Show More
@@ -1,189 +1,196 | |||
|
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 or any later version. |
|
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 | - If the repository path ends with 0x5c, .hg/hgrc cannot be read. |
|
37 | 37 | - win32mbcs is not compatible with fixutf8 extension. |
|
38 | 38 | |
|
39 | 39 | By default, win32mbcs uses encoding.encoding decided by Mercurial. |
|
40 | 40 | You can specify the encoding by config option:: |
|
41 | 41 | |
|
42 | 42 | [win32mbcs] |
|
43 | 43 | encoding = sjis |
|
44 | 44 | |
|
45 | 45 | It is useful for the users who want to commit with UTF-8 log message. |
|
46 | 46 | ''' |
|
47 | from __future__ import absolute_import | |
|
47 | 48 | |
|
48 |
import os |
|
|
49 | import os | |
|
50 | import sys | |
|
51 | ||
|
52 | from mercurial import ( | |
|
53 | encoding, | |
|
54 | error, | |
|
55 | ) | |
|
49 | 56 | from mercurial.i18n import _ |
|
50 | from mercurial import error, encoding | |
|
57 | ||
|
51 | 58 | # Note for extension authors: ONLY specify testedwith = 'internal' for |
|
52 | 59 | # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should |
|
53 | 60 | # be specifying the version(s) of Mercurial they are tested with, or |
|
54 | 61 | # leave the attribute unspecified. |
|
55 | 62 | testedwith = 'internal' |
|
56 | 63 | |
|
57 | 64 | _encoding = None # see extsetup |
|
58 | 65 | |
|
59 | 66 | def decode(arg): |
|
60 | 67 | if isinstance(arg, str): |
|
61 | 68 | uarg = arg.decode(_encoding) |
|
62 | 69 | if arg == uarg.encode(_encoding): |
|
63 | 70 | return uarg |
|
64 | 71 | raise UnicodeError("Not local encoding") |
|
65 | 72 | elif isinstance(arg, tuple): |
|
66 | 73 | return tuple(map(decode, arg)) |
|
67 | 74 | elif isinstance(arg, list): |
|
68 | 75 | return map(decode, arg) |
|
69 | 76 | elif isinstance(arg, dict): |
|
70 | 77 | for k, v in arg.items(): |
|
71 | 78 | arg[k] = decode(v) |
|
72 | 79 | return arg |
|
73 | 80 | |
|
74 | 81 | def encode(arg): |
|
75 | 82 | if isinstance(arg, unicode): |
|
76 | 83 | return arg.encode(_encoding) |
|
77 | 84 | elif isinstance(arg, tuple): |
|
78 | 85 | return tuple(map(encode, arg)) |
|
79 | 86 | elif isinstance(arg, list): |
|
80 | 87 | return map(encode, arg) |
|
81 | 88 | elif isinstance(arg, dict): |
|
82 | 89 | for k, v in arg.items(): |
|
83 | 90 | arg[k] = encode(v) |
|
84 | 91 | return arg |
|
85 | 92 | |
|
86 | 93 | def appendsep(s): |
|
87 | 94 | # ensure the path ends with os.sep, appending it if necessary. |
|
88 | 95 | try: |
|
89 | 96 | us = decode(s) |
|
90 | 97 | except UnicodeError: |
|
91 | 98 | us = s |
|
92 | 99 | if us and us[-1] not in ':/\\': |
|
93 | 100 | s += os.sep |
|
94 | 101 | return s |
|
95 | 102 | |
|
96 | 103 | |
|
97 | 104 | def basewrapper(func, argtype, enc, dec, args, kwds): |
|
98 | 105 | # check check already converted, then call original |
|
99 | 106 | for arg in args: |
|
100 | 107 | if isinstance(arg, argtype): |
|
101 | 108 | return func(*args, **kwds) |
|
102 | 109 | |
|
103 | 110 | try: |
|
104 | 111 | # convert string arguments, call func, then convert back the |
|
105 | 112 | # return value. |
|
106 | 113 | return enc(func(*dec(args), **dec(kwds))) |
|
107 | 114 | except UnicodeError: |
|
108 | 115 | raise error.Abort(_("[win32mbcs] filename conversion failed with" |
|
109 | 116 | " %s encoding\n") % (_encoding)) |
|
110 | 117 | |
|
111 | 118 | def wrapper(func, args, kwds): |
|
112 | 119 | return basewrapper(func, unicode, encode, decode, args, kwds) |
|
113 | 120 | |
|
114 | 121 | |
|
115 | 122 | def reversewrapper(func, args, kwds): |
|
116 | 123 | return basewrapper(func, str, decode, encode, args, kwds) |
|
117 | 124 | |
|
118 | 125 | def wrapperforlistdir(func, args, kwds): |
|
119 | 126 | # Ensure 'path' argument ends with os.sep to avoids |
|
120 | 127 | # misinterpreting last 0x5c of MBCS 2nd byte as path separator. |
|
121 | 128 | if args: |
|
122 | 129 | args = list(args) |
|
123 | 130 | args[0] = appendsep(args[0]) |
|
124 | 131 | if 'path' in kwds: |
|
125 | 132 | kwds['path'] = appendsep(kwds['path']) |
|
126 | 133 | return func(*args, **kwds) |
|
127 | 134 | |
|
128 | 135 | def wrapname(name, wrapper): |
|
129 | 136 | module, name = name.rsplit('.', 1) |
|
130 | 137 | module = sys.modules[module] |
|
131 | 138 | func = getattr(module, name) |
|
132 | 139 | def f(*args, **kwds): |
|
133 | 140 | return wrapper(func, args, kwds) |
|
134 | 141 | try: |
|
135 | 142 | f.__name__ = func.__name__ # fails with Python 2.3 |
|
136 | 143 | except Exception: |
|
137 | 144 | pass |
|
138 | 145 | setattr(module, name, f) |
|
139 | 146 | |
|
140 | 147 | # List of functions to be wrapped. |
|
141 | 148 | # NOTE: os.path.dirname() and os.path.basename() are safe because |
|
142 | 149 | # they use result of os.path.split() |
|
143 | 150 | funcs = '''os.path.join os.path.split os.path.splitext |
|
144 | 151 | os.path.normpath os.makedirs |
|
145 | 152 | mercurial.util.endswithsep mercurial.util.splitpath mercurial.util.checkcase |
|
146 | 153 | mercurial.util.fspath mercurial.util.pconvert mercurial.util.normpath |
|
147 | 154 | mercurial.util.checkwinfilename mercurial.util.checkosfilename |
|
148 | 155 | mercurial.util.split''' |
|
149 | 156 | |
|
150 | 157 | # These functions are required to be called with local encoded string |
|
151 | 158 | # because they expects argument is local encoded string and cause |
|
152 | 159 | # problem with unicode string. |
|
153 | 160 | rfuncs = '''mercurial.encoding.upper mercurial.encoding.lower''' |
|
154 | 161 | |
|
155 | 162 | # List of Windows specific functions to be wrapped. |
|
156 | 163 | winfuncs = '''os.path.splitunc''' |
|
157 | 164 | |
|
158 | 165 | # codec and alias names of sjis and big5 to be faked. |
|
159 | 166 | problematic_encodings = '''big5 big5-tw csbig5 big5hkscs big5-hkscs |
|
160 | 167 | hkscs cp932 932 ms932 mskanji ms-kanji shift_jis csshiftjis shiftjis |
|
161 | 168 | sjis s_jis shift_jis_2004 shiftjis2004 sjis_2004 sjis2004 |
|
162 | 169 | shift_jisx0213 shiftjisx0213 sjisx0213 s_jisx0213 950 cp950 ms950 ''' |
|
163 | 170 | |
|
164 | 171 | def extsetup(ui): |
|
165 | 172 | # TODO: decide use of config section for this extension |
|
166 | 173 | if ((not os.path.supports_unicode_filenames) and |
|
167 | 174 | (sys.platform != 'cygwin')): |
|
168 | 175 | ui.warn(_("[win32mbcs] cannot activate on this platform.\n")) |
|
169 | 176 | return |
|
170 | 177 | # determine encoding for filename |
|
171 | 178 | global _encoding |
|
172 | 179 | _encoding = ui.config('win32mbcs', 'encoding', encoding.encoding) |
|
173 | 180 | # fake is only for relevant environment. |
|
174 | 181 | if _encoding.lower() in problematic_encodings.split(): |
|
175 | 182 | for f in funcs.split(): |
|
176 | 183 | wrapname(f, wrapper) |
|
177 | 184 | if os.name == 'nt': |
|
178 | 185 | for f in winfuncs.split(): |
|
179 | 186 | wrapname(f, wrapper) |
|
180 | 187 | wrapname("mercurial.osutil.listdir", wrapperforlistdir) |
|
181 | 188 | # wrap functions to be called with local byte string arguments |
|
182 | 189 | for f in rfuncs.split(): |
|
183 | 190 | wrapname(f, reversewrapper) |
|
184 | 191 | # Check sys.args manually instead of using ui.debug() because |
|
185 | 192 | # command line options is not yet applied when |
|
186 | 193 | # extensions.loadall() is called. |
|
187 | 194 | if '--debug' in sys.argv: |
|
188 | 195 | ui.write("[win32mbcs] activated with encoding: %s\n" |
|
189 | 196 | % _encoding) |
@@ -1,130 +1,129 | |||
|
1 | 1 | #require test-repo |
|
2 | 2 | |
|
3 | 3 | $ cd "$TESTDIR"/.. |
|
4 | 4 | |
|
5 | 5 | $ hg files 'set:(**.py)' | sed 's|\\|/|g' | xargs python contrib/check-py3-compat.py |
|
6 | 6 | contrib/check-code.py not using absolute_import |
|
7 | 7 | contrib/check-code.py requires print_function |
|
8 | 8 | contrib/debugshell.py not using absolute_import |
|
9 | 9 | contrib/import-checker.py not using absolute_import |
|
10 | 10 | contrib/import-checker.py requires print_function |
|
11 | 11 | contrib/memory.py not using absolute_import |
|
12 | 12 | contrib/perf.py not using absolute_import |
|
13 | 13 | contrib/python-hook-examples.py not using absolute_import |
|
14 | 14 | contrib/revsetbenchmarks.py not using absolute_import |
|
15 | 15 | contrib/revsetbenchmarks.py requires print_function |
|
16 | 16 | contrib/showstack.py not using absolute_import |
|
17 | 17 | contrib/synthrepo.py not using absolute_import |
|
18 | 18 | contrib/win32/hgwebdir_wsgi.py not using absolute_import |
|
19 | 19 | doc/check-seclevel.py not using absolute_import |
|
20 | 20 | doc/gendoc.py not using absolute_import |
|
21 | 21 | doc/hgmanpage.py not using absolute_import |
|
22 | 22 | hgext/__init__.py not using absolute_import |
|
23 | 23 | hgext/color.py not using absolute_import |
|
24 | 24 | hgext/eol.py not using absolute_import |
|
25 | 25 | hgext/extdiff.py not using absolute_import |
|
26 | 26 | hgext/factotum.py not using absolute_import |
|
27 | 27 | hgext/fetch.py not using absolute_import |
|
28 | 28 | hgext/gpg.py not using absolute_import |
|
29 | 29 | hgext/graphlog.py not using absolute_import |
|
30 | 30 | hgext/hgcia.py not using absolute_import |
|
31 | 31 | hgext/hgk.py not using absolute_import |
|
32 | 32 | hgext/highlight/__init__.py not using absolute_import |
|
33 | 33 | hgext/highlight/highlight.py not using absolute_import |
|
34 | 34 | hgext/histedit.py not using absolute_import |
|
35 | 35 | hgext/largefiles/__init__.py not using absolute_import |
|
36 | 36 | hgext/largefiles/basestore.py not using absolute_import |
|
37 | 37 | hgext/largefiles/lfcommands.py not using absolute_import |
|
38 | 38 | hgext/largefiles/lfutil.py not using absolute_import |
|
39 | 39 | hgext/largefiles/localstore.py not using absolute_import |
|
40 | 40 | hgext/largefiles/overrides.py not using absolute_import |
|
41 | 41 | hgext/largefiles/proto.py not using absolute_import |
|
42 | 42 | hgext/largefiles/remotestore.py not using absolute_import |
|
43 | 43 | hgext/largefiles/reposetup.py not using absolute_import |
|
44 | 44 | hgext/largefiles/uisetup.py not using absolute_import |
|
45 | 45 | hgext/largefiles/wirestore.py not using absolute_import |
|
46 | 46 | hgext/mq.py not using absolute_import |
|
47 | 47 | hgext/rebase.py not using absolute_import |
|
48 | 48 | hgext/share.py not using absolute_import |
|
49 | 49 | hgext/transplant.py not using absolute_import |
|
50 | hgext/win32mbcs.py not using absolute_import | |
|
51 | 50 | hgext/win32text.py not using absolute_import |
|
52 | 51 | i18n/check-translation.py not using absolute_import |
|
53 | 52 | i18n/polib.py not using absolute_import |
|
54 | 53 | setup.py not using absolute_import |
|
55 | 54 | tests/filterpyflakes.py requires print_function |
|
56 | 55 | tests/generate-working-copy-states.py requires print_function |
|
57 | 56 | tests/get-with-headers.py requires print_function |
|
58 | 57 | tests/heredoctest.py requires print_function |
|
59 | 58 | tests/hypothesishelpers.py not using absolute_import |
|
60 | 59 | tests/hypothesishelpers.py requires print_function |
|
61 | 60 | tests/killdaemons.py not using absolute_import |
|
62 | 61 | tests/md5sum.py not using absolute_import |
|
63 | 62 | tests/mockblackbox.py not using absolute_import |
|
64 | 63 | tests/printenv.py not using absolute_import |
|
65 | 64 | tests/readlink.py not using absolute_import |
|
66 | 65 | tests/readlink.py requires print_function |
|
67 | 66 | tests/revlog-formatv0.py not using absolute_import |
|
68 | 67 | tests/run-tests.py not using absolute_import |
|
69 | 68 | tests/seq.py not using absolute_import |
|
70 | 69 | tests/seq.py requires print_function |
|
71 | 70 | tests/silenttestrunner.py not using absolute_import |
|
72 | 71 | tests/silenttestrunner.py requires print_function |
|
73 | 72 | tests/sitecustomize.py not using absolute_import |
|
74 | 73 | tests/svn-safe-append.py not using absolute_import |
|
75 | 74 | tests/svnxml.py not using absolute_import |
|
76 | 75 | tests/test-ancestor.py requires print_function |
|
77 | 76 | tests/test-atomictempfile.py not using absolute_import |
|
78 | 77 | tests/test-batching.py not using absolute_import |
|
79 | 78 | tests/test-batching.py requires print_function |
|
80 | 79 | tests/test-bdiff.py not using absolute_import |
|
81 | 80 | tests/test-bdiff.py requires print_function |
|
82 | 81 | tests/test-context.py not using absolute_import |
|
83 | 82 | tests/test-context.py requires print_function |
|
84 | 83 | tests/test-demandimport.py not using absolute_import |
|
85 | 84 | tests/test-demandimport.py requires print_function |
|
86 | 85 | tests/test-doctest.py not using absolute_import |
|
87 | 86 | tests/test-duplicateoptions.py not using absolute_import |
|
88 | 87 | tests/test-duplicateoptions.py requires print_function |
|
89 | 88 | tests/test-filecache.py not using absolute_import |
|
90 | 89 | tests/test-filecache.py requires print_function |
|
91 | 90 | tests/test-filelog.py not using absolute_import |
|
92 | 91 | tests/test-filelog.py requires print_function |
|
93 | 92 | tests/test-hg-parseurl.py not using absolute_import |
|
94 | 93 | tests/test-hg-parseurl.py requires print_function |
|
95 | 94 | tests/test-hgweb-auth.py not using absolute_import |
|
96 | 95 | tests/test-hgweb-auth.py requires print_function |
|
97 | 96 | tests/test-hgwebdir-paths.py not using absolute_import |
|
98 | 97 | tests/test-hybridencode.py not using absolute_import |
|
99 | 98 | tests/test-hybridencode.py requires print_function |
|
100 | 99 | tests/test-lrucachedict.py not using absolute_import |
|
101 | 100 | tests/test-lrucachedict.py requires print_function |
|
102 | 101 | tests/test-manifest.py not using absolute_import |
|
103 | 102 | tests/test-minirst.py not using absolute_import |
|
104 | 103 | tests/test-minirst.py requires print_function |
|
105 | 104 | tests/test-parseindex2.py not using absolute_import |
|
106 | 105 | tests/test-parseindex2.py requires print_function |
|
107 | 106 | tests/test-pathencode.py not using absolute_import |
|
108 | 107 | tests/test-pathencode.py requires print_function |
|
109 | 108 | tests/test-propertycache.py not using absolute_import |
|
110 | 109 | tests/test-propertycache.py requires print_function |
|
111 | 110 | tests/test-revlog-ancestry.py not using absolute_import |
|
112 | 111 | tests/test-revlog-ancestry.py requires print_function |
|
113 | 112 | tests/test-run-tests.py not using absolute_import |
|
114 | 113 | tests/test-simplemerge.py not using absolute_import |
|
115 | 114 | tests/test-status-inprocess.py not using absolute_import |
|
116 | 115 | tests/test-status-inprocess.py requires print_function |
|
117 | 116 | tests/test-symlink-os-yes-fs-no.py not using absolute_import |
|
118 | 117 | tests/test-trusted.py not using absolute_import |
|
119 | 118 | tests/test-trusted.py requires print_function |
|
120 | 119 | tests/test-ui-color.py not using absolute_import |
|
121 | 120 | tests/test-ui-color.py requires print_function |
|
122 | 121 | tests/test-ui-config.py not using absolute_import |
|
123 | 122 | tests/test-ui-config.py requires print_function |
|
124 | 123 | tests/test-ui-verbosity.py not using absolute_import |
|
125 | 124 | tests/test-ui-verbosity.py requires print_function |
|
126 | 125 | tests/test-url.py not using absolute_import |
|
127 | 126 | tests/test-url.py requires print_function |
|
128 | 127 | tests/test-walkrepo.py requires print_function |
|
129 | 128 | tests/test-wireproto.py requires print_function |
|
130 | 129 | tests/tinyproxy.py requires print_function |
General Comments 0
You need to be logged in to leave comments.
Login now