Bump version to 0.9.1.
[python/dscho.git] / Tools / scripts / pindent.py
blob6eba0214ca556552211a4f5c1ddbb7f95c692e33
1 #! /usr/bin/env python
3 # This file contains a class and a main program that perform three
4 # related (though complimentary) formatting operations on Python
5 # programs. When called as "pindent -c", it takes a valid Python
6 # program as input and outputs a version augmented with block-closing
7 # comments. When called as "pindent -d", it assumes its input is a
8 # Python program with block-closing comments and outputs a commentless
9 # version. When called as "pindent -r" it assumes its input is a
10 # Python program with block-closing comments but with its indentation
11 # messed up, and outputs a properly indented version.
13 # A "block-closing comment" is a comment of the form '# end <keyword>'
14 # where <keyword> is the keyword that opened the block. If the
15 # opening keyword is 'def' or 'class', the function or class name may
16 # be repeated in the block-closing comment as well. Here is an
17 # example of a program fully augmented with block-closing comments:
19 # def foobar(a, b):
20 # if a == b:
21 # a = a+1
22 # elif a < b:
23 # b = b-1
24 # if b > a: a = a-1
25 # # end if
26 # else:
27 # print 'oops!'
28 # # end if
29 # # end def foobar
31 # Note that only the last part of an if...elif...else... block needs a
32 # block-closing comment; the same is true for other compound
33 # statements (e.g. try...except). Also note that "short-form" blocks
34 # like the second 'if' in the example must be closed as well;
35 # otherwise the 'else' in the example would be ambiguous (remember
36 # that indentation is not significant when interpreting block-closing
37 # comments).
39 # The operations are idempotent (i.e. applied to their own output
40 # they yield an identical result). Running first "pindent -c" and
41 # then "pindent -r" on a valid Python program produces a program that
42 # is semantically identical to the input (though its indentation may
43 # be different). Running "pindent -e" on that output produces a
44 # program that only differs from the original in indentation.
46 # Other options:
47 # -s stepsize: set the indentation step size (default 8)
48 # -t tabsize : set the number of spaces a tab character is worth (default 8)
49 # -e : expand TABs into spaces
50 # file ... : input file(s) (default standard input)
51 # The results always go to standard output
53 # Caveats:
54 # - comments ending in a backslash will be mistaken for continued lines
55 # - continuations using backslash are always left unchanged
56 # - continuations inside parentheses are not extra indented by -r
57 # but must be indented for -c to work correctly (this breaks
58 # idempotency!)
59 # - continued lines inside triple-quoted strings are totally garbled
61 # Secret feature:
62 # - On input, a block may also be closed with an "end statement" --
63 # this is a block-closing comment without the '#' sign.
65 # Possible improvements:
66 # - check syntax based on transitions in 'next' table
67 # - better error reporting
68 # - better error recovery
69 # - check identifier after class/def
71 # The following wishes need a more complete tokenization of the source:
72 # - Don't get fooled by comments ending in backslash
73 # - reindent continuation lines indicated by backslash
74 # - handle continuation lines inside parentheses/braces/brackets
75 # - handle triple quoted strings spanning lines
76 # - realign comments
77 # - optionally do much more thorough reformatting, a la C indent
79 # Defaults
80 STEPSIZE = 8
81 TABSIZE = 8
82 EXPANDTABS = 0
84 import os
85 import re
86 import string
87 import sys
89 next = {}
90 next['if'] = next['elif'] = 'elif', 'else', 'end'
91 next['while'] = next['for'] = 'else', 'end'
92 next['try'] = 'except', 'finally'
93 next['except'] = 'except', 'else', 'end'
94 next['else'] = next['finally'] = next['def'] = next['class'] = 'end'
95 next['end'] = ()
96 start = 'if', 'while', 'for', 'try', 'def', 'class'
98 class PythonIndenter:
100 def __init__(self, fpi = sys.stdin, fpo = sys.stdout,
101 indentsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
102 self.fpi = fpi
103 self.fpo = fpo
104 self.indentsize = indentsize
105 self.tabsize = tabsize
106 self.lineno = 0
107 self.expandtabs = expandtabs
108 self._write = fpo.write
109 self.kwprog = re.compile(
110 r'^\s*(?P<kw>[a-z]+)'
111 r'(\s+(?P<id>[a-zA-Z_]\w*))?'
112 r'[^\w]')
113 self.endprog = re.compile(
114 r'^\s*#?\s*end\s+(?P<kw>[a-z]+)'
115 r'(\s+(?P<id>[a-zA-Z_]\w*))?'
116 r'[^\w]')
117 self.wsprog = re.compile(r'^[ \t]*')
118 # end def __init__
120 def write(self, line):
121 if self.expandtabs:
122 self._write(string.expandtabs(line, self.tabsize))
123 else:
124 self._write(line)
125 # end if
126 # end def write
128 def readline(self):
129 line = self.fpi.readline()
130 if line: self.lineno = self.lineno + 1
131 # end if
132 return line
133 # end def readline
135 def error(self, fmt, *args):
136 if args: fmt = fmt % args
137 # end if
138 sys.stderr.write('Error at line %d: %s\n' % (self.lineno, fmt))
139 self.write('### %s ###\n' % fmt)
140 # end def error
142 def getline(self):
143 line = self.readline()
144 while line[-2:] == '\\\n':
145 line2 = self.readline()
146 if not line2: break
147 # end if
148 line = line + line2
149 # end while
150 return line
151 # end def getline
153 def putline(self, line, indent = None):
154 if indent is None:
155 self.write(line)
156 return
157 # end if
158 tabs, spaces = divmod(indent*self.indentsize, self.tabsize)
159 i = 0
160 m = self.wsprog.match(line)
161 if m: i = m.end()
162 # end if
163 self.write('\t'*tabs + ' '*spaces + line[i:])
164 # end def putline
166 def reformat(self):
167 stack = []
168 while 1:
169 line = self.getline()
170 if not line: break # EOF
171 # end if
172 m = self.endprog.match(line)
173 if m:
174 kw = 'end'
175 kw2 = m.group('kw')
176 if not stack:
177 self.error('unexpected end')
178 elif stack[-1][0] != kw2:
179 self.error('unmatched end')
180 # end if
181 del stack[-1:]
182 self.putline(line, len(stack))
183 continue
184 # end if
185 m = self.kwprog.match(line)
186 if m:
187 kw = m.group('kw')
188 if kw in start:
189 self.putline(line, len(stack))
190 stack.append((kw, kw))
191 continue
192 # end if
193 if next.has_key(kw) and stack:
194 self.putline(line, len(stack)-1)
195 kwa, kwb = stack[-1]
196 stack[-1] = kwa, kw
197 continue
198 # end if
199 # end if
200 self.putline(line, len(stack))
201 # end while
202 if stack:
203 self.error('unterminated keywords')
204 for kwa, kwb in stack:
205 self.write('\t%s\n' % kwa)
206 # end for
207 # end if
208 # end def reformat
210 def delete(self):
211 begin_counter = 0
212 end_counter = 0
213 while 1:
214 line = self.getline()
215 if not line: break # EOF
216 # end if
217 m = self.endprog.match(line)
218 if m:
219 end_counter = end_counter + 1
220 continue
221 # end if
222 m = self.kwprog.match(line)
223 if m:
224 kw = m.group('kw')
225 if kw in start:
226 begin_counter = begin_counter + 1
227 # end if
228 # end if
229 self.putline(line)
230 # end while
231 if begin_counter - end_counter < 0:
232 sys.stderr.write('Warning: input contained more end tags than expected\n')
233 elif begin_counter - end_counter > 0:
234 sys.stderr.write('Warning: input contained less end tags than expected\n')
235 # end if
236 # end def delete
238 def complete(self):
239 self.indentsize = 1
240 stack = []
241 todo = []
242 current, firstkw, lastkw, topid = 0, '', '', ''
243 while 1:
244 line = self.getline()
245 i = 0
246 m = self.wsprog.match(line)
247 if m: i = m.end()
248 # end if
249 m = self.endprog.match(line)
250 if m:
251 thiskw = 'end'
252 endkw = m.group('kw')
253 thisid = m.group('id')
254 else:
255 m = self.kwprog.match(line)
256 if m:
257 thiskw = m.group('kw')
258 if not next.has_key(thiskw):
259 thiskw = ''
260 # end if
261 if thiskw in ('def', 'class'):
262 thisid = m.group('id')
263 else:
264 thisid = ''
265 # end if
266 elif line[i:i+1] in ('\n', '#'):
267 todo.append(line)
268 continue
269 else:
270 thiskw = ''
271 # end if
272 # end if
273 indent = len(string.expandtabs(line[:i], self.tabsize))
274 while indent < current:
275 if firstkw:
276 if topid:
277 s = '# end %s %s\n' % (
278 firstkw, topid)
279 else:
280 s = '# end %s\n' % firstkw
281 # end if
282 self.putline(s, current)
283 firstkw = lastkw = ''
284 # end if
285 current, firstkw, lastkw, topid = stack[-1]
286 del stack[-1]
287 # end while
288 if indent == current and firstkw:
289 if thiskw == 'end':
290 if endkw != firstkw:
291 self.error('mismatched end')
292 # end if
293 firstkw = lastkw = ''
294 elif not thiskw or thiskw in start:
295 if topid:
296 s = '# end %s %s\n' % (
297 firstkw, topid)
298 else:
299 s = '# end %s\n' % firstkw
300 # end if
301 self.putline(s, current)
302 firstkw = lastkw = topid = ''
303 # end if
304 # end if
305 if indent > current:
306 stack.append((current, firstkw, lastkw, topid))
307 if thiskw and thiskw not in start:
308 # error
309 thiskw = ''
310 # end if
311 current, firstkw, lastkw, topid = \
312 indent, thiskw, thiskw, thisid
313 # end if
314 if thiskw:
315 if thiskw in start:
316 firstkw = lastkw = thiskw
317 topid = thisid
318 else:
319 lastkw = thiskw
320 # end if
321 # end if
322 for l in todo: self.write(l)
323 # end for
324 todo = []
325 if not line: break
326 # end if
327 self.write(line)
328 # end while
329 # end def complete
331 # end class PythonIndenter
333 # Simplified user interface
334 # - xxx_filter(input, output): read and write file objects
335 # - xxx_string(s): take and return string object
336 # - xxx_file(filename): process file in place, return true iff changed
338 def complete_filter(input = sys.stdin, output = sys.stdout,
339 stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
340 pi = PythonIndenter(input, output, stepsize, tabsize, expandtabs)
341 pi.complete()
342 # end def complete_filter
344 def delete_filter(input= sys.stdin, output = sys.stdout,
345 stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
346 pi = PythonIndenter(input, output, stepsize, tabsize, expandtabs)
347 pi.delete()
348 # end def delete_filter
350 def reformat_filter(input = sys.stdin, output = sys.stdout,
351 stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
352 pi = PythonIndenter(input, output, stepsize, tabsize, expandtabs)
353 pi.reformat()
354 # end def reformat_filter
356 class StringReader:
357 def __init__(self, buf):
358 self.buf = buf
359 self.pos = 0
360 self.len = len(self.buf)
361 # end def __init__
362 def read(self, n = 0):
363 if n <= 0:
364 n = self.len - self.pos
365 else:
366 n = min(n, self.len - self.pos)
367 # end if
368 r = self.buf[self.pos : self.pos + n]
369 self.pos = self.pos + n
370 return r
371 # end def read
372 def readline(self):
373 i = string.find(self.buf, '\n', self.pos)
374 return self.read(i + 1 - self.pos)
375 # end def readline
376 def readlines(self):
377 lines = []
378 line = self.readline()
379 while line:
380 lines.append(line)
381 line = self.readline()
382 # end while
383 return lines
384 # end def readlines
385 # seek/tell etc. are left as an exercise for the reader
386 # end class StringReader
388 class StringWriter:
389 def __init__(self):
390 self.buf = ''
391 # end def __init__
392 def write(self, s):
393 self.buf = self.buf + s
394 # end def write
395 def getvalue(self):
396 return self.buf
397 # end def getvalue
398 # end class StringWriter
400 def complete_string(source, stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
401 input = StringReader(source)
402 output = StringWriter()
403 pi = PythonIndenter(input, output, stepsize, tabsize, expandtabs)
404 pi.complete()
405 return output.getvalue()
406 # end def complete_string
408 def delete_string(source, stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
409 input = StringReader(source)
410 output = StringWriter()
411 pi = PythonIndenter(input, output, stepsize, tabsize, expandtabs)
412 pi.delete()
413 return output.getvalue()
414 # end def delete_string
416 def reformat_string(source, stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
417 input = StringReader(source)
418 output = StringWriter()
419 pi = PythonIndenter(input, output, stepsize, tabsize, expandtabs)
420 pi.reformat()
421 return output.getvalue()
422 # end def reformat_string
424 def complete_file(filename, stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
425 source = open(filename, 'r').read()
426 result = complete_string(source, stepsize, tabsize, expandtabs)
427 if source == result: return 0
428 # end if
429 import os
430 try: os.rename(filename, filename + '~')
431 except os.error: pass
432 # end try
433 f = open(filename, 'w')
434 f.write(result)
435 f.close()
436 return 1
437 # end def complete_file
439 def delete_file(filename, stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
440 source = open(filename, 'r').read()
441 result = delete_string(source, stepsize, tabsize, expandtabs)
442 if source == result: return 0
443 # end if
444 import os
445 try: os.rename(filename, filename + '~')
446 except os.error: pass
447 # end try
448 f = open(filename, 'w')
449 f.write(result)
450 f.close()
451 return 1
452 # end def delete_file
454 def reformat_file(filename, stepsize = STEPSIZE, tabsize = TABSIZE, expandtabs = EXPANDTABS):
455 source = open(filename, 'r').read()
456 result = reformat_string(source, stepsize, tabsize, expandtabs)
457 if source == result: return 0
458 # end if
459 import os
460 try: os.rename(filename, filename + '~')
461 except os.error: pass
462 # end try
463 f = open(filename, 'w')
464 f.write(result)
465 f.close()
466 return 1
467 # end def reformat_file
469 # Test program when called as a script
471 usage = """
472 usage: pindent (-c|-d|-r) [-s stepsize] [-t tabsize] [-e] [file] ...
473 -c : complete a correctly indented program (add #end directives)
474 -d : delete #end directives
475 -r : reformat a completed program (use #end directives)
476 -s stepsize: indentation step (default %(STEPSIZE)d)
477 -t tabsize : the worth in spaces of a tab (default %(TABSIZE)d)
478 -e : expand TABs into spaces (defailt OFF)
479 [file] ... : files are changed in place, with backups in file~
480 If no files are specified or a single - is given,
481 the program acts as a filter (reads stdin, writes stdout).
482 """ % vars()
484 def error_both(op1, op2):
485 sys.stderr.write('Error: You can not specify both '+op1+' and -'+op2[0]+' at the same time\n')
486 sys.stderr.write(usage)
487 sys.exit(2)
488 # end def error_both
490 def test():
491 import getopt
492 try:
493 opts, args = getopt.getopt(sys.argv[1:], 'cdrs:t:e')
494 except getopt.error, msg:
495 sys.stderr.write('Error: %s\n' % msg)
496 sys.stderr.write(usage)
497 sys.exit(2)
498 # end try
499 action = None
500 stepsize = STEPSIZE
501 tabsize = TABSIZE
502 expandtabs = EXPANDTABS
503 for o, a in opts:
504 if o == '-c':
505 if action: error_both(o, action)
506 # end if
507 action = 'complete'
508 elif o == '-d':
509 if action: error_both(o, action)
510 # end if
511 action = 'delete'
512 elif o == '-r':
513 if action: error_both(o, action)
514 # end if
515 action = 'reformat'
516 elif o == '-s':
517 stepsize = string.atoi(a)
518 elif o == '-t':
519 tabsize = string.atoi(a)
520 elif o == '-e':
521 expandtabs = 1
522 # end if
523 # end for
524 if not action:
525 sys.stderr.write(
526 'You must specify -c(omplete), -d(elete) or -r(eformat)\n')
527 sys.stderr.write(usage)
528 sys.exit(2)
529 # end if
530 if not args or args == ['-']:
531 action = eval(action + '_filter')
532 action(sys.stdin, sys.stdout, stepsize, tabsize, expandtabs)
533 else:
534 action = eval(action + '_file')
535 for file in args:
536 action(file, stepsize, tabsize, expandtabs)
537 # end for
538 # end if
539 # end def test
541 if __name__ == '__main__':
542 test()
543 # end if