|
1 # win32mbcs.py -- MBCS filename support for Mercurial |
|
2 # |
|
3 # Copyright (c) 2008 Shun-ichi Goto <shunichi.goto@gmail.com> |
|
4 # |
|
5 # Version: 0.3 |
|
6 # Author: Shun-ichi Goto <shunichi.goto@gmail.com> |
|
7 # |
|
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. |
|
10 # |
|
11 |
|
12 '''allow the use of MBCS paths with problematic encodings |
|
13 |
|
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 |
|
19 operation. |
|
20 |
|
21 This extension is useful for: |
|
22 |
|
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. |
|
27 |
|
28 This extension is not needed for: |
|
29 |
|
30 - Any user who use only ASCII chars in path. |
|
31 - Any user who do not use any of problematic encodings. |
|
32 |
|
33 Note that there are some limitations on using this extension: |
|
34 |
|
35 - You should use single encoding in one repository. |
|
36 - If the repository path ends with 0x5c, .hg/hgrc cannot be read. |
|
37 - win32mbcs is not compatible with fixutf8 extention. |
|
38 |
|
39 By default, win32mbcs uses encoding.encoding decided by Mercurial. |
|
40 You can specify the encoding by config option:: |
|
41 |
|
42 [win32mbcs] |
|
43 encoding = sjis |
|
44 |
|
45 It is useful for the users who want to commit with UTF-8 log message. |
|
46 ''' |
|
47 |
|
48 import os, sys |
|
49 from mercurial.i18n import _ |
|
50 from mercurial import util, encoding |
|
51 |
|
52 _encoding = None # see extsetup |
|
53 |
|
54 def decode(arg): |
|
55 if isinstance(arg, str): |
|
56 uarg = arg.decode(_encoding) |
|
57 if arg == uarg.encode(_encoding): |
|
58 return uarg |
|
59 raise UnicodeError("Not local encoding") |
|
60 elif isinstance(arg, tuple): |
|
61 return tuple(map(decode, arg)) |
|
62 elif isinstance(arg, list): |
|
63 return map(decode, arg) |
|
64 elif isinstance(arg, dict): |
|
65 for k, v in arg.items(): |
|
66 arg[k] = decode(v) |
|
67 return arg |
|
68 |
|
69 def encode(arg): |
|
70 if isinstance(arg, unicode): |
|
71 return arg.encode(_encoding) |
|
72 elif isinstance(arg, tuple): |
|
73 return tuple(map(encode, arg)) |
|
74 elif isinstance(arg, list): |
|
75 return map(encode, arg) |
|
76 elif isinstance(arg, dict): |
|
77 for k, v in arg.items(): |
|
78 arg[k] = encode(v) |
|
79 return arg |
|
80 |
|
81 def appendsep(s): |
|
82 # ensure the path ends with os.sep, appending it if necessary. |
|
83 try: |
|
84 us = decode(s) |
|
85 except UnicodeError: |
|
86 us = s |
|
87 if us and us[-1] not in ':/\\': |
|
88 s += os.sep |
|
89 return s |
|
90 |
|
91 def wrapper(func, args, kwds): |
|
92 # check argument is unicode, then call original |
|
93 for arg in args: |
|
94 if isinstance(arg, unicode): |
|
95 return func(*args, **kwds) |
|
96 |
|
97 try: |
|
98 # convert arguments to unicode, call func, then convert back |
|
99 return encode(func(*decode(args), **decode(kwds))) |
|
100 except UnicodeError: |
|
101 raise util.Abort(_("[win32mbcs] filename conversion failed with" |
|
102 " %s encoding\n") % (_encoding)) |
|
103 |
|
104 def wrapperforlistdir(func, args, kwds): |
|
105 # Ensure 'path' argument ends with os.sep to avoids |
|
106 # misinterpreting last 0x5c of MBCS 2nd byte as path separator. |
|
107 if args: |
|
108 args = list(args) |
|
109 args[0] = appendsep(args[0]) |
|
110 if 'path' in kwds: |
|
111 kwds['path'] = appendsep(kwds['path']) |
|
112 return func(*args, **kwds) |
|
113 |
|
114 def wrapname(name, wrapper): |
|
115 module, name = name.rsplit('.', 1) |
|
116 module = sys.modules[module] |
|
117 func = getattr(module, name) |
|
118 def f(*args, **kwds): |
|
119 return wrapper(func, args, kwds) |
|
120 try: |
|
121 f.__name__ = func.__name__ # fail with python23 |
|
122 except Exception: |
|
123 pass |
|
124 setattr(module, name, f) |
|
125 |
|
126 # List of functions to be wrapped. |
|
127 # NOTE: os.path.dirname() and os.path.basename() are safe because |
|
128 # they use result of os.path.split() |
|
129 funcs = '''os.path.join os.path.split os.path.splitext |
|
130 os.path.splitunc os.path.normpath os.path.normcase os.makedirs |
|
131 mercurial.util.endswithsep mercurial.util.splitpath mercurial.util.checkcase |
|
132 mercurial.util.fspath mercurial.util.pconvert mercurial.util.normpath''' |
|
133 |
|
134 # codec and alias names of sjis and big5 to be faked. |
|
135 problematic_encodings = '''big5 big5-tw csbig5 big5hkscs big5-hkscs |
|
136 hkscs cp932 932 ms932 mskanji ms-kanji shift_jis csshiftjis shiftjis |
|
137 sjis s_jis shift_jis_2004 shiftjis2004 sjis_2004 sjis2004 |
|
138 shift_jisx0213 shiftjisx0213 sjisx0213 s_jisx0213 950 cp950 ms950 ''' |
|
139 |
|
140 def extsetup(ui): |
|
141 # TODO: decide use of config section for this extension |
|
142 if not os.path.supports_unicode_filenames: |
|
143 ui.warn(_("[win32mbcs] cannot activate on this platform.\n")) |
|
144 return |
|
145 # determine encoding for filename |
|
146 global _encoding |
|
147 _encoding = ui.config('win32mbcs', 'encoding', encoding.encoding) |
|
148 # fake is only for relevant environment. |
|
149 if _encoding.lower() in problematic_encodings.split(): |
|
150 for f in funcs.split(): |
|
151 wrapname(f, wrapper) |
|
152 wrapname("mercurial.osutil.listdir", wrapperforlistdir) |
|
153 # Check sys.args manually instead of using ui.debug() because |
|
154 # command line options is not yet applied when |
|
155 # extensions.loadall() is called. |
|
156 if '--debug' in sys.argv: |
|
157 ui.write("[win32mbcs] activated with encoding: %s\n" |
|
158 % _encoding) |
|
159 |