##// END OF EJS Templates
win32mbcs: wrapper supports keyword arguments and dict result....
Shun-ichi GOTO -
r9131:2bbb8419 default
parent child Browse files
Show More
@@ -1,119 +1,125 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 * Japanese Windows users using shift_jis encoding.
23 23 * Chinese Windows users using big5 encoding.
24 24 * All users who use a repository with one of problematic encodings on
25 25 case-insensitive file system.
26 26
27 27 This extension is not needed for:
28 28 * Any user who use only ASCII chars in path.
29 29 * Any user who do not use any of problematic encodings.
30 30
31 31 Note that there are some limitations on using this extension:
32 32 * You should use single encoding in one repository.
33 33 * You should set same encoding for the repository by locale or
34 34 HGENCODING.
35 35
36 36 Path encoding conversion are done between Unicode and
37 37 encoding.encoding which is decided by Mercurial from current locale
38 38 setting or HGENCODING.
39 39 '''
40 40
41 41 import os, sys
42 42 from mercurial.i18n import _
43 43 from mercurial import util, encoding
44 44
45 45 def decode(arg):
46 46 if isinstance(arg, str):
47 47 uarg = arg.decode(encoding.encoding)
48 48 if arg == uarg.encode(encoding.encoding):
49 49 return uarg
50 50 raise UnicodeError("Not local encoding")
51 51 elif isinstance(arg, tuple):
52 52 return tuple(map(decode, arg))
53 53 elif isinstance(arg, list):
54 54 return map(decode, arg)
55 elif isinstance(arg, dict):
56 for k, v in arg.items():
57 arg[k] = decode(v)
55 58 return arg
56 59
57 60 def encode(arg):
58 61 if isinstance(arg, unicode):
59 62 return arg.encode(encoding.encoding)
60 63 elif isinstance(arg, tuple):
61 64 return tuple(map(encode, arg))
62 65 elif isinstance(arg, list):
63 66 return map(encode, arg)
67 elif isinstance(arg, dict):
68 for k, v in arg.items():
69 arg[k] = encode(v)
64 70 return arg
65 71
66 def wrapper(func, args):
72 def wrapper(func, args, kwds):
67 73 # check argument is unicode, then call original
68 74 for arg in args:
69 75 if isinstance(arg, unicode):
70 return func(*args)
76 return func(*args, **kwds)
71 77
72 78 try:
73 79 # convert arguments to unicode, call func, then convert back
74 return encode(func(*decode(args)))
80 return encode(func(*decode(args), **decode(kwds)))
75 81 except UnicodeError:
76 82 # If not encoded with encoding.encoding, report it then
77 83 # continue with calling original function.
78 84 raise util.Abort(_("[win32mbcs] filename conversion fail with"
79 85 " %s encoding\n") % (encoding.encoding))
80 86
81 87 def wrapname(name):
82 88 module, name = name.rsplit('.', 1)
83 89 module = sys.modules[module]
84 90 func = getattr(module, name)
85 def f(*args):
86 return wrapper(func, args)
91 def f(*args, **kwds):
92 return wrapper(func, args, kwds)
87 93 try:
88 94 f.__name__ = func.__name__ # fail with python23
89 95 except Exception:
90 96 pass
91 97 setattr(module, name, f)
92 98
93 99 # List of functions to be wrapped.
94 100 # NOTE: os.path.dirname() and os.path.basename() are safe because
95 101 # they use result of os.path.split()
96 102 funcs = '''os.path.join os.path.split os.path.splitext
97 103 os.path.splitunc os.path.normpath os.path.normcase os.makedirs
98 104 mercurial.util.endswithsep mercurial.util.splitpath mercurial.util.checkcase
99 105 mercurial.util.fspath mercurial.windows.pconvert'''
100 106
101 107 # codec and alias names of sjis and big5 to be faked.
102 108 problematic_encodings = '''big5 big5-tw csbig5 big5hkscs big5-hkscs
103 109 hkscs cp932 932 ms932 mskanji ms-kanji shift_jis csshiftjis shiftjis
104 110 sjis s_jis shift_jis_2004 shiftjis2004 sjis_2004 sjis2004
105 111 shift_jisx0213 shiftjisx0213 sjisx0213 s_jisx0213 950 cp950 ms950 '''
106 112
107 113 def reposetup(ui, repo):
108 114 # TODO: decide use of config section for this extension
109 115 if not os.path.supports_unicode_filenames:
110 116 ui.warn(_("[win32mbcs] cannot activate on this platform.\n"))
111 117 return
112 118
113 119 # fake is only for relevant environment.
114 120 if encoding.encoding.lower() in problematic_encodings.split():
115 121 for f in funcs.split():
116 122 wrapname(f)
117 123 ui.debug(_("[win32mbcs] activated with encoding: %s\n")
118 124 % encoding.encoding)
119 125
General Comments 0
You need to be logged in to leave comments. Login now