1#!/usr/bin/env python
2#
3# Copyright 2008, Google Inc.
4# All rights reserved.
5#
6# Redistribution and use in source and binary forms, with or without
7# modification, are permitted provided that the following conditions are
8# met:
9#
10#     * Redistributions of source code must retain the above copyright
11# notice, this list of conditions and the following disclaimer.
12#     * Redistributions in binary form must reproduce the above
13# copyright notice, this list of conditions and the following disclaimer
14# in the documentation and/or other materials provided with the
15# distribution.
16#     * Neither the name of Google Inc. nor the names of its
17# contributors may be used to endorse or promote products derived from
18# this software without specific prior written permission.
19#
20# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
23# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
24# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
25# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
26# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
27# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
28# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
29# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
30# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31
32"""pump v0.2.0 - Pretty Useful for Meta Programming.
33
34A tool for preprocessor meta programming.  Useful for generating
35repetitive boilerplate code.  Especially useful for writing C++
36classes, functions, macros, and templates that need to work with
37various number of arguments.
38
39USAGE:
40       pump.py SOURCE_FILE
41
42EXAMPLES:
43       pump.py foo.cc.pump
44         Converts foo.cc.pump to foo.cc.
45
46GRAMMAR:
47       CODE ::= ATOMIC_CODE*
48       ATOMIC_CODE ::= $var ID = EXPRESSION
49           | $var ID = [[ CODE ]]
50           | $range ID EXPRESSION..EXPRESSION
51           | $for ID SEPARATOR [[ CODE ]]
52           | $($)
53           | $ID
54           | $(EXPRESSION)
55           | $if EXPRESSION [[ CODE ]] ELSE_BRANCH
56           | [[ CODE ]]
57           | RAW_CODE
58       SEPARATOR ::= RAW_CODE | EMPTY
59       ELSE_BRANCH ::= $else [[ CODE ]]
60           | $elif EXPRESSION [[ CODE ]] ELSE_BRANCH
61           | EMPTY
62       EXPRESSION has Python syntax.
63"""
64
65from __future__ import print_function
66
67import io
68import os
69import re
70import sys
71
72
73TOKEN_TABLE = [
74    (re.compile(r'\$var\s+'), '$var'),
75    (re.compile(r'\$elif\s+'), '$elif'),
76    (re.compile(r'\$else\s+'), '$else'),
77    (re.compile(r'\$for\s+'), '$for'),
78    (re.compile(r'\$if\s+'), '$if'),
79    (re.compile(r'\$range\s+'), '$range'),
80    (re.compile(r'\$[_A-Za-z]\w*'), '$id'),
81    (re.compile(r'\$\(\$\)'), '$($)'),
82    (re.compile(r'\$'), '$'),
83    (re.compile(r'\[\[\n?'), '[['),
84    (re.compile(r'\]\]\n?'), ']]'),
85    ]
86
87
88class Cursor:
89  """Represents a position (line and column) in a text file."""
90
91  def __init__(self, line=-1, column=-1):
92    self.line = line
93    self.column = column
94
95  def __eq__(self, rhs):
96    return self.line == rhs.line and self.column == rhs.column
97
98  def __ne__(self, rhs):
99    return not self == rhs
100
101  def __lt__(self, rhs):
102    return self.line < rhs.line or (
103        self.line == rhs.line and self.column < rhs.column)
104
105  def __le__(self, rhs):
106    return self < rhs or self == rhs
107
108  def __gt__(self, rhs):
109    return rhs < self
110
111  def __ge__(self, rhs):
112    return rhs <= self
113
114  def __str__(self):
115    if self == Eof():
116      return 'EOF'
117    else:
118      return '%s(%s)' % (self.line + 1, self.column)
119
120  def __add__(self, offset):
121    return Cursor(self.line, self.column + offset)
122
123  def __sub__(self, offset):
124    return Cursor(self.line, self.column - offset)
125
126  def Clone(self):
127    """Returns a copy of self."""
128
129    return Cursor(self.line, self.column)
130
131
132# Special cursor to indicate the end-of-file.
133def Eof():
134  """Returns the special cursor to denote the end-of-file."""
135  return Cursor(-1, -1)
136
137
138class Token:
139  """Represents a token in a Pump source file."""
140
141  def __init__(self, start=None, end=None, value=None, token_type=None):
142    if start is None:
143      self.start = Eof()
144    else:
145      self.start = start
146    if end is None:
147      self.end = Eof()
148    else:
149      self.end = end
150    self.value = value
151    self.token_type = token_type
152
153  def __str__(self):
154    return 'Token @%s: \'%s\' type=%s' % (
155        self.start, self.value, self.token_type)
156
157  def Clone(self):
158    """Returns a copy of self."""
159
160    return Token(self.start.Clone(), self.end.Clone(), self.value,
161                 self.token_type)
162
163
164def StartsWith(lines, pos, string):
165  """Returns True iff the given position in lines starts with 'string'."""
166
167  return lines[pos.line][pos.column:].startswith(string)
168
169
170def FindFirstInLine(line, token_table):
171  best_match_start = -1
172  for (regex, token_type) in token_table:
173    m = regex.search(line)
174    if m:
175      # We found regex in lines
176      if best_match_start < 0 or m.start() < best_match_start:
177        best_match_start = m.start()
178        best_match_length = m.end() - m.start()
179        best_match_token_type = token_type
180
181  if best_match_start < 0:
182    return None
183
184  return (best_match_start, best_match_length, best_match_token_type)
185
186
187def FindFirst(lines, token_table, cursor):
188  """Finds the first occurrence of any string in strings in lines."""
189
190  start = cursor.Clone()
191  cur_line_number = cursor.line
192  for line in lines[start.line:]:
193    if cur_line_number == start.line:
194      line = line[start.column:]
195    m = FindFirstInLine(line, token_table)
196    if m:
197      # We found a regex in line.
198      (start_column, length, token_type) = m
199      if cur_line_number == start.line:
200        start_column += start.column
201      found_start = Cursor(cur_line_number, start_column)
202      found_end = found_start + length
203      return MakeToken(lines, found_start, found_end, token_type)
204    cur_line_number += 1
205  # We failed to find str in lines
206  return None
207
208
209def SubString(lines, start, end):
210  """Returns a substring in lines."""
211
212  if end == Eof():
213    end = Cursor(len(lines) - 1, len(lines[-1]))
214
215  if start >= end:
216    return ''
217
218  if start.line == end.line:
219    return lines[start.line][start.column:end.column]
220
221  result_lines = ([lines[start.line][start.column:]] +
222                  lines[start.line + 1:end.line] +
223                  [lines[end.line][:end.column]])
224  return ''.join(result_lines)
225
226
227def StripMetaComments(str):
228  """Strip meta comments from each line in the given string."""
229
230  # First, completely remove lines containing nothing but a meta
231  # comment, including the trailing \n.
232  str = re.sub(r'^\s*\$\$.*\n', '', str)
233
234  # Then, remove meta comments from contentful lines.
235  return re.sub(r'\s*\$\$.*', '', str)
236
237
238def MakeToken(lines, start, end, token_type):
239  """Creates a new instance of Token."""
240
241  return Token(start, end, SubString(lines, start, end), token_type)
242
243
244def ParseToken(lines, pos, regex, token_type):
245  line = lines[pos.line][pos.column:]
246  m = regex.search(line)
247  if m and not m.start():
248    return MakeToken(lines, pos, pos + m.end(), token_type)
249  else:
250    print('ERROR: %s expected at %s.' % (token_type, pos))
251    sys.exit(1)
252
253
254ID_REGEX = re.compile(r'[_A-Za-z]\w*')
255EQ_REGEX = re.compile(r'=')
256REST_OF_LINE_REGEX = re.compile(r'.*?(?=$|\$\$)')
257OPTIONAL_WHITE_SPACES_REGEX = re.compile(r'\s*')
258WHITE_SPACE_REGEX = re.compile(r'\s')
259DOT_DOT_REGEX = re.compile(r'\.\.')
260
261
262def Skip(lines, pos, regex):
263  line = lines[pos.line][pos.column:]
264  m = re.search(regex, line)
265  if m and not m.start():
266    return pos + m.end()
267  else:
268    return pos
269
270
271def SkipUntil(lines, pos, regex, token_type):
272  line = lines[pos.line][pos.column:]
273  m = re.search(regex, line)
274  if m:
275    return pos + m.start()
276  else:
277    print ('ERROR: %s expected on line %s after column %s.' %
278           (token_type, pos.line + 1, pos.column))
279    sys.exit(1)
280
281
282def ParseExpTokenInParens(lines, pos):
283  def ParseInParens(pos):
284    pos = Skip(lines, pos, OPTIONAL_WHITE_SPACES_REGEX)
285    pos = Skip(lines, pos, r'\(')
286    pos = Parse(pos)
287    pos = Skip(lines, pos, r'\)')
288    return pos
289
290  def Parse(pos):
291    pos = SkipUntil(lines, pos, r'\(|\)', ')')
292    if SubString(lines, pos, pos + 1) == '(':
293      pos = Parse(pos + 1)
294      pos = Skip(lines, pos, r'\)')
295      return Parse(pos)
296    else:
297      return pos
298
299  start = pos.Clone()
300  pos = ParseInParens(pos)
301  return MakeToken(lines, start, pos, 'exp')
302
303
304def RStripNewLineFromToken(token):
305  if token.value.endswith('\n'):
306    return Token(token.start, token.end, token.value[:-1], token.token_type)
307  else:
308    return token
309
310
311def TokenizeLines(lines, pos):
312  while True:
313    found = FindFirst(lines, TOKEN_TABLE, pos)
314    if not found:
315      yield MakeToken(lines, pos, Eof(), 'code')
316      return
317
318    if found.start == pos:
319      prev_token = None
320      prev_token_rstripped = None
321    else:
322      prev_token = MakeToken(lines, pos, found.start, 'code')
323      prev_token_rstripped = RStripNewLineFromToken(prev_token)
324
325    if found.token_type == '$var':
326      if prev_token_rstripped:
327        yield prev_token_rstripped
328      yield found
329      id_token = ParseToken(lines, found.end, ID_REGEX, 'id')
330      yield id_token
331      pos = Skip(lines, id_token.end, OPTIONAL_WHITE_SPACES_REGEX)
332
333      eq_token = ParseToken(lines, pos, EQ_REGEX, '=')
334      yield eq_token
335      pos = Skip(lines, eq_token.end, r'\s*')
336
337      if SubString(lines, pos, pos + 2) != '[[':
338        exp_token = ParseToken(lines, pos, REST_OF_LINE_REGEX, 'exp')
339        yield exp_token
340        pos = Cursor(exp_token.end.line + 1, 0)
341    elif found.token_type == '$for':
342      if prev_token_rstripped:
343        yield prev_token_rstripped
344      yield found
345      id_token = ParseToken(lines, found.end, ID_REGEX, 'id')
346      yield id_token
347      pos = Skip(lines, id_token.end, WHITE_SPACE_REGEX)
348    elif found.token_type == '$range':
349      if prev_token_rstripped:
350        yield prev_token_rstripped
351      yield found
352      id_token = ParseToken(lines, found.end, ID_REGEX, 'id')
353      yield id_token
354      pos = Skip(lines, id_token.end, OPTIONAL_WHITE_SPACES_REGEX)
355
356      dots_pos = SkipUntil(lines, pos, DOT_DOT_REGEX, '..')
357      yield MakeToken(lines, pos, dots_pos, 'exp')
358      yield MakeToken(lines, dots_pos, dots_pos + 2, '..')
359      pos = dots_pos + 2
360      new_pos = Cursor(pos.line + 1, 0)
361      yield MakeToken(lines, pos, new_pos, 'exp')
362      pos = new_pos
363    elif found.token_type == '$':
364      if prev_token:
365        yield prev_token
366      yield found
367      exp_token = ParseExpTokenInParens(lines, found.end)
368      yield exp_token
369      pos = exp_token.end
370    elif (found.token_type == ']]' or found.token_type == '$if' or
371          found.token_type == '$elif' or found.token_type == '$else'):
372      if prev_token_rstripped:
373        yield prev_token_rstripped
374      yield found
375      pos = found.end
376    else:
377      if prev_token:
378        yield prev_token
379      yield found
380      pos = found.end
381
382
383def Tokenize(s):
384  """A generator that yields the tokens in the given string."""
385  if s != '':
386    lines = s.splitlines(True)
387    for token in TokenizeLines(lines, Cursor(0, 0)):
388      yield token
389
390
391class CodeNode:
392  def __init__(self, atomic_code_list=None):
393    self.atomic_code = atomic_code_list
394
395
396class VarNode:
397  def __init__(self, identifier=None, atomic_code=None):
398    self.identifier = identifier
399    self.atomic_code = atomic_code
400
401
402class RangeNode:
403  def __init__(self, identifier=None, exp1=None, exp2=None):
404    self.identifier = identifier
405    self.exp1 = exp1
406    self.exp2 = exp2
407
408
409class ForNode:
410  def __init__(self, identifier=None, sep=None, code=None):
411    self.identifier = identifier
412    self.sep = sep
413    self.code = code
414
415
416class ElseNode:
417  def __init__(self, else_branch=None):
418    self.else_branch = else_branch
419
420
421class IfNode:
422  def __init__(self, exp=None, then_branch=None, else_branch=None):
423    self.exp = exp
424    self.then_branch = then_branch
425    self.else_branch = else_branch
426
427
428class RawCodeNode:
429  def __init__(self, token=None):
430    self.raw_code = token
431
432
433class LiteralDollarNode:
434  def __init__(self, token):
435    self.token = token
436
437
438class ExpNode:
439  def __init__(self, token, python_exp):
440    self.token = token
441    self.python_exp = python_exp
442
443
444def PopFront(a_list):
445  head = a_list[0]
446  a_list[:1] = []
447  return head
448
449
450def PushFront(a_list, elem):
451  a_list[:0] = [elem]
452
453
454def PopToken(a_list, token_type=None):
455  token = PopFront(a_list)
456  if token_type is not None and token.token_type != token_type:
457    print('ERROR: %s expected at %s' % (token_type, token.start))
458    print('ERROR: %s found instead' % (token,))
459    sys.exit(1)
460
461  return token
462
463
464def PeekToken(a_list):
465  if not a_list:
466    return None
467
468  return a_list[0]
469
470
471def ParseExpNode(token):
472  python_exp = re.sub(r'([_A-Za-z]\w*)', r'self.GetValue("\1")', token.value)
473  return ExpNode(token, python_exp)
474
475
476def ParseElseNode(tokens):
477  def Pop(token_type=None):
478    return PopToken(tokens, token_type)
479
480  next = PeekToken(tokens)
481  if not next:
482    return None
483  if next.token_type == '$else':
484    Pop('$else')
485    Pop('[[')
486    code_node = ParseCodeNode(tokens)
487    Pop(']]')
488    return code_node
489  elif next.token_type == '$elif':
490    Pop('$elif')
491    exp = Pop('code')
492    Pop('[[')
493    code_node = ParseCodeNode(tokens)
494    Pop(']]')
495    inner_else_node = ParseElseNode(tokens)
496    return CodeNode([IfNode(ParseExpNode(exp), code_node, inner_else_node)])
497  elif not next.value.strip():
498    Pop('code')
499    return ParseElseNode(tokens)
500  else:
501    return None
502
503
504def ParseAtomicCodeNode(tokens):
505  def Pop(token_type=None):
506    return PopToken(tokens, token_type)
507
508  head = PopFront(tokens)
509  t = head.token_type
510  if t == 'code':
511    return RawCodeNode(head)
512  elif t == '$var':
513    id_token = Pop('id')
514    Pop('=')
515    next = PeekToken(tokens)
516    if next.token_type == 'exp':
517      exp_token = Pop()
518      return VarNode(id_token, ParseExpNode(exp_token))
519    Pop('[[')
520    code_node = ParseCodeNode(tokens)
521    Pop(']]')
522    return VarNode(id_token, code_node)
523  elif t == '$for':
524    id_token = Pop('id')
525    next_token = PeekToken(tokens)
526    if next_token.token_type == 'code':
527      sep_token = next_token
528      Pop('code')
529    else:
530      sep_token = None
531    Pop('[[')
532    code_node = ParseCodeNode(tokens)
533    Pop(']]')
534    return ForNode(id_token, sep_token, code_node)
535  elif t == '$if':
536    exp_token = Pop('code')
537    Pop('[[')
538    code_node = ParseCodeNode(tokens)
539    Pop(']]')
540    else_node = ParseElseNode(tokens)
541    return IfNode(ParseExpNode(exp_token), code_node, else_node)
542  elif t == '$range':
543    id_token = Pop('id')
544    exp1_token = Pop('exp')
545    Pop('..')
546    exp2_token = Pop('exp')
547    return RangeNode(id_token, ParseExpNode(exp1_token),
548                     ParseExpNode(exp2_token))
549  elif t == '$id':
550    return ParseExpNode(Token(head.start + 1, head.end, head.value[1:], 'id'))
551  elif t == '$($)':
552    return LiteralDollarNode(head)
553  elif t == '$':
554    exp_token = Pop('exp')
555    return ParseExpNode(exp_token)
556  elif t == '[[':
557    code_node = ParseCodeNode(tokens)
558    Pop(']]')
559    return code_node
560  else:
561    PushFront(tokens, head)
562    return None
563
564
565def ParseCodeNode(tokens):
566  atomic_code_list = []
567  while True:
568    if not tokens:
569      break
570    atomic_code_node = ParseAtomicCodeNode(tokens)
571    if atomic_code_node:
572      atomic_code_list.append(atomic_code_node)
573    else:
574      break
575  return CodeNode(atomic_code_list)
576
577
578def ParseToAST(pump_src_text):
579  """Convert the given Pump source text into an AST."""
580  tokens = list(Tokenize(pump_src_text))
581  code_node = ParseCodeNode(tokens)
582  return code_node
583
584
585class Env:
586  def __init__(self):
587    self.variables = []
588    self.ranges = []
589
590  def Clone(self):
591    clone = Env()
592    clone.variables = self.variables[:]
593    clone.ranges = self.ranges[:]
594    return clone
595
596  def PushVariable(self, var, value):
597    # If value looks like an int, store it as an int.
598    try:
599      int_value = int(value)
600      if ('%s' % int_value) == value:
601        value = int_value
602    except Exception:
603      pass
604    self.variables[:0] = [(var, value)]
605
606  def PopVariable(self):
607    self.variables[:1] = []
608
609  def PushRange(self, var, lower, upper):
610    self.ranges[:0] = [(var, lower, upper)]
611
612  def PopRange(self):
613    self.ranges[:1] = []
614
615  def GetValue(self, identifier):
616    for (var, value) in self.variables:
617      if identifier == var:
618        return value
619
620    print('ERROR: meta variable %s is undefined.' % (identifier,))
621    sys.exit(1)
622
623  def EvalExp(self, exp):
624    try:
625      result = eval(exp.python_exp)
626    except Exception as e:  # pylint: disable=broad-except
627      print('ERROR: caught exception %s: %s' % (e.__class__.__name__, e))
628      print('ERROR: failed to evaluate meta expression %s at %s' %
629            (exp.python_exp, exp.token.start))
630      sys.exit(1)
631    return result
632
633  def GetRange(self, identifier):
634    for (var, lower, upper) in self.ranges:
635      if identifier == var:
636        return (lower, upper)
637
638    print('ERROR: range %s is undefined.' % (identifier,))
639    sys.exit(1)
640
641
642class Output:
643  def __init__(self):
644    self.string = ''
645
646  def GetLastLine(self):
647    index = self.string.rfind('\n')
648    if index < 0:
649      return ''
650
651    return self.string[index + 1:]
652
653  def Append(self, s):
654    self.string += s
655
656
657def RunAtomicCode(env, node, output):
658  if isinstance(node, VarNode):
659    identifier = node.identifier.value.strip()
660    result = Output()
661    RunAtomicCode(env.Clone(), node.atomic_code, result)
662    value = result.string
663    env.PushVariable(identifier, value)
664  elif isinstance(node, RangeNode):
665    identifier = node.identifier.value.strip()
666    lower = int(env.EvalExp(node.exp1))
667    upper = int(env.EvalExp(node.exp2))
668    env.PushRange(identifier, lower, upper)
669  elif isinstance(node, ForNode):
670    identifier = node.identifier.value.strip()
671    if node.sep is None:
672      sep = ''
673    else:
674      sep = node.sep.value
675    (lower, upper) = env.GetRange(identifier)
676    for i in range(lower, upper + 1):
677      new_env = env.Clone()
678      new_env.PushVariable(identifier, i)
679      RunCode(new_env, node.code, output)
680      if i != upper:
681        output.Append(sep)
682  elif isinstance(node, RawCodeNode):
683    output.Append(node.raw_code.value)
684  elif isinstance(node, IfNode):
685    cond = env.EvalExp(node.exp)
686    if cond:
687      RunCode(env.Clone(), node.then_branch, output)
688    elif node.else_branch is not None:
689      RunCode(env.Clone(), node.else_branch, output)
690  elif isinstance(node, ExpNode):
691    value = env.EvalExp(node)
692    output.Append('%s' % (value,))
693  elif isinstance(node, LiteralDollarNode):
694    output.Append('$')
695  elif isinstance(node, CodeNode):
696    RunCode(env.Clone(), node, output)
697  else:
698    print('BAD')
699    print(node)
700    sys.exit(1)
701
702
703def RunCode(env, code_node, output):
704  for atomic_code in code_node.atomic_code:
705    RunAtomicCode(env, atomic_code, output)
706
707
708def IsSingleLineComment(cur_line):
709  return '//' in cur_line
710
711
712def IsInPreprocessorDirective(prev_lines, cur_line):
713  if cur_line.lstrip().startswith('#'):
714    return True
715  return prev_lines and prev_lines[-1].endswith('\\')
716
717
718def WrapComment(line, output):
719  loc = line.find('//')
720  before_comment = line[:loc].rstrip()
721  if before_comment == '':
722    indent = loc
723  else:
724    output.append(before_comment)
725    indent = len(before_comment) - len(before_comment.lstrip())
726  prefix = indent*' ' + '// '
727  max_len = 80 - len(prefix)
728  comment = line[loc + 2:].strip()
729  segs = [seg for seg in re.split(r'(\w+\W*)', comment) if seg != '']
730  cur_line = ''
731  for seg in segs:
732    if len((cur_line + seg).rstrip()) < max_len:
733      cur_line += seg
734    else:
735      if cur_line.strip() != '':
736        output.append(prefix + cur_line.rstrip())
737      cur_line = seg.lstrip()
738  if cur_line.strip() != '':
739    output.append(prefix + cur_line.strip())
740
741
742def WrapCode(line, line_concat, output):
743  indent = len(line) - len(line.lstrip())
744  prefix = indent*' '  # Prefix of the current line
745  max_len = 80 - indent - len(line_concat)  # Maximum length of the current line
746  new_prefix = prefix + 4*' '  # Prefix of a continuation line
747  new_max_len = max_len - 4  # Maximum length of a continuation line
748  # Prefers to wrap a line after a ',' or ';'.
749  segs = [seg for seg in re.split(r'([^,;]+[,;]?)', line.strip()) if seg != '']
750  cur_line = ''  # The current line without leading spaces.
751  for seg in segs:
752    # If the line is still too long, wrap at a space.
753    while cur_line == '' and len(seg.strip()) > max_len:
754      seg = seg.lstrip()
755      split_at = seg.rfind(' ', 0, max_len)
756      output.append(prefix + seg[:split_at].strip() + line_concat)
757      seg = seg[split_at + 1:]
758      prefix = new_prefix
759      max_len = new_max_len
760
761    if len((cur_line + seg).rstrip()) < max_len:
762      cur_line = (cur_line + seg).lstrip()
763    else:
764      output.append(prefix + cur_line.rstrip() + line_concat)
765      prefix = new_prefix
766      max_len = new_max_len
767      cur_line = seg.lstrip()
768  if cur_line.strip() != '':
769    output.append(prefix + cur_line.strip())
770
771
772def WrapPreprocessorDirective(line, output):
773  WrapCode(line, ' \\', output)
774
775
776def WrapPlainCode(line, output):
777  WrapCode(line, '', output)
778
779
780def IsMultiLineIWYUPragma(line):
781  return re.search(r'/\* IWYU pragma: ', line)
782
783
784def IsHeaderGuardIncludeOrOneLineIWYUPragma(line):
785  return (re.match(r'^#(ifndef|define|endif\s*//)\s*[\w_]+\s*$', line) or
786          re.match(r'^#include\s', line) or
787          # Don't break IWYU pragmas, either; that causes iwyu.py problems.
788          re.search(r'// IWYU pragma: ', line))
789
790
791def WrapLongLine(line, output):
792  line = line.rstrip()
793  if len(line) <= 80:
794    output.append(line)
795  elif IsSingleLineComment(line):
796    if IsHeaderGuardIncludeOrOneLineIWYUPragma(line):
797      # The style guide made an exception to allow long header guard lines,
798      # includes and IWYU pragmas.
799      output.append(line)
800    else:
801      WrapComment(line, output)
802  elif IsInPreprocessorDirective(output, line):
803    if IsHeaderGuardIncludeOrOneLineIWYUPragma(line):
804      # The style guide made an exception to allow long header guard lines,
805      # includes and IWYU pragmas.
806      output.append(line)
807    else:
808      WrapPreprocessorDirective(line, output)
809  elif IsMultiLineIWYUPragma(line):
810    output.append(line)
811  else:
812    WrapPlainCode(line, output)
813
814
815def BeautifyCode(string):
816  lines = string.splitlines()
817  output = []
818  for line in lines:
819    WrapLongLine(line, output)
820  output2 = [line.rstrip() for line in output]
821  return '\n'.join(output2) + '\n'
822
823
824def ConvertFromPumpSource(src_text):
825  """Return the text generated from the given Pump source text."""
826  ast = ParseToAST(StripMetaComments(src_text))
827  output = Output()
828  RunCode(Env(), ast, output)
829  return BeautifyCode(output.string)
830
831
832def main(argv):
833  if len(argv) == 1:
834    print(__doc__)
835    sys.exit(1)
836
837  file_path = argv[-1]
838  output_str = ConvertFromPumpSource(io.open(file_path, 'r').read())
839  if file_path.endswith('.pump'):
840    output_file_path = file_path[:-5]
841  else:
842    output_file_path = '-'
843  if output_file_path == '-':
844    print(output_str,)
845  else:
846    output_file = io.open(output_file_path, 'w')
847    output_file.write(u'// This file was GENERATED by command:\n')
848    output_file.write(u'//     %s %s\n' %
849                      (os.path.basename(__file__), os.path.basename(file_path)))
850    output_file.write(u'// DO NOT EDIT BY HAND!!!\n\n')
851    output_file.write(output_str)
852    output_file.close()
853
854
855if __name__ == '__main__':
856  main(sys.argv)
857