1 #!/usr/bin/env python 2 # pep8.py - Check Python source code formatting, according to PEP 8 3 # Copyright (C) 2006-2009 Johann C. Rocholl <johann (at] rocholl.net> 4 # Copyright (C) 2009-2014 Florent Xicluna <florent.xicluna (at] gmail.com> 5 # 6 # Permission is hereby granted, free of charge, to any person 7 # obtaining a copy of this software and associated documentation files 8 # (the "Software"), to deal in the Software without restriction, 9 # including without limitation the rights to use, copy, modify, merge, 10 # publish, distribute, sublicense, and/or sell copies of the Software, 11 # and to permit persons to whom the Software is furnished to do so, 12 # subject to the following conditions: 13 # 14 # The above copyright notice and this permission notice shall be 15 # included in all copies or substantial portions of the Software. 16 # 17 # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 18 # EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 19 # MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND 20 # NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS 21 # BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN 22 # ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN 23 # CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 24 # SOFTWARE. 25 26 r""" 27 Check Python source code formatting, according to PEP 8. 28 29 For usage and a list of options, try this: 30 $ python pep8.py -h 31 32 This program and its regression test suite live here: 33 http://github.com/jcrocholl/pep8 34 35 Groups of errors and warnings: 36 E errors 37 W warnings 38 100 indentation 39 200 whitespace 40 300 blank lines 41 400 imports 42 500 line length 43 600 deprecation 44 700 statements 45 900 syntax error 46 """ 47 from __future__ import with_statement 48 49 __version__ = '1.5.7' 50 51 import os 52 import sys 53 import re 54 import time 55 import inspect 56 import keyword 57 import tokenize 58 from optparse import OptionParser 59 from fnmatch import fnmatch 60 try: 61 from configparser import RawConfigParser 62 from io import TextIOWrapper 63 except ImportError: 64 from ConfigParser import RawConfigParser 65 66 DEFAULT_EXCLUDE = '.svn,CVS,.bzr,.hg,.git,__pycache__' 67 DEFAULT_IGNORE = 'E123,E226,E24' 68 if sys.platform == 'win32': 69 DEFAULT_CONFIG = os.path.expanduser(r'~\.pep8') 70 else: 71 DEFAULT_CONFIG = os.path.join(os.getenv('XDG_CONFIG_HOME') or 72 os.path.expanduser('~/.config'), 'pep8') 73 PROJECT_CONFIG = ('setup.cfg', 'tox.ini', '.pep8') 74 TESTSUITE_PATH = os.path.join(os.path.dirname(__file__), 'testsuite') 75 MAX_LINE_LENGTH = 79 76 REPORT_FORMAT = { 77 'default': '%(path)s:%(row)d:%(col)d: %(code)s %(text)s', 78 'pylint': '%(path)s:%(row)d: [%(code)s] %(text)s', 79 } 80 81 PyCF_ONLY_AST = 1024 82 SINGLETONS = frozenset(['False', 'None', 'True']) 83 KEYWORDS = frozenset(keyword.kwlist + ['print']) - SINGLETONS 84 UNARY_OPERATORS = frozenset(['>>', '**', '*', '+', '-']) 85 ARITHMETIC_OP = frozenset(['**', '*', '/', '//', '+', '-']) 86 WS_OPTIONAL_OPERATORS = ARITHMETIC_OP.union(['^', '&', '|', '<<', '>>', '%']) 87 WS_NEEDED_OPERATORS = frozenset([ 88 '**=', '*=', '/=', '//=', '+=', '-=', '!=', '<>', '<', '>', 89 '%=', '^=', '&=', '|=', '==', '<=', '>=', '<<=', '>>=', '=']) 90 WHITESPACE = frozenset(' \t') 91 NEWLINE = frozenset([tokenize.NL, tokenize.NEWLINE]) 92 SKIP_TOKENS = NEWLINE.union([tokenize.INDENT, tokenize.DEDENT]) 93 # ERRORTOKEN is triggered by backticks in Python 3 94 SKIP_COMMENTS = SKIP_TOKENS.union([tokenize.COMMENT, tokenize.ERRORTOKEN]) 95 BENCHMARK_KEYS = ['directories', 'files', 'logical lines', 'physical lines'] 96 97 INDENT_REGEX = re.compile(r'([ \t]*)') 98 RAISE_COMMA_REGEX = re.compile(r'raise\s+\w+\s*,') 99 RERAISE_COMMA_REGEX = re.compile(r'raise\s+\w+\s*,.*,\s*\w+\s*$') 100 ERRORCODE_REGEX = re.compile(r'\b[A-Z]\d{3}\b') 101 DOCSTRING_REGEX = re.compile(r'u?r?["\']') 102 EXTRANEOUS_WHITESPACE_REGEX = re.compile(r'[[({] | []}),;:]') 103 WHITESPACE_AFTER_COMMA_REGEX = re.compile(r'[,;:]\s*(?: |\t)') 104 COMPARE_SINGLETON_REGEX = re.compile(r'([=!]=)\s*(None|False|True)') 105 COMPARE_NEGATIVE_REGEX = re.compile(r'\b(not)\s+[^[({ ]+\s+(in|is)\s') 106 COMPARE_TYPE_REGEX = re.compile(r'(?:[=!]=|is(?:\s+not)?)\s*type(?:s.\w+Type' 107 r'|\s*\(\s*([^)]*[^ )])\s*\))') 108 KEYWORD_REGEX = re.compile(r'(\s*)\b(?:%s)\b(\s*)' % r'|'.join(KEYWORDS)) 109 OPERATOR_REGEX = re.compile(r'(?:[^,\s])(\s*)(?:[-+*/|!<=>%&^]+)(\s*)') 110 LAMBDA_REGEX = re.compile(r'\blambda\b') 111 HUNK_REGEX = re.compile(r'^@@ -\d+(?:,\d+)? \+(\d+)(?:,(\d+))? @@.*$') 112 113 # Work around Python < 2.6 behaviour, which does not generate NL after 114 # a comment which is on a line by itself. 115 COMMENT_WITH_NL = tokenize.generate_tokens(['#\n'].pop).send(None)[1] == '#\n' 116 117 118 ############################################################################## 119 # Plugins (check functions) for physical lines 120 ############################################################################## 121 122 123 def tabs_or_spaces(physical_line, indent_char): 124 r"""Never mix tabs and spaces. 125 126 The most popular way of indenting Python is with spaces only. The 127 second-most popular way is with tabs only. Code indented with a mixture 128 of tabs and spaces should be converted to using spaces exclusively. When 129 invoking the Python command line interpreter with the -t option, it issues 130 warnings about code that illegally mixes tabs and spaces. When using -tt 131 these warnings become errors. These options are highly recommended! 132 133 Okay: if a == 0:\n a = 1\n b = 1 134 E101: if a == 0:\n a = 1\n\tb = 1 135 """ 136 indent = INDENT_REGEX.match(physical_line).group(1) 137 for offset, char in enumerate(indent): 138 if char != indent_char: 139 return offset, "E101 indentation contains mixed spaces and tabs" 140 141 142 def tabs_obsolete(physical_line): 143 r"""For new projects, spaces-only are strongly recommended over tabs. 144 145 Okay: if True:\n return 146 W191: if True:\n\treturn 147 """ 148 indent = INDENT_REGEX.match(physical_line).group(1) 149 if '\t' in indent: 150 return indent.index('\t'), "W191 indentation contains tabs" 151 152 153 def trailing_whitespace(physical_line): 154 r"""Trailing whitespace is superfluous. 155 156 The warning returned varies on whether the line itself is blank, for easier 157 filtering for those who want to indent their blank lines. 158 159 Okay: spam(1)\n# 160 W291: spam(1) \n# 161 W293: class Foo(object):\n \n bang = 12 162 """ 163 physical_line = physical_line.rstrip('\n') # chr(10), newline 164 physical_line = physical_line.rstrip('\r') # chr(13), carriage return 165 physical_line = physical_line.rstrip('\x0c') # chr(12), form feed, ^L 166 stripped = physical_line.rstrip(' \t\v') 167 if physical_line != stripped: 168 if stripped: 169 return len(stripped), "W291 trailing whitespace" 170 else: 171 return 0, "W293 blank line contains whitespace" 172 173 174 def trailing_blank_lines(physical_line, lines, line_number, total_lines): 175 r"""Trailing blank lines are superfluous. 176 177 Okay: spam(1) 178 W391: spam(1)\n 179 180 However the last line should end with a new line (warning W292). 181 """ 182 if line_number == total_lines: 183 stripped_last_line = physical_line.rstrip() 184 if not stripped_last_line: 185 return 0, "W391 blank line at end of file" 186 if stripped_last_line == physical_line: 187 return len(physical_line), "W292 no newline at end of file" 188 189 190 def maximum_line_length(physical_line, max_line_length, multiline): 191 r"""Limit all lines to a maximum of 79 characters. 192 193 There are still many devices around that are limited to 80 character 194 lines; plus, limiting windows to 80 characters makes it possible to have 195 several windows side-by-side. The default wrapping on such devices looks 196 ugly. Therefore, please limit all lines to a maximum of 79 characters. 197 For flowing long blocks of text (docstrings or comments), limiting the 198 length to 72 characters is recommended. 199 200 Reports error E501. 201 """ 202 line = physical_line.rstrip() 203 length = len(line) 204 if length > max_line_length and not noqa(line): 205 # Special case for long URLs in multi-line docstrings or comments, 206 # but still report the error when the 72 first chars are whitespaces. 207 chunks = line.split() 208 if ((len(chunks) == 1 and multiline) or 209 (len(chunks) == 2 and chunks[0] == '#')) and \ 210 len(line) - len(chunks[-1]) < max_line_length - 7: 211 return 212 if hasattr(line, 'decode'): # Python 2 213 # The line could contain multi-byte characters 214 try: 215 length = len(line.decode('utf-8')) 216 except UnicodeError: 217 pass 218 if length > max_line_length: 219 return (max_line_length, "E501 line too long " 220 "(%d > %d characters)" % (length, max_line_length)) 221 222 223 ############################################################################## 224 # Plugins (check functions) for logical lines 225 ############################################################################## 226 227 228 def blank_lines(logical_line, blank_lines, indent_level, line_number, 229 blank_before, previous_logical, previous_indent_level): 230 r"""Separate top-level function and class definitions with two blank lines. 231 232 Method definitions inside a class are separated by a single blank line. 233 234 Extra blank lines may be used (sparingly) to separate groups of related 235 functions. Blank lines may be omitted between a bunch of related 236 one-liners (e.g. a set of dummy implementations). 237 238 Use blank lines in functions, sparingly, to indicate logical sections. 239 240 Okay: def a():\n pass\n\n\ndef b():\n pass 241 Okay: def a():\n pass\n\n\n# Foo\n# Bar\n\ndef b():\n pass 242 243 E301: class Foo:\n b = 0\n def bar():\n pass 244 E302: def a():\n pass\n\ndef b(n):\n pass 245 E303: def a():\n pass\n\n\n\ndef b(n):\n pass 246 E303: def a():\n\n\n\n pass 247 E304: @decorator\n\ndef a():\n pass 248 """ 249 if line_number < 3 and not previous_logical: 250 return # Don't expect blank lines before the first line 251 if previous_logical.startswith('@'): 252 if blank_lines: 253 yield 0, "E304 blank lines found after function decorator" 254 elif blank_lines > 2 or (indent_level and blank_lines == 2): 255 yield 0, "E303 too many blank lines (%d)" % blank_lines 256 elif logical_line.startswith(('def ', 'class ', '@')): 257 if indent_level: 258 if not (blank_before or previous_indent_level < indent_level or 259 DOCSTRING_REGEX.match(previous_logical)): 260 yield 0, "E301 expected 1 blank line, found 0" 261 elif blank_before != 2: 262 yield 0, "E302 expected 2 blank lines, found %d" % blank_before 263 264 265 def extraneous_whitespace(logical_line): 266 r"""Avoid extraneous whitespace. 267 268 Avoid extraneous whitespace in these situations: 269 - Immediately inside parentheses, brackets or braces. 270 - Immediately before a comma, semicolon, or colon. 271 272 Okay: spam(ham[1], {eggs: 2}) 273 E201: spam( ham[1], {eggs: 2}) 274 E201: spam(ham[ 1], {eggs: 2}) 275 E201: spam(ham[1], { eggs: 2}) 276 E202: spam(ham[1], {eggs: 2} ) 277 E202: spam(ham[1 ], {eggs: 2}) 278 E202: spam(ham[1], {eggs: 2 }) 279 280 E203: if x == 4: print x, y; x, y = y , x 281 E203: if x == 4: print x, y ; x, y = y, x 282 E203: if x == 4 : print x, y; x, y = y, x 283 """ 284 line = logical_line 285 for match in EXTRANEOUS_WHITESPACE_REGEX.finditer(line): 286 text = match.group() 287 char = text.strip() 288 found = match.start() 289 if text == char + ' ': 290 # assert char in '([{' 291 yield found + 1, "E201 whitespace after '%s'" % char 292 elif line[found - 1] != ',': 293 code = ('E202' if char in '}])' else 'E203') # if char in ',;:' 294 yield found, "%s whitespace before '%s'" % (code, char) 295 296 297 def whitespace_around_keywords(logical_line): 298 r"""Avoid extraneous whitespace around keywords. 299 300 Okay: True and False 301 E271: True and False 302 E272: True and False 303 E273: True and\tFalse 304 E274: True\tand False 305 """ 306 for match in KEYWORD_REGEX.finditer(logical_line): 307 before, after = match.groups() 308 309 if '\t' in before: 310 yield match.start(1), "E274 tab before keyword" 311 elif len(before) > 1: 312 yield match.start(1), "E272 multiple spaces before keyword" 313 314 if '\t' in after: 315 yield match.start(2), "E273 tab after keyword" 316 elif len(after) > 1: 317 yield match.start(2), "E271 multiple spaces after keyword" 318 319 320 def missing_whitespace(logical_line): 321 r"""Each comma, semicolon or colon should be followed by whitespace. 322 323 Okay: [a, b] 324 Okay: (3,) 325 Okay: a[1:4] 326 Okay: a[:4] 327 Okay: a[1:] 328 Okay: a[1:4:2] 329 E231: ['a','b'] 330 E231: foo(bar,baz) 331 E231: [{'a':'b'}] 332 """ 333 line = logical_line 334 for index in range(len(line) - 1): 335 char = line[index] 336 if char in ',;:' and line[index + 1] not in WHITESPACE: 337 before = line[:index] 338 if char == ':' and before.count('[') > before.count(']') and \ 339 before.rfind('{') < before.rfind('['): 340 continue # Slice syntax, no space required 341 if char == ',' and line[index + 1] == ')': 342 continue # Allow tuple with only one element: (3,) 343 yield index, "E231 missing whitespace after '%s'" % char 344 345 346 def indentation(logical_line, previous_logical, indent_char, 347 indent_level, previous_indent_level): 348 r"""Use 4 spaces per indentation level. 349 350 For really old code that you don't want to mess up, you can continue to 351 use 8-space tabs. 352 353 Okay: a = 1 354 Okay: if a == 0:\n a = 1 355 E111: a = 1 356 357 Okay: for item in items:\n pass 358 E112: for item in items:\npass 359 360 Okay: a = 1\nb = 2 361 E113: a = 1\n b = 2 362 """ 363 if indent_char == ' ' and indent_level % 4: 364 yield 0, "E111 indentation is not a multiple of four" 365 indent_expect = previous_logical.endswith(':') 366 if indent_expect and indent_level <= previous_indent_level: 367 yield 0, "E112 expected an indented block" 368 if indent_level > previous_indent_level and not indent_expect: 369 yield 0, "E113 unexpected indentation" 370 371 372 def continued_indentation(logical_line, tokens, indent_level, hang_closing, 373 indent_char, noqa, verbose): 374 r"""Continuation lines indentation. 375 376 Continuation lines should align wrapped elements either vertically 377 using Python's implicit line joining inside parentheses, brackets 378 and braces, or using a hanging indent. 379 380 When using a hanging indent these considerations should be applied: 381 - there should be no arguments on the first line, and 382 - further indentation should be used to clearly distinguish itself as a 383 continuation line. 384 385 Okay: a = (\n) 386 E123: a = (\n ) 387 388 Okay: a = (\n 42) 389 E121: a = (\n 42) 390 E122: a = (\n42) 391 E123: a = (\n 42\n ) 392 E124: a = (24,\n 42\n) 393 E125: if (\n b):\n pass 394 E126: a = (\n 42) 395 E127: a = (24,\n 42) 396 E128: a = (24,\n 42) 397 E129: if (a or\n b):\n pass 398 E131: a = (\n 42\n 24) 399 """ 400 first_row = tokens[0][2][0] 401 nrows = 1 + tokens[-1][2][0] - first_row 402 if noqa or nrows == 1: 403 return 404 405 # indent_next tells us whether the next block is indented; assuming 406 # that it is indented by 4 spaces, then we should not allow 4-space 407 # indents on the final continuation line; in turn, some other 408 # indents are allowed to have an extra 4 spaces. 409 indent_next = logical_line.endswith(':') 410 411 row = depth = 0 412 valid_hangs = (4,) if indent_char != '\t' else (4, 8) 413 # remember how many brackets were opened on each line 414 parens = [0] * nrows 415 # relative indents of physical lines 416 rel_indent = [0] * nrows 417 # for each depth, collect a list of opening rows 418 open_rows = [[0]] 419 # for each depth, memorize the hanging indentation 420 hangs = [None] 421 # visual indents 422 indent_chances = {} 423 last_indent = tokens[0][2] 424 visual_indent = None 425 # for each depth, memorize the visual indent column 426 indent = [last_indent[1]] 427 if verbose >= 3: 428 print(">>> " + tokens[0][4].rstrip()) 429 430 for token_type, text, start, end, line in tokens: 431 432 newline = row < start[0] - first_row 433 if newline: 434 row = start[0] - first_row 435 newline = not last_token_multiline and token_type not in NEWLINE 436 437 if newline: 438 # this is the beginning of a continuation line. 439 last_indent = start 440 if verbose >= 3: 441 print("... " + line.rstrip()) 442 443 # record the initial indent. 444 rel_indent[row] = expand_indent(line) - indent_level 445 446 # identify closing bracket 447 close_bracket = (token_type == tokenize.OP and text in ']})') 448 449 # is the indent relative to an opening bracket line? 450 for open_row in reversed(open_rows[depth]): 451 hang = rel_indent[row] - rel_indent[open_row] 452 hanging_indent = hang in valid_hangs 453 if hanging_indent: 454 break 455 if hangs[depth]: 456 hanging_indent = (hang == hangs[depth]) 457 # is there any chance of visual indent? 458 visual_indent = (not close_bracket and hang > 0 and 459 indent_chances.get(start[1])) 460 461 if close_bracket and indent[depth]: 462 # closing bracket for visual indent 463 if start[1] != indent[depth]: 464 yield (start, "E124 closing bracket does not match " 465 "visual indentation") 466 elif close_bracket and not hang: 467 # closing bracket matches indentation of opening bracket's line 468 if hang_closing: 469 yield start, "E133 closing bracket is missing indentation" 470 elif indent[depth] and start[1] < indent[depth]: 471 if visual_indent is not True: 472 # visual indent is broken 473 yield (start, "E128 continuation line " 474 "under-indented for visual indent") 475 elif hanging_indent or (indent_next and rel_indent[row] == 8): 476 # hanging indent is verified 477 if close_bracket and not hang_closing: 478 yield (start, "E123 closing bracket does not match " 479 "indentation of opening bracket's line") 480 hangs[depth] = hang 481 elif visual_indent is True: 482 # visual indent is verified 483 indent[depth] = start[1] 484 elif visual_indent in (text, str): 485 # ignore token lined up with matching one from a previous line 486 pass 487 else: 488 # indent is broken 489 if hang <= 0: 490 error = "E122", "missing indentation or outdented" 491 elif indent[depth]: 492 error = "E127", "over-indented for visual indent" 493 elif not close_bracket and hangs[depth]: 494 error = "E131", "unaligned for hanging indent" 495 else: 496 hangs[depth] = hang 497 if hang > 4: 498 error = "E126", "over-indented for hanging indent" 499 else: 500 error = "E121", "under-indented for hanging indent" 501 yield start, "%s continuation line %s" % error 502 503 # look for visual indenting 504 if (parens[row] and token_type not in (tokenize.NL, tokenize.COMMENT) 505 and not indent[depth]): 506 indent[depth] = start[1] 507 indent_chances[start[1]] = True 508 if verbose >= 4: 509 print("bracket depth %s indent to %s" % (depth, start[1])) 510 # deal with implicit string concatenation 511 elif (token_type in (tokenize.STRING, tokenize.COMMENT) or 512 text in ('u', 'ur', 'b', 'br')): 513 indent_chances[start[1]] = str 514 # special case for the "if" statement because len("if (") == 4 515 elif not indent_chances and not row and not depth and text == 'if': 516 indent_chances[end[1] + 1] = True 517 elif text == ':' and line[end[1]:].isspace(): 518 open_rows[depth].append(row) 519 520 # keep track of bracket depth 521 if token_type == tokenize.OP: 522 if text in '([{': 523 depth += 1 524 indent.append(0) 525 hangs.append(None) 526 if len(open_rows) == depth: 527 open_rows.append([]) 528 open_rows[depth].append(row) 529 parens[row] += 1 530 if verbose >= 4: 531 print("bracket depth %s seen, col %s, visual min = %s" % 532 (depth, start[1], indent[depth])) 533 elif text in ')]}' and depth > 0: 534 # parent indents should not be more than this one 535 prev_indent = indent.pop() or last_indent[1] 536 hangs.pop() 537 for d in range(depth): 538 if indent[d] > prev_indent: 539 indent[d] = 0 540 for ind in list(indent_chances): 541 if ind >= prev_indent: 542 del indent_chances[ind] 543 del open_rows[depth + 1:] 544 depth -= 1 545 if depth: 546 indent_chances[indent[depth]] = True 547 for idx in range(row, -1, -1): 548 if parens[idx]: 549 parens[idx] -= 1 550 break 551 assert len(indent) == depth + 1 552 if start[1] not in indent_chances: 553 # allow to line up tokens 554 indent_chances[start[1]] = text 555 556 last_token_multiline = (start[0] != end[0]) 557 if last_token_multiline: 558 rel_indent[end[0] - first_row] = rel_indent[row] 559 560 if indent_next and expand_indent(line) == indent_level + 4: 561 pos = (start[0], indent[0] + 4) 562 if visual_indent: 563 code = "E129 visually indented line" 564 else: 565 code = "E125 continuation line" 566 yield pos, "%s with same indent as next logical line" % code 567 568 569 def whitespace_before_parameters(logical_line, tokens): 570 r"""Avoid extraneous whitespace. 571 572 Avoid extraneous whitespace in the following situations: 573 - before the open parenthesis that starts the argument list of a 574 function call. 575 - before the open parenthesis that starts an indexing or slicing. 576 577 Okay: spam(1) 578 E211: spam (1) 579 580 Okay: dict['key'] = list[index] 581 E211: dict ['key'] = list[index] 582 E211: dict['key'] = list [index] 583 """ 584 prev_type, prev_text, __, prev_end, __ = tokens[0] 585 for index in range(1, len(tokens)): 586 token_type, text, start, end, __ = tokens[index] 587 if (token_type == tokenize.OP and 588 text in '([' and 589 start != prev_end and 590 (prev_type == tokenize.NAME or prev_text in '}])') and 591 # Syntax "class A (B):" is allowed, but avoid it 592 (index < 2 or tokens[index - 2][1] != 'class') and 593 # Allow "return (a.foo for a in range(5))" 594 not keyword.iskeyword(prev_text)): 595 yield prev_end, "E211 whitespace before '%s'" % text 596 prev_type = token_type 597 prev_text = text 598 prev_end = end 599 600 601 def whitespace_around_operator(logical_line): 602 r"""Avoid extraneous whitespace around an operator. 603 604 Okay: a = 12 + 3 605 E221: a = 4 + 5 606 E222: a = 4 + 5 607 E223: a = 4\t+ 5 608 E224: a = 4 +\t5 609 """ 610 for match in OPERATOR_REGEX.finditer(logical_line): 611 before, after = match.groups() 612 613 if '\t' in before: 614 yield match.start(1), "E223 tab before operator" 615 elif len(before) > 1: 616 yield match.start(1), "E221 multiple spaces before operator" 617 618 if '\t' in after: 619 yield match.start(2), "E224 tab after operator" 620 elif len(after) > 1: 621 yield match.start(2), "E222 multiple spaces after operator" 622 623 624 def missing_whitespace_around_operator(logical_line, tokens): 625 r"""Surround operators with a single space on either side. 626 627 - Always surround these binary operators with a single space on 628 either side: assignment (=), augmented assignment (+=, -= etc.), 629 comparisons (==, <, >, !=, <=, >=, in, not in, is, is not), 630 Booleans (and, or, not). 631 632 - If operators with different priorities are used, consider adding 633 whitespace around the operators with the lowest priorities. 634 635 Okay: i = i + 1 636 Okay: submitted += 1 637 Okay: x = x * 2 - 1 638 Okay: hypot2 = x * x + y * y 639 Okay: c = (a + b) * (a - b) 640 Okay: foo(bar, key='word', *args, **kwargs) 641 Okay: alpha[:-i] 642 643 E225: i=i+1 644 E225: submitted +=1 645 E225: x = x /2 - 1 646 E225: z = x **y 647 E226: c = (a+b) * (a-b) 648 E226: hypot2 = x*x + y*y 649 E227: c = a|b 650 E228: msg = fmt%(errno, errmsg) 651 """ 652 parens = 0 653 need_space = False 654 prev_type = tokenize.OP 655 prev_text = prev_end = None 656 for token_type, text, start, end, line in tokens: 657 if token_type in SKIP_COMMENTS: 658 continue 659 if text in ('(', 'lambda'): 660 parens += 1 661 elif text == ')': 662 parens -= 1 663 if need_space: 664 if start != prev_end: 665 # Found a (probably) needed space 666 if need_space is not True and not need_space[1]: 667 yield (need_space[0], 668 "E225 missing whitespace around operator") 669 need_space = False 670 elif text == '>' and prev_text in ('<', '-'): 671 # Tolerate the "<>" operator, even if running Python 3 672 # Deal with Python 3's annotated return value "->" 673 pass 674 else: 675 if need_space is True or need_space[1]: 676 # A needed trailing space was not found 677 yield prev_end, "E225 missing whitespace around operator" 678 else: 679 code, optype = 'E226', 'arithmetic' 680 if prev_text == '%': 681 code, optype = 'E228', 'modulo' 682 elif prev_text not in ARITHMETIC_OP: 683 code, optype = 'E227', 'bitwise or shift' 684 yield (need_space[0], "%s missing whitespace " 685 "around %s operator" % (code, optype)) 686 need_space = False 687 elif token_type == tokenize.OP and prev_end is not None: 688 if text == '=' and parens: 689 # Allow keyword args or defaults: foo(bar=None). 690 pass 691 elif text in WS_NEEDED_OPERATORS: 692 need_space = True 693 elif text in UNARY_OPERATORS: 694 # Check if the operator is being used as a binary operator 695 # Allow unary operators: -123, -x, +1. 696 # Allow argument unpacking: foo(*args, **kwargs). 697 if (prev_text in '}])' if prev_type == tokenize.OP 698 else prev_text not in KEYWORDS): 699 need_space = None 700 elif text in WS_OPTIONAL_OPERATORS: 701 need_space = None 702 703 if need_space is None: 704 # Surrounding space is optional, but ensure that 705 # trailing space matches opening space 706 need_space = (prev_end, start != prev_end) 707 elif need_space and start == prev_end: 708 # A needed opening space was not found 709 yield prev_end, "E225 missing whitespace around operator" 710 need_space = False 711 prev_type = token_type 712 prev_text = text 713 prev_end = end 714 715 716 def whitespace_around_comma(logical_line): 717 r"""Avoid extraneous whitespace after a comma or a colon. 718 719 Note: these checks are disabled by default 720 721 Okay: a = (1, 2) 722 E241: a = (1, 2) 723 E242: a = (1,\t2) 724 """ 725 line = logical_line 726 for m in WHITESPACE_AFTER_COMMA_REGEX.finditer(line): 727 found = m.start() + 1 728 if '\t' in m.group(): 729 yield found, "E242 tab after '%s'" % m.group()[0] 730 else: 731 yield found, "E241 multiple spaces after '%s'" % m.group()[0] 732 733 734 def whitespace_around_named_parameter_equals(logical_line, tokens): 735 r"""Don't use spaces around the '=' sign in function arguments. 736 737 Don't use spaces around the '=' sign when used to indicate a 738 keyword argument or a default parameter value. 739 740 Okay: def complex(real, imag=0.0): 741 Okay: return magic(r=real, i=imag) 742 Okay: boolean(a == b) 743 Okay: boolean(a != b) 744 Okay: boolean(a <= b) 745 Okay: boolean(a >= b) 746 747 E251: def complex(real, imag = 0.0): 748 E251: return magic(r = real, i = imag) 749 """ 750 parens = 0 751 no_space = False 752 prev_end = None 753 message = "E251 unexpected spaces around keyword / parameter equals" 754 for token_type, text, start, end, line in tokens: 755 if token_type == tokenize.NL: 756 continue 757 if no_space: 758 no_space = False 759 if start != prev_end: 760 yield (prev_end, message) 761 elif token_type == tokenize.OP: 762 if text == '(': 763 parens += 1 764 elif text == ')': 765 parens -= 1 766 elif parens and text == '=': 767 no_space = True 768 if start != prev_end: 769 yield (prev_end, message) 770 prev_end = end 771 772 773 def whitespace_before_comment(logical_line, tokens): 774 r"""Separate inline comments by at least two spaces. 775 776 An inline comment is a comment on the same line as a statement. Inline 777 comments should be separated by at least two spaces from the statement. 778 They should start with a # and a single space. 779 780 Each line of a block comment starts with a # and a single space 781 (unless it is indented text inside the comment). 782 783 Okay: x = x + 1 # Increment x 784 Okay: x = x + 1 # Increment x 785 Okay: # Block comment 786 E261: x = x + 1 # Increment x 787 E262: x = x + 1 #Increment x 788 E262: x = x + 1 # Increment x 789 E265: #Block comment 790 """ 791 prev_end = (0, 0) 792 for token_type, text, start, end, line in tokens: 793 if token_type == tokenize.COMMENT: 794 inline_comment = line[:start[1]].strip() 795 if inline_comment: 796 if prev_end[0] == start[0] and start[1] < prev_end[1] + 2: 797 yield (prev_end, 798 "E261 at least two spaces before inline comment") 799 symbol, sp, comment = text.partition(' ') 800 bad_prefix = symbol not in ('#', '#:') 801 if inline_comment: 802 if bad_prefix or comment[:1].isspace(): 803 yield start, "E262 inline comment should start with '# '" 804 elif bad_prefix: 805 if text.rstrip('#') and (start[0] > 1 or symbol[1] != '!'): 806 yield start, "E265 block comment should start with '# '" 807 elif token_type != tokenize.NL: 808 prev_end = end 809 810 811 def imports_on_separate_lines(logical_line): 812 r"""Imports should usually be on separate lines. 813 814 Okay: import os\nimport sys 815 E401: import sys, os 816 817 Okay: from subprocess import Popen, PIPE 818 Okay: from myclas import MyClass 819 Okay: from foo.bar.yourclass import YourClass 820 Okay: import myclass 821 Okay: import foo.bar.yourclass 822 """ 823 line = logical_line 824 if line.startswith('import '): 825 found = line.find(',') 826 if -1 < found and ';' not in line[:found]: 827 yield found, "E401 multiple imports on one line" 828 829 830 def compound_statements(logical_line): 831 r"""Compound statements (on the same line) are generally discouraged. 832 833 While sometimes it's okay to put an if/for/while with a small body 834 on the same line, never do this for multi-clause statements. 835 Also avoid folding such long lines! 836 837 Okay: if foo == 'blah':\n do_blah_thing() 838 Okay: do_one() 839 Okay: do_two() 840 Okay: do_three() 841 842 E701: if foo == 'blah': do_blah_thing() 843 E701: for x in lst: total += x 844 E701: while t < 10: t = delay() 845 E701: if foo == 'blah': do_blah_thing() 846 E701: else: do_non_blah_thing() 847 E701: try: something() 848 E701: finally: cleanup() 849 E701: if foo == 'blah': one(); two(); three() 850 851 E702: do_one(); do_two(); do_three() 852 E703: do_four(); # useless semicolon 853 """ 854 line = logical_line 855 last_char = len(line) - 1 856 found = line.find(':') 857 while -1 < found < last_char: 858 before = line[:found] 859 if (before.count('{') <= before.count('}') and # {'a': 1} (dict) 860 before.count('[') <= before.count(']') and # [1:2] (slice) 861 before.count('(') <= before.count(')') and # (Python 3 annotation) 862 not LAMBDA_REGEX.search(before)): # lambda x: x 863 yield found, "E701 multiple statements on one line (colon)" 864 found = line.find(':', found + 1) 865 found = line.find(';') 866 while -1 < found: 867 if found < last_char: 868 yield found, "E702 multiple statements on one line (semicolon)" 869 else: 870 yield found, "E703 statement ends with a semicolon" 871 found = line.find(';', found + 1) 872 873 874 def explicit_line_join(logical_line, tokens): 875 r"""Avoid explicit line join between brackets. 876 877 The preferred way of wrapping long lines is by using Python's implied line 878 continuation inside parentheses, brackets and braces. Long lines can be 879 broken over multiple lines by wrapping expressions in parentheses. These 880 should be used in preference to using a backslash for line continuation. 881 882 E502: aaa = [123, \\n 123] 883 E502: aaa = ("bbb " \\n "ccc") 884 885 Okay: aaa = [123,\n 123] 886 Okay: aaa = ("bbb "\n "ccc") 887 Okay: aaa = "bbb " \\n "ccc" 888 """ 889 prev_start = prev_end = parens = 0 890 for token_type, text, start, end, line in tokens: 891 if start[0] != prev_start and parens and backslash: 892 yield backslash, "E502 the backslash is redundant between brackets" 893 if end[0] != prev_end: 894 if line.rstrip('\r\n').endswith('\\'): 895 backslash = (end[0], len(line.splitlines()[-1]) - 1) 896 else: 897 backslash = None 898 prev_start = prev_end = end[0] 899 else: 900 prev_start = start[0] 901 if token_type == tokenize.OP: 902 if text in '([{': 903 parens += 1 904 elif text in ')]}': 905 parens -= 1 906 907 908 def comparison_to_singleton(logical_line, noqa): 909 r"""Comparison to singletons should use "is" or "is not". 910 911 Comparisons to singletons like None should always be done 912 with "is" or "is not", never the equality operators. 913 914 Okay: if arg is not None: 915 E711: if arg != None: 916 E712: if arg == True: 917 918 Also, beware of writing if x when you really mean if x is not None -- 919 e.g. when testing whether a variable or argument that defaults to None was 920 set to some other value. The other value might have a type (such as a 921 container) that could be false in a boolean context! 922 """ 923 match = not noqa and COMPARE_SINGLETON_REGEX.search(logical_line) 924 if match: 925 same = (match.group(1) == '==') 926 singleton = match.group(2) 927 msg = "'if cond is %s:'" % (('' if same else 'not ') + singleton) 928 if singleton in ('None',): 929 code = 'E711' 930 else: 931 code = 'E712' 932 nonzero = ((singleton == 'True' and same) or 933 (singleton == 'False' and not same)) 934 msg += " or 'if %scond:'" % ('' if nonzero else 'not ') 935 yield match.start(1), ("%s comparison to %s should be %s" % 936 (code, singleton, msg)) 937 938 939 def comparison_negative(logical_line): 940 r"""Negative comparison should be done using "not in" and "is not". 941 942 Okay: if x not in y:\n pass 943 Okay: assert (X in Y or X is Z) 944 Okay: if not (X in Y):\n pass 945 Okay: zz = x is not y 946 E713: Z = not X in Y 947 E713: if not X.B in Y:\n pass 948 E714: if not X is Y:\n pass 949 E714: Z = not X.B is Y 950 """ 951 match = COMPARE_NEGATIVE_REGEX.search(logical_line) 952 if match: 953 pos = match.start(1) 954 if match.group(2) == 'in': 955 yield pos, "E713 test for membership should be 'not in'" 956 else: 957 yield pos, "E714 test for object identity should be 'is not'" 958 959 960 def comparison_type(logical_line): 961 r"""Object type comparisons should always use isinstance(). 962 963 Do not compare types directly. 964 965 Okay: if isinstance(obj, int): 966 E721: if type(obj) is type(1): 967 968 When checking if an object is a string, keep in mind that it might be a 969 unicode string too! In Python 2.3, str and unicode have a common base 970 class, basestring, so you can do: 971 972 Okay: if isinstance(obj, basestring): 973 Okay: if type(a1) is type(b1): 974 """ 975 match = COMPARE_TYPE_REGEX.search(logical_line) 976 if match: 977 inst = match.group(1) 978 if inst and isidentifier(inst) and inst not in SINGLETONS: 979 return # Allow comparison for types which are not obvious 980 yield match.start(), "E721 do not compare types, use 'isinstance()'" 981 982 983 def python_3000_has_key(logical_line, noqa): 984 r"""The {}.has_key() method is removed in Python 3: use the 'in' operator. 985 986 Okay: if "alph" in d:\n print d["alph"] 987 W601: assert d.has_key('alph') 988 """ 989 pos = logical_line.find('.has_key(') 990 if pos > -1 and not noqa: 991 yield pos, "W601 .has_key() is deprecated, use 'in'" 992 993 994 def python_3000_raise_comma(logical_line): 995 r"""When raising an exception, use "raise ValueError('message')". 996 997 The older form is removed in Python 3. 998 999 Okay: raise DummyError("Message") 1000 W602: raise DummyError, "Message" 1001 """ 1002 match = RAISE_COMMA_REGEX.match(logical_line) 1003 if match and not RERAISE_COMMA_REGEX.match(logical_line): 1004 yield match.end() - 1, "W602 deprecated form of raising exception" 1005 1006 1007 def python_3000_not_equal(logical_line): 1008 r"""New code should always use != instead of <>. 1009 1010 The older syntax is removed in Python 3. 1011 1012 Okay: if a != 'no': 1013 W603: if a <> 'no': 1014 """ 1015 pos = logical_line.find('<>') 1016 if pos > -1: 1017 yield pos, "W603 '<>' is deprecated, use '!='" 1018 1019 1020 def python_3000_backticks(logical_line): 1021 r"""Backticks are removed in Python 3: use repr() instead. 1022 1023 Okay: val = repr(1 + 2) 1024 W604: val = `1 + 2` 1025 """ 1026 pos = logical_line.find('`') 1027 if pos > -1: 1028 yield pos, "W604 backticks are deprecated, use 'repr()'" 1029 1030 1031 ############################################################################## 1032 # Helper functions 1033 ############################################################################## 1034 1035 1036 if '' == ''.encode(): 1037 # Python 2: implicit encoding. 1038 def readlines(filename): 1039 """Read the source code.""" 1040 with open(filename, 'rU') as f: 1041 return f.readlines() 1042 isidentifier = re.compile(r'[a-zA-Z_]\w*').match 1043 stdin_get_value = sys.stdin.read 1044 else: 1045 # Python 3 1046 def readlines(filename): 1047 """Read the source code.""" 1048 try: 1049 with open(filename, 'rb') as f: 1050 (coding, lines) = tokenize.detect_encoding(f.readline) 1051 f = TextIOWrapper(f, coding, line_buffering=True) 1052 return [l.decode(coding) for l in lines] + f.readlines() 1053 except (LookupError, SyntaxError, UnicodeError): 1054 # Fall back if file encoding is improperly declared 1055 with open(filename, encoding='latin-1') as f: 1056 return f.readlines() 1057 isidentifier = str.isidentifier 1058 1059 def stdin_get_value(): 1060 return TextIOWrapper(sys.stdin.buffer, errors='ignore').read() 1061 noqa = re.compile(r'# no(?:qa|pep8)\b', re.I).search 1062 1063 1064 def expand_indent(line): 1065 r"""Return the amount of indentation. 1066 1067 Tabs are expanded to the next multiple of 8. 1068 1069 >>> expand_indent(' ') 1070 4 1071 >>> expand_indent('\t') 1072 8 1073 >>> expand_indent(' \t') 1074 8 1075 >>> expand_indent(' \t') 1076 16 1077 """ 1078 if '\t' not in line: 1079 return len(line) - len(line.lstrip()) 1080 result = 0 1081 for char in line: 1082 if char == '\t': 1083 result = result // 8 * 8 + 8 1084 elif char == ' ': 1085 result += 1 1086 else: 1087 break 1088 return result 1089 1090 1091 def mute_string(text): 1092 """Replace contents with 'xxx' to prevent syntax matching. 1093 1094 >>> mute_string('"abc"') 1095 '"xxx"' 1096 >>> mute_string("'''abc'''") 1097 "'''xxx'''" 1098 >>> mute_string("r'abc'") 1099 "r'xxx'" 1100 """ 1101 # String modifiers (e.g. u or r) 1102 start = text.index(text[-1]) + 1 1103 end = len(text) - 1 1104 # Triple quotes 1105 if text[-3:] in ('"""', "'''"): 1106 start += 2 1107 end -= 2 1108 return text[:start] + 'x' * (end - start) + text[end:] 1109 1110 1111 def parse_udiff(diff, patterns=None, parent='.'): 1112 """Return a dictionary of matching lines.""" 1113 # For each file of the diff, the entry key is the filename, 1114 # and the value is a set of row numbers to consider. 1115 rv = {} 1116 path = nrows = None 1117 for line in diff.splitlines(): 1118 if nrows: 1119 if line[:1] != '-': 1120 nrows -= 1 1121 continue 1122 if line[:3] == '@@ ': 1123 hunk_match = HUNK_REGEX.match(line) 1124 (row, nrows) = [int(g or '1') for g in hunk_match.groups()] 1125 rv[path].update(range(row, row + nrows)) 1126 elif line[:3] == '+++': 1127 path = line[4:].split('\t', 1)[0] 1128 if path[:2] == 'b/': 1129 path = path[2:] 1130 rv[path] = set() 1131 return dict([(os.path.join(parent, path), rows) 1132 for (path, rows) in rv.items() 1133 if rows and filename_match(path, patterns)]) 1134 1135 1136 def normalize_paths(value, parent=os.curdir): 1137 """Parse a comma-separated list of paths. 1138 1139 Return a list of absolute paths. 1140 """ 1141 if not value or isinstance(value, list): 1142 return value 1143 paths = [] 1144 for path in value.split(','): 1145 if '/' in path: 1146 path = os.path.abspath(os.path.join(parent, path)) 1147 paths.append(path.rstrip('/')) 1148 return paths 1149 1150 1151 def filename_match(filename, patterns, default=True): 1152 """Check if patterns contains a pattern that matches filename. 1153 1154 If patterns is unspecified, this always returns True. 1155 """ 1156 if not patterns: 1157 return default 1158 return any(fnmatch(filename, pattern) for pattern in patterns) 1159 1160 1161 if COMMENT_WITH_NL: 1162 def _is_eol_token(token): 1163 return (token[0] in NEWLINE or 1164 (token[0] == tokenize.COMMENT and token[1] == token[4])) 1165 else: 1166 def _is_eol_token(token): 1167 return token[0] in NEWLINE 1168 1169 1170 ############################################################################## 1171 # Framework to run all checks 1172 ############################################################################## 1173 1174 1175 _checks = {'physical_line': {}, 'logical_line': {}, 'tree': {}} 1176 1177 1178 def register_check(check, codes=None): 1179 """Register a new check object.""" 1180 def _add_check(check, kind, codes, args): 1181 if check in _checks[kind]: 1182 _checks[kind][check][0].extend(codes or []) 1183 else: 1184 _checks[kind][check] = (codes or [''], args) 1185 if inspect.isfunction(check): 1186 args = inspect.getargspec(check)[0] 1187 if args and args[0] in ('physical_line', 'logical_line'): 1188 if codes is None: 1189 codes = ERRORCODE_REGEX.findall(check.__doc__ or '') 1190 _add_check(check, args[0], codes, args) 1191 elif inspect.isclass(check): 1192 if inspect.getargspec(check.__init__)[0][:2] == ['self', 'tree']: 1193 _add_check(check, 'tree', codes, None) 1194 1195 1196 def init_checks_registry(): 1197 """Register all globally visible functions. 1198 1199 The first argument name is either 'physical_line' or 'logical_line'. 1200 """ 1201 mod = inspect.getmodule(register_check) 1202 for (name, function) in inspect.getmembers(mod, inspect.isfunction): 1203 register_check(function) 1204 init_checks_registry() 1205 1206 1207 class Checker(object): 1208 """Load a Python source file, tokenize it, check coding style.""" 1209 1210 def __init__(self, filename=None, lines=None, 1211 options=None, report=None, **kwargs): 1212 if options is None: 1213 options = StyleGuide(kwargs).options 1214 else: 1215 assert not kwargs 1216 self._io_error = None 1217 self._physical_checks = options.physical_checks 1218 self._logical_checks = options.logical_checks 1219 self._ast_checks = options.ast_checks 1220 self.max_line_length = options.max_line_length 1221 self.multiline = False # in a multiline string? 1222 self.hang_closing = options.hang_closing 1223 self.verbose = options.verbose 1224 self.filename = filename 1225 if filename is None: 1226 self.filename = 'stdin' 1227 self.lines = lines or [] 1228 elif filename == '-': 1229 self.filename = 'stdin' 1230 self.lines = stdin_get_value().splitlines(True) 1231 elif lines is None: 1232 try: 1233 self.lines = readlines(filename) 1234 except IOError: 1235 (exc_type, exc) = sys.exc_info()[:2] 1236 self._io_error = '%s: %s' % (exc_type.__name__, exc) 1237 self.lines = [] 1238 else: 1239 self.lines = lines 1240 if self.lines: 1241 ord0 = ord(self.lines[0][0]) 1242 if ord0 in (0xef, 0xfeff): # Strip the UTF-8 BOM 1243 if ord0 == 0xfeff: 1244 self.lines[0] = self.lines[0][1:] 1245 elif self.lines[0][:3] == '\xef\xbb\xbf': 1246 self.lines[0] = self.lines[0][3:] 1247 self.report = report or options.report 1248 self.report_error = self.report.error 1249 1250 def report_invalid_syntax(self): 1251 """Check if the syntax is valid.""" 1252 (exc_type, exc) = sys.exc_info()[:2] 1253 if len(exc.args) > 1: 1254 offset = exc.args[1] 1255 if len(offset) > 2: 1256 offset = offset[1:3] 1257 else: 1258 offset = (1, 0) 1259 self.report_error(offset[0], offset[1] or 0, 1260 'E901 %s: %s' % (exc_type.__name__, exc.args[0]), 1261 self.report_invalid_syntax) 1262 1263 def readline(self): 1264 """Get the next line from the input buffer.""" 1265 if self.line_number >= self.total_lines: 1266 return '' 1267 line = self.lines[self.line_number] 1268 self.line_number += 1 1269 if self.indent_char is None and line[:1] in WHITESPACE: 1270 self.indent_char = line[0] 1271 return line 1272 1273 def run_check(self, check, argument_names): 1274 """Run a check plugin.""" 1275 arguments = [] 1276 for name in argument_names: 1277 arguments.append(getattr(self, name)) 1278 return check(*arguments) 1279 1280 def check_physical(self, line): 1281 """Run all physical checks on a raw input line.""" 1282 self.physical_line = line 1283 for name, check, argument_names in self._physical_checks: 1284 result = self.run_check(check, argument_names) 1285 if result is not None: 1286 (offset, text) = result 1287 self.report_error(self.line_number, offset, text, check) 1288 if text[:4] == 'E101': 1289 self.indent_char = line[0] 1290 1291 def build_tokens_line(self): 1292 """Build a logical line from tokens.""" 1293 logical = [] 1294 comments = [] 1295 length = 0 1296 prev_row = prev_col = mapping = None 1297 for token_type, text, start, end, line in self.tokens: 1298 if token_type in SKIP_TOKENS: 1299 continue 1300 if not mapping: 1301 mapping = [(0, start)] 1302 if token_type == tokenize.COMMENT: 1303 comments.append(text) 1304 continue 1305 if token_type == tokenize.STRING: 1306 text = mute_string(text) 1307 if prev_row: 1308 (start_row, start_col) = start 1309 if prev_row != start_row: # different row 1310 prev_text = self.lines[prev_row - 1][prev_col - 1] 1311 if prev_text == ',' or (prev_text not in '{[(' 1312 and text not in '}])'): 1313 text = ' ' + text 1314 elif prev_col != start_col: # different column 1315 text = line[prev_col:start_col] + text 1316 logical.append(text) 1317 length += len(text) 1318 mapping.append((length, end)) 1319 (prev_row, prev_col) = end 1320 self.logical_line = ''.join(logical) 1321 self.noqa = comments and noqa(''.join(comments)) 1322 return mapping 1323 1324 def check_logical(self): 1325 """Build a line from tokens and run all logical checks on it.""" 1326 self.report.increment_logical_line() 1327 mapping = self.build_tokens_line() 1328 (start_row, start_col) = mapping[0][1] 1329 start_line = self.lines[start_row - 1] 1330 self.indent_level = expand_indent(start_line[:start_col]) 1331 if self.blank_before < self.blank_lines: 1332 self.blank_before = self.blank_lines 1333 if self.verbose >= 2: 1334 print(self.logical_line[:80].rstrip()) 1335 for name, check, argument_names in self._logical_checks: 1336 if self.verbose >= 4: 1337 print(' ' + name) 1338 for offset, text in self.run_check(check, argument_names) or (): 1339 if not isinstance(offset, tuple): 1340 for token_offset, pos in mapping: 1341 if offset <= token_offset: 1342 break 1343 offset = (pos[0], pos[1] + offset - token_offset) 1344 self.report_error(offset[0], offset[1], text, check) 1345 if self.logical_line: 1346 self.previous_indent_level = self.indent_level 1347 self.previous_logical = self.logical_line 1348 self.blank_lines = 0 1349 self.tokens = [] 1350 1351 def check_ast(self): 1352 """Build the file's AST and run all AST checks.""" 1353 try: 1354 tree = compile(''.join(self.lines), '', 'exec', PyCF_ONLY_AST) 1355 except (SyntaxError, TypeError): 1356 return self.report_invalid_syntax() 1357 for name, cls, __ in self._ast_checks: 1358 checker = cls(tree, self.filename) 1359 for lineno, offset, text, check in checker.run(): 1360 if not self.lines or not noqa(self.lines[lineno - 1]): 1361 self.report_error(lineno, offset, text, check) 1362 1363 def generate_tokens(self): 1364 """Tokenize the file, run physical line checks and yield tokens.""" 1365 if self._io_error: 1366 self.report_error(1, 0, 'E902 %s' % self._io_error, readlines) 1367 tokengen = tokenize.generate_tokens(self.readline) 1368 try: 1369 for token in tokengen: 1370 if token[2][0] > self.total_lines: 1371 return 1372 self.maybe_check_physical(token) 1373 yield token 1374 except (SyntaxError, tokenize.TokenError): 1375 self.report_invalid_syntax() 1376 1377 def maybe_check_physical(self, token): 1378 """If appropriate (based on token), check current physical line(s).""" 1379 # Called after every token, but act only on end of line. 1380 if _is_eol_token(token): 1381 # Obviously, a newline token ends a single physical line. 1382 self.check_physical(token[4]) 1383 elif token[0] == tokenize.STRING and '\n' in token[1]: 1384 # Less obviously, a string that contains newlines is a 1385 # multiline string, either triple-quoted or with internal 1386 # newlines backslash-escaped. Check every physical line in the 1387 # string *except* for the last one: its newline is outside of 1388 # the multiline string, so we consider it a regular physical 1389 # line, and will check it like any other physical line. 1390 # 1391 # Subtleties: 1392 # - we don't *completely* ignore the last line; if it contains 1393 # the magical "# noqa" comment, we disable all physical 1394 # checks for the entire multiline string 1395 # - have to wind self.line_number back because initially it 1396 # points to the last line of the string, and we want 1397 # check_physical() to give accurate feedback 1398 if noqa(token[4]): 1399 return 1400 self.multiline = True 1401 self.line_number = token[2][0] 1402 for line in token[1].split('\n')[:-1]: 1403 self.check_physical(line + '\n') 1404 self.line_number += 1 1405 self.multiline = False 1406 1407 def check_all(self, expected=None, line_offset=0): 1408 """Run all checks on the input file.""" 1409 self.report.init_file(self.filename, self.lines, expected, line_offset) 1410 self.total_lines = len(self.lines) 1411 if self._ast_checks: 1412 self.check_ast() 1413 self.line_number = 0 1414 self.indent_char = None 1415 self.indent_level = self.previous_indent_level = 0 1416 self.previous_logical = '' 1417 self.tokens = [] 1418 self.blank_lines = self.blank_before = 0 1419 parens = 0 1420 for token in self.generate_tokens(): 1421 self.tokens.append(token) 1422 token_type, text = token[0:2] 1423 if self.verbose >= 3: 1424 if token[2][0] == token[3][0]: 1425 pos = '[%s:%s]' % (token[2][1] or '', token[3][1]) 1426 else: 1427 pos = 'l.%s' % token[3][0] 1428 print('l.%s\t%s\t%s\t%r' % 1429 (token[2][0], pos, tokenize.tok_name[token[0]], text)) 1430 if token_type == tokenize.OP: 1431 if text in '([{': 1432 parens += 1 1433 elif text in '}])': 1434 parens -= 1 1435 elif not parens: 1436 if token_type in NEWLINE: 1437 if token_type == tokenize.NEWLINE: 1438 self.check_logical() 1439 self.blank_before = 0 1440 elif len(self.tokens) == 1: 1441 # The physical line contains only this token. 1442 self.blank_lines += 1 1443 del self.tokens[0] 1444 else: 1445 self.check_logical() 1446 elif COMMENT_WITH_NL and token_type == tokenize.COMMENT: 1447 if len(self.tokens) == 1: 1448 # The comment also ends a physical line 1449 token = list(token) 1450 token[1] = text.rstrip('\r\n') 1451 token[3] = (token[2][0], token[2][1] + len(token[1])) 1452 self.tokens = [tuple(token)] 1453 self.check_logical() 1454 if self.tokens: 1455 self.check_physical(self.lines[-1]) 1456 self.check_logical() 1457 return self.report.get_file_results() 1458 1459 1460 class BaseReport(object): 1461 """Collect the results of the checks.""" 1462 1463 print_filename = False 1464 1465 def __init__(self, options): 1466 self._benchmark_keys = options.benchmark_keys 1467 self._ignore_code = options.ignore_code 1468 # Results 1469 self.elapsed = 0 1470 self.total_errors = 0 1471 self.counters = dict.fromkeys(self._benchmark_keys, 0) 1472 self.messages = {} 1473 1474 def start(self): 1475 """Start the timer.""" 1476 self._start_time = time.time() 1477 1478 def stop(self): 1479 """Stop the timer.""" 1480 self.elapsed = time.time() - self._start_time 1481 1482 def init_file(self, filename, lines, expected, line_offset): 1483 """Signal a new file.""" 1484 self.filename = filename 1485 self.lines = lines 1486 self.expected = expected or () 1487 self.line_offset = line_offset 1488 self.file_errors = 0 1489 self.counters['files'] += 1 1490 self.counters['physical lines'] += len(lines) 1491 1492 def increment_logical_line(self): 1493 """Signal a new logical line.""" 1494 self.counters['logical lines'] += 1 1495 1496 def error(self, line_number, offset, text, check): 1497 """Report an error, according to options.""" 1498 code = text[:4] 1499 if self._ignore_code(code): 1500 return 1501 if code in self.counters: 1502 self.counters[code] += 1 1503 else: 1504 self.counters[code] = 1 1505 self.messages[code] = text[5:] 1506 # Don't care about expected errors or warnings 1507 if code in self.expected: 1508 return 1509 if self.print_filename and not self.file_errors: 1510 print(self.filename) 1511 self.file_errors += 1 1512 self.total_errors += 1 1513 return code 1514 1515 def get_file_results(self): 1516 """Return the count of errors and warnings for this file.""" 1517 return self.file_errors 1518 1519 def get_count(self, prefix=''): 1520 """Return the total count of errors and warnings.""" 1521 return sum([self.counters[key] 1522 for key in self.messages if key.startswith(prefix)]) 1523 1524 def get_statistics(self, prefix=''): 1525 """Get statistics for message codes that start with the prefix. 1526 1527 prefix='' matches all errors and warnings 1528 prefix='E' matches all errors 1529 prefix='W' matches all warnings 1530 prefix='E4' matches all errors that have to do with imports 1531 """ 1532 return ['%-7s %s %s' % (self.counters[key], key, self.messages[key]) 1533 for key in sorted(self.messages) if key.startswith(prefix)] 1534 1535 def print_statistics(self, prefix=''): 1536 """Print overall statistics (number of errors and warnings).""" 1537 for line in self.get_statistics(prefix): 1538 print(line) 1539 1540 def print_benchmark(self): 1541 """Print benchmark numbers.""" 1542 print('%-7.2f %s' % (self.elapsed, 'seconds elapsed')) 1543 if self.elapsed: 1544 for key in self._benchmark_keys: 1545 print('%-7d %s per second (%d total)' % 1546 (self.counters[key] / self.elapsed, key, 1547 self.counters[key])) 1548 1549 1550 class FileReport(BaseReport): 1551 """Collect the results of the checks and print only the filenames.""" 1552 print_filename = True 1553 1554 1555 class StandardReport(BaseReport): 1556 """Collect and print the results of the checks.""" 1557 1558 def __init__(self, options): 1559 super(StandardReport, self).__init__(options) 1560 self._fmt = REPORT_FORMAT.get(options.format.lower(), 1561 options.format) 1562 self._repeat = options.repeat 1563 self._show_source = options.show_source 1564 self._show_pep8 = options.show_pep8 1565 1566 def init_file(self, filename, lines, expected, line_offset): 1567 """Signal a new file.""" 1568 self._deferred_print = [] 1569 return super(StandardReport, self).init_file( 1570 filename, lines, expected, line_offset) 1571 1572 def error(self, line_number, offset, text, check): 1573 """Report an error, according to options.""" 1574 code = super(StandardReport, self).error(line_number, offset, 1575 text, check) 1576 if code and (self.counters[code] == 1 or self._repeat): 1577 self._deferred_print.append( 1578 (line_number, offset, code, text[5:], check.__doc__)) 1579 return code 1580 1581 def get_file_results(self): 1582 """Print the result and return the overall count for this file.""" 1583 self._deferred_print.sort() 1584 for line_number, offset, code, text, doc in self._deferred_print: 1585 print(self._fmt % { 1586 'path': self.filename, 1587 'row': self.line_offset + line_number, 'col': offset + 1, 1588 'code': code, 'text': text, 1589 }) 1590 if self._show_source: 1591 if line_number > len(self.lines): 1592 line = '' 1593 else: 1594 line = self.lines[line_number - 1] 1595 print(line.rstrip()) 1596 print(re.sub(r'\S', ' ', line[:offset]) + '^') 1597 if self._show_pep8 and doc: 1598 print(' ' + doc.strip()) 1599 return self.file_errors 1600 1601 1602 class DiffReport(StandardReport): 1603 """Collect and print the results for the changed lines only.""" 1604 1605 def __init__(self, options): 1606 super(DiffReport, self).__init__(options) 1607 self._selected = options.selected_lines 1608 1609 def error(self, line_number, offset, text, check): 1610 if line_number not in self._selected[self.filename]: 1611 return 1612 return super(DiffReport, self).error(line_number, offset, text, check) 1613 1614 1615 class StyleGuide(object): 1616 """Initialize a PEP-8 instance with few options.""" 1617 1618 def __init__(self, *args, **kwargs): 1619 # build options from the command line 1620 self.checker_class = kwargs.pop('checker_class', Checker) 1621 parse_argv = kwargs.pop('parse_argv', False) 1622 config_file = kwargs.pop('config_file', None) 1623 parser = kwargs.pop('parser', None) 1624 # build options from dict 1625 options_dict = dict(*args, **kwargs) 1626 arglist = None if parse_argv else options_dict.get('paths', None) 1627 options, self.paths = process_options( 1628 arglist, parse_argv, config_file, parser) 1629 if options_dict: 1630 options.__dict__.update(options_dict) 1631 if 'paths' in options_dict: 1632 self.paths = options_dict['paths'] 1633 1634 self.runner = self.input_file 1635 self.options = options 1636 1637 if not options.reporter: 1638 options.reporter = BaseReport if options.quiet else StandardReport 1639 1640 options.select = tuple(options.select or ()) 1641 if not (options.select or options.ignore or 1642 options.testsuite or options.doctest) and DEFAULT_IGNORE: 1643 # The default choice: ignore controversial checks 1644 options.ignore = tuple(DEFAULT_IGNORE.split(',')) 1645 else: 1646 # Ignore all checks which are not explicitly selected 1647 options.ignore = ('',) if options.select else tuple(options.ignore) 1648 options.benchmark_keys = BENCHMARK_KEYS[:] 1649 options.ignore_code = self.ignore_code 1650 options.physical_checks = self.get_checks('physical_line') 1651 options.logical_checks = self.get_checks('logical_line') 1652 options.ast_checks = self.get_checks('tree') 1653 self.init_report() 1654 1655 def init_report(self, reporter=None): 1656 """Initialize the report instance.""" 1657 self.options.report = (reporter or self.options.reporter)(self.options) 1658 return self.options.report 1659 1660 def check_files(self, paths=None): 1661 """Run all checks on the paths.""" 1662 if paths is None: 1663 paths = self.paths 1664 report = self.options.report 1665 runner = self.runner 1666 report.start() 1667 try: 1668 for path in paths: 1669 if os.path.isdir(path): 1670 self.input_dir(path) 1671 elif not self.excluded(path): 1672 runner(path) 1673 except KeyboardInterrupt: 1674 print('... stopped') 1675 report.stop() 1676 return report 1677 1678 def input_file(self, filename, lines=None, expected=None, line_offset=0): 1679 """Run all checks on a Python source file.""" 1680 if self.options.verbose: 1681 print('checking %s' % filename) 1682 fchecker = self.checker_class( 1683 filename, lines=lines, options=self.options) 1684 return fchecker.check_all(expected=expected, line_offset=line_offset) 1685 1686 def input_dir(self, dirname): 1687 """Check all files in this directory and all subdirectories.""" 1688 dirname = dirname.rstrip('/') 1689 if self.excluded(dirname): 1690 return 0 1691 counters = self.options.report.counters 1692 verbose = self.options.verbose 1693 filepatterns = self.options.filename 1694 runner = self.runner 1695 for root, dirs, files in os.walk(dirname): 1696 if verbose: 1697 print('directory ' + root) 1698 counters['directories'] += 1 1699 for subdir in sorted(dirs): 1700 if self.excluded(subdir, root): 1701 dirs.remove(subdir) 1702 for filename in sorted(files): 1703 # contain a pattern that matches? 1704 if ((filename_match(filename, filepatterns) and 1705 not self.excluded(filename, root))): 1706 runner(os.path.join(root, filename)) 1707 1708 def excluded(self, filename, parent=None): 1709 """Check if the file should be excluded. 1710 1711 Check if 'options.exclude' contains a pattern that matches filename. 1712 """ 1713 if not self.options.exclude: 1714 return False 1715 basename = os.path.basename(filename) 1716 if filename_match(basename, self.options.exclude): 1717 return True 1718 if parent: 1719 filename = os.path.join(parent, filename) 1720 filename = os.path.abspath(filename) 1721 return filename_match(filename, self.options.exclude) 1722 1723 def ignore_code(self, code): 1724 """Check if the error code should be ignored. 1725 1726 If 'options.select' contains a prefix of the error code, 1727 return False. Else, if 'options.ignore' contains a prefix of 1728 the error code, return True. 1729 """ 1730 if len(code) < 4 and any(s.startswith(code) 1731 for s in self.options.select): 1732 return False 1733 return (code.startswith(self.options.ignore) and 1734 not code.startswith(self.options.select)) 1735 1736 def get_checks(self, argument_name): 1737 """Get all the checks for this category. 1738 1739 Find all globally visible functions where the first argument name 1740 starts with argument_name and which contain selected tests. 1741 """ 1742 checks = [] 1743 for check, attrs in _checks[argument_name].items(): 1744 (codes, args) = attrs 1745 if any(not (code and self.ignore_code(code)) for code in codes): 1746 checks.append((check.__name__, check, args)) 1747 return sorted(checks) 1748 1749 1750 def get_parser(prog='pep8', version=__version__): 1751 parser = OptionParser(prog=prog, version=version, 1752 usage="%prog [options] input ...") 1753 parser.config_options = [ 1754 'exclude', 'filename', 'select', 'ignore', 'max-line-length', 1755 'hang-closing', 'count', 'format', 'quiet', 'show-pep8', 1756 'show-source', 'statistics', 'verbose'] 1757 parser.add_option('-v', '--verbose', default=0, action='count', 1758 help="print status messages, or debug with -vv") 1759 parser.add_option('-q', '--quiet', default=0, action='count', 1760 help="report only file names, or nothing with -qq") 1761 parser.add_option('-r', '--repeat', default=True, action='store_true', 1762 help="(obsolete) show all occurrences of the same error") 1763 parser.add_option('--first', action='store_false', dest='repeat', 1764 help="show first occurrence of each error") 1765 parser.add_option('--exclude', metavar='patterns', default=DEFAULT_EXCLUDE, 1766 help="exclude files or directories which match these " 1767 "comma separated patterns (default: %default)") 1768 parser.add_option('--filename', metavar='patterns', default='*.py', 1769 help="when parsing directories, only check filenames " 1770 "matching these comma separated patterns " 1771 "(default: %default)") 1772 parser.add_option('--select', metavar='errors', default='', 1773 help="select errors and warnings (e.g. E,W6)") 1774 parser.add_option('--ignore', metavar='errors', default='', 1775 help="skip errors and warnings (e.g. E4,W)") 1776 parser.add_option('--show-source', action='store_true', 1777 help="show source code for each error") 1778 parser.add_option('--show-pep8', action='store_true', 1779 help="show text of PEP 8 for each error " 1780 "(implies --first)") 1781 parser.add_option('--statistics', action='store_true', 1782 help="count errors and warnings") 1783 parser.add_option('--count', action='store_true', 1784 help="print total number of errors and warnings " 1785 "to standard error and set exit code to 1 if " 1786 "total is not null") 1787 parser.add_option('--max-line-length', type='int', metavar='n', 1788 default=MAX_LINE_LENGTH, 1789 help="set maximum allowed line length " 1790 "(default: %default)") 1791 parser.add_option('--hang-closing', action='store_true', 1792 help="hang closing bracket instead of matching " 1793 "indentation of opening bracket's line") 1794 parser.add_option('--format', metavar='format', default='default', 1795 help="set the error format [default|pylint|<custom>]") 1796 parser.add_option('--diff', action='store_true', 1797 help="report only lines changed according to the " 1798 "unified diff received on STDIN") 1799 group = parser.add_option_group("Testing Options") 1800 if os.path.exists(TESTSUITE_PATH): 1801 group.add_option('--testsuite', metavar='dir', 1802 help="run regression tests from dir") 1803 group.add_option('--doctest', action='store_true', 1804 help="run doctest on myself") 1805 group.add_option('--benchmark', action='store_true', 1806 help="measure processing speed") 1807 return parser 1808 1809 1810 def read_config(options, args, arglist, parser): 1811 """Read both user configuration and local configuration.""" 1812 config = RawConfigParser() 1813 1814 user_conf = options.config 1815 if user_conf and os.path.isfile(user_conf): 1816 if options.verbose: 1817 print('user configuration: %s' % user_conf) 1818 config.read(user_conf) 1819 1820 local_dir = os.curdir 1821 parent = tail = args and os.path.abspath(os.path.commonprefix(args)) 1822 while tail: 1823 if config.read([os.path.join(parent, fn) for fn in PROJECT_CONFIG]): 1824 local_dir = parent 1825 if options.verbose: 1826 print('local configuration: in %s' % parent) 1827 break 1828 (parent, tail) = os.path.split(parent) 1829 1830 pep8_section = parser.prog 1831 if config.has_section(pep8_section): 1832 option_list = dict([(o.dest, o.type or o.action) 1833 for o in parser.option_list]) 1834 1835 # First, read the default values 1836 (new_options, __) = parser.parse_args([]) 1837 1838 # Second, parse the configuration 1839 for opt in config.options(pep8_section): 1840 if opt.replace('_', '-') not in parser.config_options: 1841 print(" unknown option '%s' ignored" % opt) 1842 continue 1843 if options.verbose > 1: 1844 print(" %s = %s" % (opt, config.get(pep8_section, opt))) 1845 normalized_opt = opt.replace('-', '_') 1846 opt_type = option_list[normalized_opt] 1847 if opt_type in ('int', 'count'): 1848 value = config.getint(pep8_section, opt) 1849 elif opt_type == 'string': 1850 value = config.get(pep8_section, opt) 1851 if normalized_opt == 'exclude': 1852 value = normalize_paths(value, local_dir) 1853 else: 1854 assert opt_type in ('store_true', 'store_false') 1855 value = config.getboolean(pep8_section, opt) 1856 setattr(new_options, normalized_opt, value) 1857 1858 # Third, overwrite with the command-line options 1859 (options, __) = parser.parse_args(arglist, values=new_options) 1860 options.doctest = options.testsuite = False 1861 return options 1862 1863 1864 def process_options(arglist=None, parse_argv=False, config_file=None, 1865 parser=None): 1866 """Process options passed either via arglist or via command line args.""" 1867 if not parser: 1868 parser = get_parser() 1869 if not parser.has_option('--config'): 1870 if config_file is True: 1871 config_file = DEFAULT_CONFIG 1872 group = parser.add_option_group("Configuration", description=( 1873 "The project options are read from the [%s] section of the " 1874 "tox.ini file or the setup.cfg file located in any parent folder " 1875 "of the path(s) being processed. Allowed options are: %s." % 1876 (parser.prog, ', '.join(parser.config_options)))) 1877 group.add_option('--config', metavar='path', default=config_file, 1878 help="user config file location (default: %default)") 1879 # Don't read the command line if the module is used as a library. 1880 if not arglist and not parse_argv: 1881 arglist = [] 1882 # If parse_argv is True and arglist is None, arguments are 1883 # parsed from the command line (sys.argv) 1884 (options, args) = parser.parse_args(arglist) 1885 options.reporter = None 1886 1887 if options.ensure_value('testsuite', False): 1888 args.append(options.testsuite) 1889 elif not options.ensure_value('doctest', False): 1890 if parse_argv and not args: 1891 if options.diff or any(os.path.exists(name) 1892 for name in PROJECT_CONFIG): 1893 args = ['.'] 1894 else: 1895 parser.error('input not specified') 1896 options = read_config(options, args, arglist, parser) 1897 options.reporter = parse_argv and options.quiet == 1 and FileReport 1898 1899 options.filename = options.filename and options.filename.split(',') 1900 options.exclude = normalize_paths(options.exclude) 1901 options.select = options.select and options.select.split(',') 1902 options.ignore = options.ignore and options.ignore.split(',') 1903 1904 if options.diff: 1905 options.reporter = DiffReport 1906 stdin = stdin_get_value() 1907 options.selected_lines = parse_udiff(stdin, options.filename, args[0]) 1908 args = sorted(options.selected_lines) 1909 1910 return options, args 1911 1912 1913 def _main(): 1914 """Parse options and run checks on Python source.""" 1915 import signal 1916 1917 # Handle "Broken pipe" gracefully 1918 try: 1919 signal.signal(signal.SIGPIPE, lambda signum, frame: sys.exit(1)) 1920 except AttributeError: 1921 pass # not supported on Windows 1922 1923 pep8style = StyleGuide(parse_argv=True, config_file=True) 1924 options = pep8style.options 1925 if options.doctest or options.testsuite: 1926 from testsuite.support import run_tests 1927 report = run_tests(pep8style) 1928 else: 1929 report = pep8style.check_files() 1930 if options.statistics: 1931 report.print_statistics() 1932 if options.benchmark: 1933 report.print_benchmark() 1934 if options.testsuite and not options.quiet: 1935 report.print_results() 1936 if report.total_errors: 1937 if options.count: 1938 sys.stderr.write(str(report.total_errors) + '\n') 1939 sys.exit(1) 1940 1941 if __name__ == '__main__': 1942 _main() 1943