##// END OF EJS Templates
win32mbcs: util.normpath should be wrapped....
Shun-ichi GOTO -
r9463:bd017f35 default
parent child Browse files
Show More
@@ -1,147 +1,147 b''
1 1 # win32mbcs.py -- MBCS filename support for Mercurial
2 2 #
3 3 # Copyright (c) 2008 Shun-ichi Goto <shunichi.goto@gmail.com>
4 4 #
5 5 # Version: 0.2
6 6 # Author: Shun-ichi Goto <shunichi.goto@gmail.com>
7 7 #
8 8 # This software may be used and distributed according to the terms of the
9 9 # GNU General Public License version 2, incorporated herein by reference.
10 10 #
11 11
12 12 '''allow the use of MBCS paths with problematic encodings
13 13
14 14 Some MBCS encodings are not good for some path operations (i.e.
15 15 splitting path, case conversion, etc.) with its encoded bytes. We call
16 16 such a encoding (i.e. shift_jis and big5) as "problematic encoding".
17 17 This extension can be used to fix the issue with those encodings by
18 18 wrapping some functions to convert to Unicode string before path
19 19 operation.
20 20
21 21 This extension is useful for:
22 22
23 23 - Japanese Windows users using shift_jis encoding.
24 24 - Chinese Windows users using big5 encoding.
25 25 - All users who use a repository with one of problematic encodings on
26 26 case-insensitive file system.
27 27
28 28 This extension is not needed for:
29 29
30 30 - Any user who use only ASCII chars in path.
31 31 - Any user who do not use any of problematic encodings.
32 32
33 33 Note that there are some limitations on using this extension:
34 34
35 35 - You should use single encoding in one repository.
36 36 - You should set same encoding for the repository by locale or
37 37 HGENCODING.
38 38
39 39 Path encoding conversion are done between Unicode and
40 40 encoding.encoding which is decided by Mercurial from current locale
41 41 setting or HGENCODING.
42 42 '''
43 43
44 44 import os, sys
45 45 from mercurial.i18n import _
46 46 from mercurial import util, encoding
47 47
48 48 def decode(arg):
49 49 if isinstance(arg, str):
50 50 uarg = arg.decode(encoding.encoding)
51 51 if arg == uarg.encode(encoding.encoding):
52 52 return uarg
53 53 raise UnicodeError("Not local encoding")
54 54 elif isinstance(arg, tuple):
55 55 return tuple(map(decode, arg))
56 56 elif isinstance(arg, list):
57 57 return map(decode, arg)
58 58 elif isinstance(arg, dict):
59 59 for k, v in arg.items():
60 60 arg[k] = decode(v)
61 61 return arg
62 62
63 63 def encode(arg):
64 64 if isinstance(arg, unicode):
65 65 return arg.encode(encoding.encoding)
66 66 elif isinstance(arg, tuple):
67 67 return tuple(map(encode, arg))
68 68 elif isinstance(arg, list):
69 69 return map(encode, arg)
70 70 elif isinstance(arg, dict):
71 71 for k, v in arg.items():
72 72 arg[k] = encode(v)
73 73 return arg
74 74
75 75 def appendsep(s):
76 76 # ensure the path ends with os.sep, appending it if necessary.
77 77 try:
78 78 us = decode(s)
79 79 except UnicodeError:
80 80 us = s
81 81 if us and us[-1] not in ':/\\':
82 82 s += os.sep
83 83 return s
84 84
85 85 def wrapper(func, args, kwds):
86 86 # check argument is unicode, then call original
87 87 for arg in args:
88 88 if isinstance(arg, unicode):
89 89 return func(*args, **kwds)
90 90
91 91 try:
92 92 # convert arguments to unicode, call func, then convert back
93 93 return encode(func(*decode(args), **decode(kwds)))
94 94 except UnicodeError:
95 95 raise util.Abort(_("[win32mbcs] filename conversion failed with"
96 96 " %s encoding\n") % (encoding.encoding))
97 97
98 98 def wrapperforlistdir(func, args, kwds):
99 99 # Ensure 'path' argument ends with os.sep to avoids
100 100 # misinterpreting last 0x5c of MBCS 2nd byte as path separator.
101 101 if args:
102 102 args = list(args)
103 103 args[0] = appendsep(args[0])
104 104 if 'path' in kwds:
105 105 kwds['path'] = appendsep(kwds['path'])
106 106 return func(*args, **kwds)
107 107
108 108 def wrapname(name, wrapper):
109 109 module, name = name.rsplit('.', 1)
110 110 module = sys.modules[module]
111 111 func = getattr(module, name)
112 112 def f(*args, **kwds):
113 113 return wrapper(func, args, kwds)
114 114 try:
115 115 f.__name__ = func.__name__ # fail with python23
116 116 except Exception:
117 117 pass
118 118 setattr(module, name, f)
119 119
120 120 # List of functions to be wrapped.
121 121 # NOTE: os.path.dirname() and os.path.basename() are safe because
122 122 # they use result of os.path.split()
123 123 funcs = '''os.path.join os.path.split os.path.splitext
124 124 os.path.splitunc os.path.normpath os.path.normcase os.makedirs
125 125 mercurial.util.endswithsep mercurial.util.splitpath mercurial.util.checkcase
126 mercurial.util.fspath mercurial.util.pconvert'''
126 mercurial.util.fspath mercurial.util.pconvert mercurial.util.normpath'''
127 127
128 128 # codec and alias names of sjis and big5 to be faked.
129 129 problematic_encodings = '''big5 big5-tw csbig5 big5hkscs big5-hkscs
130 130 hkscs cp932 932 ms932 mskanji ms-kanji shift_jis csshiftjis shiftjis
131 131 sjis s_jis shift_jis_2004 shiftjis2004 sjis_2004 sjis2004
132 132 shift_jisx0213 shiftjisx0213 sjisx0213 s_jisx0213 950 cp950 ms950 '''
133 133
134 134 def reposetup(ui, repo):
135 135 # TODO: decide use of config section for this extension
136 136 if not os.path.supports_unicode_filenames:
137 137 ui.warn(_("[win32mbcs] cannot activate on this platform.\n"))
138 138 return
139 139
140 140 # fake is only for relevant environment.
141 141 if encoding.encoding.lower() in problematic_encodings.split():
142 142 for f in funcs.split():
143 143 wrapname(f, wrapper)
144 144 wrapname("mercurial.osutil.listdir", wrapperforlistdir)
145 145 ui.debug(_("[win32mbcs] activated with encoding: %s\n")
146 146 % encoding.encoding)
147 147
General Comments 0
You need to be logged in to leave comments. Login now