1#!/usr/bin/env python
2
3"""This is for cleaning up binary files improperly added to CVS. This script
4scans the given path to find binary files; checks with CVS to see if the sticky
5options are set to -kb; finally if sticky options are not -kb then uses 'cvs
6admin' to set the -kb option.
7
8This script ignores CVS directories, symbolic links, and files not known under
9CVS control (cvs status is 'Unknown').
10
11Run this on a CHECKED OUT module sandbox, not on the repository itself. After
12if fixes the sticky options on any files you should manually do a 'cvs commit'
13to accept the changes. Then be sure to have all users do a 'cvs up -A' to
14update the Sticky Option status.
15
16Noah Spurrier
1720030426
18"""
19
20import os, sys, time
21import pexpect
22
23VERBOSE = 1
24
25def is_binary (filename):
26
27    """Assume that any file with a character where the 8th bit is set is
28    binary. """
29
30	fin = open(filename, 'rb')
31	wholething = fin.read()
32	fin.close()
33	for c in wholething:
34		if ord(c) & 0x80:
35			return 1
36	return 0
37
38def is_kb_sticky (filename):
39
40    """This checks if 'cvs status' reports '-kb' for Sticky options. If the
41    Sticky Option status is '-ks' then this returns 1. If the status is
42    'Unknown' then it returns 1. Otherwise 0 is returned. """
43
44	try:
45		s = pexpect.spawn ('cvs status %s' % filename)
46		i = s.expect (['Sticky Options:\s*(.*)\r\n',  'Status: Unknown'])
47		if i==1 and VERBOSE:
48			print 'File not part of CVS repository:', filename
49			return 1 # Pretend it's OK.
50		if s.match.group(1) == '-kb':
51			return 1
52		s = None
53	except:
54		print 'Something went wrong trying to run external cvs command.'
55		print '    cvs status %s' % filename
56		print 'The cvs command returned:'
57		print s.before
58	return 0
59
60def cvs_admin_kb (filename):
61
62    """This uses 'cvs admin' to set the '-kb' sticky option. """
63
64	s = pexpect.run ('cvs admin -kb %s' % filename)
65	# There is a timing issue. If I run 'cvs admin' too quickly
66	# cvs sometimes has trouble obtaining the directory lock.
67	time.sleep(1)
68
69def walk_and_clean_cvs_binaries (arg, dirname, names):
70
71    """This contains the logic for processing files. This is the os.path.walk
72    callback. This skips dirnames that end in CVS. """
73
74	if len(dirname)>3 and dirname[-3:]=='CVS':
75		return
76	for n in names:
77		fullpath = os.path.join (dirname, n)
78		if os.path.isdir(fullpath) or os.path.islink(fullpath):
79			continue
80		if is_binary(fullpath):
81			if not is_kb_sticky (fullpath):
82				if VERBOSE: print fullpath
83				cvs_admin_kb (fullpath)
84
85def main ():
86
87	if len(sys.argv) == 1:
88		root = '.'
89	else:
90		root = sys.argv[1]
91	os.path.walk (root, walk_and_clean_cvs_binaries, None)
92
93if __name__ == '__main__':
94	main ()
95
96