Tools/idle/PyParse.py

   1 import string
   2 import re
   3 import sys
   4
   5 # Reason last stmt is continued (or C_NONE if it's not).
   6 C_NONE, C_BACKSLASH, C_STRING, C_BRACKET = range(4)
   7
   8 if 0:   # for throwaway debugging output
   9     def dump(*stuff):
  10         sys.__stdout__.write(string.join(map(str, stuff), " ") + "\n")
  11
  12 # Find what looks like the start of a popular stmt.
  13
  14 _synchre = re.compile(r"""
  15     ^
  16     [ \t]*
  17     (?: if
  18     |   for
  19     |   while
  20     |   else
  21     |   def
  22     |   return
  23     |   assert
  24     |   break
  25     |   class
  26     |   continue
  27     |   elif
  28     |   try
  29     |   except
  30     |   raise
  31     |   import
  32     )
  33     \b
  34 """, re.VERBOSE | re.MULTILINE).search
  35
  36 # Match blank line or non-indenting comment line.
  37
  38 _junkre = re.compile(r"""
  39     [ \t]*
  40     (?: \# \S .* )?
  41     \n
  42 """, re.VERBOSE).match
  43
  44 # Match any flavor of string; the terminating quote is optional
  45 # so that we're robust in the face of incomplete program text.
  46
  47 _match_stringre = re.compile(r"""
  48     \""" [^"\\]* (?:
  49                      (?: \\. | "(?!"") )
  50                      [^"\\]*
  51                  )*
  52     (?: \""" )?
  53
  54 |   " [^"\\\n]* (?: \\. [^"\\\n]* )* "?
  55
  56 |   ''' [^'\\]* (?:
  57                    (?: \\. | '(?!'') )
  58                    [^'\\]*
  59                 )*
  60     (?: ''' )?
  61
  62 |   ' [^'\\\n]* (?: \\. [^'\\\n]* )* '?
  63 """, re.VERBOSE | re.DOTALL).match
  64
  65 # Match a line that starts with something interesting;
  66 # used to find the first item of a bracket structure.
  67
  68 _itemre = re.compile(r"""
  69     [ \t]*
  70     [^\s#\\]    # if we match, m.end()-1 is the interesting char
  71 """, re.VERBOSE).match
  72
  73 # Match start of stmts that should be followed by a dedent.
  74
  75 _closere = re.compile(r"""
  76     \s*
  77     (?: return
  78     |   break
  79     |   continue
  80     |   raise
  81     |   pass
  82     )
  83     \b
  84 """, re.VERBOSE).match
  85
  86 # Chew up non-special chars as quickly as possible.  If match is
  87 # successful, m.end() less 1 is the index of the last boring char
  88 # matched.  If match is unsuccessful, the string starts with an
  89 # interesting char.
  90
  91 _chew_ordinaryre = re.compile(r"""
  92     [^[\](){}#'"\\]+
  93 """, re.VERBOSE).match
  94
  95 # Build translation table to map uninteresting chars to "x", open
  96 # brackets to "(", and close brackets to ")".
  97
  98 _tran = ['x'] * 256
  99 for ch in "({[":
 100     _tran[ord(ch)] = '('
 101 for ch in ")}]":
 102     _tran[ord(ch)] = ')'
 103 for ch in "\"'\\\n#":
 104     _tran[ord(ch)] = ch
 105 _tran = string.join(_tran, '')
 106 del ch
 107
 108 try:
 109     UnicodeType = type(unicode(""))
 110 except NameError:
 111     UnicodeType = None
 112
 113 class Parser:
 114
 115     def __init__(self, indentwidth, tabwidth):
 116         self.indentwidth = indentwidth
 117         self.tabwidth = tabwidth
 118
 119     def set_str(self, str):
 120         assert len(str) == 0 or str[-1] == '\n'
 121         if type(str) is UnicodeType:
 122             # The parse functions have no idea what to do with Unicode, so
 123             # replace all Unicode characters with "x".  This is "safe"
 124             # so long as the only characters germane to parsing the structure
 125             # of Python are 7-bit ASCII.  It's *necessary* because Unicode
 126             # strings don't have a .translate() method that supports
 127             # deletechars.
 128             uniphooey = str
 129             str = []
 130             push = str.append
 131             for raw in map(ord, uniphooey):
 132                 push(raw < 127 and chr(raw) or "x")
 133             str = "".join(str)
 134         self.str = str
 135         self.study_level = 0
 136
 137     # Return index of a good place to begin parsing, as close to the
 138     # end of the string as possible.  This will be the start of some
 139     # popular stmt like "if" or "def".  Return None if none found:
 140     # the caller should pass more prior context then, if possible, or
 141     # if not (the entire program text up until the point of interest
 142     # has already been tried) pass 0 to set_lo.
 143     #
 144     # This will be reliable iff given a reliable is_char_in_string
 145     # function, meaning that when it says "no", it's absolutely
 146     # guaranteed that the char is not in a string.
 147     #
 148     # Ack, hack: in the shell window this kills us, because there's
 149     # no way to tell the differences between output, >>> etc and
 150     # user input.  Indeed, IDLE's first output line makes the rest
 151     # look like it's in an unclosed paren!:
 152     # Python 1.5.2 (#0, Apr 13 1999, ...
 153
 154     def find_good_parse_start(self, use_ps1, is_char_in_string=None,
 155                               _rfind=string.rfind,
 156                               _synchre=_synchre):
 157         str, pos = self.str, None
 158         if use_ps1:
 159             # shell window
 160             ps1 = '\n' + sys.ps1
 161             i = _rfind(str, ps1)
 162             if i >= 0:
 163                 pos = i + len(ps1)
 164                 # make it look like there's a newline instead
 165                 # of ps1 at the start -- hacking here once avoids
 166                 # repeated hackery later
 167                 self.str = str[:pos-1] + '\n' + str[pos:]
 168             return pos
 169
 170         # File window -- real work.
 171         if not is_char_in_string:
 172             # no clue -- make the caller pass everything
 173             return None
 174
 175         # Peek back from the end for a good place to start,
 176         # but don't try too often; pos will be left None, or
 177         # bumped to a legitimate synch point.
 178         limit = len(str)
 179         for tries in range(5):
 180             i = _rfind(str, ":\n", 0, limit)
 181             if i < 0:
 182                 break
 183             i = _rfind(str, '\n', 0, i) + 1  # start of colon line
 184             m = _synchre(str, i, limit)
 185             if m and not is_char_in_string(m.start()):
 186                 pos = m.start()
 187                 break
 188             limit = i
 189         if pos is None:
 190             # Nothing looks like a block-opener, or stuff does
 191             # but is_char_in_string keeps returning true; most likely
 192             # we're in or near a giant string, the colorizer hasn't
 193             # caught up enough to be helpful, or there simply *aren't*
 194             # any interesting stmts.  In any of these cases we're
 195             # going to have to parse the whole thing to be sure, so
 196             # give it one last try from the start, but stop wasting
 197             # time here regardless of the outcome.
 198             m = _synchre(str)
 199             if m and not is_char_in_string(m.start()):
 200                 pos = m.start()
 201             return pos
 202
 203         # Peeking back worked; look forward until _synchre no longer
 204         # matches.
 205         i = pos + 1
 206         while 1:
 207             m = _synchre(str, i)
 208             if m:
 209                 s, i = m.span()
 210                 if not is_char_in_string(s):
 211                     pos = s
 212             else:
 213                 break
 214         return pos
 215
 216     # Throw away the start of the string.  Intended to be called with
 217     # find_good_parse_start's result.
 218
 219     def set_lo(self, lo):
 220         assert lo == 0 or self.str[lo-1] == '\n'
 221         if lo > 0:
 222             self.str = self.str[lo:]
 223
 224     # As quickly as humanly possible <wink>, find the line numbers (0-
 225     # based) of the non-continuation lines.
 226     # Creates self.{goodlines, continuation}.
 227
 228     def _study1(self, _replace=string.replace, _find=string.find):
 229         if self.study_level >= 1:
 230             return
 231         self.study_level = 1
 232
 233         # Map all uninteresting characters to "x", all open brackets
 234         # to "(", all close brackets to ")", then collapse runs of
 235         # uninteresting characters.  This can cut the number of chars
 236         # by a factor of 10-40, and so greatly speed the following loop.
 237         str = self.str
 238         str = string.translate(str, _tran)
 239         str = _replace(str, 'xxxxxxxx', 'x')
 240         str = _replace(str, 'xxxx', 'x')
 241         str = _replace(str, 'xx', 'x')
 242         str = _replace(str, 'xx', 'x')
 243         str = _replace(str, '\nx', '\n')
 244         # note that replacing x\n with \n would be incorrect, because
 245         # x may be preceded by a backslash
 246
 247         # March over the squashed version of the program, accumulating
 248         # the line numbers of non-continued stmts, and determining
 249         # whether & why the last stmt is a continuation.
 250         continuation = C_NONE
 251         level = lno = 0     # level is nesting level; lno is line number
 252         self.goodlines = goodlines = [0]
 253         push_good = goodlines.append
 254         i, n = 0, len(str)
 255         while i < n:
 256             ch = str[i]
 257             i = i+1
 258
 259             # cases are checked in decreasing order of frequency
 260             if ch == 'x':
 261                 continue
 262
 263             if ch == '\n':
 264                 lno = lno + 1
 265                 if level == 0:
 266                     push_good(lno)
 267                     # else we're in an unclosed bracket structure
 268                 continue
 269
 270             if ch == '(':
 271                 level = level + 1
 272                 continue
 273
 274             if ch == ')':
 275                 if level:
 276                     level = level - 1
 277                     # else the program is invalid, but we can't complain
 278                 continue
 279
 280             if ch == '"' or ch == "'":
 281                 # consume the string
 282                 quote = ch
 283                 if str[i-1:i+2] == quote * 3:
 284                     quote = quote * 3
 285                 w = len(quote) - 1
 286                 i = i+w
 287                 while i < n:
 288                     ch = str[i]
 289                     i = i+1
 290
 291                     if ch == 'x':
 292                         continue
 293
 294                     if str[i-1:i+w] == quote:
 295                         i = i+w
 296                         break
 297
 298                     if ch == '\n':
 299                         lno = lno + 1
 300                         if w == 0:
 301                             # unterminated single-quoted string
 302                             if level == 0:
 303                                 push_good(lno)
 304                             break
 305                         continue
 306
 307                     if ch == '\\':
 308                         assert i < n
 309                         if str[i] == '\n':
 310                             lno = lno + 1
 311                         i = i+1
 312                         continue
 313
 314                     # else comment char or paren inside string
 315
 316                 else:
 317                     # didn't break out of the loop, so we're still
 318                     # inside a string
 319                     continuation = C_STRING
 320                 continue    # with outer loop
 321
 322             if ch == '#':
 323                 # consume the comment
 324                 i = _find(str, '\n', i)
 325                 assert i >= 0
 326                 continue
 327
 328             assert ch == '\\'
 329             assert i < n
 330             if str[i] == '\n':
 331                 lno = lno + 1
 332                 if i+1 == n:
 333                     continuation = C_BACKSLASH
 334             i = i+1
 335
 336         # The last stmt may be continued for all 3 reasons.
 337         # String continuation takes precedence over bracket
 338         # continuation, which beats backslash continuation.
 339         if continuation != C_STRING and level > 0:
 340             continuation = C_BRACKET
 341         self.continuation = continuation
 342
 343         # Push the final line number as a sentinel value, regardless of
 344         # whether it's continued.
 345         assert (continuation == C_NONE) == (goodlines[-1] == lno)
 346         if goodlines[-1] != lno:
 347             push_good(lno)
 348
 349     def get_continuation_type(self):
 350         self._study1()
 351         return self.continuation
 352
 353     # study1 was sufficient to determine the continuation status,
 354     # but doing more requires looking at every character.  study2
 355     # does this for the last interesting statement in the block.
 356     # Creates:
 357     #     self.stmt_start, stmt_end
 358     #         slice indices of last interesting stmt
 359     #     self.lastch
 360     #         last non-whitespace character before optional trailing
 361     #         comment
 362     #     self.lastopenbracketpos
 363     #         if continuation is C_BRACKET, index of last open bracket
 364
 365     def _study2(self, _rfind=string.rfind, _find=string.find,
 366                       _ws=string.whitespace):
 367         if self.study_level >= 2:
 368             return
 369         self._study1()
 370         self.study_level = 2
 371
 372         # Set p and q to slice indices of last interesting stmt.
 373         str, goodlines = self.str, self.goodlines
 374         i = len(goodlines) - 1
 375         p = len(str)    # index of newest line
 376         while i:
 377             assert p
 378             # p is the index of the stmt at line number goodlines[i].
 379             # Move p back to the stmt at line number goodlines[i-1].
 380             q = p
 381             for nothing in range(goodlines[i-1], goodlines[i]):
 382                 # tricky: sets p to 0 if no preceding newline
 383                 p = _rfind(str, '\n', 0, p-1) + 1
 384             # The stmt str[p:q] isn't a continuation, but may be blank
 385             # or a non-indenting comment line.
 386             if  _junkre(str, p):
 387                 i = i-1
 388             else:
 389                 break
 390         if i == 0:
 391             # nothing but junk!
 392             assert p == 0
 393             q = p
 394         self.stmt_start, self.stmt_end = p, q
 395
 396         # Analyze this stmt, to find the last open bracket (if any)
 397         # and last interesting character (if any).
 398         lastch = ""
 399         stack = []  # stack of open bracket indices
 400         push_stack = stack.append
 401         while p < q:
 402             # suck up all except ()[]{}'"#\\
 403             m = _chew_ordinaryre(str, p, q)
 404             if m:
 405                 # we skipped at least one boring char
 406                 newp = m.end()
 407                 # back up over totally boring whitespace
 408                 i = newp - 1    # index of last boring char
 409                 while i >= p and str[i] in " \t\n":
 410                     i = i-1
 411                 if i >= p:
 412                     lastch = str[i]
 413                 p = newp
 414                 if p >= q:
 415                     break
 416
 417             ch = str[p]
 418
 419             if ch in "([{":
 420                 push_stack(p)
 421                 lastch = ch
 422                 p = p+1
 423                 continue
 424
 425             if ch in ")]}":
 426                 if stack:
 427                     del stack[-1]
 428                 lastch = ch
 429                 p = p+1
 430                 continue
 431
 432             if ch == '"' or ch == "'":
 433                 # consume string
 434                 # Note that study1 did this with a Python loop, but
 435                 # we use a regexp here; the reason is speed in both
 436                 # cases; the string may be huge, but study1 pre-squashed
 437                 # strings to a couple of characters per line.  study1
 438                 # also needed to keep track of newlines, and we don't
 439                 # have to.
 440                 lastch = ch
 441                 p = _match_stringre(str, p, q).end()
 442                 continue
 443
 444             if ch == '#':
 445                 # consume comment and trailing newline
 446                 p = _find(str, '\n', p, q) + 1
 447                 assert p > 0
 448                 continue
 449
 450             assert ch == '\\'
 451             p = p+1     # beyond backslash
 452             assert p < q
 453             if str[p] != '\n':
 454                 # the program is invalid, but can't complain
 455                 lastch = ch + str[p]
 456             p = p+1     # beyond escaped char
 457
 458         # end while p < q:
 459
 460         self.lastch = lastch
 461         if stack:
 462             self.lastopenbracketpos = stack[-1]
 463
 464     # Assuming continuation is C_BRACKET, return the number
 465     # of spaces the next line should be indented.
 466
 467     def compute_bracket_indent(self, _find=string.find):
 468         self._study2()
 469         assert self.continuation == C_BRACKET
 470         j = self.lastopenbracketpos
 471         str = self.str
 472         n = len(str)
 473         origi = i = string.rfind(str, '\n', 0, j) + 1
 474         j = j+1     # one beyond open bracket
 475         # find first list item; set i to start of its line
 476         while j < n:
 477             m = _itemre(str, j)
 478             if m:
 479                 j = m.end() - 1     # index of first interesting char
 480                 extra = 0
 481                 break
 482             else:
 483                 # this line is junk; advance to next line
 484                 i = j = _find(str, '\n', j) + 1
 485         else:
 486             # nothing interesting follows the bracket;
 487             # reproduce the bracket line's indentation + a level
 488             j = i = origi
 489             while str[j] in " \t":
 490                 j = j+1
 491             extra = self.indentwidth
 492         return len(string.expandtabs(str[i:j],
 493                                      self.tabwidth)) + extra
 494
 495     # Return number of physical lines in last stmt (whether or not
 496     # it's an interesting stmt!  this is intended to be called when
 497     # continuation is C_BACKSLASH).
 498
 499     def get_num_lines_in_stmt(self):
 500         self._study1()
 501         goodlines = self.goodlines
 502         return goodlines[-1] - goodlines[-2]
 503
 504     # Assuming continuation is C_BACKSLASH, return the number of spaces
 505     # the next line should be indented.  Also assuming the new line is
 506     # the first one following the initial line of the stmt.
 507
 508     def compute_backslash_indent(self):
 509         self._study2()
 510         assert self.continuation == C_BACKSLASH
 511         str = self.str
 512         i = self.stmt_start
 513         while str[i] in " \t":
 514             i = i+1
 515         startpos = i
 516
 517         # See whether the initial line starts an assignment stmt; i.e.,
 518         # look for an = operator
 519         endpos = string.find(str, '\n', startpos) + 1
 520         found = level = 0
 521         while i < endpos:
 522             ch = str[i]
 523             if ch in "([{":
 524                 level = level + 1
 525                 i = i+1
 526             elif ch in ")]}":
 527                 if level:
 528                     level = level - 1
 529                 i = i+1
 530             elif ch == '"' or ch == "'":
 531                 i = _match_stringre(str, i, endpos).end()
 532             elif ch == '#':
 533                 break
 534             elif level == 0 and ch == '=' and \
 535                    (i == 0 or str[i-1] not in "=<>!") and \
 536                    str[i+1] != '=':
 537                 found = 1
 538                 break
 539             else:
 540                 i = i+1
 541
 542         if found:
 543             # found a legit =, but it may be the last interesting
 544             # thing on the line
 545             i = i+1     # move beyond the =
 546             found = re.match(r"\s*\\", str[i:endpos]) is None
 547
 548         if not found:
 549             # oh well ... settle for moving beyond the first chunk
 550             # of non-whitespace chars
 551             i = startpos
 552             while str[i] not in " \t\n":
 553                 i = i+1
 554
 555         return len(string.expandtabs(str[self.stmt_start :
 556                                          i],
 557                                      self.tabwidth)) + 1
 558
 559     # Return the leading whitespace on the initial line of the last
 560     # interesting stmt.
 561
 562     def get_base_indent_string(self):
 563         self._study2()
 564         i, n = self.stmt_start, self.stmt_end
 565         j = i
 566         str = self.str
 567         while j < n and str[j] in " \t":
 568             j = j + 1
 569         return str[i:j]
 570
 571     # Did the last interesting stmt open a block?
 572
 573     def is_block_opener(self):
 574         self._study2()
 575         return self.lastch == ':'
 576
 577     # Did the last interesting stmt close a block?
 578
 579     def is_block_closer(self):
 580         self._study2()
 581         return _closere(self.str, self.stmt_start) is not None
 582
 583     # index of last open bracket ({[, or None if none
 584     lastopenbracketpos = None
 585
 586     def get_last_open_bracket_pos(self):
 587         self._study2()
 588         return self.lastopenbracketpos