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