Show More
@@ -1,156 +1,154 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: 0.3 |
|
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 | |
|
37 | 37 | |
|
38 |
By default, win32mbcs uses encoding.encoding decided by |
|
|
39 |
You can specify the encoding by config option |
|
|
38 | By default, win32mbcs uses encoding.encoding decided by Mercurial. | |
|
39 | You can specify the encoding by config option:: | |
|
40 | 40 | |
|
41 | ex.) | |
|
42 | 41 | [win32mbcs] |
|
43 | 42 | encoding = sjis |
|
44 | 43 | |
|
45 |
It is useful |
|
|
46 | ||
|
44 | It is useful for the users who want to commit with UTF-8 log message. | |
|
47 | 45 | ''' |
|
48 | 46 | |
|
49 | 47 | import os, sys |
|
50 | 48 | from mercurial.i18n import _ |
|
51 | 49 | from mercurial import util, encoding |
|
52 | 50 | |
|
53 | 51 | _encoding = None # see reposetup() |
|
54 | 52 | |
|
55 | 53 | def decode(arg): |
|
56 | 54 | if isinstance(arg, str): |
|
57 | 55 | uarg = arg.decode(_encoding) |
|
58 | 56 | if arg == uarg.encode(_encoding): |
|
59 | 57 | return uarg |
|
60 | 58 | raise UnicodeError("Not local encoding") |
|
61 | 59 | elif isinstance(arg, tuple): |
|
62 | 60 | return tuple(map(decode, arg)) |
|
63 | 61 | elif isinstance(arg, list): |
|
64 | 62 | return map(decode, arg) |
|
65 | 63 | elif isinstance(arg, dict): |
|
66 | 64 | for k, v in arg.items(): |
|
67 | 65 | arg[k] = decode(v) |
|
68 | 66 | return arg |
|
69 | 67 | |
|
70 | 68 | def encode(arg): |
|
71 | 69 | if isinstance(arg, unicode): |
|
72 | 70 | return arg.encode(_encoding) |
|
73 | 71 | elif isinstance(arg, tuple): |
|
74 | 72 | return tuple(map(encode, arg)) |
|
75 | 73 | elif isinstance(arg, list): |
|
76 | 74 | return map(encode, arg) |
|
77 | 75 | elif isinstance(arg, dict): |
|
78 | 76 | for k, v in arg.items(): |
|
79 | 77 | arg[k] = encode(v) |
|
80 | 78 | return arg |
|
81 | 79 | |
|
82 | 80 | def appendsep(s): |
|
83 | 81 | # ensure the path ends with os.sep, appending it if necessary. |
|
84 | 82 | try: |
|
85 | 83 | us = decode(s) |
|
86 | 84 | except UnicodeError: |
|
87 | 85 | us = s |
|
88 | 86 | if us and us[-1] not in ':/\\': |
|
89 | 87 | s += os.sep |
|
90 | 88 | return s |
|
91 | 89 | |
|
92 | 90 | def wrapper(func, args, kwds): |
|
93 | 91 | # check argument is unicode, then call original |
|
94 | 92 | for arg in args: |
|
95 | 93 | if isinstance(arg, unicode): |
|
96 | 94 | return func(*args, **kwds) |
|
97 | 95 | |
|
98 | 96 | try: |
|
99 | 97 | # convert arguments to unicode, call func, then convert back |
|
100 | 98 | return encode(func(*decode(args), **decode(kwds))) |
|
101 | 99 | except UnicodeError: |
|
102 | 100 | raise util.Abort(_("[win32mbcs] filename conversion failed with" |
|
103 | 101 | " %s encoding\n") % (_encoding)) |
|
104 | 102 | |
|
105 | 103 | def wrapperforlistdir(func, args, kwds): |
|
106 | 104 | # Ensure 'path' argument ends with os.sep to avoids |
|
107 | 105 | # misinterpreting last 0x5c of MBCS 2nd byte as path separator. |
|
108 | 106 | if args: |
|
109 | 107 | args = list(args) |
|
110 | 108 | args[0] = appendsep(args[0]) |
|
111 | 109 | if 'path' in kwds: |
|
112 | 110 | kwds['path'] = appendsep(kwds['path']) |
|
113 | 111 | return func(*args, **kwds) |
|
114 | 112 | |
|
115 | 113 | def wrapname(name, wrapper): |
|
116 | 114 | module, name = name.rsplit('.', 1) |
|
117 | 115 | module = sys.modules[module] |
|
118 | 116 | func = getattr(module, name) |
|
119 | 117 | def f(*args, **kwds): |
|
120 | 118 | return wrapper(func, args, kwds) |
|
121 | 119 | try: |
|
122 | 120 | f.__name__ = func.__name__ # fail with python23 |
|
123 | 121 | except Exception: |
|
124 | 122 | pass |
|
125 | 123 | setattr(module, name, f) |
|
126 | 124 | |
|
127 | 125 | # List of functions to be wrapped. |
|
128 | 126 | # NOTE: os.path.dirname() and os.path.basename() are safe because |
|
129 | 127 | # they use result of os.path.split() |
|
130 | 128 | funcs = '''os.path.join os.path.split os.path.splitext |
|
131 | 129 | os.path.splitunc os.path.normpath os.path.normcase os.makedirs |
|
132 | 130 | mercurial.util.endswithsep mercurial.util.splitpath mercurial.util.checkcase |
|
133 | 131 | mercurial.util.fspath mercurial.util.pconvert mercurial.util.normpath''' |
|
134 | 132 | |
|
135 | 133 | # codec and alias names of sjis and big5 to be faked. |
|
136 | 134 | problematic_encodings = '''big5 big5-tw csbig5 big5hkscs big5-hkscs |
|
137 | 135 | hkscs cp932 932 ms932 mskanji ms-kanji shift_jis csshiftjis shiftjis |
|
138 | 136 | sjis s_jis shift_jis_2004 shiftjis2004 sjis_2004 sjis2004 |
|
139 | 137 | shift_jisx0213 shiftjisx0213 sjisx0213 s_jisx0213 950 cp950 ms950 ''' |
|
140 | 138 | |
|
141 | 139 | def reposetup(ui, repo): |
|
142 | 140 | # TODO: decide use of config section for this extension |
|
143 | 141 | if not os.path.supports_unicode_filenames: |
|
144 | 142 | ui.warn(_("[win32mbcs] cannot activate on this platform.\n")) |
|
145 | 143 | return |
|
146 | 144 | # determine encoding for filename |
|
147 | 145 | global _encoding |
|
148 | 146 | _encoding = ui.config('win32mbcs', 'encoding', encoding.encoding) |
|
149 | 147 | # fake is only for relevant environment. |
|
150 | 148 | if _encoding.lower() in problematic_encodings.split(): |
|
151 | 149 | for f in funcs.split(): |
|
152 | 150 | wrapname(f, wrapper) |
|
153 | 151 | wrapname("mercurial.osutil.listdir", wrapperforlistdir) |
|
154 | 152 | ui.debug("[win32mbcs] activated with encoding: %s\n" |
|
155 | 153 | % _encoding) |
|
156 | 154 |
General Comments 0
You need to be logged in to leave comments.
Login now