cpplint.py 179 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
#!/usr/bin/python
#
# Copyright (c) 2009 Google Inc. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
#    * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
#    * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
#    * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

"""Does google-lint on c++ files.

The goal of this script is to identify places in the code that *may*
be in non-compliance with google style.  It does not attempt to fix
up these problems -- the point is to educate.  It does also not
attempt to find all problems, or to ensure that everything it does
find is legitimately a problem.

In particular, we can get very confused by /* and // inside strings!
We do a small hack, which is to ignore //'s with "'s after them on the
same line, but it is far from perfect (in either direction).
"""

import codecs
James Zern's avatar
James Zern committed
45
import copy
46 47 48 49 50 51 52 53 54 55 56 57
import getopt
import math  # for log
import os
import re
import sre_compile
import string
import sys
import unicodedata


_USAGE = """
Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...]
James Zern's avatar
James Zern committed
58 59
                   [--counting=total|toplevel|detailed] [--root=subdir]
                   [--linelength=digits]
60 61 62 63 64 65 66 67 68 69 70 71 72 73
        <file> [file] ...

  The style guidelines this tries to follow are those in
    http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml

  Every problem is given a confidence score from 1-5, with 5 meaning we are
  certain of the problem, and 1 meaning it could be a legitimate construct.
  This will miss some errors, and is not a substitute for a code review.

  To suppress false-positive errors of a certain category, add a
  'NOLINT(category)' comment to the line.  NOLINT or NOLINT(*)
  suppresses errors of all categories on that line.

  The files passed in will be linted; at least one file must be provided.
James Zern's avatar
James Zern committed
74 75
  Default linted extensions are .cc, .cpp, .cu, .cuh and .h.  Change the
  extensions with the --extensions flag.
76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106

  Flags:

    output=vs7
      By default, the output is formatted to ease emacs parsing.  Visual Studio
      compatible output (vs7) may also be used.  Other formats are unsupported.

    verbose=#
      Specify a number 0-5 to restrict errors to certain verbosity levels.

    filter=-x,+y,...
      Specify a comma-separated list of category-filters to apply: only
      error messages whose category names pass the filters will be printed.
      (Category names are printed with the message and look like
      "[whitespace/indent]".)  Filters are evaluated left to right.
      "-FOO" and "FOO" means "do not print categories that start with FOO".
      "+FOO" means "do print categories that start with FOO".

      Examples: --filter=-whitespace,+whitespace/braces
                --filter=whitespace,runtime/printf,+runtime/printf_format
                --filter=-,+build/include_what_you_use

      To see a list of all the categories used in cpplint, pass no arg:
         --filter=

    counting=total|toplevel|detailed
      The total number of errors found is always printed. If
      'toplevel' is provided, then the count of errors in each of
      the top-level categories like 'build' and 'whitespace' will
      also be printed. If 'detailed' is provided, then a count
      is provided for each category like 'build/class'.
James Zern's avatar
James Zern committed
107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122

    root=subdir
      The root directory used for deriving header guard CPP variable.
      By default, the header guard CPP variable is calculated as the relative
      path to the directory that contains .git, .hg, or .svn.  When this flag
      is specified, the relative path is calculated from the specified
      directory. If the specified directory does not exist, this flag is
      ignored.

      Examples:
        Assuing that src/.git exists, the header guard CPP variables for
        src/chrome/browser/ui/browser.h are:

        No flag => CHROME_BROWSER_UI_BROWSER_H_
        --root=chrome => BROWSER_UI_BROWSER_H_
        --root=chrome/browser => UI_BROWSER_H_
James Zern's avatar
James Zern committed
123 124 125 126 127 128 129 130 131 132 133 134 135

    linelength=digits
      This is the allowed line length for the project. The default value is
      80 characters.

      Examples:
        --linelength=120

    extensions=extension,extension,...
      The allowed file extensions that cpplint will check

      Examples:
        --extensions=hpp,cpp
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156
"""

# We categorize each error message we print.  Here are the categories.
# We want an explicit list so we can list them all in cpplint --filter=.
# If you add a new error message with a new category, add it to the list
# here!  cpplint_unittest.py should tell you if you forget to do this.
_ERROR_CATEGORIES = [
  'build/class',
  'build/deprecated',
  'build/endif_comment',
  'build/explicit_make_pair',
  'build/forward_decl',
  'build/header_guard',
  'build/include',
  'build/include_alpha',
  'build/include_order',
  'build/include_what_you_use',
  'build/namespaces',
  'build/printf_format',
  'build/storage_class',
  'legal/copyright',
James Zern's avatar
James Zern committed
157
  'readability/alt_tokens',
158 159 160 161 162 163 164 165
  'readability/braces',
  'readability/casting',
  'readability/check',
  'readability/constructors',
  'readability/fn_size',
  'readability/function',
  'readability/multiline_comment',
  'readability/multiline_string',
James Zern's avatar
James Zern committed
166
  'readability/namespace',
167
  'readability/nolint',
James Zern's avatar
James Zern committed
168
  'readability/nul',
169 170 171 172 173 174 175 176 177 178 179 180 181 182 183
  'readability/streams',
  'readability/todo',
  'readability/utf8',
  'runtime/arrays',
  'runtime/casting',
  'runtime/explicit',
  'runtime/int',
  'runtime/init',
  'runtime/invalid_increment',
  'runtime/member_string_references',
  'runtime/memset',
  'runtime/operator',
  'runtime/printf',
  'runtime/printf_format',
  'runtime/references',
184
  'runtime/sizeof',
185 186
  'runtime/string',
  'runtime/threadsafe_fn',
James Zern's avatar
James Zern committed
187
  'runtime/vlog',
188 189 190 191
  'whitespace/blank_line',
  'whitespace/braces',
  'whitespace/comma',
  'whitespace/comments',
James Zern's avatar
James Zern committed
192
  'whitespace/empty_conditional_body',
James Zern's avatar
James Zern committed
193
  'whitespace/empty_loop_body',
194 195
  'whitespace/end_of_line',
  'whitespace/ending_newline',
James Zern's avatar
James Zern committed
196
  'whitespace/forcolon',
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
  'whitespace/indent',
  'whitespace/line_length',
  'whitespace/newline',
  'whitespace/operators',
  'whitespace/parens',
  'whitespace/semicolon',
  'whitespace/tab',
  'whitespace/todo'
  ]

# The default state of the category filter. This is overrided by the --filter=
# flag. By default all errors are on, so only add here categories that should be
# off by default (i.e., categories that must be enabled by the --filter= flags).
# All entries here should start with a '-' or '+', as in the --filter= flag.
_DEFAULT_FILTERS = ['-build/include_alpha']

# We used to check for high-bit characters, but after much discussion we
# decided those were OK, as long as they were in UTF-8 and didn't represent
# hard-coded international strings, which belong in a separate i18n file.


James Zern's avatar
James Zern committed
218
# C++ headers
219
_CPP_HEADERS = frozenset([
James Zern's avatar
James Zern committed
220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351
    # Legacy
    'algobase.h',
    'algo.h',
    'alloc.h',
    'builtinbuf.h',
    'bvector.h',
    'complex.h',
    'defalloc.h',
    'deque.h',
    'editbuf.h',
    'fstream.h',
    'function.h',
    'hash_map',
    'hash_map.h',
    'hash_set',
    'hash_set.h',
    'hashtable.h',
    'heap.h',
    'indstream.h',
    'iomanip.h',
    'iostream.h',
    'istream.h',
    'iterator.h',
    'list.h',
    'map.h',
    'multimap.h',
    'multiset.h',
    'ostream.h',
    'pair.h',
    'parsestream.h',
    'pfstream.h',
    'procbuf.h',
    'pthread_alloc',
    'pthread_alloc.h',
    'rope',
    'rope.h',
    'ropeimpl.h',
    'set.h',
    'slist',
    'slist.h',
    'stack.h',
    'stdiostream.h',
    'stl_alloc.h',
    'stl_relops.h',
    'streambuf.h',
    'stream.h',
    'strfile.h',
    'strstream.h',
    'tempbuf.h',
    'tree.h',
    'type_traits.h',
    'vector.h',
    # 17.6.1.2 C++ library headers
    'algorithm',
    'array',
    'atomic',
    'bitset',
    'chrono',
    'codecvt',
    'complex',
    'condition_variable',
    'deque',
    'exception',
    'forward_list',
    'fstream',
    'functional',
    'future',
    'initializer_list',
    'iomanip',
    'ios',
    'iosfwd',
    'iostream',
    'istream',
    'iterator',
    'limits',
    'list',
    'locale',
    'map',
    'memory',
    'mutex',
    'new',
    'numeric',
    'ostream',
    'queue',
    'random',
    'ratio',
    'regex',
    'set',
    'sstream',
    'stack',
    'stdexcept',
    'streambuf',
    'string',
    'strstream',
    'system_error',
    'thread',
    'tuple',
    'typeindex',
    'typeinfo',
    'type_traits',
    'unordered_map',
    'unordered_set',
    'utility',
    'valarray',
    'vector',
    # 17.6.1.2 C++ headers for C library facilities
    'cassert',
    'ccomplex',
    'cctype',
    'cerrno',
    'cfenv',
    'cfloat',
    'cinttypes',
    'ciso646',
    'climits',
    'clocale',
    'cmath',
    'csetjmp',
    'csignal',
    'cstdalign',
    'cstdarg',
    'cstdbool',
    'cstddef',
    'cstdint',
    'cstdio',
    'cstdlib',
    'cstring',
    'ctgmath',
    'ctime',
    'cuchar',
    'cwchar',
    'cwctype',
352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385
    ])

# Assertion macros.  These are defined in base/logging.h and
# testing/base/gunit.h.  Note that the _M versions need to come first
# for substring matching to work.
_CHECK_MACROS = [
    'DCHECK', 'CHECK',
    'EXPECT_TRUE_M', 'EXPECT_TRUE',
    'ASSERT_TRUE_M', 'ASSERT_TRUE',
    'EXPECT_FALSE_M', 'EXPECT_FALSE',
    'ASSERT_FALSE_M', 'ASSERT_FALSE',
    ]

# Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE
_CHECK_REPLACEMENT = dict([(m, {}) for m in _CHECK_MACROS])

for op, replacement in [('==', 'EQ'), ('!=', 'NE'),
                        ('>=', 'GE'), ('>', 'GT'),
                        ('<=', 'LE'), ('<', 'LT')]:
  _CHECK_REPLACEMENT['DCHECK'][op] = 'DCHECK_%s' % replacement
  _CHECK_REPLACEMENT['CHECK'][op] = 'CHECK_%s' % replacement
  _CHECK_REPLACEMENT['EXPECT_TRUE'][op] = 'EXPECT_%s' % replacement
  _CHECK_REPLACEMENT['ASSERT_TRUE'][op] = 'ASSERT_%s' % replacement
  _CHECK_REPLACEMENT['EXPECT_TRUE_M'][op] = 'EXPECT_%s_M' % replacement
  _CHECK_REPLACEMENT['ASSERT_TRUE_M'][op] = 'ASSERT_%s_M' % replacement

for op, inv_replacement in [('==', 'NE'), ('!=', 'EQ'),
                            ('>=', 'LT'), ('>', 'LE'),
                            ('<=', 'GT'), ('<', 'GE')]:
  _CHECK_REPLACEMENT['EXPECT_FALSE'][op] = 'EXPECT_%s' % inv_replacement
  _CHECK_REPLACEMENT['ASSERT_FALSE'][op] = 'ASSERT_%s' % inv_replacement
  _CHECK_REPLACEMENT['EXPECT_FALSE_M'][op] = 'EXPECT_%s_M' % inv_replacement
  _CHECK_REPLACEMENT['ASSERT_FALSE_M'][op] = 'ASSERT_%s_M' % inv_replacement

James Zern's avatar
James Zern committed
386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407
# Alternative tokens and their replacements.  For full list, see section 2.5
# Alternative tokens [lex.digraph] in the C++ standard.
#
# Digraphs (such as '%:') are not included here since it's a mess to
# match those on a word boundary.
_ALT_TOKEN_REPLACEMENT = {
    'and': '&&',
    'bitor': '|',
    'or': '||',
    'xor': '^',
    'compl': '~',
    'bitand': '&',
    'and_eq': '&=',
    'or_eq': '|=',
    'xor_eq': '^=',
    'not': '!',
    'not_eq': '!='
    }

# Compile regular expression that matches all the above keywords.  The "[ =()]"
# bit is meant to avoid matching these keywords outside of boolean expressions.
#
James Zern's avatar
James Zern committed
408 409
# False positives include C-style multi-line comments and multi-line strings
# but those have always been troublesome for cpplint.
James Zern's avatar
James Zern committed
410 411 412
_ALT_TOKEN_REPLACEMENT_PATTERN = re.compile(
    r'[ =()](' + ('|'.join(_ALT_TOKEN_REPLACEMENT.keys())) + r')(?=[ (]|$)')

413 414 415 416 417 418 419 420 421

# These constants define types of headers for use with
# _IncludeState.CheckNextIncludeOrder().
_C_SYS_HEADER = 1
_CPP_SYS_HEADER = 2
_LIKELY_MY_HEADER = 3
_POSSIBLE_MY_HEADER = 4
_OTHER_HEADER = 5

James Zern's avatar
James Zern committed
422 423 424 425 426 427 428 429 430 431 432
# These constants define the current inline assembly state
_NO_ASM = 0       # Outside of inline assembly block
_INSIDE_ASM = 1   # Inside inline assembly block
_END_ASM = 2      # Last line of inline assembly block
_BLOCK_ASM = 3    # The whole block is an inline assembly block

# Match start of assembly blocks
_MATCH_ASM = re.compile(r'^\s*(?:asm|_asm|__asm|__asm__)'
                        r'(?:\s+(volatile|__volatile__))?'
                        r'\s*[{(]')

433 434 435 436 437 438 439 440 441 442

_regexp_compile_cache = {}

# Finds occurrences of NOLINT or NOLINT(...).
_RE_SUPPRESSION = re.compile(r'\bNOLINT\b(\([^)]*\))?')

# {str, set(int)}: a map from error categories to sets of linenumbers
# on which those errors are expected and should be suppressed.
_error_suppressions = {}

James Zern's avatar
James Zern committed
443 444 445 446
# The root directory used for deriving header guard CPP variable.
# This is set by --root flag.
_root = None

James Zern's avatar
James Zern committed
447 448 449 450 451 452 453 454
# The allowed line length of files.
# This is set by --linelength flag.
_line_length = 80

# The allowed extensions for file names
# This is set by --extensions flag.
_valid_extensions = set(['cc', 'h', 'cpp', 'cu', 'cuh'])

455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508
def ParseNolintSuppressions(filename, raw_line, linenum, error):
  """Updates the global list of error-suppressions.

  Parses any NOLINT comments on the current line, updating the global
  error_suppressions store.  Reports an error if the NOLINT comment
  was malformed.

  Args:
    filename: str, the name of the input file.
    raw_line: str, the line of input text, with comments.
    linenum: int, the number of the current line.
    error: function, an error handler.
  """
  # FIXME(adonovan): "NOLINT(" is misparsed as NOLINT(*).
  matched = _RE_SUPPRESSION.search(raw_line)
  if matched:
    category = matched.group(1)
    if category in (None, '(*)'):  # => "suppress all"
      _error_suppressions.setdefault(None, set()).add(linenum)
    else:
      if category.startswith('(') and category.endswith(')'):
        category = category[1:-1]
        if category in _ERROR_CATEGORIES:
          _error_suppressions.setdefault(category, set()).add(linenum)
        else:
          error(filename, linenum, 'readability/nolint', 5,
                'Unknown NOLINT error category: %s' % category)


def ResetNolintSuppressions():
  "Resets the set of NOLINT suppressions to empty."
  _error_suppressions.clear()


def IsErrorSuppressedByNolint(category, linenum):
  """Returns true if the specified error category is suppressed on this line.

  Consults the global error_suppressions map populated by
  ParseNolintSuppressions/ResetNolintSuppressions.

  Args:
    category: str, the category of the error.
    linenum: int, the current line number.
  Returns:
    bool, True iff the error should be suppressed due to a NOLINT comment.
  """
  return (linenum in _error_suppressions.get(category, set()) or
          linenum in _error_suppressions.get(None, set()))

def Match(pattern, s):
  """Matches the string with the pattern, caching the compiled regexp."""
  # The regexp compilation caching is inlined in both Match and Search for
  # performance reasons; factoring it out into a separate function turns out
  # to be noticeably expensive.
James Zern's avatar
James Zern committed
509
  if pattern not in _regexp_compile_cache:
510 511 512 513
    _regexp_compile_cache[pattern] = sre_compile.compile(pattern)
  return _regexp_compile_cache[pattern].match(s)


James Zern's avatar
James Zern committed
514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
def ReplaceAll(pattern, rep, s):
  """Replaces instances of pattern in a string with a replacement.

  The compiled regex is kept in a cache shared by Match and Search.

  Args:
    pattern: regex pattern
    rep: replacement text
    s: search string

  Returns:
    string with replacements made (or original string if no replacements)
  """
  if pattern not in _regexp_compile_cache:
    _regexp_compile_cache[pattern] = sre_compile.compile(pattern)
  return _regexp_compile_cache[pattern].sub(rep, s)


532 533
def Search(pattern, s):
  """Searches the string for the pattern, caching the compiled regexp."""
James Zern's avatar
James Zern committed
534
  if pattern not in _regexp_compile_cache:
535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
    _regexp_compile_cache[pattern] = sre_compile.compile(pattern)
  return _regexp_compile_cache[pattern].search(s)


class _IncludeState(dict):
  """Tracks line numbers for includes, and the order in which includes appear.

  As a dict, an _IncludeState object serves as a mapping between include
  filename and line number on which that file was included.

  Call CheckNextIncludeOrder() once for each header in the file, passing
  in the type constants defined above. Calls in an illegal order will
  raise an _IncludeError with an appropriate error message.

  """
  # self._section will move monotonically through this set. If it ever
  # needs to move backwards, CheckNextIncludeOrder will raise an error.
  _INITIAL_SECTION = 0
  _MY_H_SECTION = 1
  _C_SECTION = 2
  _CPP_SECTION = 3
  _OTHER_H_SECTION = 4

  _TYPE_NAMES = {
      _C_SYS_HEADER: 'C system header',
      _CPP_SYS_HEADER: 'C++ system header',
      _LIKELY_MY_HEADER: 'header this file implements',
      _POSSIBLE_MY_HEADER: 'header this file may implement',
      _OTHER_HEADER: 'other header',
      }
  _SECTION_NAMES = {
      _INITIAL_SECTION: "... nothing. (This can't be an error.)",
      _MY_H_SECTION: 'a header this file implements',
      _C_SECTION: 'C system header',
      _CPP_SECTION: 'C++ system header',
      _OTHER_H_SECTION: 'other header',
      }

  def __init__(self):
    dict.__init__(self)
James Zern's avatar
James Zern committed
575 576 577
    self.ResetSection()

  def ResetSection(self):
578 579 580 581 582
    # The name of the current section.
    self._section = self._INITIAL_SECTION
    # The path of last found header.
    self._last_header = ''

James Zern's avatar
James Zern committed
583 584 585
  def SetLastHeader(self, header_path):
    self._last_header = header_path

586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
  def CanonicalizeAlphabeticalOrder(self, header_path):
    """Returns a path canonicalized for alphabetical comparison.

    - replaces "-" with "_" so they both cmp the same.
    - removes '-inl' since we don't require them to be after the main header.
    - lowercase everything, just in case.

    Args:
      header_path: Path to be canonicalized.

    Returns:
      Canonicalized path.
    """
    return header_path.replace('-inl.h', '.h').replace('-', '_').lower()

James Zern's avatar
James Zern committed
601
  def IsInAlphabeticalOrder(self, clean_lines, linenum, header_path):
602 603 604
    """Check if a header is in alphabetical order with the previous header.

    Args:
James Zern's avatar
James Zern committed
605 606 607
      clean_lines: A CleansedLines instance containing the file.
      linenum: The number of the line to check.
      header_path: Canonicalized header to be checked.
608 609 610 611

    Returns:
      Returns true if the header is in alphabetical order.
    """
James Zern's avatar
James Zern committed
612 613 614 615 616 617 618
    # If previous section is different from current section, _last_header will
    # be reset to empty string, so it's always less than current header.
    #
    # If previous line was a blank line, assume that the headers are
    # intentionally sorted the way they are.
    if (self._last_header > header_path and
        not Match(r'^\s*$', clean_lines.elided[linenum - 1])):
619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003
      return False
    return True

  def CheckNextIncludeOrder(self, header_type):
    """Returns a non-empty error message if the next header is out of order.

    This function also updates the internal state to be ready to check
    the next include.

    Args:
      header_type: One of the _XXX_HEADER constants defined above.

    Returns:
      The empty string if the header is in the right order, or an
      error message describing what's wrong.

    """
    error_message = ('Found %s after %s' %
                     (self._TYPE_NAMES[header_type],
                      self._SECTION_NAMES[self._section]))

    last_section = self._section

    if header_type == _C_SYS_HEADER:
      if self._section <= self._C_SECTION:
        self._section = self._C_SECTION
      else:
        self._last_header = ''
        return error_message
    elif header_type == _CPP_SYS_HEADER:
      if self._section <= self._CPP_SECTION:
        self._section = self._CPP_SECTION
      else:
        self._last_header = ''
        return error_message
    elif header_type == _LIKELY_MY_HEADER:
      if self._section <= self._MY_H_SECTION:
        self._section = self._MY_H_SECTION
      else:
        self._section = self._OTHER_H_SECTION
    elif header_type == _POSSIBLE_MY_HEADER:
      if self._section <= self._MY_H_SECTION:
        self._section = self._MY_H_SECTION
      else:
        # This will always be the fallback because we're not sure
        # enough that the header is associated with this file.
        self._section = self._OTHER_H_SECTION
    else:
      assert header_type == _OTHER_HEADER
      self._section = self._OTHER_H_SECTION

    if last_section != self._section:
      self._last_header = ''

    return ''


class _CppLintState(object):
  """Maintains module-wide state.."""

  def __init__(self):
    self.verbose_level = 1  # global setting.
    self.error_count = 0    # global count of reported errors
    # filters to apply when emitting error messages
    self.filters = _DEFAULT_FILTERS[:]
    self.counting = 'total'  # In what way are we counting errors?
    self.errors_by_category = {}  # string to int dict storing error counts

    # output format:
    # "emacs" - format that emacs can parse (default)
    # "vs7" - format that Microsoft Visual Studio 7 can parse
    self.output_format = 'emacs'

  def SetOutputFormat(self, output_format):
    """Sets the output format for errors."""
    self.output_format = output_format

  def SetVerboseLevel(self, level):
    """Sets the module's verbosity, and returns the previous setting."""
    last_verbose_level = self.verbose_level
    self.verbose_level = level
    return last_verbose_level

  def SetCountingStyle(self, counting_style):
    """Sets the module's counting options."""
    self.counting = counting_style

  def SetFilters(self, filters):
    """Sets the error-message filters.

    These filters are applied when deciding whether to emit a given
    error message.

    Args:
      filters: A string of comma-separated filters (eg "+whitespace/indent").
               Each filter should start with + or -; else we die.

    Raises:
      ValueError: The comma-separated filters did not all start with '+' or '-'.
                  E.g. "-,+whitespace,-whitespace/indent,whitespace/badfilter"
    """
    # Default filters always have less priority than the flag ones.
    self.filters = _DEFAULT_FILTERS[:]
    for filt in filters.split(','):
      clean_filt = filt.strip()
      if clean_filt:
        self.filters.append(clean_filt)
    for filt in self.filters:
      if not (filt.startswith('+') or filt.startswith('-')):
        raise ValueError('Every filter in --filters must start with + or -'
                         ' (%s does not)' % filt)

  def ResetErrorCounts(self):
    """Sets the module's error statistic back to zero."""
    self.error_count = 0
    self.errors_by_category = {}

  def IncrementErrorCount(self, category):
    """Bumps the module's error statistic."""
    self.error_count += 1
    if self.counting in ('toplevel', 'detailed'):
      if self.counting != 'detailed':
        category = category.split('/')[0]
      if category not in self.errors_by_category:
        self.errors_by_category[category] = 0
      self.errors_by_category[category] += 1

  def PrintErrorCounts(self):
    """Print a summary of errors by category, and the total."""
    for category, count in self.errors_by_category.iteritems():
      sys.stderr.write('Category \'%s\' errors found: %d\n' %
                       (category, count))
    sys.stderr.write('Total errors found: %d\n' % self.error_count)

_cpplint_state = _CppLintState()


def _OutputFormat():
  """Gets the module's output format."""
  return _cpplint_state.output_format


def _SetOutputFormat(output_format):
  """Sets the module's output format."""
  _cpplint_state.SetOutputFormat(output_format)


def _VerboseLevel():
  """Returns the module's verbosity setting."""
  return _cpplint_state.verbose_level


def _SetVerboseLevel(level):
  """Sets the module's verbosity, and returns the previous setting."""
  return _cpplint_state.SetVerboseLevel(level)


def _SetCountingStyle(level):
  """Sets the module's counting options."""
  _cpplint_state.SetCountingStyle(level)


def _Filters():
  """Returns the module's list of output filters, as a list."""
  return _cpplint_state.filters


def _SetFilters(filters):
  """Sets the module's error-message filters.

  These filters are applied when deciding whether to emit a given
  error message.

  Args:
    filters: A string of comma-separated filters (eg "whitespace/indent").
             Each filter should start with + or -; else we die.
  """
  _cpplint_state.SetFilters(filters)


class _FunctionState(object):
  """Tracks current function name and the number of lines in its body."""

  _NORMAL_TRIGGER = 250  # for --v=0, 500 for --v=1, etc.
  _TEST_TRIGGER = 400    # about 50% more than _NORMAL_TRIGGER.

  def __init__(self):
    self.in_a_function = False
    self.lines_in_function = 0
    self.current_function = ''

  def Begin(self, function_name):
    """Start analyzing function body.

    Args:
      function_name: The name of the function being tracked.
    """
    self.in_a_function = True
    self.lines_in_function = 0
    self.current_function = function_name

  def Count(self):
    """Count line in current function body."""
    if self.in_a_function:
      self.lines_in_function += 1

  def Check(self, error, filename, linenum):
    """Report if too many lines in function body.

    Args:
      error: The function to call with any errors found.
      filename: The name of the current file.
      linenum: The number of the line to check.
    """
    if Match(r'T(EST|est)', self.current_function):
      base_trigger = self._TEST_TRIGGER
    else:
      base_trigger = self._NORMAL_TRIGGER
    trigger = base_trigger * 2**_VerboseLevel()

    if self.lines_in_function > trigger:
      error_level = int(math.log(self.lines_in_function / base_trigger, 2))
      # 50 => 0, 100 => 1, 200 => 2, 400 => 3, 800 => 4, 1600 => 5, ...
      if error_level > 5:
        error_level = 5
      error(filename, linenum, 'readability/fn_size', error_level,
            'Small and focused functions are preferred:'
            ' %s has %d non-comment lines'
            ' (error triggered by exceeding %d lines).'  % (
                self.current_function, self.lines_in_function, trigger))

  def End(self):
    """Stop analyzing function body."""
    self.in_a_function = False


class _IncludeError(Exception):
  """Indicates a problem with the include order in a file."""
  pass


class FileInfo:
  """Provides utility functions for filenames.

  FileInfo provides easy access to the components of a file's path
  relative to the project root.
  """

  def __init__(self, filename):
    self._filename = filename

  def FullName(self):
    """Make Windows paths like Unix."""
    return os.path.abspath(self._filename).replace('\\', '/')

  def RepositoryName(self):
    """FullName after removing the local path to the repository.

    If we have a real absolute path name here we can try to do something smart:
    detecting the root of the checkout and truncating /path/to/checkout from
    the name so that we get header guards that don't include things like
    "C:\Documents and Settings\..." or "/home/username/..." in them and thus
    people on different computers who have checked the source out to different
    locations won't see bogus errors.
    """
    fullname = self.FullName()

    if os.path.exists(fullname):
      project_dir = os.path.dirname(fullname)

      if os.path.exists(os.path.join(project_dir, ".svn")):
        # If there's a .svn file in the current directory, we recursively look
        # up the directory tree for the top of the SVN checkout
        root_dir = project_dir
        one_up_dir = os.path.dirname(root_dir)
        while os.path.exists(os.path.join(one_up_dir, ".svn")):
          root_dir = os.path.dirname(root_dir)
          one_up_dir = os.path.dirname(one_up_dir)

        prefix = os.path.commonprefix([root_dir, project_dir])
        return fullname[len(prefix) + 1:]

      # Not SVN <= 1.6? Try to find a git, hg, or svn top level directory by
      # searching up from the current path.
      root_dir = os.path.dirname(fullname)
      while (root_dir != os.path.dirname(root_dir) and
             not os.path.exists(os.path.join(root_dir, ".git")) and
             not os.path.exists(os.path.join(root_dir, ".hg")) and
             not os.path.exists(os.path.join(root_dir, ".svn"))):
        root_dir = os.path.dirname(root_dir)

      if (os.path.exists(os.path.join(root_dir, ".git")) or
          os.path.exists(os.path.join(root_dir, ".hg")) or
          os.path.exists(os.path.join(root_dir, ".svn"))):
        prefix = os.path.commonprefix([root_dir, project_dir])
        return fullname[len(prefix) + 1:]

    # Don't know what to do; header guard warnings may be wrong...
    return fullname

  def Split(self):
    """Splits the file into the directory, basename, and extension.

    For 'chrome/browser/browser.cc', Split() would
    return ('chrome/browser', 'browser', '.cc')

    Returns:
      A tuple of (directory, basename, extension).
    """

    googlename = self.RepositoryName()
    project, rest = os.path.split(googlename)
    return (project,) + os.path.splitext(rest)

  def BaseName(self):
    """File base name - text after the final slash, before the final period."""
    return self.Split()[1]

  def Extension(self):
    """File extension - text following the final period."""
    return self.Split()[2]

  def NoExtension(self):
    """File has no source file extension."""
    return '/'.join(self.Split()[0:2])

  def IsSource(self):
    """File has a source file extension."""
    return self.Extension()[1:] in ('c', 'cc', 'cpp', 'cxx')


def _ShouldPrintError(category, confidence, linenum):
  """If confidence >= verbose, category passes filter and is not suppressed."""

  # There are three ways we might decide not to print an error message:
  # a "NOLINT(category)" comment appears in the source,
  # the verbosity level isn't high enough, or the filters filter it out.
  if IsErrorSuppressedByNolint(category, linenum):
    return False
  if confidence < _cpplint_state.verbose_level:
    return False

  is_filtered = False
  for one_filter in _Filters():
    if one_filter.startswith('-'):
      if category.startswith(one_filter[1:]):
        is_filtered = True
    elif one_filter.startswith('+'):
      if category.startswith(one_filter[1:]):
        is_filtered = False
    else:
      assert False  # should have been checked for in SetFilter.
  if is_filtered:
    return False

  return True


def Error(filename, linenum, category, confidence, message):
  """Logs the fact we've found a lint error.

  We log where the error was found, and also our confidence in the error,
  that is, how certain we are this is a legitimate style regression, and
  not a misidentification or a use that's sometimes justified.

  False positives can be suppressed by the use of
  "cpplint(category)"  comments on the offending line.  These are
  parsed into _error_suppressions.

  Args:
    filename: The name of the file containing the error.
    linenum: The number of the line containing the error.
    category: A string used to describe the "category" this bug
      falls under: "whitespace", say, or "runtime".  Categories
      may have a hierarchy separated by slashes: "whitespace/indent".
    confidence: A number from 1-5 representing a confidence score for
      the error, with 5 meaning that we are certain of the problem,
      and 1 meaning that it could be a legitimate construct.
    message: The error message.
  """
  if _ShouldPrintError(category, confidence, linenum):
    _cpplint_state.IncrementErrorCount(category)
    if _cpplint_state.output_format == 'vs7':
      sys.stderr.write('%s(%s):  %s  [%s] [%d]\n' % (
          filename, linenum, message, category, confidence))
James Zern's avatar
James Zern committed
1004 1005 1006
    elif _cpplint_state.output_format == 'eclipse':
      sys.stderr.write('%s:%s: warning: %s  [%s] [%d]\n' % (
          filename, linenum, message, category, confidence))
1007 1008 1009 1010 1011
    else:
      sys.stderr.write('%s:%s:  %s  [%s] [%d]\n' % (
          filename, linenum, message, category, confidence))


James Zern's avatar
James Zern committed
1012
# Matches standard C++ escape sequences per 2.13.2.3 of the C++ standard.
1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050
_RE_PATTERN_CLEANSE_LINE_ESCAPES = re.compile(
    r'\\([abfnrtv?"\\\']|\d+|x[0-9a-fA-F]+)')
# Matches strings.  Escape codes should already be removed by ESCAPES.
_RE_PATTERN_CLEANSE_LINE_DOUBLE_QUOTES = re.compile(r'"[^"]*"')
# Matches characters.  Escape codes should already be removed by ESCAPES.
_RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES = re.compile(r"'.'")
# Matches multi-line C++ comments.
# This RE is a little bit more complicated than one might expect, because we
# have to take care of space removals tools so we can handle comments inside
# statements better.
# The current rule is: We only clear spaces from both sides when we're at the
# end of the line. Otherwise, we try to remove spaces from the right side,
# if this doesn't work we try on left side but only if there's a non-character
# on the right.
_RE_PATTERN_CLEANSE_LINE_C_COMMENTS = re.compile(
    r"""(\s*/\*.*\*/\s*$|
            /\*.*\*/\s+|
         \s+/\*.*\*/(?=\W)|
            /\*.*\*/)""", re.VERBOSE)


def IsCppString(line):
  """Does line terminate so, that the next symbol is in string constant.

  This function does not consider single-line nor multi-line comments.

  Args:
    line: is a partial line of code starting from the 0..n.

  Returns:
    True, if next character appended to 'line' is inside a
    string constant.
  """

  line = line.replace(r'\\', 'XX')  # after this, \\" does not match to \"
  return ((line.count('"') - line.count(r'\"') - line.count("'\"'")) & 1) == 1


James Zern's avatar
James Zern committed
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111
def CleanseRawStrings(raw_lines):
  """Removes C++11 raw strings from lines.

    Before:
      static const char kData[] = R"(
          multi-line string
          )";

    After:
      static const char kData[] = ""
          (replaced by blank line)
          "";

  Args:
    raw_lines: list of raw lines.

  Returns:
    list of lines with C++11 raw strings replaced by empty strings.
  """

  delimiter = None
  lines_without_raw_strings = []
  for line in raw_lines:
    if delimiter:
      # Inside a raw string, look for the end
      end = line.find(delimiter)
      if end >= 0:
        # Found the end of the string, match leading space for this
        # line and resume copying the original lines, and also insert
        # a "" on the last line.
        leading_space = Match(r'^(\s*)\S', line)
        line = leading_space.group(1) + '""' + line[end + len(delimiter):]
        delimiter = None
      else:
        # Haven't found the end yet, append a blank line.
        line = ''

    else:
      # Look for beginning of a raw string.
      # See 2.14.15 [lex.string] for syntax.
      matched = Match(r'^(.*)\b(?:R|u8R|uR|UR|LR)"([^\s\\()]*)\((.*)$', line)
      if matched:
        delimiter = ')' + matched.group(2) + '"'

        end = matched.group(3).find(delimiter)
        if end >= 0:
          # Raw string ended on same line
          line = (matched.group(1) + '""' +
                  matched.group(3)[end + len(delimiter):])
          delimiter = None
        else:
          # Start of a multi-line raw string
          line = matched.group(1) + '""'

    lines_without_raw_strings.append(line)

  # TODO(unknown): if delimiter is not None here, we might want to
  # emit a warning for unterminated string.
  return lines_without_raw_strings


1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176
def FindNextMultiLineCommentStart(lines, lineix):
  """Find the beginning marker for a multiline comment."""
  while lineix < len(lines):
    if lines[lineix].strip().startswith('/*'):
      # Only return this marker if the comment goes beyond this line
      if lines[lineix].strip().find('*/', 2) < 0:
        return lineix
    lineix += 1
  return len(lines)


def FindNextMultiLineCommentEnd(lines, lineix):
  """We are inside a comment, find the end marker."""
  while lineix < len(lines):
    if lines[lineix].strip().endswith('*/'):
      return lineix
    lineix += 1
  return len(lines)


def RemoveMultiLineCommentsFromRange(lines, begin, end):
  """Clears a range of lines for multi-line comments."""
  # Having // dummy comments makes the lines non-empty, so we will not get
  # unnecessary blank line warnings later in the code.
  for i in range(begin, end):
    lines[i] = '// dummy'


def RemoveMultiLineComments(filename, lines, error):
  """Removes multiline (c-style) comments from lines."""
  lineix = 0
  while lineix < len(lines):
    lineix_begin = FindNextMultiLineCommentStart(lines, lineix)
    if lineix_begin >= len(lines):
      return
    lineix_end = FindNextMultiLineCommentEnd(lines, lineix_begin)
    if lineix_end >= len(lines):
      error(filename, lineix_begin + 1, 'readability/multiline_comment', 5,
            'Could not find end of multi-line comment')
      return
    RemoveMultiLineCommentsFromRange(lines, lineix_begin, lineix_end + 1)
    lineix = lineix_end + 1


def CleanseComments(line):
  """Removes //-comments and single-line C-style /* */ comments.

  Args:
    line: A line of C++ source.

  Returns:
    The line with single-line comments removed.
  """
  commentpos = line.find('//')
  if commentpos != -1 and not IsCppString(line[:commentpos]):
    line = line[:commentpos].rstrip()
  # get rid of /* ... */
  return _RE_PATTERN_CLEANSE_LINE_C_COMMENTS.sub('', line)


class CleansedLines(object):
  """Holds 3 copies of all lines with different preprocessing applied to them.

  1) elided member contains lines without strings and comments,
  2) lines member contains lines without comments, and
James Zern's avatar
James Zern committed
1177
  3) raw_lines member contains all the lines without processing.
1178 1179 1180 1181 1182 1183 1184 1185
  All these three members are of <type 'list'>, and of the same length.
  """

  def __init__(self, lines):
    self.elided = []
    self.lines = []
    self.raw_lines = lines
    self.num_lines = len(lines)
James Zern's avatar
James Zern committed
1186 1187 1188 1189 1190
    self.lines_without_raw_strings = CleanseRawStrings(lines)
    for linenum in range(len(self.lines_without_raw_strings)):
      self.lines.append(CleanseComments(
          self.lines_without_raw_strings[linenum]))
      elided = self._CollapseStrings(self.lines_without_raw_strings[linenum])
1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218
      self.elided.append(CleanseComments(elided))

  def NumLines(self):
    """Returns the number of lines represented."""
    return self.num_lines

  @staticmethod
  def _CollapseStrings(elided):
    """Collapses strings and chars on a line to simple "" or '' blocks.

    We nix strings first so we're not fooled by text like '"http://"'

    Args:
      elided: The line being processed.

    Returns:
      The line with collapsed strings.
    """
    if not _RE_PATTERN_INCLUDE.match(elided):
      # Remove escaped characters first to make quote/single quote collapsing
      # basic.  Things that look like escaped characters shouldn't occur
      # outside of strings and chars.
      elided = _RE_PATTERN_CLEANSE_LINE_ESCAPES.sub('', elided)
      elided = _RE_PATTERN_CLEANSE_LINE_SINGLE_QUOTES.sub("''", elided)
      elided = _RE_PATTERN_CLEANSE_LINE_DOUBLE_QUOTES.sub('""', elided)
    return elided


James Zern's avatar
James Zern committed
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
def FindEndOfExpressionInLine(line, startpos, depth, startchar, endchar):
  """Find the position just after the matching endchar.

  Args:
    line: a CleansedLines line.
    startpos: start searching at this position.
    depth: nesting level at startpos.
    startchar: expression opening character.
    endchar: expression closing character.

  Returns:
James Zern's avatar
James Zern committed
1230 1231
    On finding matching endchar: (index just after matching endchar, 0)
    Otherwise: (-1, new depth at end of this line)
James Zern's avatar
James Zern committed
1232 1233 1234 1235 1236 1237 1238
  """
  for i in xrange(startpos, len(line)):
    if line[i] == startchar:
      depth += 1
    elif line[i] == endchar:
      depth -= 1
      if depth == 0:
James Zern's avatar
James Zern committed
1239 1240
        return (i + 1, 0)
  return (-1, depth)
James Zern's avatar
James Zern committed
1241 1242


1243
def CloseExpression(clean_lines, linenum, pos):
James Zern's avatar
James Zern committed
1244
  """If input points to ( or { or [ or <, finds the position that closes it.
1245

James Zern's avatar
James Zern committed
1246
  If lines[linenum][pos] points to a '(' or '{' or '[' or '<', finds the
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
  linenum/pos that correspond to the closing of the expression.

  Args:
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
    pos: A position on the line.

  Returns:
    A tuple (line, linenum, pos) pointer *past* the closing brace, or
    (line, len(lines), -1) if we never find a close.  Note we ignore
    strings and comments when matching; and the line we return is the
    'cleansed' line at linenum.
  """

  line = clean_lines.elided[linenum]
  startchar = line[pos]
James Zern's avatar
James Zern committed
1263
  if startchar not in '({[<':
1264 1265 1266 1267
    return (line, clean_lines.NumLines(), -1)
  if startchar == '(': endchar = ')'
  if startchar == '[': endchar = ']'
  if startchar == '{': endchar = '}'
James Zern's avatar
James Zern committed
1268
  if startchar == '<': endchar = '>'
1269

James Zern's avatar
James Zern committed
1270
  # Check first line
James Zern's avatar
James Zern committed
1271 1272
  (end_pos, num_open) = FindEndOfExpressionInLine(
      line, pos, 0, startchar, endchar)
James Zern's avatar
James Zern committed
1273 1274
  if end_pos > -1:
    return (line, linenum, end_pos)
James Zern's avatar
James Zern committed
1275 1276

  # Continue scanning forward
James Zern's avatar
James Zern committed
1277
  while linenum < clean_lines.NumLines() - 1:
1278 1279
    linenum += 1
    line = clean_lines.elided[linenum]
James Zern's avatar
James Zern committed
1280 1281 1282 1283
    (end_pos, num_open) = FindEndOfExpressionInLine(
        line, 0, num_open, startchar, endchar)
    if end_pos > -1:
      return (line, linenum, end_pos)
1284

James Zern's avatar
James Zern committed
1285 1286
  # Did not find endchar before end of file, give up
  return (line, clean_lines.NumLines(), -1)
1287

James Zern's avatar
James Zern committed
1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360

def FindStartOfExpressionInLine(line, endpos, depth, startchar, endchar):
  """Find position at the matching startchar.

  This is almost the reverse of FindEndOfExpressionInLine, but note
  that the input position and returned position differs by 1.

  Args:
    line: a CleansedLines line.
    endpos: start searching at this position.
    depth: nesting level at endpos.
    startchar: expression opening character.
    endchar: expression closing character.

  Returns:
    On finding matching startchar: (index at matching startchar, 0)
    Otherwise: (-1, new depth at beginning of this line)
  """
  for i in xrange(endpos, -1, -1):
    if line[i] == endchar:
      depth += 1
    elif line[i] == startchar:
      depth -= 1
      if depth == 0:
        return (i, 0)
  return (-1, depth)


def ReverseCloseExpression(clean_lines, linenum, pos):
  """If input points to ) or } or ] or >, finds the position that opens it.

  If lines[linenum][pos] points to a ')' or '}' or ']' or '>', finds the
  linenum/pos that correspond to the opening of the expression.

  Args:
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
    pos: A position on the line.

  Returns:
    A tuple (line, linenum, pos) pointer *at* the opening brace, or
    (line, 0, -1) if we never find the matching opening brace.  Note
    we ignore strings and comments when matching; and the line we
    return is the 'cleansed' line at linenum.
  """
  line = clean_lines.elided[linenum]
  endchar = line[pos]
  if endchar not in ')}]>':
    return (line, 0, -1)
  if endchar == ')': startchar = '('
  if endchar == ']': startchar = '['
  if endchar == '}': startchar = '{'
  if endchar == '>': startchar = '<'

  # Check last line
  (start_pos, num_open) = FindStartOfExpressionInLine(
      line, pos, 0, startchar, endchar)
  if start_pos > -1:
    return (line, linenum, start_pos)

  # Continue scanning backward
  while linenum > 0:
    linenum -= 1
    line = clean_lines.elided[linenum]
    (start_pos, num_open) = FindStartOfExpressionInLine(
        line, len(line) - 1, num_open, startchar, endchar)
    if start_pos > -1:
      return (line, linenum, start_pos)

  # Did not find startchar before beginning of file, give up
  return (line, 0, -1)


1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388
def CheckForCopyright(filename, lines, error):
  """Logs an error if no Copyright message appears at the top of the file."""

  # We'll say it should occur by line 10. Don't forget there's a
  # dummy line at the front.
  for line in xrange(1, min(len(lines), 11)):
    if re.search(r'Copyright', lines[line], re.I): break
  else:                       # means no copyright line was found
    error(filename, 0, 'legal/copyright', 5,
          'No copyright message found.  '
          'You should have a line: "Copyright [year] <Copyright Owner>"')


def GetHeaderGuardCPPVariable(filename):
  """Returns the CPP variable that should be used as a header guard.

  Args:
    filename: The name of a C++ header file.

  Returns:
    The CPP variable that should be used as a header guard in the
    named file.

  """

  # Restores original filename in case that cpplint is invoked from Emacs's
  # flymake.
  filename = re.sub(r'_flymake\.h$', '.h', filename)
James Zern's avatar
James Zern committed
1389
  filename = re.sub(r'/\.flymake/([^/]*)$', r'/\1', filename)
1390 1391

  fileinfo = FileInfo(filename)
James Zern's avatar
James Zern committed
1392 1393 1394 1395
  file_path_from_root = fileinfo.RepositoryName()
  if _root:
    file_path_from_root = re.sub('^' + _root + os.sep, '', file_path_from_root)
  return re.sub(r'[-./\s]', '_', file_path_from_root).upper() + '_'
1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472


def CheckForHeaderGuard(filename, lines, error):
  """Checks that the file contains a header guard.

  Logs an error if no #ifndef header guard is present.  For other
  headers, checks that the full pathname is used.

  Args:
    filename: The name of the C++ header file.
    lines: An array of strings, each representing a line of the file.
    error: The function to call with any errors found.
  """

  cppvar = GetHeaderGuardCPPVariable(filename)

  ifndef = None
  ifndef_linenum = 0
  define = None
  endif = None
  endif_linenum = 0
  for linenum, line in enumerate(lines):
    linesplit = line.split()
    if len(linesplit) >= 2:
      # find the first occurrence of #ifndef and #define, save arg
      if not ifndef and linesplit[0] == '#ifndef':
        # set ifndef to the header guard presented on the #ifndef line.
        ifndef = linesplit[1]
        ifndef_linenum = linenum
      if not define and linesplit[0] == '#define':
        define = linesplit[1]
    # find the last occurrence of #endif, save entire line
    if line.startswith('#endif'):
      endif = line
      endif_linenum = linenum

  if not ifndef:
    error(filename, 0, 'build/header_guard', 5,
          'No #ifndef header guard found, suggested CPP variable is: %s' %
          cppvar)
    return

  if not define:
    error(filename, 0, 'build/header_guard', 5,
          'No #define header guard found, suggested CPP variable is: %s' %
          cppvar)
    return

  # The guard should be PATH_FILE_H_, but we also allow PATH_FILE_H__
  # for backward compatibility.
  if ifndef != cppvar:
    error_level = 0
    if ifndef != cppvar + '_':
      error_level = 5

    ParseNolintSuppressions(filename, lines[ifndef_linenum], ifndef_linenum,
                            error)
    error(filename, ifndef_linenum, 'build/header_guard', error_level,
          '#ifndef header guard has wrong style, please use: %s' % cppvar)

  if define != ifndef:
    error(filename, 0, 'build/header_guard', 5,
          '#ifndef and #define don\'t match, suggested CPP variable is: %s' %
          cppvar)
    return

  if endif != ('#endif  // %s' % cppvar):
    error_level = 0
    if endif != ('#endif  // %s' % (cppvar + '_')):
      error_level = 5

    ParseNolintSuppressions(filename, lines[endif_linenum], endif_linenum,
                            error)
    error(filename, endif_linenum, 'build/header_guard', error_level,
          '#endif line should be "#endif  // %s"' % cppvar)


James Zern's avatar
James Zern committed
1473 1474 1475 1476 1477 1478 1479 1480 1481
def CheckForBadCharacters(filename, lines, error):
  """Logs an error for each line containing bad characters.

  Two kinds of bad characters:

  1. Unicode replacement characters: These indicate that either the file
  contained invalid UTF-8 (likely) or Unicode replacement characters (which
  it shouldn't).  Note that it's possible for this to throw off line
  numbering if the invalid UTF-8 occurred adjacent to a newline.
1482

James Zern's avatar
James Zern committed
1483
  2. NUL bytes.  These are problematic for some tools.
1484 1485 1486 1487 1488 1489 1490 1491 1492 1493

  Args:
    filename: The name of the current file.
    lines: An array of strings, each representing a line of the file.
    error: The function to call with any errors found.
  """
  for linenum, line in enumerate(lines):
    if u'\ufffd' in line:
      error(filename, linenum, 'readability/utf8', 5,
            'Line contains invalid UTF-8 (or Unicode replacement character).')
James Zern's avatar
James Zern committed
1494 1495
    if '\0' in line:
      error(filename, linenum, 'readability/nul', 5, 'Line contains NUL byte.')
1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549


def CheckForNewlineAtEOF(filename, lines, error):
  """Logs an error if there is no newline char at the end of the file.

  Args:
    filename: The name of the current file.
    lines: An array of strings, each representing a line of the file.
    error: The function to call with any errors found.
  """

  # The array lines() was created by adding two newlines to the
  # original file (go figure), then splitting on \n.
  # To verify that the file ends in \n, we just have to make sure the
  # last-but-two element of lines() exists and is empty.
  if len(lines) < 3 or lines[-2]:
    error(filename, len(lines) - 2, 'whitespace/ending_newline', 5,
          'Could not find a newline character at the end of the file.')


def CheckForMultilineCommentsAndStrings(filename, clean_lines, linenum, error):
  """Logs an error if we see /* ... */ or "..." that extend past one line.

  /* ... */ comments are legit inside macros, for one line.
  Otherwise, we prefer // comments, so it's ok to warn about the
  other.  Likewise, it's ok for strings to extend across multiple
  lines, as long as a line continuation character (backslash)
  terminates each line. Although not currently prohibited by the C++
  style guide, it's ugly and unnecessary. We don't do well with either
  in this lint program, so we warn about both.

  Args:
    filename: The name of the current file.
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
    error: The function to call with any errors found.
  """
  line = clean_lines.elided[linenum]

  # Remove all \\ (escaped backslashes) from the line. They are OK, and the
  # second (escaped) slash may trigger later \" detection erroneously.
  line = line.replace('\\\\', '')

  if line.count('/*') > line.count('*/'):
    error(filename, linenum, 'readability/multiline_comment', 5,
          'Complex multi-line /*...*/-style comment found. '
          'Lint may give bogus warnings.  '
          'Consider replacing these with //-style comments, '
          'with #if 0...#endif, '
          'or with more clearly structured multi-line comments.')

  if (line.count('"') - line.count('\\"')) % 2:
    error(filename, linenum, 'readability/multiline_string', 5,
          'Multi-line string ("...") found.  This lint script doesn\'t '
James Zern's avatar
James Zern committed
1550 1551
          'do well with such strings, and may give bogus warnings.  '
          'Use C++11 raw strings or concatenation instead.')
1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587


threading_list = (
    ('asctime(', 'asctime_r('),
    ('ctime(', 'ctime_r('),
    ('getgrgid(', 'getgrgid_r('),
    ('getgrnam(', 'getgrnam_r('),
    ('getlogin(', 'getlogin_r('),
    ('getpwnam(', 'getpwnam_r('),
    ('getpwuid(', 'getpwuid_r('),
    ('gmtime(', 'gmtime_r('),
    ('localtime(', 'localtime_r('),
    ('rand(', 'rand_r('),
    ('strtok(', 'strtok_r('),
    ('ttyname(', 'ttyname_r('),
    )


def CheckPosixThreading(filename, clean_lines, linenum, error):
  """Checks for calls to thread-unsafe functions.

  Much code has been originally written without consideration of
  multi-threading. Also, engineers are relying on their old experience;
  they have learned posix before threading extensions were added. These
  tests guide the engineers to use thread-safe functions (when using
  posix directly).

  Args:
    filename: The name of the current file.
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
    error: The function to call with any errors found.
  """
  line = clean_lines.elided[linenum]
  for single_thread_function, multithread_safe_function in threading_list:
    ix = line.find(single_thread_function)
James Zern's avatar
James Zern committed
1588
    # Comparisons made explicit for clarity -- pylint: disable=g-explicit-bool-comparison
1589 1590 1591 1592 1593 1594 1595 1596
    if ix >= 0 and (ix == 0 or (not line[ix - 1].isalnum() and
                                line[ix - 1] not in ('_', '.', '>'))):
      error(filename, linenum, 'runtime/threadsafe_fn', 2,
            'Consider using ' + multithread_safe_function +
            '...) instead of ' + single_thread_function +
            '...) for improved thread safety.')


James Zern's avatar
James Zern committed
1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615
def CheckVlogArguments(filename, clean_lines, linenum, error):
  """Checks that VLOG() is only used for defining a logging level.

  For example, VLOG(2) is correct. VLOG(INFO), VLOG(WARNING), VLOG(ERROR), and
  VLOG(FATAL) are not.

  Args:
    filename: The name of the current file.
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
    error: The function to call with any errors found.
  """
  line = clean_lines.elided[linenum]
  if Search(r'\bVLOG\((INFO|ERROR|WARNING|DFATAL|FATAL)\)', line):
    error(filename, linenum, 'runtime/vlog', 5,
          'VLOG() should be used with numeric verbosity level.  '
          'Use LOG() if you want symbolic severity levels.')


1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643
# Matches invalid increment: *count++, which moves pointer instead of
# incrementing a value.
_RE_PATTERN_INVALID_INCREMENT = re.compile(
    r'^\s*\*\w+(\+\+|--);')


def CheckInvalidIncrement(filename, clean_lines, linenum, error):
  """Checks for invalid increment *count++.

  For example following function:
  void increment_counter(int* count) {
    *count++;
  }
  is invalid, because it effectively does count++, moving pointer, and should
  be replaced with ++*count, (*count)++ or *count += 1.

  Args:
    filename: The name of the current file.
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
    error: The function to call with any errors found.
  """
  line = clean_lines.elided[linenum]
  if _RE_PATTERN_INVALID_INCREMENT.match(line):
    error(filename, linenum, 'runtime/invalid_increment', 5,
          'Changing pointer instead of value (or unused value of operator*).')


James Zern's avatar
James Zern committed
1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681
class _BlockInfo(object):
  """Stores information about a generic block of code."""

  def __init__(self, seen_open_brace):
    self.seen_open_brace = seen_open_brace
    self.open_parentheses = 0
    self.inline_asm = _NO_ASM

  def CheckBegin(self, filename, clean_lines, linenum, error):
    """Run checks that applies to text up to the opening brace.

    This is mostly for checking the text after the class identifier
    and the "{", usually where the base class is specified.  For other
    blocks, there isn't much to check, so we always pass.

    Args:
      filename: The name of the current file.
      clean_lines: A CleansedLines instance containing the file.
      linenum: The number of the line to check.
      error: The function to call with any errors found.
    """
    pass

  def CheckEnd(self, filename, clean_lines, linenum, error):
    """Run checks that applies to text after the closing brace.

    This is mostly used for checking end of namespace comments.

    Args:
      filename: The name of the current file.
      clean_lines: A CleansedLines instance containing the file.
      linenum: The number of the line to check.
      error: The function to call with any errors found.
    """
    pass


class _ClassInfo(_BlockInfo):
1682 1683
  """Stores information about a class."""

James Zern's avatar
James Zern committed
1684 1685
  def __init__(self, name, class_or_struct, clean_lines, linenum):
    _BlockInfo.__init__(self, False)
1686
    self.name = name
James Zern's avatar
James Zern committed
1687
    self.starting_linenum = linenum
1688
    self.is_derived = False
James Zern's avatar
James Zern committed
1689 1690
    if class_or_struct == 'struct':
      self.access = 'public'
James Zern's avatar
James Zern committed
1691
      self.is_struct = True
James Zern's avatar
James Zern committed
1692 1693
    else:
      self.access = 'private'
James Zern's avatar
James Zern committed
1694 1695 1696 1697 1698 1699 1700 1701 1702
      self.is_struct = False

    # Remember initial indentation level for this class.  Using raw_lines here
    # instead of elided to account for leading comments.
    initial_indent = Match(r'^( *)\S', clean_lines.raw_lines[linenum])
    if initial_indent:
      self.class_indent = len(initial_indent.group(1))
    else:
      self.class_indent = 0
1703 1704 1705 1706 1707 1708 1709 1710 1711

    # Try to find the end of the class.  This will be confused by things like:
    #   class A {
    #   } *x = { ...
    #
    # But it's still good enough for CheckSectionSpacing.
    self.last_line = 0
    depth = 0
    for i in range(linenum, clean_lines.NumLines()):
James Zern's avatar
James Zern committed
1712
      line = clean_lines.elided[i]
1713 1714 1715 1716 1717
      depth += line.count('{') - line.count('}')
      if not depth:
        self.last_line = i
        break

James Zern's avatar
James Zern committed
1718 1719 1720 1721
  def CheckBegin(self, filename, clean_lines, linenum, error):
    # Look for a bare ':'
    if Search('(^|[^:]):($|[^:])', clean_lines.elided[linenum]):
      self.is_derived = True
1722

James Zern's avatar
James Zern committed
1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735
  def CheckEnd(self, filename, clean_lines, linenum, error):
    # Check that closing brace is aligned with beginning of the class.
    # Only do this if the closing brace is indented by only whitespaces.
    # This means we will not check single-line class definitions.
    indent = Match(r'^( *)\}', clean_lines.elided[linenum])
    if indent and len(indent.group(1)) != self.class_indent:
      if self.is_struct:
        parent = 'struct ' + self.name
      else:
        parent = 'class ' + self.name
      error(filename, linenum, 'whitespace/indent', 3,
            'Closing brace should be aligned with beginning of %s' % parent)

1736

James Zern's avatar
James Zern committed
1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767
class _NamespaceInfo(_BlockInfo):
  """Stores information about a namespace."""

  def __init__(self, name, linenum):
    _BlockInfo.__init__(self, False)
    self.name = name or ''
    self.starting_linenum = linenum

  def CheckEnd(self, filename, clean_lines, linenum, error):
    """Check end of namespace comments."""
    line = clean_lines.raw_lines[linenum]

    # Check how many lines is enclosed in this namespace.  Don't issue
    # warning for missing namespace comments if there aren't enough
    # lines.  However, do apply checks if there is already an end of
    # namespace comment and it's incorrect.
    #
    # TODO(unknown): We always want to check end of namespace comments
    # if a namespace is large, but sometimes we also want to apply the
    # check if a short namespace contained nontrivial things (something
    # other than forward declarations).  There is currently no logic on
    # deciding what these nontrivial things are, so this check is
    # triggered by namespace size only, which works most of the time.
    if (linenum - self.starting_linenum < 10
        and not Match(r'};*\s*(//|/\*).*\bnamespace\b', line)):
      return

    # Look for matching comment at end of namespace.
    #
    # Note that we accept C style "/* */" comments for terminating
    # namespaces, so that code that terminate namespaces inside
James Zern's avatar
James Zern committed
1768
    # preprocessor macros can be cpplint clean.
James Zern's avatar
James Zern committed
1769 1770 1771 1772 1773 1774
    #
    # We also accept stuff like "// end of namespace <name>." with the
    # period at the end.
    #
    # Besides these, we don't accept anything else, otherwise we might
    # get false negatives when existing comment is a substring of the
James Zern's avatar
James Zern committed
1775
    # expected namespace.
James Zern's avatar
James Zern committed
1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806
    if self.name:
      # Named namespace
      if not Match((r'};*\s*(//|/\*).*\bnamespace\s+' + re.escape(self.name) +
                    r'[\*/\.\\\s]*$'),
                   line):
        error(filename, linenum, 'readability/namespace', 5,
              'Namespace should be terminated with "// namespace %s"' %
              self.name)
    else:
      # Anonymous namespace
      if not Match(r'};*\s*(//|/\*).*\bnamespace[\*/\.\\\s]*$', line):
        error(filename, linenum, 'readability/namespace', 5,
              'Namespace should be terminated with "// namespace"')


class _PreprocessorInfo(object):
  """Stores checkpoints of nesting stacks when #if/#else is seen."""

  def __init__(self, stack_before_if):
    # The entire nesting stack before #if
    self.stack_before_if = stack_before_if

    # The entire nesting stack up to #else
    self.stack_before_else = []

    # Whether we have already seen #else or #elif
    self.seen_else = False


class _NestingState(object):
  """Holds states related to parsing braces."""
1807 1808

  def __init__(self):
James Zern's avatar
James Zern committed
1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821
    # Stack for tracking all braces.  An object is pushed whenever we
    # see a "{", and popped when we see a "}".  Only 3 types of
    # objects are possible:
    # - _ClassInfo: a class or struct.
    # - _NamespaceInfo: a namespace.
    # - _BlockInfo: some other type of block.
    self.stack = []

    # Stack of _PreprocessorInfo objects.
    self.pp_stack = []

  def SeenOpenBrace(self):
    """Check if we have seen the opening brace for the innermost block.
1822

James Zern's avatar
James Zern committed
1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972
    Returns:
      True if we have seen the opening brace, False if the innermost
      block is still expecting an opening brace.
    """
    return (not self.stack) or self.stack[-1].seen_open_brace

  def InNamespaceBody(self):
    """Check if we are currently one level inside a namespace body.

    Returns:
      True if top of the stack is a namespace block, False otherwise.
    """
    return self.stack and isinstance(self.stack[-1], _NamespaceInfo)

  def UpdatePreprocessor(self, line):
    """Update preprocessor stack.

    We need to handle preprocessors due to classes like this:
      #ifdef SWIG
      struct ResultDetailsPageElementExtensionPoint {
      #else
      struct ResultDetailsPageElementExtensionPoint : public Extension {
      #endif

    We make the following assumptions (good enough for most files):
    - Preprocessor condition evaluates to true from #if up to first
      #else/#elif/#endif.

    - Preprocessor condition evaluates to false from #else/#elif up
      to #endif.  We still perform lint checks on these lines, but
      these do not affect nesting stack.

    Args:
      line: current line to check.
    """
    if Match(r'^\s*#\s*(if|ifdef|ifndef)\b', line):
      # Beginning of #if block, save the nesting stack here.  The saved
      # stack will allow us to restore the parsing state in the #else case.
      self.pp_stack.append(_PreprocessorInfo(copy.deepcopy(self.stack)))
    elif Match(r'^\s*#\s*(else|elif)\b', line):
      # Beginning of #else block
      if self.pp_stack:
        if not self.pp_stack[-1].seen_else:
          # This is the first #else or #elif block.  Remember the
          # whole nesting stack up to this point.  This is what we
          # keep after the #endif.
          self.pp_stack[-1].seen_else = True
          self.pp_stack[-1].stack_before_else = copy.deepcopy(self.stack)

        # Restore the stack to how it was before the #if
        self.stack = copy.deepcopy(self.pp_stack[-1].stack_before_if)
      else:
        # TODO(unknown): unexpected #else, issue warning?
        pass
    elif Match(r'^\s*#\s*endif\b', line):
      # End of #if or #else blocks.
      if self.pp_stack:
        # If we saw an #else, we will need to restore the nesting
        # stack to its former state before the #else, otherwise we
        # will just continue from where we left off.
        if self.pp_stack[-1].seen_else:
          # Here we can just use a shallow copy since we are the last
          # reference to it.
          self.stack = self.pp_stack[-1].stack_before_else
        # Drop the corresponding #if
        self.pp_stack.pop()
      else:
        # TODO(unknown): unexpected #endif, issue warning?
        pass

  def Update(self, filename, clean_lines, linenum, error):
    """Update nesting state with current line.

    Args:
      filename: The name of the current file.
      clean_lines: A CleansedLines instance containing the file.
      linenum: The number of the line to check.
      error: The function to call with any errors found.
    """
    line = clean_lines.elided[linenum]

    # Update pp_stack first
    self.UpdatePreprocessor(line)

    # Count parentheses.  This is to avoid adding struct arguments to
    # the nesting stack.
    if self.stack:
      inner_block = self.stack[-1]
      depth_change = line.count('(') - line.count(')')
      inner_block.open_parentheses += depth_change

      # Also check if we are starting or ending an inline assembly block.
      if inner_block.inline_asm in (_NO_ASM, _END_ASM):
        if (depth_change != 0 and
            inner_block.open_parentheses == 1 and
            _MATCH_ASM.match(line)):
          # Enter assembly block
          inner_block.inline_asm = _INSIDE_ASM
        else:
          # Not entering assembly block.  If previous line was _END_ASM,
          # we will now shift to _NO_ASM state.
          inner_block.inline_asm = _NO_ASM
      elif (inner_block.inline_asm == _INSIDE_ASM and
            inner_block.open_parentheses == 0):
        # Exit assembly block
        inner_block.inline_asm = _END_ASM

    # Consume namespace declaration at the beginning of the line.  Do
    # this in a loop so that we catch same line declarations like this:
    #   namespace proto2 { namespace bridge { class MessageSet; } }
    while True:
      # Match start of namespace.  The "\b\s*" below catches namespace
      # declarations even if it weren't followed by a whitespace, this
      # is so that we don't confuse our namespace checker.  The
      # missing spaces will be flagged by CheckSpacing.
      namespace_decl_match = Match(r'^\s*namespace\b\s*([:\w]+)?(.*)$', line)
      if not namespace_decl_match:
        break

      new_namespace = _NamespaceInfo(namespace_decl_match.group(1), linenum)
      self.stack.append(new_namespace)

      line = namespace_decl_match.group(2)
      if line.find('{') != -1:
        new_namespace.seen_open_brace = True
        line = line[line.find('{') + 1:]

    # Look for a class declaration in whatever is left of the line
    # after parsing namespaces.  The regexp accounts for decorated classes
    # such as in:
    #   class LOCKABLE API Object {
    #   };
    #
    # Templates with class arguments may confuse the parser, for example:
    #   template <class T
    #             class Comparator = less<T>,
    #             class Vector = vector<T> >
    #   class HeapQueue {
    #
    # Because this parser has no nesting state about templates, by the
    # time it saw "class Comparator", it may think that it's a new class.
    # Nested templates have a similar problem:
    #   template <
    #       typename ExportedType,
    #       typename TupleType,
    #       template <typename, typename> class ImplTemplate>
    #
    # To avoid these cases, we ignore classes that are followed by '=' or '>'
    class_decl_match = Match(
        r'\s*(template\s*<[\w\s<>,:]*>\s*)?'
James Zern's avatar
James Zern committed
1973 1974
        r'(class|struct)\s+([A-Z_]+\s+)*(\w+(?:::\w+)*)'
        r'(([^=>]|<[^<>]*>|<[^<>]*<[^<>]*>\s*>)*)$', line)
James Zern's avatar
James Zern committed
1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988
    if (class_decl_match and
        (not self.stack or self.stack[-1].open_parentheses == 0)):
      self.stack.append(_ClassInfo(
          class_decl_match.group(4), class_decl_match.group(2),
          clean_lines, linenum))
      line = class_decl_match.group(5)

    # If we have not yet seen the opening brace for the innermost block,
    # run checks here.
    if not self.SeenOpenBrace():
      self.stack[-1].CheckBegin(filename, clean_lines, linenum, error)

    # Update access control if we are inside a class/struct
    if self.stack and isinstance(self.stack[-1], _ClassInfo):
James Zern's avatar
James Zern committed
1989 1990 1991 1992 1993
      classinfo = self.stack[-1]
      access_match = Match(
          r'^(.*)\b(public|private|protected|signals)(\s+(?:slots\s*)?)?'
          r':(?:[^:]|$)',
          line)
James Zern's avatar
James Zern committed
1994
      if access_match:
James Zern's avatar
James Zern committed
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
        classinfo.access = access_match.group(2)

        # Check that access keywords are indented +1 space.  Skip this
        # check if the keywords are not preceded by whitespaces.
        indent = access_match.group(1)
        if (len(indent) != classinfo.class_indent + 1 and
            Match(r'^\s*$', indent)):
          if classinfo.is_struct:
            parent = 'struct ' + classinfo.name
          else:
            parent = 'class ' + classinfo.name
          slots = ''
          if access_match.group(3):
            slots = access_match.group(3)
          error(filename, linenum, 'whitespace/indent', 3,
                '%s%s: should be indented +1 space inside %s' % (
                    access_match.group(2), slots, parent))
James Zern's avatar
James Zern committed
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060

    # Consume braces or semicolons from what's left of the line
    while True:
      # Match first brace, semicolon, or closed parenthesis.
      matched = Match(r'^[^{;)}]*([{;)}])(.*)$', line)
      if not matched:
        break

      token = matched.group(1)
      if token == '{':
        # If namespace or class hasn't seen a opening brace yet, mark
        # namespace/class head as complete.  Push a new block onto the
        # stack otherwise.
        if not self.SeenOpenBrace():
          self.stack[-1].seen_open_brace = True
        else:
          self.stack.append(_BlockInfo(True))
          if _MATCH_ASM.match(line):
            self.stack[-1].inline_asm = _BLOCK_ASM
      elif token == ';' or token == ')':
        # If we haven't seen an opening brace yet, but we already saw
        # a semicolon, this is probably a forward declaration.  Pop
        # the stack for these.
        #
        # Similarly, if we haven't seen an opening brace yet, but we
        # already saw a closing parenthesis, then these are probably
        # function arguments with extra "class" or "struct" keywords.
        # Also pop these stack for these.
        if not self.SeenOpenBrace():
          self.stack.pop()
      else:  # token == '}'
        # Perform end of block checks and pop the stack.
        if self.stack:
          self.stack[-1].CheckEnd(filename, clean_lines, linenum, error)
          self.stack.pop()
      line = matched.group(2)

  def InnermostClass(self):
    """Get class info on the top of the stack.

    Returns:
      A _ClassInfo object if we are inside a class, or None otherwise.
    """
    for i in range(len(self.stack), 0, -1):
      classinfo = self.stack[i - 1]
      if isinstance(classinfo, _ClassInfo):
        return classinfo
    return None

James Zern's avatar
James Zern committed
2061 2062
  def CheckCompletedBlocks(self, filename, error):
    """Checks that all classes and namespaces have been completely parsed.
2063 2064 2065 2066 2067 2068

    Call this when all lines in a file have been processed.
    Args:
      filename: The name of the current file.
      error: The function to call with any errors found.
    """
James Zern's avatar
James Zern committed
2069 2070 2071 2072 2073 2074 2075 2076
    # Note: This test can result in false positives if #ifdef constructs
    # get in the way of brace matching. See the testBuildClass test in
    # cpplint_unittest.py for an example of this.
    for obj in self.stack:
      if isinstance(obj, _ClassInfo):
        error(filename, obj.starting_linenum, 'build/class', 5,
              'Failed to find complete declaration of class %s' %
              obj.name)
James Zern's avatar
James Zern committed
2077 2078 2079 2080
      elif isinstance(obj, _NamespaceInfo):
        error(filename, obj.starting_linenum, 'build/namespaces', 5,
              'Failed to find complete declaration of namespace %s' %
              obj.name)
2081 2082 2083


def CheckForNonStandardConstructs(filename, clean_lines, linenum,
James Zern's avatar
James Zern committed
2084
                                  nesting_state, error):
James Zern's avatar
James Zern committed
2085
  r"""Logs an error if we see certain non-ANSI constructs ignored by gcc-2.
2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105

  Complain about several constructs which gcc-2 accepts, but which are
  not standard C++.  Warning about these in lint is one way to ease the
  transition to new compilers.
  - put storage class first (e.g. "static const" instead of "const static").
  - "%lld" instead of %qd" in printf-type functions.
  - "%1$d" is non-standard in printf-type functions.
  - "\%" is an undefined character escape sequence.
  - text after #endif is not allowed.
  - invalid inner-style forward declaration.
  - >? and <? operators, and their >?= and <?= cousins.

  Additionally, check for constructor/destructor style violations and reference
  members, as it is very convenient to do so while checking for
  gcc-2 compliance.

  Args:
    filename: The name of the current file.
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
James Zern's avatar
James Zern committed
2106 2107
    nesting_state: A _NestingState instance which maintains information about
                   the current stack of nested blocks being parsed.
2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135
    error: A callable to which errors are reported, which takes 4 arguments:
           filename, line number, error level, and message
  """

  # Remove comments from the line, but leave in strings for now.
  line = clean_lines.lines[linenum]

  if Search(r'printf\s*\(.*".*%[-+ ]?\d*q', line):
    error(filename, linenum, 'runtime/printf_format', 3,
          '%q in format strings is deprecated.  Use %ll instead.')

  if Search(r'printf\s*\(.*".*%\d+\$', line):
    error(filename, linenum, 'runtime/printf_format', 2,
          '%N$ formats are unconventional.  Try rewriting to avoid them.')

  # Remove escaped backslashes before looking for undefined escapes.
  line = line.replace('\\\\', '')

  if Search(r'("|\').*\\(%|\[|\(|{)', line):
    error(filename, linenum, 'build/printf_format', 3,
          '%, [, (, and { are undefined character escapes.  Unescape them.')

  # For the rest, work with both comments and strings removed.
  line = clean_lines.elided[linenum]

  if Search(r'\b(const|volatile|void|char|short|int|long'
            r'|float|double|signed|unsigned'
            r'|schar|u?int8|u?int16|u?int32|u?int64)'
James Zern's avatar
James Zern committed
2136
            r'\s+(register|static|extern|typedef)\b',
2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165
            line):
    error(filename, linenum, 'build/storage_class', 5,
          'Storage class (static, extern, typedef, etc) should be first.')

  if Match(r'\s*#\s*endif\s*[^/\s]+', line):
    error(filename, linenum, 'build/endif_comment', 5,
          'Uncommented text after #endif is non-standard.  Use a comment.')

  if Match(r'\s*class\s+(\w+\s*::\s*)+\w+\s*;', line):
    error(filename, linenum, 'build/forward_decl', 5,
          'Inner-style forward declarations are invalid.  Remove this line.')

  if Search(r'(\w+|[+-]?\d+(\.\d*)?)\s*(<|>)\?=?\s*(\w+|[+-]?\d+)(\.\d*)?',
            line):
    error(filename, linenum, 'build/deprecated', 3,
          '>? and <? (max and min) operators are non-standard and deprecated.')

  if Search(r'^\s*const\s*string\s*&\s*\w+\s*;', line):
    # TODO(unknown): Could it be expanded safely to arbitrary references,
    # without triggering too many false positives? The first
    # attempt triggered 5 warnings for mostly benign code in the regtest, hence
    # the restriction.
    # Here's the original regexp, for the reference:
    # type_name = r'\w+((\s*::\s*\w+)|(\s*<\s*\w+?\s*>))?'
    # r'\s*const\s*' + type_name + '\s*&\s*\w+\s*;'
    error(filename, linenum, 'runtime/member_string_references', 2,
          'const string& members are dangerous. It is much better to use '
          'alternatives, such as pointers or simple constants.')

James Zern's avatar
James Zern committed
2166 2167 2168 2169 2170
  # Everything else in this function operates on class declarations.
  # Return early if the top of the nesting stack is not a class, or if
  # the class head is not completed yet.
  classinfo = nesting_state.InnermostClass()
  if not classinfo or not classinfo.seen_open_brace:
2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183
    return

  # The class may have been declared with namespace or classname qualifiers.
  # The constructor and destructor will not have those qualifiers.
  base_classname = classinfo.name.split('::')[-1]

  # Look for single-argument constructors that aren't marked explicit.
  # Technically a valid construct, but against style.
  args = Match(r'\s+(?:inline\s+)?%s\s*\(([^,()]+)\)'
               % re.escape(base_classname),
               line)
  if (args and
      args.group(1) != 'void' and
James Zern's avatar
James Zern committed
2184 2185
      not Match(r'(const\s+)?%s(\s+const)?\s*(?:<\w+>\s*)?&'
                % re.escape(base_classname), args.group(1).strip())):
2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227
    error(filename, linenum, 'runtime/explicit', 5,
          'Single-argument constructors should be marked explicit.')


def CheckSpacingForFunctionCall(filename, line, linenum, error):
  """Checks for the correctness of various spacing around function calls.

  Args:
    filename: The name of the current file.
    line: The text of the line to check.
    linenum: The number of the line to check.
    error: The function to call with any errors found.
  """

  # Since function calls often occur inside if/for/while/switch
  # expressions - which have their own, more liberal conventions - we
  # first see if we should be looking inside such an expression for a
  # function call, to which we can apply more strict standards.
  fncall = line    # if there's no control flow construct, look at whole line
  for pattern in (r'\bif\s*\((.*)\)\s*{',
                  r'\bfor\s*\((.*)\)\s*{',
                  r'\bwhile\s*\((.*)\)\s*[{;]',
                  r'\bswitch\s*\((.*)\)\s*{'):
    match = Search(pattern, line)
    if match:
      fncall = match.group(1)    # look inside the parens for function calls
      break

  # Except in if/for/while/switch, there should never be space
  # immediately inside parens (eg "f( 3, 4 )").  We make an exception
  # for nested parens ( (a+b) + c ).  Likewise, there should never be
  # a space before a ( when it's a function argument.  I assume it's a
  # function argument when the char before the whitespace is legal in
  # a function name (alnum + _) and we're not starting a macro. Also ignore
  # pointers and references to arrays and functions coz they're too tricky:
  # we use a very simple way to recognize these:
  # " (something)(maybe-something)" or
  # " (something)(maybe-something," or
  # " (something)[something]"
  # Note that we assume the contents of [] to be short enough that
  # they'll never need to wrap.
  if (  # Ignore control structures.
James Zern's avatar
James Zern committed
2228 2229
      not Search(r'\b(if|for|while|switch|return|new|delete|catch|sizeof)\b',
                 fncall) and
2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240
      # Ignore pointers/references to functions.
      not Search(r' \([^)]+\)\([^)]*(\)|,$)', fncall) and
      # Ignore pointers/references to arrays.
      not Search(r' \([^)]+\)\[[^\]]+\]', fncall)):
    if Search(r'\w\s*\(\s(?!\s*\\$)', fncall):      # a ( used for a fn call
      error(filename, linenum, 'whitespace/parens', 4,
            'Extra space after ( in function call')
    elif Search(r'\(\s+(?!(\s*\\)|\()', fncall):
      error(filename, linenum, 'whitespace/parens', 2,
            'Extra space after (')
    if (Search(r'\w\s+\(', fncall) and
James Zern's avatar
James Zern committed
2241
        not Search(r'#\s*define|typedef', fncall) and
James Zern's avatar
James Zern committed
2242
        not Search(r'\w\s+\((\w+::)*\*\w+\)\(', fncall)):
2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369
      error(filename, linenum, 'whitespace/parens', 4,
            'Extra space before ( in function call')
    # If the ) is followed only by a newline or a { + newline, assume it's
    # part of a control statement (if/while/etc), and don't complain
    if Search(r'[^)]\s+\)\s*[^{\s]', fncall):
      # If the closing parenthesis is preceded by only whitespaces,
      # try to give a more descriptive error message.
      if Search(r'^\s+\)', fncall):
        error(filename, linenum, 'whitespace/parens', 2,
              'Closing ) should be moved to the previous line')
      else:
        error(filename, linenum, 'whitespace/parens', 2,
              'Extra space before )')


def IsBlankLine(line):
  """Returns true if the given line is blank.

  We consider a line to be blank if the line is empty or consists of
  only white spaces.

  Args:
    line: A line of a string.

  Returns:
    True, if the given line is blank.
  """
  return not line or line.isspace()


def CheckForFunctionLengths(filename, clean_lines, linenum,
                            function_state, error):
  """Reports for long function bodies.

  For an overview why this is done, see:
  http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions

  Uses a simplistic algorithm assuming other style guidelines
  (especially spacing) are followed.
  Only checks unindented functions, so class members are unchecked.
  Trivial bodies are unchecked, so constructors with huge initializer lists
  may be missed.
  Blank/comment lines are not counted so as to avoid encouraging the removal
  of vertical space and comments just to get through a lint check.
  NOLINT *on the last line of a function* disables this check.

  Args:
    filename: The name of the current file.
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
    function_state: Current function name and lines in body so far.
    error: The function to call with any errors found.
  """
  lines = clean_lines.lines
  line = lines[linenum]
  raw = clean_lines.raw_lines
  raw_line = raw[linenum]
  joined_line = ''

  starting_func = False
  regexp = r'(\w(\w|::|\*|\&|\s)*)\('  # decls * & space::name( ...
  match_result = Match(regexp, line)
  if match_result:
    # If the name is all caps and underscores, figure it's a macro and
    # ignore it, unless it's TEST or TEST_F.
    function_name = match_result.group(1).split()[-1]
    if function_name == 'TEST' or function_name == 'TEST_F' or (
        not Match(r'[A-Z_]+$', function_name)):
      starting_func = True

  if starting_func:
    body_found = False
    for start_linenum in xrange(linenum, clean_lines.NumLines()):
      start_line = lines[start_linenum]
      joined_line += ' ' + start_line.lstrip()
      if Search(r'(;|})', start_line):  # Declarations and trivial functions
        body_found = True
        break                              # ... ignore
      elif Search(r'{', start_line):
        body_found = True
        function = Search(r'((\w|:)*)\(', line).group(1)
        if Match(r'TEST', function):    # Handle TEST... macros
          parameter_regexp = Search(r'(\(.*\))', joined_line)
          if parameter_regexp:             # Ignore bad syntax
            function += parameter_regexp.group(1)
        else:
          function += '()'
        function_state.Begin(function)
        break
    if not body_found:
      # No body for the function (or evidence of a non-function) was found.
      error(filename, linenum, 'readability/fn_size', 5,
            'Lint failed to find start of function body.')
  elif Match(r'^\}\s*$', line):  # function end
    function_state.Check(error, filename, linenum)
    function_state.End()
  elif not Match(r'^\s*$', line):
    function_state.Count()  # Count non-blank/non-comment lines.


_RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?')


def CheckComment(comment, filename, linenum, error):
  """Checks for common mistakes in TODO comments.

  Args:
    comment: The text of the comment from the line in question.
    filename: The name of the current file.
    linenum: The number of the line to check.
    error: The function to call with any errors found.
  """
  match = _RE_PATTERN_TODO.match(comment)
  if match:
    # One whitespace is correct; zero whitespace is handled elsewhere.
    leading_whitespace = match.group(1)
    if len(leading_whitespace) > 1:
      error(filename, linenum, 'whitespace/todo', 2,
            'Too many spaces before TODO')

    username = match.group(2)
    if not username:
      error(filename, linenum, 'readability/todo', 2,
            'Missing username in TODO; it should look like '
            '"// TODO(my_username): Stuff."')

    middle_whitespace = match.group(3)
James Zern's avatar
James Zern committed
2370
    # Comparisons made explicit for correctness -- pylint: disable=g-explicit-bool-comparison
2371 2372 2373 2374
    if middle_whitespace != ' ' and middle_whitespace != '':
      error(filename, linenum, 'whitespace/todo', 2,
            'TODO(my_username) should be followed by a space')

James Zern's avatar
James Zern committed
2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426
def CheckAccess(filename, clean_lines, linenum, nesting_state, error):
  """Checks for improper use of DISALLOW* macros.

  Args:
    filename: The name of the current file.
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
    nesting_state: A _NestingState instance which maintains information about
                   the current stack of nested blocks being parsed.
    error: The function to call with any errors found.
  """
  line = clean_lines.elided[linenum]  # get rid of comments and strings

  matched = Match((r'\s*(DISALLOW_COPY_AND_ASSIGN|'
                   r'DISALLOW_EVIL_CONSTRUCTORS|'
                   r'DISALLOW_IMPLICIT_CONSTRUCTORS)'), line)
  if not matched:
    return
  if nesting_state.stack and isinstance(nesting_state.stack[-1], _ClassInfo):
    if nesting_state.stack[-1].access != 'private':
      error(filename, linenum, 'readability/constructors', 3,
            '%s must be in the private: section' % matched.group(1))

  else:
    # Found DISALLOW* macro outside a class declaration, or perhaps it
    # was used inside a function when it should have been part of the
    # class declaration.  We could issue a warning here, but it
    # probably resulted in a compiler error already.
    pass


def FindNextMatchingAngleBracket(clean_lines, linenum, init_suffix):
  """Find the corresponding > to close a template.

  Args:
    clean_lines: A CleansedLines instance containing the file.
    linenum: Current line number.
    init_suffix: Remainder of the current line after the initial <.

  Returns:
    True if a matching bracket exists.
  """
  line = init_suffix
  nesting_stack = ['<']
  while True:
    # Find the next operator that can tell us whether < is used as an
    # opening bracket or as a less-than operator.  We only want to
    # warn on the latter case.
    #
    # We could also check all other operators and terminate the search
    # early, e.g. if we got something like this "a<b+c", the "<" is
    # most likely a less-than operator, but then we will get false
James Zern's avatar
James Zern committed
2427
    # positives for default arguments and other template expressions.
James Zern's avatar
James Zern committed
2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519
    match = Search(r'^[^<>(),;\[\]]*([<>(),;\[\]])(.*)$', line)
    if match:
      # Found an operator, update nesting stack
      operator = match.group(1)
      line = match.group(2)

      if nesting_stack[-1] == '<':
        # Expecting closing angle bracket
        if operator in ('<', '(', '['):
          nesting_stack.append(operator)
        elif operator == '>':
          nesting_stack.pop()
          if not nesting_stack:
            # Found matching angle bracket
            return True
        elif operator == ',':
          # Got a comma after a bracket, this is most likely a template
          # argument.  We have not seen a closing angle bracket yet, but
          # it's probably a few lines later if we look for it, so just
          # return early here.
          return True
        else:
          # Got some other operator.
          return False

      else:
        # Expecting closing parenthesis or closing bracket
        if operator in ('<', '(', '['):
          nesting_stack.append(operator)
        elif operator in (')', ']'):
          # We don't bother checking for matching () or [].  If we got
          # something like (] or [), it would have been a syntax error.
          nesting_stack.pop()

    else:
      # Scan the next line
      linenum += 1
      if linenum >= len(clean_lines.elided):
        break
      line = clean_lines.elided[linenum]

  # Exhausted all remaining lines and still no matching angle bracket.
  # Most likely the input was incomplete, otherwise we should have
  # seen a semicolon and returned early.
  return True


def FindPreviousMatchingAngleBracket(clean_lines, linenum, init_prefix):
  """Find the corresponding < that started a template.

  Args:
    clean_lines: A CleansedLines instance containing the file.
    linenum: Current line number.
    init_prefix: Part of the current line before the initial >.

  Returns:
    True if a matching bracket exists.
  """
  line = init_prefix
  nesting_stack = ['>']
  while True:
    # Find the previous operator
    match = Search(r'^(.*)([<>(),;\[\]])[^<>(),;\[\]]*$', line)
    if match:
      # Found an operator, update nesting stack
      operator = match.group(2)
      line = match.group(1)

      if nesting_stack[-1] == '>':
        # Expecting opening angle bracket
        if operator in ('>', ')', ']'):
          nesting_stack.append(operator)
        elif operator == '<':
          nesting_stack.pop()
          if not nesting_stack:
            # Found matching angle bracket
            return True
        elif operator == ',':
          # Got a comma before a bracket, this is most likely a
          # template argument.  The opening angle bracket is probably
          # there if we look for it, so just return early here.
          return True
        else:
          # Got some other operator.
          return False

      else:
        # Expecting opening parenthesis or opening bracket
        if operator in ('>', ')', ']'):
          nesting_stack.append(operator)
        elif operator in ('(', '['):
          nesting_stack.pop()
2520

James Zern's avatar
James Zern committed
2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532
    else:
      # Scan the previous line
      linenum -= 1
      if linenum < 0:
        break
      line = clean_lines.elided[linenum]

  # Exhausted all earlier lines and still no matching angle bracket.
  return False


def CheckSpacing(filename, clean_lines, linenum, nesting_state, error):
2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544
  """Checks for the correctness of various spacing issues in the code.

  Things we check for: spaces around operators, spaces after
  if/for/while/switch, no spaces around parens in function calls, two
  spaces between code and comment, don't start a block with a blank
  line, don't end a function with a blank line, don't add a blank line
  after public/protected/private, don't have too many blank lines in a row.

  Args:
    filename: The name of the current file.
    clean_lines: A CleansedLines instance containing the file.
    linenum: The number of the line to check.
James Zern's avatar
James Zern committed
2545 2546
    nesting_state: A _NestingState instance which maintains information about
                   the current stack of nested blocks being parsed.
2547 2548 2549
    error: The function to call with any errors found.
  """

James Zern's avatar
James Zern committed
2550 2551 2552 2553
  # Don't use "elided" lines here, otherwise we can't check commented lines.
  # Don't want to use "raw" either, because we don't want to check inside C++11
  # raw strings,
  raw = clean_lines.lines_without_raw_strings
2554 2555 2556 2557 2558
  line = raw[linenum]

  # Before nixing comments, check if the line is blank for no good
  # reason.  This includes the first line after a block is opened, and
  # blank lines at the end of a function (ie, right before a line like '}'
James Zern's avatar
James Zern committed
2559 2560 2561 2562 2563 2564 2565 2566 2567 2568
  #
  # Skip all the blank line checks if we are immediately inside a
  # namespace body.  In other words, don't issue blank line warnings
  # for this block:
  #   namespace {
  #
  #   }
  #
  # A warning about missing end of namespace comments will be issued instead.
  if IsBlankLine(line) and not nesting_state.InNamespaceBody():
2569 2570 2571 2572 2573 2574 2575
    elided = clean_lines.elided
    prev_line = elided[linenum - 1]
    prevbrace = prev_line.rfind('{')
    # TODO(unknown): Don't complain if line before blank line, and line after,
    #                both start with alnums and are indented the same amount.
    #                This ignores whitespace at the start of a namespace block
    #                because those are not usually indented.
James Zern's avatar
James Zern committed
2576
    if prevbrace != -1 and prev_line[prevbrace:].find('}') == -1:
2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606
      # OK, we have a blank line at the start of a code block.  Before we
      # complain, we check if it is an exception to the rule: The previous
      # non-empty line has the parameters of a function header that are indented
      # 4 spaces (because they did not fit in a 80 column line when placed on
      # the same line as the function name).  We also check for the case where
      # the previous line is indented 6 spaces, which may happen when the
      # initializers of a constructor do not fit into a 80 column line.
      exception = False
      if Match(r' {6}\w', prev_line):  # Initializer list?
        # We are looking for the opening column of initializer list, which
        # should be indented 4 spaces to cause 6 space indentation afterwards.
        search_position = linenum-2
        while (search_position >= 0
               and Match(r' {6}\w', elided[search_position])):
          search_position -= 1
        exception = (search_position >= 0
                     and elided[search_position][:5] == '    :')
      else:
        # Search for the function arguments or an initializer list.  We use a
        # simple heuristic here: If the line is indented 4 spaces; and we have a
        # closing paren, without the opening paren, followed by an opening brace
        # or colon (for initializer lists) we assume that it is the last line of
        # a function header.  If we have a colon indented 4 spaces, it is an
        # initializer list.
        exception = (Match(r' {4}\w[^\(]*\)\s*(const\s*)?(\{\s*$|:)',
                           prev_line)
                     or Match(r' {4}:', prev_line))

      if not exception:
        error(filename, linenum, 'whitespace/blank_line', 2,
James Zern's avatar
James Zern committed
2607 2608
              'Redundant blank line at the start of a code block '
              'should be deleted.')
James Zern's avatar
James Zern committed
2609
    # Ignore blank lines at the end of a block in a long if-else
2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622
    # chain, like this:
    #   if (condition1) {
    #     // Something followed by a blank line
    #
    #   } else if (condition2) {
    #     // Something else
    #   }
    if linenum + 1 < clean_lines.NumLines():
      next_line = raw[linenum + 1]
      if (next_line
          and Match(r'\s*}', next_line)
          and next_line.find('} else ') == -1):
        error(filename, linenum, 'whitespace/blank_line', 3,
James Zern's avatar
James Zern committed
2623 2624
              'Redundant blank line at the end of a code block '
              'should be deleted.')
2625 2626 2627 2628 2629 2630 2631