max(0, linenum - 10), -1): | |
5648 | previous_line = clean_lines.elided[i] | |
5649 | if not Search(r'[),]\s*$', previous_line): | |
5650 | break | |
5651 | if Match(r'^\s*:\s+\S', previous_line): | |
5652 | return | |
5653 | ||
5654 | # Avoid preprocessors | |
5655 | if Search(r'\\\s*$', line): | |
5656 | return | |
5657 | ||
5658 | # Avoid constructor initializer lists | |
5659 | if IsInitializerList(clean_lines, linenum): | |
5660 | return | |
5661 | ||
5662 | # We allow non-const references in a few standard places, like functions | |
5663 | # called "swap()" or iostream operators like "<<" or ">>". Do not check | |
5664 | # those function parameters. | |
5665 | # | |
5666 | # We also accept & in static_assert, which looks like a function but | |
5667 | # it's actually a declaration expression. | |
5668 | allowed_functions = (r'(?:[sS]wap(?:<\w:+>)?|' | |
5669 | r'operator\s*[<>][<>]|' | |
5670 | r'static_assert|COMPILE_ASSERT' | |
5671 | r')\s*\(') | |
5672 | if Search(allowed_functions, line): | |
5673 | return | |
5674 | elif not Search(r'\S+\([^)]*$', line): | |
5675 | # Don't see an allowed function on this line. Actually we | |
5676 | # didn't see any function name on this line, so this is likely a | |
5677 | # multi-line parameter list. Try a bit harder to catch this case. | |
5678 | for i in xrange(2): | |
5679 | if (linenum > i and | |
5680 | Search(allowed_functions, clean_lines.elided[linenum - i - 1])): | |
5681 | return | |
5682 | ||
5683 | decls = ReplaceAll(r'{[^}]*}', ' ', line) # exclude function body | |
5684 | for parameter in re.findall(_RE_PATTERN_REF_PARAM, decls): | |
5685 | if (not Match(_RE_PATTERN_CONST_REF_PARAM, parameter) and | |
5686 | not Match(_RE_PATTERN_REF_STREAM_PARAM, parameter)): | |
5687 | error(filename, linenum, 'runtime/references', 2, | |
5688 | 'Is this a non-const reference? ' | |
5689 | 'If so, make const or use a pointer: ' + | |
5690 | ReplaceAll(' *<', '<', parameter)) | |
5691 | ||
5692 | ||
5693 | def CheckCasts(filename, clean_lines, linenum, error): | |
5694 | """Various cast related checks. | |
5695 | ||
5696 | Args: | |
5697 | filename: The name of the current file. | |
5698 | clean_lines: A CleansedLines instance containing the file. | |
5699 | linenum: The number of the line to check. | |
5700 | error: The function to call with any errors found. | |
5701 | """ | |
5702 | line = clean_lines.elided[linenum] | |
5703 | ||
5704 | # Check to see if they're using an conversion function cast. | |
5705 | # I just try to capture the most common basic types, though there are more. | |
5706 | # Parameterless conversion functions, such as bool(), are allowed as they are | |
5707 | # probably a member operator declaration or default constructor. | |
5708 | match = Search( | |
5709 | r'(\bnew\s+(?:const\s+)?|\S<\s*(?:const\s+)?)?\b' | |
5710 | r'(int|float|double|bool|char|int32|uint32|int64|uint64)' | |
5711 | r'(\([^)].*)', line) | |
5712 | expecting_function = ExpectingFunctionArgs(clean_lines, linenum) | |
5713 | if match and not expecting_function: | |
5714 | matched_type = match.group(2) | |
5715 | ||
5716 | # matched_new_or_template is used to silence two false positives: | |
5717 | # - New operators | |
5718 | # - Template arguments with function types | |
5719 | # | |
5720 | # For template arguments, we match on types immediately following | |
5721 | # an opening bracket without any spaces. This is a fast way to | |
5722 | # silence the common case where the function type is the first | |
5723 | # template argument. False negative with less-than comparison is | |
5724 | # avoided because those operators are usually followed by a space. | |
5725 | # | |
5726 | # function<double(double)> // bracket + no space = false positive | |
5727 | # value < double(42) // bracket + space = true positive | |
5728 | matched_new_or_template = match.group(1) | |
5729 | ||
5730 | # Avoid arrays by looking for brackets that come after the closing | |
5731 | # parenthesis. | |
5732 | if Match(r'\([^()]+\)\s*\[', match.group(3)): | |
5733 | return | |
5734 | ||
5735 | # Other things to ignore: | |
5736 | # - Function pointers | |
5737 | # - Casts to pointer types | |
5738 | # - Placement new | |
5739 | # - Alias declarations | |
5740 | matched_funcptr = match.group(3) | |
5741 | if (matched_new_or_template is None and | |
5742 | not (matched_funcptr and | |
5743 | (Match(r'\((?:[^() ]+::\s*\*\s*)?[^() ]+\)\s*\(', | |
5744 | matched_funcptr) or | |
5745 | matched_funcptr.startswith('(*)'))) and | |
5746 | not Match(r'\s*using\s+\S+\s*=\s*' + matched_type, line) and | |
5747 | not Search(r'new\(\S+\)\s*' + matched_type, line)): | |
5748 | error(filename, linenum, 'readability/casting', 4, | |
5749 | 'Using deprecated casting style. ' | |
5750 | 'Use static_cast<%s>(...) instead' % | |
5751 | matched_type) | |
5752 | ||
5753 | if not expecting_function: | |
5754 | CheckCStyleCast(filename, clean_lines, linenum, 'static_cast', | |
5755 | r'\((int|float|double|bool|char|u?int(16|32|64)|size_t)\)', error) | |
5756 | ||
5757 | # This doesn't catch all cases. Consider (const char * const)"hello". | |
5758 | # | |
5759 | # (char *) "foo" should always be a const_cast (reinterpret_cast won't | |
5760 | # compile). | |
5761 | if CheckCStyleCast(filename, clean_lines, linenum, 'const_cast', | |
5762 | r'\((char\s?\*+\s?)\)\s*"', error): | |
5763 | pass | |
5764 | else: | |
5765 | # Check pointer casts for other than string constants | |
5766 | CheckCStyleCast(filename, clean_lines, linenum, 'reinterpret_cast', | |
5767 | r'\((\w+\s?\*+\s?)\)', error) | |
5768 | ||
5769 | # In addition, we look for people taking the address of a cast. This | |
5770 | # is dangerous -- casts can assign to temporaries, so the pointer doesn't | |
5771 | # point where you think. | |
5772 | # | |
5773 | # Some non-identifier character is required before the '&' for the | |
5774 | # expression to be recognized as a cast. These are casts: | |
5775 | # expression = &static_cast<int*>(temporary()); | |
5776 | # function(&(int*)(temporary())); | |
5777 | # | |
5778 | # This is not a cast: | |
5779 | # reference_type&(int* function_param); | |
5780 | match = Search( | |
5781 | r'(?:[^\w]&\(([^)*][^)]*)\)[\w(])|' | |
5782 | r'(?:[^\w]&(static|dynamic|down|reinterpret)_cast\b)', line) | |
5783 | if match: | |
5784 | # Try a better error message when the & is bound to something | |
5785 | # dereferenced by the casted pointer, as opposed to the casted | |
5786 | # pointer itself. | |
5787 | parenthesis_error = False | |
5788 | match = Match(r'^(.*&(?:static|dynamic|down|reinterpret)_cast\b)<', line) | |
5789 | if match: | |
5790 | _, y1, x1 = CloseExpression(clean_lines, linenum, len(match.group(1))) | |
5791 | if x1 >= 0 and clean_lines.elided[y1][x1] == '(': | |
5792 | _, y2, x2 = CloseExpression(clean_lines, y1, x1) | |
5793 | if x2 >= 0: | |
5794 | extended_line = clean_lines.elided[y2][x2:] | |
5795 | if y2 < clean_lines.NumLines() - 1: | |
5796 | extended_line += clean_lines.elided[y2 + 1] | |
5797 | if Match(r'\s*(?:->|\[)', extended_line): | |
5798 | parenthesis_error = True | |
5799 | ||
5800 | if parenthesis_error: | |
5801 | error(filename, linenum, 'readability/casting', 4, | |
5802 | ('Are you taking an address of something dereferenced ' | |
5803 | 'from a cast? Wrapping the dereferenced expression in ' | |
5804 | 'parentheses will make the binding more obvious')) | |
5805 | else: | |
5806 | error(filename, linenum, 'runtime/casting', 4, | |
5807 | ('Are you taking an address of a cast? ' | |
5808 | 'This is dangerous: could be a temp var. ' | |
5809 | 'Take the address before doing the cast, rather than after')) | |
5810 | ||
5811 | ||
5812 | def CheckCStyleCast(filename, clean_lines, linenum, cast_type, pattern, error): | |
5813 | """Checks for a C-style cast by looking for the pattern. | |
5814 | ||
5815 | Args: | |
5816 | filename: The name of the current file. | |
5817 | clean_lines: A CleansedLines instance containing the file. | |
5818 | linenum: The number of the line to check. | |
5819 | cast_type: The string for the C++ cast to recommend. This is either | |
5820 | reinterpret_cast, static_cast, or const_cast, depending. | |
5821 | pattern: The regular expression used to find C-style casts. | |
5822 | error: The function to call with any errors found. | |
5823 | ||
5824 | Returns: | |
5825 | True if an error was emitted. | |
5826 | False otherwise. | |
5827 | """ | |
5828 | line = clean_lines.elided[linenum] | |
5829 | match = Search(pattern, line) | |
5830 | if not match: | |
5831 | return False | |
5832 | ||
5833 | # Exclude lines with keywords that tend to look like casts | |
5834 | context = line[0:match.start(1) - 1] | |
5835 | if Match(r'.*\b(?:sizeof|alignof|alignas|[_A-Z][_A-Z0-9]*)\s*$', context): | |
5836 | return False | |
5837 | ||
5838 | # Try expanding current context to see if we one level of | |
5839 | # parentheses inside a macro. | |
5840 | if linenum > 0: | |
5841 | for i in xrange(linenum - 1, max(0, linenum - 5), -1): | |
5842 | context = clean_lines.elided[i] + context | |
5843 | if Match(r'.*\b[_A-Z][_A-Z0-9]*\s*\((?:\([^()]*\)|[^()])*$', context): | |
5844 | return False | |
5845 | ||
5846 | # operator++(int) and operator--(int) | |
5847 | if context.endswith(' operator++') or context.endswith(' operator--'): | |
5848 | return False | |
5849 | ||
5850 | # A single unnamed argument for a function tends to look like old style cast. | |
5851 | # If we see those, don't issue warnings for deprecated casts. | |
5852 | remainder = line[match.end(0):] | |
5853 | if Match(r'^\s*(?:;|const\b|throw\b|final\b|override\b|[=>{),]|->)', | |
5854 | remainder): | |
5855 | return False | |
5856 | ||
5857 | # At this point, all that should be left is actual casts. | |
5858 | error(filename, linenum, 'readability/casting', 4, | |
5859 | 'Using C-style cast. Use %s<%s>(...) instead' % | |
5860 | (cast_type, match.group(1))) | |
5861 | ||
5862 | return True | |
5863 | ||
5864 | ||
5865 | def ExpectingFunctionArgs(clean_lines, linenum): | |
5866 | """Checks whether where function type arguments are expected. | |
5867 | ||
5868 | Args: | |
5869 | clean_lines: A CleansedLines instance containing the file. | |
5870 | linenum: The number of the line to check. | |
5871 | ||
5872 | Returns: | |
5873 | True if the line at 'linenum' is inside something that expects arguments | |
5874 | of function types. | |
5875 | """ | |
5876 | line = clean_lines.elided[linenum] | |
5877 | return (Match(r'^\s*MOCK_(CONST_)?METHOD\d+(_T)?\(', line) or | |
5878 | (linenum >= 2 and | |
5879 | (Match(r'^\s*MOCK_(?:CONST_)?METHOD\d+(?:_T)?\((?:\S+,)?\s*$', | |
5880 | clean_lines.elided[linenum - 1]) or | |
5881 | Match(r'^\s*MOCK_(?:CONST_)?METHOD\d+(?:_T)?\(\s*$', | |
5882 | clean_lines.elided[linenum - 2]) or | |
5883 | Search(r'\bstd::m?function\s*\<\s*$', | |
5884 | clean_lines.elided[linenum - 1])))) | |
5885 | ||
5886 | ||
5887 | _HEADERS_CONTAINING_TEMPLATES = ( | |
5888 | ('<deque>', ('deque',)), | |
5889 | ('<functional>', ('unary_function', 'binary_function', | |
5890 | 'plus', 'minus', 'multiplies', 'divides', 'modulus', | |
5891 | 'negate', | |
5892 | 'equal_to', 'not_equal_to', 'greater', 'less', | |
5893 | 'greater_equal', 'less_equal', | |
5894 | 'logical_and', 'logical_or', 'logical_not', | |
5895 | 'unary_negate', 'not1', 'binary_negate', 'not2', | |
5896 | 'bind1st', 'bind2nd', | |
5897 | 'pointer_to_unary_function', | |
5898 | 'pointer_to_binary_function', | |
5899 | 'ptr_fun', | |
5900 | 'mem_fun_t', 'mem_fun', 'mem_fun1_t', 'mem_fun1_ref_t', | |
5901 | 'mem_fun_ref_t', | |
5902 | 'const_mem_fun_t', 'const_mem_fun1_t', | |
5903 | 'const_mem_fun_ref_t', 'const_mem_fun1_ref_t', | |
5904 | 'mem_fun_ref', | |
5905 | )), | |
5906 | ('<limits>', ('numeric_limits',)), | |
5907 | ('<list>', ('list',)), | |
5908 | ('<map>', ('multimap',)), | |
5909 | ('<memory>', ('allocator', 'make_shared', 'make_unique', 'shared_ptr', | |
5910 | 'unique_ptr', 'weak_ptr')), | |
5911 | ('<queue>', ('queue', 'priority_queue',)), | |
5912 | ('<set>', ('multiset',)), | |
5913 | ('<stack>', ('stack',)), | |
5914 | ('<string>', ('char_traits', 'basic_string',)), | |
5915 | ('<tuple>', ('tuple',)), | |
5916 | ('<unordered_map>', ('unordered_map', 'unordered_multimap')), | |
5917 | ('<unordered_set>', ('unordered_set', 'unordered_multiset')), | |
5918 | ('<utility>', ('pair',)), | |
5919 | ('<vector>', ('vector',)), | |
5920 | ||
5921 | # gcc extensions. | |
5922 | # Note: std::hash is their hash, ::hash is our hash | |
5923 | ('<hash_map>', ('hash_map', 'hash_multimap',)), | |
5924 | ('<hash_set>', ('hash_set', 'hash_multiset',)), | |
5925 | ('<slist>', ('slist',)), | |
5926 | ) | |
5927 | ||
5928 | _HEADERS_MAYBE_TEMPLATES = ( | |
5929 | ('<algorithm>', ('copy', 'max', 'min', 'min_element', 'sort', | |
5930 | 'transform', | |
5931 | )), | |
5932 | ('<utility>', ('forward', 'make_pair', 'move', 'swap')), | |
5933 | ) | |
5934 | ||
5935 | _RE_PATTERN_STRING = re.compile(r'\bstring\b') | |
5936 | ||
5937 | _re_pattern_headers_maybe_templates = [] | |
5938 | for _header, _templates in _HEADERS_MAYBE_TEMPLATES: | |
5939 | for _template in _templates: | |
5940 | # Match max<type>(..., ...), max(..., ...), but not foo->max, foo.max or | |
5941 | # 'type::max()'. | |
5942 | _re_pattern_headers_maybe_templates.append( | |
5943 | (re.compile(r'[^>.]\b' + _template + r'(<.*?>)?\([^\)]'), | |
5944 | _template, | |
5945 | _header)) | |
5946 | # Match set<type>, but not foo->set<type>, foo.set<type> | |
5947 | _re_pattern_headers_maybe_templates.append( | |
5948 | (re.compile(r'[^>.]\bset\s*\<'), | |
5949 | 'set<>', | |
5950 | '<set>')) | |
5951 | # Match 'map<type> var' and 'std::map<type>(...)', but not 'map<type>(...)'' | |
5952 | _re_pattern_headers_maybe_templates.append( | |
5953 | (re.compile(r'(std\b::\bmap\s*\<)|(^(std\b::\b)map\b\(\s*\<)'), | |
5954 | 'map<>', | |
5955 | '<map>')) | |
5956 | ||
5957 | # Other scripts may reach in and modify this pattern. | |
5958 | _re_pattern_templates = [] | |
5959 | for _header, _templates in _HEADERS_CONTAINING_TEMPLATES: | |
5960 | for _template in _templates: | |
5961 | _re_pattern_templates.append( | |
5962 | (re.compile(r'(\<|\b)' + _template + r'\s*\<'), | |
5963 | _template + '<>', | |
5964 | _header)) | |
5965 | ||
5966 | ||
5967 | def FilesBelongToSameModule(filename_cc, filename_h): | |
5968 | """Check if these two filenames belong to the same module. | |
5969 | ||
5970 | The concept of a 'module' here is a as follows: | |
5971 | foo.h, foo-inl.h, foo.cc, foo_test.cc and foo_unittest.cc belong to the | |
5972 | same 'module' if they are in the same directory. | |
5973 | some/path/public/xyzzy and some/path/internal/xyzzy are also considered | |
5974 | to belong to the same module here. | |
5975 | ||
5976 | If the filename_cc contains a longer path than the filename_h, for example, | |
5977 | '/absolute/path/to/base/sysinfo.cc', and this file would include | |
5978 | 'base/sysinfo.h', this function also produces the prefix needed to open the | |
5979 | header. This is used by the caller of this function to more robustly open the | |
5980 | header file. We don't have access to the real include paths in this context, | |
5981 | so we need this guesswork here. | |
5982 | ||
5983 | Known bugs: tools/base/bar.cc and base/bar.h belong to the same module | |
5984 | according to this implementation. Because of this, this function gives | |
5985 | some false positives. This should be sufficiently rare in practice. | |
5986 | ||
5987 | Args: | |
5988 | filename_cc: is the path for the source (e.g. .cc) file | |
5989 | filename_h: is the path for the header path | |
5990 | ||
5991 | Returns: | |
5992 | Tuple with a bool and a string: | |
5993 | bool: True if filename_cc and filename_h belong to the same module. | |
5994 | string: the additional prefix needed to open the header file. | |
5995 | """ | |
5996 | fileinfo_cc = FileInfo(filename_cc) | |
5997 | if not fileinfo_cc.Extension().lstrip('.') in GetNonHeaderExtensions(): | |
5998 | return (False, '') | |
5999 | ||
6000 | fileinfo_h = FileInfo(filename_h) | |
6001 | if not IsHeaderExtension(fileinfo_h.Extension().lstrip('.')): | |
6002 | return (False, '') | |
6003 | ||
6004 | filename_cc = filename_cc[:-(len(fileinfo_cc.Extension()))] | |
6005 | matched_test_suffix = Search(_TEST_FILE_SUFFIX, fileinfo_cc.BaseName()) | |
6006 | if matched_test_suffix: | |
6007 | filename_cc = filename_cc[:-len(matched_test_suffix.group(1))] | |
6008 | ||
6009 | filename_cc = filename_cc.replace('/public/', '/') | |
6010 | filename_cc = filename_cc.replace('/internal/', '/') | |
6011 | ||
6012 | filename_h = filename_h[:-(len(fileinfo_h.Extension()))] | |
6013 | if filename_h.endswith('-inl'): | |
6014 | filename_h = filename_h[:-len('-inl')] | |
6015 | filename_h = filename_h.replace('/public/', '/') | |
6016 | filename_h = filename_h.replace('/internal/', '/') | |
6017 | ||
6018 | files_belong_to_same_module = filename_cc.endswith(filename_h) | |
6019 | common_path = '' | |
6020 | if files_belong_to_same_module: | |
6021 | common_path = filename_cc[:-len(filename_h)] | |
6022 | return files_belong_to_same_module, common_path | |
6023 | ||
6024 | ||
6025 | def UpdateIncludeState(filename, include_dict, io=codecs): | |
6026 | """Fill up the include_dict with new includes found from the file. | |
6027 | ||
6028 | Args: | |
6029 | filename: the name of the header to read. | |
6030 | include_dict: a dictionary in which the headers are inserted. | |
6031 | io: The io factory to use to read the file. Provided for testability. | |
6032 | ||
6033 | Returns: | |
6034 | True if a header was successfully added. False otherwise. | |
6035 | """ | |
6036 | headerfile = None | |
6037 | try: | |
6038 | with io.open(filename, 'r', 'utf8', 'replace') as headerfile: | |
6039 | linenum = 0 | |
6040 | for line in headerfile: | |
6041 | linenum += 1 | |
6042 | clean_line = CleanseComments(line) | |
6043 | match = _RE_PATTERN_INCLUDE.search(clean_line) | |
6044 | if match: | |
6045 | include = match.group(2) | |
6046 | include_dict.setdefault(include, linenum) | |
6047 | return True | |
6048 | except IOError: | |
6049 | return False | |
6050 | ||
6051 | ||
6052 | ||
6053 | def CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error, | |
6054 | io=codecs): | |
6055 | """Reports for missing stl includes. | |
6056 | ||
6057 | This function will output warnings to make sure you are including the headers | |
6058 | necessary for the stl containers and functions that you use. We only give one | |
6059 | reason to include a header. For example, if you use both equal_to<> and | |
6060 | less<> in a .h file, only one (the latter in the file) of these will be | |
6061 | reported as a reason to include the <functional>. | |
6062 | ||
6063 | Args: | |
6064 | filename: The name of the current file. | |
6065 | clean_lines: A CleansedLines instance containing the file. | |
6066 | include_state: An _IncludeState instance. | |
6067 | error: The function to call with any errors found. | |
6068 | io: The IO factory to use to read the header file. Provided for unittest | |
6069 | injection. | |
6070 | """ | |
6071 | required = {} # A map of header name to linenumber and the template entity. | |
6072 | # Example of required: { '<functional>': (1219, 'less<>') } | |
6073 | ||
6074 | for linenum in xrange(clean_lines.NumLines()): | |
6075 | line = clean_lines.elided[linenum] | |
6076 | if not line or line[0] == '#': | |
6077 | continue | |
6078 | ||
6079 | # String is special -- it is a non-templatized type in STL. | |
6080 | matched = _RE_PATTERN_STRING.search(line) | |
6081 | if matched: | |
6082 | # Don't warn about strings in non-STL namespaces: | |
6083 | # (We check only the first match per line; good enough.) | |
6084 | prefix = line[:matched.start()] | |
6085 | if prefix.endswith('std::') or not prefix.endswith('::'): | |
6086 | required['<string>'] = (linenum, 'string') | |
6087 | ||
6088 | for pattern, template, header in _re_pattern_headers_maybe_templates: | |
6089 | if pattern.search(line): | |
6090 | required[header] = (linenum, template) | |
6091 | ||
6092 | # The following function is just a speed up, no semantics are changed. | |
6093 | if not '<' in line: # Reduces the cpu time usage by skipping lines. | |
6094 | continue | |
6095 | ||
6096 | for pattern, template, header in _re_pattern_templates: | |
6097 | matched = pattern.search(line) | |
6098 | if matched: | |
6099 | # Don't warn about IWYU in non-STL namespaces: | |
6100 | # (We check only the first match per line; good enough.) | |
6101 | prefix = line[:matched.start()] | |
6102 | if prefix.endswith('std::') or not prefix.endswith('::'): | |
6103 | required[header] = (linenum, template) | |
6104 | ||
6105 | # The policy is that if you #include something in foo.h you don't need to | |
6106 | # include it again in foo.cc. Here, we will look at possible includes. | |
6107 | # Let's flatten the include_state include_list and copy it into a dictionary. | |
6108 | include_dict = dict([item for sublist in include_state.include_list | |
6109 | for item in sublist]) | |
6110 | ||
6111 | # Did we find the header for this file (if any) and successfully load it? | |
6112 | header_found = False | |
6113 | ||
6114 | # Use the absolute path so that matching works properly. | |
6115 | abs_filename = FileInfo(filename).FullName() | |
6116 | ||
6117 | # For Emacs's flymake. | |
6118 | # If cpplint is invoked from Emacs's flymake, a temporary file is generated | |
6119 | # by flymake and that file name might end with '_flymake.cc'. In that case, | |
6120 | # restore original file name here so that the corresponding header file can be | |
6121 | # found. | |
6122 | # e.g. If the file name is 'foo_flymake.cc', we should search for 'foo.h' | |
6123 | # instead of 'foo_flymake.h' | |
6124 | abs_filename = re.sub(r'_flymake\.cc$', '.cc', abs_filename) | |
6125 | ||
6126 | # include_dict is modified during iteration, so we iterate over a copy of | |
6127 | # the keys. | |
6128 | header_keys = list(include_dict.keys()) | |
6129 | for header in header_keys: | |
6130 | (same_module, common_path) = FilesBelongToSameModule(abs_filename, header) | |
6131 | fullpath = common_path + header | |
6132 | if same_module and UpdateIncludeState(fullpath, include_dict, io): | |
6133 | header_found = True | |
6134 | ||
6135 | # If we can't find the header file for a .cc, assume it's because we don't | |
6136 | # know where to look. In that case we'll give up as we're not sure they | |
6137 | # didn't include it in the .h file. | |
6138 | # TODO(unknown): Do a better job of finding .h files so we are confident that | |
6139 | # not having the .h file means there isn't one. | |
6140 | if not header_found: | |
6141 | for extension in GetNonHeaderExtensions(): | |
6142 | if filename.endswith('.' + extension): | |
6143 | return | |
6144 | ||
6145 | # All the lines have been processed, report the errors found. | |
6146 | for required_header_unstripped in sorted(required, key=required.__getitem__): | |
6147 | template = required[required_header_unstripped][1] | |
6148 | if required_header_unstripped.strip('<>"') not in include_dict: | |
6149 | error(filename, required[required_header_unstripped][0], | |
6150 | 'build/include_what_you_use', 4, | |
6151 | 'Add #include ' + required_header_unstripped + ' for ' + template) | |
6152 | ||
6153 | ||
6154 | _RE_PATTERN_EXPLICIT_MAKEPAIR = re.compile(r'\bmake_pair\s*<') | |
6155 | ||
6156 | ||
6157 | def CheckMakePairUsesDeduction(filename, clean_lines, linenum, error): | |
6158 | """Check that make_pair's template arguments are deduced. | |
6159 | ||
6160 | G++ 4.6 in C++11 mode fails badly if make_pair's template arguments are | |
6161 | specified explicitly, and such use isn't intended in any case. | |
6162 | ||
6163 | Args: | |
6164 | filename: The name of the current file. | |
6165 | clean_lines: A CleansedLines instance containing the file. | |
6166 | linenum: The number of the line to check. | |
6167 | error: The function to call with any errors found. | |
6168 | """ | |
6169 | line = clean_lines.elided[linenum] | |
6170 | match = _RE_PATTERN_EXPLICIT_MAKEPAIR.search(line) | |
6171 | if match: | |
6172 | error(filename, linenum, 'build/explicit_make_pair', | |
6173 | 4, # 4 = high confidence | |
6174 | 'For C++11-compatibility, omit template arguments from make_pair' | |
6175 | ' OR use pair directly OR if appropriate, construct a pair directly') | |
6176 | ||
6177 | ||
6178 | def CheckRedundantVirtual(filename, clean_lines, linenum, error): | |
6179 | """Check if line contains a redundant "virtual" function-specifier. | |
6180 | ||
6181 | Args: | |
6182 | filename: The name of the current file. | |
6183 | clean_lines: A CleansedLines instance containing the file. | |
6184 | linenum: The number of the line to check. | |
6185 | error: The function to call with any errors found. | |
6186 | """ | |
6187 | # Look for "virtual" on current line. | |
6188 | line = clean_lines.elided[linenum] | |
6189 | virtual = Match(r'^(.*)(\bvirtual\b)(.*)$', line) | |
6190 | if not virtual: return | |
6191 | ||
6192 | # Ignore "virtual" keywords that are near access-specifiers. These | |
6193 | # are only used in class base-specifier and do not apply to member | |
6194 | # functions. | |
6195 | if (Search(r'\b(public|protected|private)\s+$', virtual.group(1)) or | |
6196 | Match(r'^\s+(public|protected|private)\b', virtual.group(3))): | |
6197 | return | |
6198 | ||
6199 | # Ignore the "virtual" keyword from virtual base classes. Usually | |
6200 | # there is a column on the same line in these cases (virtual base | |
6201 | # classes are rare in google3 because multiple inheritance is rare). | |
6202 | if Match(r'^.*[^:]:[^:].*$', line): return | |
6203 | ||
6204 | # Look for the next opening parenthesis. This is the start of the | |
6205 | # parameter list (possibly on the next line shortly after virtual). | |
6206 | # TODO(unknown): doesn't work if there are virtual functions with | |
6207 | # decltype() or other things that use parentheses, but csearch suggests | |
6208 | # that this is rare. | |
6209 | end_col = -1 | |
6210 | end_line = -1 | |
6211 | start_col = len(virtual.group(2)) | |
6212 | for start_line in xrange(linenum, min(linenum + 3, clean_lines.NumLines())): | |
6213 | line = clean_lines.elided[start_line][start_col:] | |
6214 | parameter_list = Match(r'^([^(]*)\(', line) | |
6215 | if parameter_list: | |
6216 | # Match parentheses to find the end of the parameter list | |
6217 | (_, end_line, end_col) = CloseExpression( | |
6218 | clean_lines, start_line, start_col + len(parameter_list.group(1))) | |
6219 | break | |
6220 | start_col = 0 | |
6221 | ||
6222 | if end_col < 0: | |
6223 | return # Couldn't find end of parameter list, give up | |
6224 | ||
6225 | # Look for "override" or "final" after the parameter list | |
6226 | # (possibly on the next few lines). | |
6227 | for i in xrange(end_line, min(end_line + 3, clean_lines.NumLines())): | |
6228 | line = clean_lines.elided[i][end_col:] | |
6229 | match = Search(r'\b(override|final)\b', line) | |
6230 | if match: | |
6231 | error(filename, linenum, 'readability/inheritance', 4, | |
6232 | ('"virtual" is redundant since function is ' | |
6233 | 'already declared as "%s"' % match.group(1))) | |
6234 | ||
6235 | # Set end_col to check whole lines after we are done with the | |
6236 | # first line. | |
6237 | end_col = 0 | |
6238 | if Search(r'[^\w]\s*$', line): | |
6239 | break | |
6240 | ||
6241 | ||
6242 | def CheckRedundantOverrideOrFinal(filename, clean_lines, linenum, error): | |
6243 | """Check if line contains a redundant "override" or "final" virt-specifier. | |
6244 | ||
6245 | Args: | |
6246 | filename: The name of the current file. | |
6247 | clean_lines: A CleansedLines instance containing the file. | |
6248 | linenum: The number of the line to check. | |
6249 | error: The function to call with any errors found. | |
6250 | """ | |
6251 | # Look for closing parenthesis nearby. We need one to confirm where | |
6252 | # the declarator ends and where the virt-specifier starts to avoid | |
6253 | # false positives. | |
6254 | line = clean_lines.elided[linenum] | |
6255 | declarator_end = line.rfind(')') | |
6256 | if declarator_end >= 0: | |
6257 | fragment = line[declarator_end:] | |
6258 | else: | |
6259 | if linenum > 1 and clean_lines.elided[linenum - 1].rfind(')') >= 0: | |
6260 | fragment = line | |
6261 | else: | |
6262 | return | |
6263 | ||
6264 | # Check that at most one of "override" or "final" is present, not both | |
6265 | if Search(r'\boverride\b', fragment) and Search(r'\bfinal\b', fragment): | |
6266 | error(filename, linenum, 'readability/inheritance', 4, | |
6267 | ('"override" is redundant since function is ' | |
6268 | 'already declared as "final"')) | |
6269 | ||
6270 | ||
6271 | ||
6272 | ||
6273 | # Returns true if we are at a new block, and it is directly | |
6274 | # inside of a namespace. | |
6275 | def IsBlockInNameSpace(nesting_state, is_forward_declaration): | |
6276 | """Checks that the new block is directly in a namespace. | |
6277 | ||
6278 | Args: | |
6279 | nesting_state: The _NestingState object that contains info about our state. | |
6280 | is_forward_declaration: If the class is a forward declared class. | |
6281 | Returns: | |
6282 | Whether or not the new block is directly in a namespace. | |
6283 | """ | |
6284 | if is_forward_declaration: | |
6285 | return len(nesting_state.stack) >= 1 and ( | |
6286 | isinstance(nesting_state.stack[-1], _NamespaceInfo)) | |
6287 | ||
6288 | ||
6289 | return (len(nesting_state.stack) > 1 and | |
6290 | nesting_state.stack[-1].check_namespace_indentation and | |
6291 | isinstance(nesting_state.stack[-2], _NamespaceInfo)) | |
6292 | ||
6293 | ||
6294 | def ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item, | |
6295 | raw_lines_no_comments, linenum): | |
6296 | """This method determines if we should apply our namespace indentation check. | |
6297 | ||
6298 | Args: | |
6299 | nesting_state: The current nesting state. | |
6300 | is_namespace_indent_item: If we just put a new class on the stack, True. | |
6301 | If the top of the stack is not a class, or we did not recently | |
6302 | add the class, False. | |
6303 | raw_lines_no_comments: The lines without the comments. | |
6304 | linenum: The current line number we are processing. | |
6305 | ||
6306 | Returns: | |
6307 | True if we should apply our namespace indentation check. Currently, it | |
6308 | only works for classes and namespaces inside of a namespace. | |
6309 | """ | |
6310 | ||
6311 | is_forward_declaration = IsForwardClassDeclaration(raw_lines_no_comments, | |
6312 | linenum) | |
6313 | ||
6314 | if not (is_namespace_indent_item or is_forward_declaration): | |
6315 | return False | |
6316 | ||
6317 | # If we are in a macro, we do not want to check the namespace indentation. | |
6318 | if IsMacroDefinition(raw_lines_no_comments, linenum): | |
6319 | return False | |
6320 | ||
6321 | return IsBlockInNameSpace(nesting_state, is_forward_declaration) | |
6322 | ||
6323 | ||
6324 | # Call this method if the line is directly inside of a namespace. | |
6325 | # If the line above is blank (excluding comments) or the start of | |
6326 | # an inner namespace, it cannot be indented. | |
6327 | def CheckItemIndentationInNamespace(filename, raw_lines_no_comments, linenum, | |
6328 | error): | |
6329 | line = raw_lines_no_comments[linenum] | |
6330 | if Match(r'^\s+', line): | |
6331 | error(filename, linenum, 'runtime/indentation_namespace', 4, | |
6332 | 'Do not indent within a namespace') | |
6333 | ||
6334 | ||
6335 | def ProcessLine(filename, file_extension, clean_lines, line, | |
6336 | include_state, function_state, nesting_state, error, | |
6337 | extra_check_functions=None): | |
6338 | """Processes a single line in the file. | |
6339 | ||
6340 | Args: | |
6341 | filename: Filename of the file that is being processed. | |
6342 | file_extension: The extension (dot not included) of the file. | |
6343 | clean_lines: An array of strings, each representing a line of the file, | |
6344 | with comments stripped. | |
6345 | line: Number of line being processed. | |
6346 | include_state: An _IncludeState instance in which the headers are inserted. | |
6347 | function_state: A _FunctionState instance which counts function lines, etc. | |
6348 | nesting_state: A NestingState instance which maintains information about | |
6349 | the current stack of nested blocks being parsed. | |
6350 | error: A callable to which errors are reported, which takes 4 arguments: | |
6351 | filename, line number, error level, and message | |
6352 | extra_check_functions: An array of additional check functions that will be | |
6353 | run on each source line. Each function takes 4 | |
6354 | arguments: filename, clean_lines, line, error | |
6355 | """ | |
6356 | raw_lines = clean_lines.raw_lines | |
6357 | ParseNolintSuppressions(filename, raw_lines[line], line, error) | |
6358 | nesting_state.Update(filename, clean_lines, line, error) | |
6359 | CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line, | |
6360 | error) | |
6361 | if nesting_state.InAsmBlock(): return | |
6362 | CheckForFunctionLengths(filename, clean_lines, line, function_state, error) | |
6363 | CheckForMultilineCommentsAndStrings(filename, clean_lines, line, error) | |
6364 | CheckStyle(filename, clean_lines, line, file_extension, nesting_state, error) | |
6365 | CheckLanguage(filename, clean_lines, line, file_extension, include_state, | |
6366 | nesting_state, error) | |
6367 | CheckForNonConstReference(filename, clean_lines, line, nesting_state, error) | |
6368 | CheckForNonStandardConstructs(filename, clean_lines, line, | |
6369 | nesting_state, error) | |
6370 | CheckVlogArguments(filename, clean_lines, line, error) | |
6371 | CheckPosixThreading(filename, clean_lines, line, error) | |
6372 | CheckInvalidIncrement(filename, clean_lines, line, error) | |
6373 | CheckMakePairUsesDeduction(filename, clean_lines, line, error) | |
6374 | CheckRedundantVirtual(filename, clean_lines, line, error) | |
6375 | CheckRedundantOverrideOrFinal(filename, clean_lines, line, error) | |
6376 | if extra_check_functions: | |
6377 | for check_fn in extra_check_functions: | |
6378 | check_fn(filename, clean_lines, line, error) | |
6379 | ||
6380 | def FlagCxx11Features(filename, clean_lines, linenum, error): | |
6381 | """Flag those c++11 features that we only allow in certain places. | |
6382 | ||
6383 | Args: | |
6384 | filename: The name of the current file. | |
6385 | clean_lines: A CleansedLines instance containing the file. | |
6386 | linenum: The number of the line to check. | |
6387 | error: The function to call with any errors found. | |
6388 | """ | |
6389 | line = clean_lines.elided[linenum] | |
6390 | ||
6391 | include = Match(r'\s*#\s*include\s+[<"]([^<"]+)[">]', line) | |
6392 | ||
6393 | # Flag unapproved C++ TR1 headers. | |
6394 | if include and include.group(1).startswith('tr1/'): | |
6395 | error(filename, linenum, 'build/c++tr1', 5, | |
6396 | ('C++ TR1 headers such as <%s> are unapproved.') % include.group(1)) | |
6397 | ||
6398 | # Flag unapproved C++11 headers. | |
6399 | if include and include.group(1) in ('cfenv', | |
6400 | 'condition_variable', | |
6401 | 'fenv.h', | |
6402 | 'future', | |
6403 | 'mutex', | |
6404 | 'thread', | |
6405 | 'chrono', | |
6406 | 'ratio', | |
6407 | 'regex', | |
6408 | 'system_error', | |
6409 | ): | |
6410 | error(filename, linenum, 'build/c++11', 5, | |
6411 | ('<%s> is an unapproved C++11 header.') % include.group(1)) | |
6412 | ||
6413 | # The only place where we need to worry about C++11 keywords and library | |
6414 | # features in preprocessor directives is in macro definitions. | |
6415 | if Match(r'\s*#', line) and not Match(r'\s*#\s*define\b', line): return | |
6416 | ||
6417 | # These are classes and free functions. The classes are always | |
6418 | # mentioned as std::*, but we only catch the free functions if | |
6419 | # they're not found by ADL. They're alphabetical by header. | |
6420 | for top_name in ( | |
6421 | # type_traits | |
6422 | 'alignment_of', | |
6423 | 'aligned_union', | |
6424 | ): | |
6425 | if Search(r'\bstd::%s\b' % top_name, line): | |
6426 | error(filename, linenum, 'build/c++11', 5, | |
6427 | ('std::%s is an unapproved C++11 class or function. Send c-style ' | |
6428 | 'an example of where it would make your code more readable, and ' | |
6429 | 'they may let you use it.') % top_name) | |
6430 | ||
6431 | ||
6432 | def FlagCxx14Features(filename, clean_lines, linenum, error): | |
6433 | """Flag those C++14 features that we restrict. | |
6434 | ||
6435 | Args: | |
6436 | filename: The name of the current file. | |
6437 | clean_lines: A CleansedLines instance containing the file. | |
6438 | linenum: The number of the line to check. | |
6439 | error: The function to call with any errors found. | |
6440 | """ | |
6441 | line = clean_lines.elided[linenum] | |
6442 | ||
6443 | include = Match(r'\s*#\s*include\s+[<"]([^<"]+)[">]', line) | |
6444 | ||
6445 | # Flag unapproved C++14 headers. | |
6446 | if include and include.group(1) in ('scoped_allocator', 'shared_mutex'): | |
6447 | error(filename, linenum, 'build/c++14', 5, | |
6448 | ('<%s> is an unapproved C++14 header.') % include.group(1)) | |
6449 | ||
6450 | ||
6451 | def ProcessFileData(filename, file_extension, lines, error, | |
6452 | extra_check_functions=None): | |
6453 | """Performs lint checks and reports any errors to the given error function. | |
6454 | ||
6455 | Args: | |
6456 | filename: Filename of the file that is being processed. | |
6457 | file_extension: The extension (dot not included) of the file. | |
6458 | lines: An array of strings, each representing a line of the file, with the | |
6459 | last element being empty if the file is terminated with a newline. | |
6460 | error: A callable to which errors are reported, which takes 4 arguments: | |
6461 | filename, line number, error level, and message | |
6462 | extra_check_functions: An array of additional check functions that will be | |
6463 | run on each source line. Each function takes 4 | |
6464 | arguments: filename, clean_lines, line, error | |
6465 | """ | |
6466 | lines = (['// marker so line numbers and indices both start at 1'] + lines + | |
6467 | ['// marker so line numbers end in a known way']) | |
6468 | ||
6469 | include_state = _IncludeState() | |
6470 | function_state = _FunctionState() | |
6471 | nesting_state = NestingState() | |
6472 | ||
6473 | ResetNolintSuppressions() | |
6474 | ||
6475 | CheckForCopyright(filename, lines, error) | |
6476 | ProcessGlobalSuppresions(lines) | |
6477 | RemoveMultiLineComments(filename, lines, error) | |
6478 | clean_lines = CleansedLines(lines) | |
6479 | ||
6480 | if IsHeaderExtension(file_extension): | |
6481 | CheckForHeaderGuard(filename, clean_lines, error) | |
6482 | ||
6483 | for line in xrange(clean_lines.NumLines()): | |
6484 | ProcessLine(filename, file_extension, clean_lines, line, | |
6485 | include_state, function_state, nesting_state, error, | |
6486 | extra_check_functions) | |
6487 | FlagCxx11Features(filename, clean_lines, line, error) | |
6488 | nesting_state.CheckCompletedBlocks(filename, error) | |
6489 | ||
6490 | CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error) | |
6491 | ||
6492 | # Check that the .cc file has included its header if it exists. | |
6493 | if _IsSourceExtension(file_extension): | |
6494 | CheckHeaderFileIncluded(filename, include_state, error) | |
6495 | ||
6496 | # We check here rather than inside ProcessLine so that we see raw | |
6497 | # lines rather than "cleaned" lines. | |
6498 | CheckForBadCharacters(filename, lines, error) | |
6499 | ||
6500 | CheckForNewlineAtEOF(filename, lines, error) | |
6501 | ||
6502 | def ProcessConfigOverrides(filename): | |
6503 | """ Loads the configuration files and processes the config overrides. | |
6504 | ||
6505 | Args: | |
6506 | filename: The name of the file being processed by the linter. | |
6507 | ||
6508 | Returns: | |
6509 | False if the current |filename| should not be processed further. | |
6510 | """ | |
6511 | ||
6512 | abs_filename = os.path.abspath(filename) | |
6513 | cfg_filters = [] | |
6514 | keep_looking = True | |
6515 | while keep_looking: | |
6516 | abs_path, base_name = os.path.split(abs_filename) | |
6517 | if not base_name: | |
6518 | break # Reached the root directory. | |
6519 | ||
6520 | cfg_file = os.path.join(abs_path, "CPPLINT.cfg") | |
6521 | abs_filename = abs_path | |
6522 | if not os.path.isfile(cfg_file): | |
6523 | continue | |
6524 | ||
6525 | try: | |
6526 | with open(cfg_file) as file_handle: | |
6527 | for line in file_handle: | |
6528 | line, _, _ = line.partition('#') # Remove comments. | |
6529 | if not line.strip(): | |
6530 | continue | |
6531 | ||
6532 | name, _, val = line.partition('=') | |
6533 | name = name.strip() | |
6534 | val = val.strip() | |
6535 | if name == 'set noparent': | |
6536 | keep_looking = False | |
6537 | elif name == 'filter': | |
6538 | cfg_filters.append(val) | |
6539 | elif name == 'exclude_files': | |
6540 | # When matching exclude_files pattern, use the base_name of | |
6541 | # the current file name or the directory name we are processing. | |
6542 | # For example, if we are checking for lint errors in /foo/bar/baz.cc | |
6543 | # and we found the .cfg file at /foo/CPPLINT.cfg, then the config | |
6544 | # file's "exclude_files" filter is meant to be checked against "bar" | |
6545 | # and not "baz" nor "bar/baz.cc". | |
6546 | if base_name: | |
6547 | pattern = re.compile(val) | |
6548 | if pattern.match(base_name): | |
6549 | if _cpplint_state.quiet: | |
6550 | # Suppress "Ignoring file" warning when using --quiet. | |
6551 | return False | |
6552 | _cpplint_state.PrintInfo('Ignoring "%s": file excluded by "%s". ' | |
6553 | 'File path component "%s" matches ' | |
6554 | 'pattern "%s"\n' % | |
6555 | (filename, cfg_file, base_name, val)) | |
6556 | return False | |
6557 | elif name == 'linelength': | |
6558 | global _line_length | |
6559 | try: | |
6560 | _line_length = int(val) | |
6561 | except ValueError: | |
6562 | _cpplint_state.PrintError('Line length must be numeric.') | |
6563 | elif name == 'extensions': | |
6564 | ProcessExtensionsOption(val) | |
6565 | elif name == 'root': | |
6566 | global _root | |
6567 | # root directories are specified relative to CPPLINT.cfg dir. | |
6568 | _root = os.path.join(os.path.dirname(cfg_file), val) | |
6569 | elif name == 'headers': | |
6570 | ProcessHppHeadersOption(val) | |
6571 | elif name == 'includeorder': | |
6572 | ProcessIncludeOrderOption(val) | |
6573 | else: | |
6574 | _cpplint_state.PrintError( | |
6575 | 'Invalid configuration option (%s) in file %s\n' % | |
6576 | (name, cfg_file)) | |
6577 | ||
6578 | except IOError: | |
6579 | _cpplint_state.PrintError( | |
6580 | "Skipping config file '%s': Can't open for reading\n" % cfg_file) | |
6581 | keep_looking = False | |
6582 | ||
6583 | # Apply all the accumulated filters in reverse order (top-level directory | |
6584 | # config options having the least priority). | |
6585 | for cfg_filter in reversed(cfg_filters): | |
6586 | _AddFilters(cfg_filter) | |
6587 | ||
6588 | return True | |
6589 | ||
6590 | ||
6591 | def ProcessFile(filename, vlevel, extra_check_functions=None): | |
6592 | """Does google-lint on a single file. | |
6593 | ||
6594 | Args: | |
6595 | filename: The name of the file to parse. | |
6596 | ||
6597 | vlevel: The level of errors to report. Every error of confidence | |
6598 | >= verbose_level will be reported. 0 is a good default. | |
6599 | ||
6600 | extra_check_functions: An array of additional check functions that will be | |
6601 | run on each source line. Each function takes 4 | |
6602 | arguments: filename, clean_lines, line, error | |
6603 | """ | |
6604 | ||
6605 | _SetVerboseLevel(vlevel) | |
6606 | _BackupFilters() | |
6607 | old_errors = _cpplint_state.error_count | |
6608 | ||
6609 | if not ProcessConfigOverrides(filename): | |
6610 | _RestoreFilters() | |
6611 | return | |
6612 | ||
6613 | lf_lines = [] | |
6614 | crlf_lines = [] | |
6615 | try: | |
6616 | # Support the UNIX convention of using "-" for stdin. Note that | |
6617 | # we are not opening the file with universal newline support | |
6618 | # (which codecs doesn't support anyway), so the resulting lines do | |
6619 | # contain trailing '\r' characters if we are reading a file that | |
6620 | # has CRLF endings. | |
6621 | # If after the split a trailing '\r' is present, it is removed | |
6622 | # below. | |
6623 | if filename == '-': | |
6624 | lines = codecs.StreamReaderWriter(sys.stdin, | |
6625 | codecs.getreader('utf8'), | |
6626 | codecs.getwriter('utf8'), | |
6627 | 'replace').read().split('\n') | |
6628 | else: | |
6629 | with codecs.open(filename, 'r', 'utf8', 'replace') as target_file: | |
6630 | lines = target_file.read().split('\n') | |
6631 | ||
6632 | # Remove trailing '\r'. | |
6633 | # The -1 accounts for the extra trailing blank line we get from split() | |
6634 | for linenum in range(len(lines) - 1): | |
6635 | if lines[linenum].endswith('\r'): | |
6636 | lines[linenum] = lines[linenum].rstrip('\r') | |
6637 | crlf_lines.append(linenum + 1) | |
6638 | else: | |
6639 | lf_lines.append(linenum + 1) | |
6640 | ||
6641 | except IOError: | |
6642 | _cpplint_state.PrintError( | |
6643 | "Skipping input '%s': Can't open for reading\n" % filename) | |
6644 | _RestoreFilters() | |
6645 | return | |
6646 | ||
6647 | # Note, if no dot is found, this will give the entire filename as the ext. | |
6648 | file_extension = filename[filename.rfind('.') + 1:] | |
6649 | ||
6650 | # When reading from stdin, the extension is unknown, so no cpplint tests | |
6651 | # should rely on the extension. | |
6652 | if filename != '-' and file_extension not in GetAllExtensions(): | |
6653 | _cpplint_state.PrintError('Ignoring %s; not a valid file name ' | |
6654 | '(%s)\n' % (filename, ', '.join(GetAllExtensions()))) | |
6655 | else: | |
6656 | ProcessFileData(filename, file_extension, lines, Error, | |
6657 | extra_check_functions) | |
6658 | ||
6659 | # If end-of-line sequences are a mix of LF and CR-LF, issue | |
6660 | # warnings on the lines with CR. | |
6661 | # | |
6662 | # Don't issue any warnings if all lines are uniformly LF or CR-LF, | |
6663 | # since critique can handle these just fine, and the style guide | |
6664 | # doesn't dictate a particular end of line sequence. | |
6665 | # | |
6666 | # We can't depend on os.linesep to determine what the desired | |
6667 | # end-of-line sequence should be, since that will return the | |
6668 | # server-side end-of-line sequence. | |
6669 | if lf_lines and crlf_lines: | |
6670 | # Warn on every line with CR. An alternative approach might be to | |
6671 | # check whether the file is mostly CRLF or just LF, and warn on the | |
6672 | # minority, we bias toward LF here since most tools prefer LF. | |
6673 | for linenum in crlf_lines: | |
6674 | Error(filename, linenum, 'whitespace/newline', 1, | |
6675 | 'Unexpected \\r (^M) found; better to use only \\n') | |
6676 | ||
6677 | # Suppress printing anything if --quiet was passed unless the error | |
6678 | # count has increased after processing this file. | |
6679 | if not _cpplint_state.quiet or old_errors != _cpplint_state.error_count: | |
6680 | _cpplint_state.PrintInfo('Done processing %s\n' % filename) | |
6681 | _RestoreFilters() | |
6682 | ||
6683 | ||
6684 | def PrintUsage(message): | |
6685 | """Prints a brief usage string and exits, optionally with an error message. | |
6686 | ||
6687 | Args: | |
6688 | message: The optional error message. | |
6689 | """ | |
6690 | sys.stderr.write(_USAGE % (sorted(list(GetAllExtensions())), | |
6691 | ','.join(sorted(list(GetAllExtensions()))), | |
6692 | sorted(GetHeaderExtensions()), | |
6693 | ','.join(sorted(GetHeaderExtensions())))) | |
6694 | ||
6695 | if message: | |
6696 | sys.exit('\nFATAL ERROR: ' + message) | |
6697 | else: | |
6698 | sys.exit(0) | |
6699 | ||
6700 | def PrintVersion(): | |
6701 | sys.stdout.write('Cpplint fork (https://github.com/cpplint/cpplint)\n') | |
6702 | sys.stdout.write('cpplint ' + __VERSION__ + '\n') | |
6703 | sys.stdout.write('Python ' + sys.version + '\n') | |
6704 | sys.exit(0) | |
6705 | ||
6706 | def PrintCategories(): | |
6707 | """Prints a list of all the error-categories used by error messages. | |
6708 | ||
6709 | These are the categories used to filter messages via --filter. | |
6710 | """ | |
6711 | sys.stderr.write(''.join(' %s\n' % cat for cat in _ERROR_CATEGORIES)) | |
6712 | sys.exit(0) | |
6713 | ||
6714 | ||
6715 | def ParseArguments(args): | |
6716 | """Parses the command line arguments. | |
6717 | ||
6718 | This may set the output format and verbosity level as side-effects. | |
6719 | ||
6720 | Args: | |
6721 | args: The command line arguments: | |
6722 | ||
6723 | Returns: | |
6724 | The list of filenames to lint. | |
6725 | """ | |
6726 | try: | |
6727 | (opts, filenames) = getopt.getopt(args, '', ['help', 'output=', 'verbose=', | |
6728 | 'v=', | |
6729 | 'version', | |
6730 | 'counting=', | |
6731 | 'filter=', | |
6732 | 'root=', | |
6733 | 'repository=', | |
6734 | 'linelength=', | |
6735 | 'extensions=', | |
6736 | 'exclude=', | |
6737 | 'recursive', | |
6738 | 'headers=', | |
6739 | 'includeorder=', | |
6740 | 'quiet']) | |
6741 | except getopt.GetoptError: | |
6742 | PrintUsage('Invalid arguments.') | |
6743 | ||
6744 | verbosity = _VerboseLevel() | |
6745 | output_format = _OutputFormat() | |
6746 | filters = '' | |
6747 | quiet = _Quiet() | |
6748 | counting_style = '' | |
6749 | recursive = False | |
6750 | ||
6751 | for (opt, val) in opts: | |
6752 | if opt == '--help': | |
6753 | PrintUsage(None) | |
6754 | if opt == '--version': | |
6755 | PrintVersion() | |
6756 | elif opt == '--output': | |
6757 | if val not in ('emacs', 'vs7', 'eclipse', 'junit', 'sed', 'gsed'): | |
6758 | PrintUsage('The only allowed output formats are emacs, vs7, eclipse ' | |
6759 | 'sed, gsed and junit.') | |
6760 | output_format = val | |
6761 | elif opt == '--quiet': | |
6762 | quiet = True | |
6763 | elif opt == '--verbose' or opt == '--v': | |
6764 | verbosity = int(val) | |
6765 | elif opt == '--filter': | |
6766 | filters = val | |
6767 | if not filters: | |
6768 | PrintCategories() | |
6769 | elif opt == '--counting': | |
6770 | if val not in ('total', 'toplevel', 'detailed'): | |
6771 | PrintUsage('Valid counting options are total, toplevel, and detailed') | |
6772 | counting_style = val | |
6773 | elif opt == '--root': | |
6774 | global _root | |
6775 | _root = val | |
6776 | elif opt == '--repository': | |
6777 | global _repository | |
6778 | _repository = val | |
6779 | elif opt == '--linelength': | |
6780 | global _line_length | |
6781 | try: | |
6782 | _line_length = int(val) | |
6783 | except ValueError: | |
6784 | PrintUsage('Line length must be digits.') | |
6785 | elif opt == '--exclude': | |
6786 | global _excludes | |
6787 | if not _excludes: | |
6788 | _excludes = set() | |
6789 | _excludes.update(glob.glob(val)) | |
6790 | elif opt == '--extensions': | |
6791 | ProcessExtensionsOption(val) | |
6792 | elif opt == '--headers': | |
6793 | ProcessHppHeadersOption(val) | |
6794 | elif opt == '--recursive': | |
6795 | recursive = True | |
6796 | elif opt == '--includeorder': | |
6797 | ProcessIncludeOrderOption(val) | |
6798 | ||
6799 | if not filenames: | |
6800 | PrintUsage('No files were specified.') | |
6801 | ||
6802 | if recursive: | |
6803 | filenames = _ExpandDirectories(filenames) | |
6804 | ||
6805 | if _excludes: | |
6806 | filenames = _FilterExcludedFiles(filenames) | |
6807 | ||
6808 | _SetOutputFormat(output_format) | |
6809 | _SetQuiet(quiet) | |
6810 | _SetVerboseLevel(verbosity) | |
6811 | _SetFilters(filters) | |
6812 | _SetCountingStyle(counting_style) | |
6813 | ||
6814 | filenames.sort() | |
6815 | return filenames | |
6816 | ||
6817 | def _ExpandDirectories(filenames): | |
6818 | """Searches a list of filenames and replaces directories in the list with | |
6819 | all files descending from those directories. Files with extensions not in | |
6820 | the valid extensions list are excluded. | |
6821 | ||
6822 | Args: | |
6823 | filenames: A list of files or directories | |
6824 | ||
6825 | Returns: | |
6826 | A list of all files that are members of filenames or descended from a | |
6827 | directory in filenames | |
6828 | """ | |
6829 | expanded = set() | |
6830 | for filename in filenames: | |
6831 | if not os.path.isdir(filename): | |
6832 | expanded.add(filename) | |
6833 | continue | |
6834 | ||
6835 | for root, _, files in os.walk(filename): | |
6836 | for loopfile in files: | |
6837 | fullname = os.path.join(root, loopfile) | |
6838 | if fullname.startswith('.' + os.path.sep): | |
6839 | fullname = fullname[len('.' + os.path.sep):] | |
6840 | expanded.add(fullname) | |
6841 | ||
6842 | filtered = [] | |
6843 | for filename in expanded: | |
6844 | if os.path.splitext(filename)[1][1:] in GetAllExtensions(): | |
6845 | filtered.append(filename) | |
6846 | return filtered | |
6847 | ||
6848 | def _FilterExcludedFiles(fnames): | |
6849 | """Filters out files listed in the --exclude command line switch. File paths | |
6850 | in the switch are evaluated relative to the current working directory | |
6851 | """ | |
6852 | exclude_paths = [os.path.abspath(f) for f in _excludes] | |
6853 | # because globbing does not work recursively, exclude all subpath of all excluded entries | |
6854 | return [f for f in fnames | |
6855 | if not any(e for e in exclude_paths | |
6856 | if _IsParentOrSame(e, os.path.abspath(f)))] | |
6857 | ||
6858 | def _IsParentOrSame(parent, child): | |
6859 | """Return true if child is subdirectory of parent. | |
6860 | Assumes both paths are absolute and don't contain symlinks. | |
6861 | """ | |
6862 | parent = os.path.normpath(parent) | |
6863 | child = os.path.normpath(child) | |
6864 | if parent == child: | |
6865 | return True | |
6866 | ||
6867 | prefix = os.path.commonprefix([parent, child]) | |
6868 | if prefix != parent: | |
6869 | return False | |
6870 | # Note: os.path.commonprefix operates on character basis, so | |
6871 | # take extra care of situations like '/foo/ba' and '/foo/bar/baz' | |
6872 | child_suffix = child[len(prefix):] | |
6873 | child_suffix = child_suffix.lstrip(os.sep) | |
6874 | return child == os.path.join(prefix, child_suffix) | |
6875 | ||
6876 | def main(): | |
6877 | filenames = ParseArguments(sys.argv[1:]) | |
6878 | backup_err = sys.stderr | |
6879 | try: | |
6880 | # Change stderr to write with replacement characters so we don't die | |
6881 | # if we try to print something containing non-ASCII characters. | |
6882 | sys.stderr = codecs.StreamReader(sys.stderr, 'replace') | |
6883 | ||
6884 | _cpplint_state.ResetErrorCounts() | |
6885 | for filename in filenames: | |
6886 | ProcessFile(filename, _cpplint_state.verbose_level) | |
6887 | # If --quiet is passed, suppress printing error count unless there are errors. | |
6888 | if not _cpplint_state.quiet or _cpplint_state.error_count > 0: | |
6889 | _cpplint_state.PrintErrorCounts() | |
6890 | ||
6891 | if _cpplint_state.output_format == 'junit': | |
6892 | sys.stderr.write(_cpplint_state.FormatJUnitXML()) | |
6893 | ||
6894 | finally: | |
6895 | sys.stderr = backup_err | |
6896 | ||
6897 | sys.exit(_cpplint_state.error_count > 0) | |
6898 | ||
6899 | ||
6900 | if __name__ == '__main__': | |
6901 | main() |