3
# Released to the public domain, by Tim Peters, 03 October 2000.
5
"""reindent [-d][-r][-v] [ path ... ]
7
-d (--dryrun) Dry run. Analyze, but don't make any changes to, files.
8
-r (--recurse) Recurse. Search for all .py files in subdirectories too.
9
-n (--nobackup) No backup. Does not make a ".bak" file before reindenting.
10
-v (--verbose) Verbose. Print informative msgs; else no output.
11
-h (--help) Help. Print this usage information and exit.
13
Change Python (.py) files to use 4-space indents and no hard tab characters.
14
Also trim excess spaces and tabs from ends of lines, and remove empty lines
15
at the end of files. Also ensure the last line ends with a newline.
17
If no paths are given on the command line, reindent operates as a filter,
18
reading a single source file from standard input and writing the transformed
19
source to standard output. In this case, the -d, -r and -v flags are
22
You can pass one or more file and/or directory paths. When a directory
23
path, all .py files within the directory will be examined, and, if the -r
24
option is given, likewise recursively for subdirectories.
26
If output is not to standard output, reindent overwrites files in place,
27
renaming the originals with a .bak extension. If it finds nothing to
28
change, the file is left alone. If reindent does change a file, the changed
29
file is a fixed-point for future runs (i.e., running reindent on the
30
resulting .py file won't change it again).
32
The hard part of reindenting is figuring out what to do with comment
33
lines. So long as the input files get a clean bill of health from
34
tabnanny.py, reindent should do a good job.
36
The backup file is a copy of the one that is being reindented. The ".bak"
37
file is generated with shutil.copy(), but some corner cases regarding
38
user/group and permissions could leave the backup file more readable that
39
you'd prefer. You can always use the --nobackup option to prevent this.
55
print >> sys.stderr, msg
56
print >> sys.stderr, __doc__
61
sys.stderr.write(sep + str(arg))
63
sys.stderr.write("\n")
67
global verbose, recurse, dryrun, makebackup
69
opts, args = getopt.getopt(sys.argv[1:], "drnvh",
70
["dryrun", "recurse", "nobackup", "verbose", "help"])
71
except getopt.error, msg:
75
if o in ('-d', '--dryrun'):
77
elif o in ('-r', '--recurse'):
79
elif o in ('-n', '--nobackup'):
81
elif o in ('-v', '--verbose'):
83
elif o in ('-h', '--help'):
87
r = Reindenter(sys.stdin)
95
if os.path.isdir(file) and not os.path.islink(file):
97
print "listing directory", file
98
names = os.listdir(file)
100
fullname = os.path.join(file, name)
101
if ((recurse and os.path.isdir(fullname) and
102
not os.path.islink(fullname) and
103
not os.path.split(fullname)[1].startswith("."))
104
or name.lower().endswith(".py")):
109
print "checking", file, "...",
113
errprint("%s: I/O Error: %s" % (file, str(msg)))
122
print "But this is a dry run, so leaving it alone."
126
shutil.copyfile(file, bak)
128
print "backed up", file, "to", bak
133
print "wrote new", file
140
def _rstrip(line, JUNK='\n \t'):
141
"""Return line stripped of trailing spaces, tabs, newlines.
143
Note that line.rstrip() instead also strips sundry control characters,
144
but at least one known Emacs user expects to keep junk like that, not
145
mentioning Barry by name or anything <wink>.
149
while i > 0 and line[i-1] in JUNK:
155
def __init__(self, f):
156
self.find_stmt = 1 # next token begins a fresh stmt?
157
self.level = 0 # current indent level
160
self.raw = f.readlines()
162
# File lines, rstripped & tab-expanded. Dummy at start is so
163
# that we can use tokenize's 1-based line numbering easily.
164
# Note that a line is all-blank iff it's "\n".
165
self.lines = [_rstrip(line).expandtabs() + "\n"
166
for line in self.raw]
167
self.lines.insert(0, None)
168
self.index = 1 # index into self.lines of next line
170
# List of (lineno, indentlevel) pairs, one for each stmt and
171
# comment line. indentlevel is -1 for comment lines, as a
172
# signal that tokenize doesn't know what to do about them;
173
# indeed, they're our headache!
177
tokenize.tokenize(self.getline, self.tokeneater)
178
# Remove trailing empty lines.
180
while lines and lines[-1] == "\n":
184
stats.append((len(lines), 0))
185
# Map count of leading spaces to # we want.
187
# Program after transformation.
188
after = self.after = []
189
# Copy over initial empty lines -- there's nothing to do until
190
# we see a line with *something* on it.
192
after.extend(lines[1:i])
193
for i in range(len(stats)-1):
194
thisstmt, thislevel = stats[i]
195
nextstmt = stats[i+1][0]
196
have = getlspace(lines[thisstmt])
201
# An indented comment line. If we saw the same
202
# indentation before, reuse what it most recently
204
want = have2want.get(have, -1)
206
# Then it probably belongs to the next real stmt.
207
for j in xrange(i+1, len(stats)-1):
208
jline, jlevel = stats[j]
210
if have == getlspace(lines[jline]):
213
if want < 0: # Maybe it's a hanging
214
# comment like this one,
215
# in which case we should shift it like its base
217
for j in xrange(i-1, -1, -1):
218
jline, jlevel = stats[j]
220
want = have + getlspace(after[jline-1]) - \
221
getlspace(lines[jline])
224
# Still no luck -- leave it alone.
229
have2want[have] = want
231
if diff == 0 or have == 0:
232
after.extend(lines[thisstmt:nextstmt])
234
for line in lines[thisstmt:nextstmt]:
239
after.append(" " * diff + line)
241
remove = min(getlspace(line), -diff)
242
after.append(line[remove:])
243
return self.raw != self.after
246
f.writelines(self.after)
248
# Line-getter for tokenize.
250
if self.index >= len(self.lines):
253
line = self.lines[self.index]
257
# Line-eater for tokenize.
258
def tokeneater(self, type, token, (sline, scol), end, line,
259
INDENT=tokenize.INDENT,
260
DEDENT=tokenize.DEDENT,
261
NEWLINE=tokenize.NEWLINE,
262
COMMENT=tokenize.COMMENT,
266
# A program statement, or ENDMARKER, will eventually follow,
267
# after some (possibly empty) run of tokens of the form
268
# (NL | COMMENT)* (INDENT | DEDENT+)?
279
elif type == COMMENT:
281
self.stats.append((sline, -1))
282
# but we're still looking for a new stmt, so leave
289
# This is the first "real token" following a NEWLINE, so it
290
# must be the first token of the next program statement, or an
293
if line: # not endmarker
294
self.stats.append((sline, self.level))
296
# Count number of leading blanks.
299
while i < n and line[i] == " ":
303
if __name__ == '__main__':