##// END OF EJS Templates
win32mbcs: add reversing wrapper for some unicode-incompatible functions....
Shun-ichi GOTO -
r17798:4091b032 default
parent child Browse files
Show More
@@ -1,167 +1,184
1 # win32mbcs.py -- MBCS filename support for Mercurial
1 # win32mbcs.py -- MBCS filename support for Mercurial
2 #
2 #
3 # Copyright (c) 2008 Shun-ichi Goto <shunichi.goto@gmail.com>
3 # Copyright (c) 2008 Shun-ichi Goto <shunichi.goto@gmail.com>
4 #
4 #
5 # Version: 0.3
5 # Version: 0.3
6 # Author: Shun-ichi Goto <shunichi.goto@gmail.com>
6 # Author: Shun-ichi Goto <shunichi.goto@gmail.com>
7 #
7 #
8 # This software may be used and distributed according to the terms of the
8 # This software may be used and distributed according to the terms of the
9 # GNU General Public License version 2 or any later version.
9 # GNU General Public License version 2 or any later version.
10 #
10 #
11
11
12 '''allow the use of MBCS paths with problematic encodings
12 '''allow the use of MBCS paths with problematic encodings
13
13
14 Some MBCS encodings are not good for some path operations (i.e.
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
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".
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
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
18 wrapping some functions to convert to Unicode string before path
19 operation.
19 operation.
20
20
21 This extension is useful for:
21 This extension is useful for:
22
22
23 - Japanese Windows users using shift_jis encoding.
23 - Japanese Windows users using shift_jis encoding.
24 - Chinese Windows users using big5 encoding.
24 - Chinese Windows users using big5 encoding.
25 - All users who use a repository with one of problematic encodings on
25 - All users who use a repository with one of problematic encodings on
26 case-insensitive file system.
26 case-insensitive file system.
27
27
28 This extension is not needed for:
28 This extension is not needed for:
29
29
30 - Any user who use only ASCII chars in path.
30 - Any user who use only ASCII chars in path.
31 - Any user who do not use any of problematic encodings.
31 - Any user who do not use any of problematic encodings.
32
32
33 Note that there are some limitations on using this extension:
33 Note that there are some limitations on using this extension:
34
34
35 - You should use single encoding in one repository.
35 - You should use single encoding in one repository.
36 - If the repository path ends with 0x5c, .hg/hgrc cannot be read.
36 - If the repository path ends with 0x5c, .hg/hgrc cannot be read.
37 - win32mbcs is not compatible with fixutf8 extension.
37 - win32mbcs is not compatible with fixutf8 extension.
38
38
39 By default, win32mbcs uses encoding.encoding decided by Mercurial.
39 By default, win32mbcs uses encoding.encoding decided by Mercurial.
40 You can specify the encoding by config option::
40 You can specify the encoding by config option::
41
41
42 [win32mbcs]
42 [win32mbcs]
43 encoding = sjis
43 encoding = sjis
44
44
45 It is useful for the users who want to commit with UTF-8 log message.
45 It is useful for the users who want to commit with UTF-8 log message.
46 '''
46 '''
47
47
48 import os, sys
48 import os, sys
49 from mercurial.i18n import _
49 from mercurial.i18n import _
50 from mercurial import util, encoding
50 from mercurial import util, encoding
51 testedwith = 'internal'
51 testedwith = 'internal'
52
52
53 _encoding = None # see extsetup
53 _encoding = None # see extsetup
54
54
55 def decode(arg):
55 def decode(arg):
56 if isinstance(arg, str):
56 if isinstance(arg, str):
57 uarg = arg.decode(_encoding)
57 uarg = arg.decode(_encoding)
58 if arg == uarg.encode(_encoding):
58 if arg == uarg.encode(_encoding):
59 return uarg
59 return uarg
60 raise UnicodeError("Not local encoding")
60 raise UnicodeError("Not local encoding")
61 elif isinstance(arg, tuple):
61 elif isinstance(arg, tuple):
62 return tuple(map(decode, arg))
62 return tuple(map(decode, arg))
63 elif isinstance(arg, list):
63 elif isinstance(arg, list):
64 return map(decode, arg)
64 return map(decode, arg)
65 elif isinstance(arg, dict):
65 elif isinstance(arg, dict):
66 for k, v in arg.items():
66 for k, v in arg.items():
67 arg[k] = decode(v)
67 arg[k] = decode(v)
68 return arg
68 return arg
69
69
70 def encode(arg):
70 def encode(arg):
71 if isinstance(arg, unicode):
71 if isinstance(arg, unicode):
72 return arg.encode(_encoding)
72 return arg.encode(_encoding)
73 elif isinstance(arg, tuple):
73 elif isinstance(arg, tuple):
74 return tuple(map(encode, arg))
74 return tuple(map(encode, arg))
75 elif isinstance(arg, list):
75 elif isinstance(arg, list):
76 return map(encode, arg)
76 return map(encode, arg)
77 elif isinstance(arg, dict):
77 elif isinstance(arg, dict):
78 for k, v in arg.items():
78 for k, v in arg.items():
79 arg[k] = encode(v)
79 arg[k] = encode(v)
80 return arg
80 return arg
81
81
82 def appendsep(s):
82 def appendsep(s):
83 # ensure the path ends with os.sep, appending it if necessary.
83 # ensure the path ends with os.sep, appending it if necessary.
84 try:
84 try:
85 us = decode(s)
85 us = decode(s)
86 except UnicodeError:
86 except UnicodeError:
87 us = s
87 us = s
88 if us and us[-1] not in ':/\\':
88 if us and us[-1] not in ':/\\':
89 s += os.sep
89 s += os.sep
90 return s
90 return s
91
91
92 def wrapper(func, args, kwds):
92
93 # check argument is unicode, then call original
93 def basewrapper(func, argtype, enc, dec, args, kwds):
94 # check check already converted, then call original
94 for arg in args:
95 for arg in args:
95 if isinstance(arg, unicode):
96 if isinstance(arg, argtype):
96 return func(*args, **kwds)
97 return func(*args, **kwds)
97
98
98 try:
99 try:
99 # convert arguments to unicode, call func, then convert back
100 # convert string arguments, call func, then convert back the
100 return encode(func(*decode(args), **decode(kwds)))
101 # return value.
102 return enc(func(*dec(args), **dec(kwds)))
101 except UnicodeError:
103 except UnicodeError:
102 raise util.Abort(_("[win32mbcs] filename conversion failed with"
104 raise util.Abort(_("[win32mbcs] filename conversion failed with"
103 " %s encoding\n") % (_encoding))
105 " %s encoding\n") % (_encoding))
104
106
107 def wrapper(func, args, kwds):
108 return basewrapper(func, unicode, encode, decode, args, kwds)
109
110
111 def reversewrapper(func, args, kwds):
112 return basewrapper(func, str, decode, encode, args, kwds)
113
105 def wrapperforlistdir(func, args, kwds):
114 def wrapperforlistdir(func, args, kwds):
106 # Ensure 'path' argument ends with os.sep to avoids
115 # Ensure 'path' argument ends with os.sep to avoids
107 # misinterpreting last 0x5c of MBCS 2nd byte as path separator.
116 # misinterpreting last 0x5c of MBCS 2nd byte as path separator.
108 if args:
117 if args:
109 args = list(args)
118 args = list(args)
110 args[0] = appendsep(args[0])
119 args[0] = appendsep(args[0])
111 if 'path' in kwds:
120 if 'path' in kwds:
112 kwds['path'] = appendsep(kwds['path'])
121 kwds['path'] = appendsep(kwds['path'])
113 return func(*args, **kwds)
122 return func(*args, **kwds)
114
123
115 def wrapname(name, wrapper):
124 def wrapname(name, wrapper):
116 module, name = name.rsplit('.', 1)
125 module, name = name.rsplit('.', 1)
117 module = sys.modules[module]
126 module = sys.modules[module]
118 func = getattr(module, name)
127 func = getattr(module, name)
119 def f(*args, **kwds):
128 def f(*args, **kwds):
120 return wrapper(func, args, kwds)
129 return wrapper(func, args, kwds)
121 try:
130 try:
122 f.__name__ = func.__name__ # fails with Python 2.3
131 f.__name__ = func.__name__ # fails with Python 2.3
123 except Exception:
132 except Exception:
124 pass
133 pass
125 setattr(module, name, f)
134 setattr(module, name, f)
126
135
127 # List of functions to be wrapped.
136 # List of functions to be wrapped.
128 # NOTE: os.path.dirname() and os.path.basename() are safe because
137 # NOTE: os.path.dirname() and os.path.basename() are safe because
129 # they use result of os.path.split()
138 # they use result of os.path.split()
130 funcs = '''os.path.join os.path.split os.path.splitext
139 funcs = '''os.path.join os.path.split os.path.splitext
131 os.path.normpath os.makedirs
140 os.path.normpath os.makedirs
132 mercurial.util.endswithsep mercurial.util.splitpath mercurial.util.checkcase
141 mercurial.util.endswithsep mercurial.util.splitpath mercurial.util.checkcase
133 mercurial.util.fspath mercurial.util.pconvert mercurial.util.normpath
142 mercurial.util.fspath mercurial.util.pconvert mercurial.util.normpath
134 mercurial.util.checkwinfilename mercurial.util.checkosfilename'''
143 mercurial.util.checkwinfilename mercurial.util.checkosfilename'''
135
144
145 # These functions are required to be called with local encoded string
146 # because they expects argument is local encoded string and cause
147 # problem with unicode string.
148 rfuncs = '''mercurial.encoding.upper mercurial.encoding.lower'''
149
136 # List of Windows specific functions to be wrapped.
150 # List of Windows specific functions to be wrapped.
137 winfuncs = '''os.path.splitunc'''
151 winfuncs = '''os.path.splitunc'''
138
152
139 # codec and alias names of sjis and big5 to be faked.
153 # codec and alias names of sjis and big5 to be faked.
140 problematic_encodings = '''big5 big5-tw csbig5 big5hkscs big5-hkscs
154 problematic_encodings = '''big5 big5-tw csbig5 big5hkscs big5-hkscs
141 hkscs cp932 932 ms932 mskanji ms-kanji shift_jis csshiftjis shiftjis
155 hkscs cp932 932 ms932 mskanji ms-kanji shift_jis csshiftjis shiftjis
142 sjis s_jis shift_jis_2004 shiftjis2004 sjis_2004 sjis2004
156 sjis s_jis shift_jis_2004 shiftjis2004 sjis_2004 sjis2004
143 shift_jisx0213 shiftjisx0213 sjisx0213 s_jisx0213 950 cp950 ms950 '''
157 shift_jisx0213 shiftjisx0213 sjisx0213 s_jisx0213 950 cp950 ms950 '''
144
158
145 def extsetup(ui):
159 def extsetup(ui):
146 # TODO: decide use of config section for this extension
160 # TODO: decide use of config section for this extension
147 if ((not os.path.supports_unicode_filenames) and
161 if ((not os.path.supports_unicode_filenames) and
148 (sys.platform != 'cygwin')):
162 (sys.platform != 'cygwin')):
149 ui.warn(_("[win32mbcs] cannot activate on this platform.\n"))
163 ui.warn(_("[win32mbcs] cannot activate on this platform.\n"))
150 return
164 return
151 # determine encoding for filename
165 # determine encoding for filename
152 global _encoding
166 global _encoding
153 _encoding = ui.config('win32mbcs', 'encoding', encoding.encoding)
167 _encoding = ui.config('win32mbcs', 'encoding', encoding.encoding)
154 # fake is only for relevant environment.
168 # fake is only for relevant environment.
155 if _encoding.lower() in problematic_encodings.split():
169 if _encoding.lower() in problematic_encodings.split():
156 for f in funcs.split():
170 for f in funcs.split():
157 wrapname(f, wrapper)
171 wrapname(f, wrapper)
158 if os.name == 'nt':
172 if os.name == 'nt':
159 for f in winfuncs.split():
173 for f in winfuncs.split():
160 wrapname(f, wrapper)
174 wrapname(f, wrapper)
161 wrapname("mercurial.osutil.listdir", wrapperforlistdir)
175 wrapname("mercurial.osutil.listdir", wrapperforlistdir)
176 # wrap functions to be called with local byte string arguments
177 for f in rfuncs.split():
178 wrapname(f, reversewrapper)
162 # Check sys.args manually instead of using ui.debug() because
179 # Check sys.args manually instead of using ui.debug() because
163 # command line options is not yet applied when
180 # command line options is not yet applied when
164 # extensions.loadall() is called.
181 # extensions.loadall() is called.
165 if '--debug' in sys.argv:
182 if '--debug' in sys.argv:
166 ui.write("[win32mbcs] activated with encoding: %s\n"
183 ui.write("[win32mbcs] activated with encoding: %s\n"
167 % _encoding)
184 % _encoding)
General Comments 0
You need to be logged in to leave comments. Login now