Mercurial Repositories > eric / file comparison
comparison: UtilitiesPython2/Tabnanny.py
UtilitiesPython2/Tabnanny.py
- changeset 805
- 83ca4d1ff648
- child 945
- 8cd4d08fa9f6
equal
deleted
inserted
replaced
| 804:3465556892de | 805:83ca4d1ff648 |
|---|---|
| 1 # -*- coding: utf-8 -*- | |
| 2 | |
| 3 """ | |
| 4 The Tab Nanny despises ambiguous indentation. She knows no mercy. | |
| 5 | |
| 6 tabnanny -- Detection of ambiguous indentation | |
| 7 | |
| 8 For the time being this module is intended to be called as a script. | |
| 9 However it is possible to import it into an IDE and use the function | |
| 10 check() described below. | |
| 11 | |
| 12 Warning: The API provided by this module is likely to change in future | |
| 13 releases; such changes may not be backward compatible. | |
| 14 | |
| 15 This is a modified version to make the original tabnanny better suitable | |
| 16 for being called from within the eric5 IDE. | |
| 17 | |
| 18 @exception ValueError The tokenize module is too old. | |
| 19 """ | |
| 20 | |
| 21 # Released to the public domain, by Tim Peters, 15 April 1998. | |
| 22 | |
| 23 # XXX Note: this is now a standard library module. | |
| 24 # XXX The API needs to undergo changes however; the current code is too | |
| 25 # XXX script-like. This will be addressed later. | |
| 26 | |
| 27 # | |
| 28 # This is a modified version to make the original tabnanny better suitable | |
| 29 # for being called from within the eric4 IDE. The modifications are as | |
| 30 # follows: | |
| 31 # | |
| 32 # - there is no main function anymore | |
| 33 # - check function has been modified to only accept a filename and return | |
| 34 # a tuple indicating status (1 = an error was found), the filename, the | |
| 35 # linenumber and the error message (boolean, string, string, string). The | |
| 36 # values are only valid, if the status equals 1. | |
| 37 # | |
| 38 # Modifications copyright (c) 2003 Detlev Offenbach <detlev@die-offenbachs.de> | |
| 39 # | |
| 40 | |
| 41 __version__ = "6_eric5" | |
| 42 | |
| 43 import tokenize | |
| 44 import cStringIO | |
| 45 | |
| 46 if not hasattr(tokenize, 'NL'): | |
| 47 raise ValueError("tokenize.NL doesn't exist -- tokenize module too old") | |
| 48 | |
| 49 __all__ = ["check", "NannyNag", "process_tokens"] | |
| 50 | |
| 51 class NannyNag(Exception): | |
| 52 """ | |
| 53 Raised by tokeneater() if detecting an ambiguous indent. | |
| 54 Captured and handled in check(). | |
| 55 """ | |
| 56 def __init__(self, lineno, msg, line): | |
| 57 """ | |
| 58 Constructor | |
| 59 | |
| 60 @param lineno Line number of the ambiguous indent. | |
| 61 @param msg Descriptive message assigned to this problem. | |
| 62 @param line The offending source line. | |
| 63 """ | |
| 64 self.lineno, self.msg, self.line = lineno, msg, line | |
| 65 | |
| 66 def get_lineno(self): | |
| 67 """ | |
| 68 Method to retrieve the line number. | |
| 69 | |
| 70 @return The line number (integer) | |
| 71 """ | |
| 72 return self.lineno | |
| 73 | |
| 74 def get_msg(self): | |
| 75 """ | |
| 76 Method to retrieve the message. | |
| 77 | |
| 78 @return The error message (string) | |
| 79 """ | |
| 80 return self.msg | |
| 81 | |
| 82 def get_line(self): | |
| 83 """ | |
| 84 Method to retrieve the offending line. | |
| 85 | |
| 86 @return The line of code (string) | |
| 87 """ | |
| 88 return self.line | |
| 89 | |
| 90 def check(filename, codestring): | |
| 91 """ | |
| 92 Private function to check one Python source file for whitespace related problems. | |
| 93 | |
| 94 @param filename source filename (string) | |
| 95 @param codestring source code (string) | |
| 96 @return A tuple indicating status (True = an error was found), the | |
| 97 filename, the linenumber and the error message | |
| 98 (boolean, string, string, string). The values are only | |
| 99 valid, if the status is True. | |
| 100 """ | |
| 101 global indents, check_equal | |
| 102 indents = [Whitespace("")] | |
| 103 check_equal = 0 | |
| 104 | |
| 105 source = cStringIO.StringIO(codestring) | |
| 106 try: | |
| 107 process_tokens(tokenize.generate_tokens(source.readline)) | |
| 108 | |
| 109 except tokenize.TokenError, msg: | |
| 110 return (True, filename, "1", "Token Error: %s" % unicode(msg)) | |
| 111 | |
| 112 except IndentationError, err: | |
| 113 return (True, filename, err.lineno, "Indentation Error: %s" % unicode(err.msg)) | |
| 114 | |
| 115 except NannyNag, nag: | |
| 116 badline = nag.get_lineno() | |
| 117 line = nag.get_line() | |
| 118 return (True, filename, str(badline), line) | |
| 119 | |
| 120 except Exception, err: | |
| 121 return (True, filename, "1", "Unspecific Error: %s" % unicode(err)) | |
| 122 | |
| 123 return (False, None, None, None) | |
| 124 | |
| 125 class Whitespace(object): | |
| 126 """ | |
| 127 Class implementing the whitespace checker. | |
| 128 """ | |
| 129 # the characters used for space and tab | |
| 130 S, T = ' \t' | |
| 131 | |
| 132 # members: | |
| 133 # raw | |
| 134 # the original string | |
| 135 # n | |
| 136 # the number of leading whitespace characters in raw | |
| 137 # nt | |
| 138 # the number of tabs in raw[:n] | |
| 139 # norm | |
| 140 # the normal form as a pair (count, trailing), where: | |
| 141 # count | |
| 142 # a tuple such that raw[:n] contains count[i] | |
| 143 # instances of S * i + T | |
| 144 # trailing | |
| 145 # the number of trailing spaces in raw[:n] | |
| 146 # It's A Theorem that m.indent_level(t) == | |
| 147 # n.indent_level(t) for all t >= 1 iff m.norm == n.norm. | |
| 148 # is_simple | |
| 149 # true iff raw[:n] is of the form (T*)(S*) | |
| 150 | |
| 151 def __init__(self, ws): | |
| 152 """ | |
| 153 Constructor | |
| 154 | |
| 155 @param ws The string to be checked. | |
| 156 """ | |
| 157 self.raw = ws | |
| 158 S, T = Whitespace.S, Whitespace.T | |
| 159 count = [] | |
| 160 b = n = nt = 0 | |
| 161 for ch in self.raw: | |
| 162 if ch == S: | |
| 163 n = n + 1 | |
| 164 b = b + 1 | |
| 165 elif ch == T: | |
| 166 n = n + 1 | |
| 167 nt = nt + 1 | |
| 168 if b >= len(count): | |
| 169 count = count + [0] * (b - len(count) + 1) | |
| 170 count[b] = count[b] + 1 | |
| 171 b = 0 | |
| 172 else: | |
| 173 break | |
| 174 self.n = n | |
| 175 self.nt = nt | |
| 176 self.norm = tuple(count), b | |
| 177 self.is_simple = len(count) <= 1 | |
| 178 | |
| 179 # return length of longest contiguous run of spaces (whether or not | |
| 180 # preceding a tab) | |
| 181 def longest_run_of_spaces(self): | |
| 182 """ | |
| 183 Method to calculate the length of longest contiguous run of spaces. | |
| 184 | |
| 185 @return The length of longest contiguous run of spaces (whether or not | |
| 186 preceding a tab) | |
| 187 """ | |
| 188 count, trailing = self.norm | |
| 189 return max(len(count)-1, trailing) | |
| 190 | |
| 191 def indent_level(self, tabsize): | |
| 192 """ | |
| 193 Method to determine the indentation level. | |
| 194 | |
| 195 @param tabsize The length of a tab stop. (integer) | |
| 196 @return indentation level (integer) | |
| 197 """ | |
| 198 # count, il = self.norm | |
| 199 # for i in range(len(count)): | |
| 200 # if count[i]: | |
| 201 # il = il + (i/tabsize + 1)*tabsize * count[i] | |
| 202 # return il | |
| 203 | |
| 204 # quicker: | |
| 205 # il = trailing + sum (i/ts + 1)*ts*count[i] = | |
| 206 # trailing + ts * sum (i/ts + 1)*count[i] = | |
| 207 # trailing + ts * sum i/ts*count[i] + count[i] = | |
| 208 # trailing + ts * [(sum i/ts*count[i]) + (sum count[i])] = | |
| 209 # trailing + ts * [(sum i/ts*count[i]) + num_tabs] | |
| 210 # and note that i/ts*count[i] is 0 when i < ts | |
| 211 | |
| 212 count, trailing = self.norm | |
| 213 il = 0 | |
| 214 for i in range(tabsize, len(count)): | |
| 215 il = il + i/tabsize * count[i] | |
| 216 return trailing + tabsize * (il + self.nt) | |
| 217 | |
| 218 # return true iff self.indent_level(t) == other.indent_level(t) | |
| 219 # for all t >= 1 | |
| 220 def equal(self, other): | |
| 221 """ | |
| 222 Method to compare the indentation levels of two Whitespace objects for equality. | |
| 223 | |
| 224 @param other Whitespace object to compare against. | |
| 225 @return True, if we compare equal against the other Whitespace object. | |
| 226 """ | |
| 227 return self.norm == other.norm | |
| 228 | |
| 229 # return a list of tuples (ts, i1, i2) such that | |
| 230 # i1 == self.indent_level(ts) != other.indent_level(ts) == i2. | |
| 231 # Intended to be used after not self.equal(other) is known, in which | |
| 232 # case it will return at least one witnessing tab size. | |
| 233 def not_equal_witness(self, other): | |
| 234 """ | |
| 235 Method to calculate a tuple of witnessing tab size. | |
| 236 | |
| 237 Intended to be used after not self.equal(other) is known, in which | |
| 238 case it will return at least one witnessing tab size. | |
| 239 | |
| 240 @param other Whitespace object to calculate against. | |
| 241 @return A list of tuples (ts, i1, i2) such that | |
| 242 i1 == self.indent_level(ts) != other.indent_level(ts) == i2. | |
| 243 """ | |
| 244 n = max(self.longest_run_of_spaces(), | |
| 245 other.longest_run_of_spaces()) + 1 | |
| 246 a = [] | |
| 247 for ts in range(1, n+1): | |
| 248 if self.indent_level(ts) != other.indent_level(ts): | |
| 249 a.append( (ts, | |
| 250 self.indent_level(ts), | |
| 251 other.indent_level(ts)) ) | |
| 252 return a | |
| 253 | |
| 254 # Return True iff self.indent_level(t) < other.indent_level(t) | |
| 255 # for all t >= 1. | |
| 256 # The algorithm is due to Vincent Broman. | |
| 257 # Easy to prove it's correct. | |
| 258 # XXXpost that. | |
| 259 # Trivial to prove n is sharp (consider T vs ST). | |
| 260 # Unknown whether there's a faster general way. I suspected so at | |
| 261 # first, but no longer. | |
| 262 # For the special (but common!) case where M and N are both of the | |
| 263 # form (T*)(S*), M.less(N) iff M.len() < N.len() and | |
| 264 # M.num_tabs() <= N.num_tabs(). Proof is easy but kinda long-winded. | |
| 265 # XXXwrite that up. | |
| 266 # Note that M is of the form (T*)(S*) iff len(M.norm[0]) <= 1. | |
| 267 def less(self, other): | |
| 268 """ | |
| 269 Method to compare the indentation level against another Whitespace objects to | |
| 270 be smaller. | |
| 271 | |
| 272 @param other Whitespace object to compare against. | |
| 273 @return True, if we compare less against the other Whitespace object. | |
| 274 """ | |
| 275 if self.n >= other.n: | |
| 276 return False | |
| 277 if self.is_simple and other.is_simple: | |
| 278 return self.nt <= other.nt | |
| 279 n = max(self.longest_run_of_spaces(), | |
| 280 other.longest_run_of_spaces()) + 1 | |
| 281 # the self.n >= other.n test already did it for ts=1 | |
| 282 for ts in range(2, n+1): | |
| 283 if self.indent_level(ts) >= other.indent_level(ts): | |
| 284 return False | |
| 285 return True | |
| 286 | |
| 287 # return a list of tuples (ts, i1, i2) such that | |
| 288 # i1 == self.indent_level(ts) >= other.indent_level(ts) == i2. | |
| 289 # Intended to be used after not self.less(other) is known, in which | |
| 290 # case it will return at least one witnessing tab size. | |
| 291 def not_less_witness(self, other): | |
| 292 """ | |
| 293 Method to calculate a tuple of witnessing tab size. | |
| 294 | |
| 295 Intended to be used after not self.less(other is known, in which | |
| 296 case it will return at least one witnessing tab size. | |
| 297 | |
| 298 @param other Whitespace object to calculate against. | |
| 299 @return A list of tuples (ts, i1, i2) such that | |
| 300 i1 == self.indent_level(ts) >= other.indent_level(ts) == i2. | |
| 301 """ | |
| 302 n = max(self.longest_run_of_spaces(), | |
| 303 other.longest_run_of_spaces()) + 1 | |
| 304 a = [] | |
| 305 for ts in range(1, n+1): | |
| 306 if self.indent_level(ts) >= other.indent_level(ts): | |
| 307 a.append( (ts, | |
| 308 self.indent_level(ts), | |
| 309 other.indent_level(ts)) ) | |
| 310 return a | |
| 311 | |
| 312 def format_witnesses(w): | |
| 313 """ | |
| 314 Function to format the witnesses as a readable string. | |
| 315 | |
| 316 @param w A list of witnesses | |
| 317 @return A formated string of the witnesses. | |
| 318 """ | |
| 319 firsts = map(lambda tup: str(tup[0]), w) | |
| 320 prefix = "at tab size" | |
| 321 if len(w) > 1: | |
| 322 prefix = prefix + "s" | |
| 323 return prefix + " " + ', '.join(firsts) | |
| 324 | |
| 325 def process_tokens(tokens): | |
| 326 """ | |
| 327 Function processing all tokens generated by a tokenizer run. | |
| 328 | |
| 329 @param tokens list of tokens | |
| 330 """ | |
| 331 INDENT = tokenize.INDENT | |
| 332 DEDENT = tokenize.DEDENT | |
| 333 NEWLINE = tokenize.NEWLINE | |
| 334 JUNK = tokenize.COMMENT, tokenize.NL | |
| 335 indents = [Whitespace("")] | |
| 336 check_equal = 0 | |
| 337 | |
| 338 for (type, token, start, end, line) in tokens: | |
| 339 if type == NEWLINE: | |
| 340 # a program statement, or ENDMARKER, will eventually follow, | |
| 341 # after some (possibly empty) run of tokens of the form | |
| 342 # (NL | COMMENT)* (INDENT | DEDENT+)? | |
| 343 # If an INDENT appears, setting check_equal is wrong, and will | |
| 344 # be undone when we see the INDENT. | |
| 345 check_equal = 1 | |
| 346 | |
| 347 elif type == INDENT: | |
| 348 check_equal = 0 | |
| 349 thisguy = Whitespace(token) | |
| 350 if not indents[-1].less(thisguy): | |
| 351 witness = indents[-1].not_less_witness(thisguy) | |
| 352 msg = "indent not greater e.g. " + format_witnesses(witness) | |
| 353 raise NannyNag(start[0], msg, line) | |
| 354 indents.append(thisguy) | |
| 355 | |
| 356 elif type == DEDENT: | |
| 357 # there's nothing we need to check here! what's important is | |
| 358 # that when the run of DEDENTs ends, the indentation of the | |
| 359 # program statement (or ENDMARKER) that triggered the run is | |
| 360 # equal to what's left at the top of the indents stack | |
| 361 | |
| 362 # Ouch! This assert triggers if the last line of the source | |
| 363 # is indented *and* lacks a newline -- then DEDENTs pop out | |
| 364 # of thin air. | |
| 365 # assert check_equal # else no earlier NEWLINE, or an earlier INDENT | |
| 366 check_equal = 1 | |
| 367 | |
| 368 del indents[-1] | |
| 369 | |
| 370 elif check_equal and type not in JUNK: | |
| 371 # this is the first "real token" following a NEWLINE, so it | |
| 372 # must be the first token of the next program statement, or an | |
| 373 # ENDMARKER; the "line" argument exposes the leading whitespace | |
| 374 # for this statement; in the case of ENDMARKER, line is an empty | |
| 375 # string, so will properly match the empty string with which the | |
| 376 # "indents" stack was seeded | |
| 377 check_equal = 0 | |
| 378 thisguy = Whitespace(line) | |
| 379 if not indents[-1].equal(thisguy): | |
| 380 witness = indents[-1].not_equal_witness(thisguy) | |
| 381 msg = "indent not equal e.g. " + format_witnesses(witness) | |
| 382 raise NannyNag(start[0], msg, line) | |
| 383 | |
| 384 # | |
| 385 # eflag: FileType = Python2 |
