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