/brz/remove-bazaar

To get this branch, use:
bzr branch http://gegoxaren.bato24.eu/bzr/brz/remove-bazaar
2095.3.1 by Martin Pool
Tests shouldn't assume os.listdir returns sorted results
1
# Copyright (C) 2005, 2006 Canonical Ltd
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
2
#
1 by mbp at sourcefrog
import from baz patch-364
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
7
#
1 by mbp at sourcefrog
import from baz patch-364
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
11
# GNU General Public License for more details.
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
12
#
1 by mbp at sourcefrog
import from baz patch-364
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
16
1390 by Robert Collins
pair programming worx... merge integration and weave
17
from cStringIO import StringIO
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
18
import os
19
import re
20
import stat
21
from stat import (S_ISREG, S_ISDIR, S_ISLNK, ST_MODE, ST_SIZE,
22
                  S_ISCHR, S_ISBLK, S_ISFIFO, S_ISSOCK)
23
import sys
24
import time
25
26
from bzrlib.lazy_import import lazy_import
27
lazy_import(globals(), """
2192.1.1 by Alexander Belchenko
Before actually using encoding need to check that Python has corresponding codec
28
import codecs
2215.6.1 by James Henstridge
Don't rely on time.timezone and time.altzone in local_time_offset(),
29
from datetime import datetime
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
30
import errno
1711.4.5 by John Arbash Meinel
the _posix_* routines should use posixpath not os.path, so tests pass on win32
31
from ntpath import (abspath as _nt_abspath,
32
                    join as _nt_join,
33
                    normpath as _nt_normpath,
34
                    realpath as _nt_realpath,
1711.5.2 by John Arbash Meinel
win32 likes to return lowercase drive letters sometimes, and uppercase at other times. normalize this
35
                    splitdrive as _nt_splitdrive,
1711.4.5 by John Arbash Meinel
the _posix_* routines should use posixpath not os.path, so tests pass on win32
36
                    )
37
import posixpath
1236 by Martin Pool
- fix up imports
38
import sha
1692.7.6 by Martin Pool
[patch] force deletion of trees containing readonly files (alexander)
39
import shutil
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
40
from shutil import (
41
    rmtree,
42
    )
1185.31.40 by John Arbash Meinel
Added osutils.mkdtemp()
43
import tempfile
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
44
from tempfile import (
45
    mkdtemp,
46
    )
1185.85.75 by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths.
47
import unicodedata
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
48
49
from bzrlib import (
50
    errors,
2245.4.6 by Alexander Belchenko
osutils.py: terminal_width() now use win32utils.get_console_size()
51
    win32utils,
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
52
    )
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
53
""")
1 by mbp at sourcefrog
import from baz patch-364
54
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
55
import bzrlib
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
56
from bzrlib.symbol_versioning import (
57
    deprecated_function,
58
    zero_nine,
59
    )
694 by Martin Pool
- weed out all remaining calls to bailout() and remove the function
60
from bzrlib.trace import mutter
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
61
1 by mbp at sourcefrog
import from baz patch-364
62
1755.3.7 by John Arbash Meinel
Clean up and write tests for permissions. Now we use fstat which should be cheap, and lets us check the permissions and the file size
63
# On win32, O_BINARY is used to indicate the file should
64
# be opened in binary mode, rather than text mode.
65
# On other platforms, O_BINARY doesn't exist, because
66
# they always open in binary mode, so it is okay to
67
# OR with 0 on those platforms
68
O_BINARY = getattr(os, 'O_BINARY', 0)
69
70
1 by mbp at sourcefrog
import from baz patch-364
71
def make_readonly(filename):
72
    """Make a filename read-only."""
73
    mod = os.stat(filename).st_mode
74
    mod = mod & 0777555
75
    os.chmod(filename, mod)
76
77
78
def make_writable(filename):
79
    mod = os.stat(filename).st_mode
80
    mod = mod | 0200
81
    os.chmod(filename, mod)
82
83
1077 by Martin Pool
- avoid compiling REs at module load time
84
_QUOTE_RE = None
969 by Martin Pool
- Add less-sucky is_within_any
85
86
1 by mbp at sourcefrog
import from baz patch-364
87
def quotefn(f):
779 by Martin Pool
- better quotefn for windows: use doublequotes for strings with
88
    """Return a quoted filename filename
89
90
    This previously used backslash quoting, but that works poorly on
91
    Windows."""
92
    # TODO: I'm not really sure this is the best format either.x
1077 by Martin Pool
- avoid compiling REs at module load time
93
    global _QUOTE_RE
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
94
    if _QUOTE_RE is None:
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
95
        _QUOTE_RE = re.compile(r'([^a-zA-Z0-9.,:/\\_~-])')
1077 by Martin Pool
- avoid compiling REs at module load time
96
        
779 by Martin Pool
- better quotefn for windows: use doublequotes for strings with
97
    if _QUOTE_RE.search(f):
98
        return '"' + f + '"'
99
    else:
100
        return f
1 by mbp at sourcefrog
import from baz patch-364
101
102
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
103
_directory_kind = 'directory'
104
1732.1.10 by John Arbash Meinel
Updated version of file_kind. Rather than multiple function calls, one mask + dictionary lookup
105
_formats = {
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
106
    stat.S_IFDIR:_directory_kind,
1732.1.10 by John Arbash Meinel
Updated version of file_kind. Rather than multiple function calls, one mask + dictionary lookup
107
    stat.S_IFCHR:'chardev',
108
    stat.S_IFBLK:'block',
109
    stat.S_IFREG:'file',
110
    stat.S_IFIFO:'fifo',
111
    stat.S_IFLNK:'symlink',
112
    stat.S_IFSOCK:'socket',
113
}
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
114
115
116
def file_kind_from_stat_mode(stat_mode, _formats=_formats, _unknown='unknown'):
117
    """Generate a file kind from a stat mode. This is used in walkdirs.
118
119
    Its performance is critical: Do not mutate without careful benchmarking.
120
    """
1732.1.12 by John Arbash Meinel
improve bzrlib.osutils.file_kind performance from 324ms => 275ms
121
    try:
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
122
        return _formats[stat_mode & 0170000]
1732.1.12 by John Arbash Meinel
improve bzrlib.osutils.file_kind performance from 324ms => 275ms
123
    except KeyError:
1732.1.30 by John Arbash Meinel
More file_kind tweaks. Use keyword parameters to make everything a local variable.
124
        return _unknown
488 by Martin Pool
- new helper function kind_marker()
125
126
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
127
def file_kind(f, _lstat=os.lstat, _mapper=file_kind_from_stat_mode):
1757.2.4 by Robert Collins
Teach file_kind about NoSuchFile, reducing duplicate code, and add user files before entering the main loop in smart_add.
128
    try:
129
        return _mapper(_lstat(f).st_mode)
130
    except OSError, e:
131
        if getattr(e, 'errno', None) == errno.ENOENT:
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
132
            raise errors.NoSuchFile(f)
1757.2.4 by Robert Collins
Teach file_kind about NoSuchFile, reducing duplicate code, and add user files before entering the main loop in smart_add.
133
        raise
134
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
135
1755.3.7 by John Arbash Meinel
Clean up and write tests for permissions. Now we use fstat which should be cheap, and lets us check the permissions and the file size
136
def get_umask():
137
    """Return the current umask"""
138
    # Assume that people aren't messing with the umask while running
139
    # XXX: This is not thread safe, but there is no way to get the
140
    #      umask without setting it
141
    umask = os.umask(0)
142
    os.umask(umask)
143
    return umask
144
145
488 by Martin Pool
- new helper function kind_marker()
146
def kind_marker(kind):
147
    if kind == 'file':
148
        return ''
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
149
    elif kind == _directory_kind:
488 by Martin Pool
- new helper function kind_marker()
150
        return '/'
151
    elif kind == 'symlink':
152
        return '@'
153
    else:
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
154
        raise errors.BzrError('invalid file kind %r' % kind)
1 by mbp at sourcefrog
import from baz patch-364
155
1732.1.2 by John Arbash Meinel
just use os.path.lexists if it exists
156
lexists = getattr(os.path, 'lexists', None)
157
if lexists is None:
158
    def lexists(f):
159
        try:
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
160
            if getattr(os, 'lstat') is not None:
1732.1.2 by John Arbash Meinel
just use os.path.lexists if it exists
161
                os.lstat(f)
162
            else:
163
                os.stat(f)
164
            return True
165
        except OSError,e:
166
            if e.errno == errno.ENOENT:
167
                return False;
168
            else:
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
169
                raise errors.BzrError("lstat/stat of (%r): %r" % (f, e))
1732.1.2 by John Arbash Meinel
just use os.path.lexists if it exists
170
1 by mbp at sourcefrog
import from baz patch-364
171
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
172
def fancy_rename(old, new, rename_func, unlink_func):
173
    """A fancy rename, when you don't have atomic rename.
174
    
175
    :param old: The old path, to rename from
176
    :param new: The new path, to rename to
177
    :param rename_func: The potentially non-atomic rename function
178
    :param unlink_func: A way to delete the target file if the full rename succeeds
179
    """
180
181
    # sftp rename doesn't allow overwriting, so play tricks:
182
    import random
183
    base = os.path.basename(new)
184
    dirname = os.path.dirname(new)
1553.5.22 by Martin Pool
Change fancy_rename to use rand_chars rather than reinvent it.
185
    tmp_name = u'tmp.%s.%.9f.%d.%s' % (base, time.time(), os.getpid(), rand_chars(10))
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
186
    tmp_name = pathjoin(dirname, tmp_name)
187
188
    # Rename the file out of the way, but keep track if it didn't exist
189
    # We don't want to grab just any exception
190
    # something like EACCES should prevent us from continuing
191
    # The downside is that the rename_func has to throw an exception
192
    # with an errno = ENOENT, or NoSuchFile
193
    file_existed = False
194
    try:
195
        rename_func(new, tmp_name)
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
196
    except (errors.NoSuchFile,), e:
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
197
        pass
1532 by Robert Collins
Merge in John Meinels integration branch.
198
    except IOError, e:
199
        # RBC 20060103 abstraction leakage: the paramiko SFTP clients rename
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
200
        # function raises an IOError with errno is None when a rename fails.
1532 by Robert Collins
Merge in John Meinels integration branch.
201
        # This then gets caught here.
1185.50.37 by John Arbash Meinel
Fixed exception handling for fancy_rename
202
        if e.errno not in (None, errno.ENOENT, errno.ENOTDIR):
1532 by Robert Collins
Merge in John Meinels integration branch.
203
            raise
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
204
    except Exception, e:
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
205
        if (getattr(e, 'errno', None) is None
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
206
            or e.errno not in (errno.ENOENT, errno.ENOTDIR)):
207
            raise
208
    else:
209
        file_existed = True
210
211
    success = False
212
    try:
213
        # This may throw an exception, in which case success will
214
        # not be set.
215
        rename_func(old, new)
216
        success = True
217
    finally:
218
        if file_existed:
219
            # If the file used to exist, rename it back into place
220
            # otherwise just delete it from the tmp location
221
            if success:
222
                unlink_func(tmp_name)
223
            else:
1185.31.49 by John Arbash Meinel
Some corrections using the new osutils.rename. **ALL TESTS PASS**
224
                rename_func(tmp_name, new)
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
225
1685.1.9 by John Arbash Meinel
Updated LocalTransport so that it's base is now a URL rather than a local path. This helps consistency with all other functions. To do so, I added local_abspath() which returns the local path, and local_path_to/from_url
226
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
227
# In Python 2.4.2 and older, os.path.abspath and os.path.realpath
228
# choke on a Unicode string containing a relative path if
229
# os.getcwd() returns a non-sys.getdefaultencoding()-encoded
230
# string.
2093.1.1 by John Arbash Meinel
(Bart Teeuwisse) if sys.getfilesystemencoding() is None, use 'utf-8'
231
_fs_enc = sys.getfilesystemencoding() or 'utf-8'
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
232
def _posix_abspath(path):
1711.4.5 by John Arbash Meinel
the _posix_* routines should use posixpath not os.path, so tests pass on win32
233
    # jam 20060426 rather than encoding to fsencoding
234
    # copy posixpath.abspath, but use os.getcwdu instead
235
    if not posixpath.isabs(path):
236
        path = posixpath.join(getcwd(), path)
237
    return posixpath.normpath(path)
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
238
239
240
def _posix_realpath(path):
1711.4.5 by John Arbash Meinel
the _posix_* routines should use posixpath not os.path, so tests pass on win32
241
    return posixpath.realpath(path.encode(_fs_enc)).decode(_fs_enc)
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
242
243
1711.5.2 by John Arbash Meinel
win32 likes to return lowercase drive letters sometimes, and uppercase at other times. normalize this
244
def _win32_fixdrive(path):
245
    """Force drive letters to be consistent.
246
247
    win32 is inconsistent whether it returns lower or upper case
248
    and even if it was consistent the user might type the other
249
    so we force it to uppercase
250
    running python.exe under cmd.exe return capital C:\\
251
    running win32 python inside a cygwin shell returns lowercase c:\\
252
    """
253
    drive, path = _nt_splitdrive(path)
254
    return drive.upper() + path
255
256
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
257
def _win32_abspath(path):
1711.4.6 by John Arbash Meinel
Removing hacks for _win32_abspath, on real win32 abspath handles unicode just fine, it doesn't handle encoding into 'mbcs'
258
    # Real _nt_abspath doesn't have a problem with a unicode cwd
1711.5.2 by John Arbash Meinel
win32 likes to return lowercase drive letters sometimes, and uppercase at other times. normalize this
259
    return _win32_fixdrive(_nt_abspath(unicode(path)).replace('\\', '/'))
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
260
261
2279.4.1 by Alexander Belchenko
Reimplementation of ntpath.abspath in Python for Windows98: unicode safe, UNC path safe
262
def _win98_abspath(path):
263
    """Return the absolute version of a path.
264
    Windows 98 safe implementation (python reimplementation
265
    of Win32 API function GetFullPathNameW)
266
    """
267
    # Corner cases:
268
    #   C:\path     => C:/path
269
    #   C:/path     => C:/path
270
    #   \\HOST\path => //HOST/path
271
    #   //HOST/path => //HOST/path
272
    #   path        => C:/cwd/path
273
    #   /path       => C:/path
274
    path = unicode(path)
275
    # check for absolute path
276
    drive = _nt_splitdrive(path)[0]
277
    if drive == '' and path[:2] not in('//','\\\\'):
278
        cwd = os.getcwdu()
279
        # we cannot simply os.path.join cwd and path
280
        # because os.path.join('C:','/path') produce '/path'
281
        # and this is incorrect
282
        if path[:1] in ('/','\\'):
283
            cwd = _nt_splitdrive(cwd)[0]
284
        path = cwd + '\\' + path
285
    return _win32_fixdrive(_nt_normpath(path).replace('\\', '/'))
286
287
if win32utils.winver == 'Windows 98':
288
    _win32_abspath = _win98_abspath
289
290
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
291
def _win32_realpath(path):
1711.4.6 by John Arbash Meinel
Removing hacks for _win32_abspath, on real win32 abspath handles unicode just fine, it doesn't handle encoding into 'mbcs'
292
    # Real _nt_realpath doesn't have a problem with a unicode cwd
1711.5.2 by John Arbash Meinel
win32 likes to return lowercase drive letters sometimes, and uppercase at other times. normalize this
293
    return _win32_fixdrive(_nt_realpath(unicode(path)).replace('\\', '/'))
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
294
295
296
def _win32_pathjoin(*args):
1685.1.31 by John Arbash Meinel
Adding tests for the rest of the _win32 functions.
297
    return _nt_join(*args).replace('\\', '/')
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
298
299
300
def _win32_normpath(path):
1711.5.2 by John Arbash Meinel
win32 likes to return lowercase drive letters sometimes, and uppercase at other times. normalize this
301
    return _win32_fixdrive(_nt_normpath(unicode(path)).replace('\\', '/'))
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
302
303
304
def _win32_getcwd():
1711.5.2 by John Arbash Meinel
win32 likes to return lowercase drive letters sometimes, and uppercase at other times. normalize this
305
    return _win32_fixdrive(os.getcwdu().replace('\\', '/'))
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
306
307
308
def _win32_mkdtemp(*args, **kwargs):
1711.5.2 by John Arbash Meinel
win32 likes to return lowercase drive letters sometimes, and uppercase at other times. normalize this
309
    return _win32_fixdrive(tempfile.mkdtemp(*args, **kwargs).replace('\\', '/'))
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
310
311
312
def _win32_rename(old, new):
1711.7.6 by John Arbash Meinel
Change _win32_rename() so that it raises ENOENT *before* it tries any renaming.
313
    """We expect to be able to atomically replace 'new' with old.
314
1711.7.17 by John Arbash Meinel
Delay the extra syscall in _win32_rename until we get a failure.
315
    On win32, if new exists, it must be moved out of the way first,
316
    and then deleted. 
1711.7.6 by John Arbash Meinel
Change _win32_rename() so that it raises ENOENT *before* it tries any renaming.
317
    """
1711.7.17 by John Arbash Meinel
Delay the extra syscall in _win32_rename until we get a failure.
318
    try:
319
        fancy_rename(old, new, rename_func=os.rename, unlink_func=os.unlink)
320
    except OSError, e:
1830.3.15 by John Arbash Meinel
On Mac we get EINVAL when renaming cwd
321
        if e.errno in (errno.EPERM, errno.EACCES, errno.EBUSY, errno.EINVAL):
322
            # If we try to rename a non-existant file onto cwd, we get 
323
            # EPERM or EACCES instead of ENOENT, this will raise ENOENT 
324
            # if the old path doesn't exist, sometimes we get EACCES
325
            # On Linux, we seem to get EBUSY, on Mac we get EINVAL
1711.7.17 by John Arbash Meinel
Delay the extra syscall in _win32_rename until we get a failure.
326
            os.lstat(old)
327
        raise
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
328
329
1830.3.11 by John Arbash Meinel
Create a mac version of 'getcwd()' which normalizes the path.
330
def _mac_getcwd():
331
    return unicodedata.normalize('NFKC', os.getcwdu())
332
333
1692.7.6 by Martin Pool
[patch] force deletion of trees containing readonly files (alexander)
334
# Default is to just use the python builtins, but these can be rebound on
335
# particular platforms.
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
336
abspath = _posix_abspath
337
realpath = _posix_realpath
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
338
pathjoin = os.path.join
339
normpath = os.path.normpath
340
getcwd = os.getcwdu
341
rename = os.rename
342
dirname = os.path.dirname
343
basename = os.path.basename
2215.4.2 by Alexander Belchenko
split and splitext now the part of osutils
344
split = os.path.split
345
splitext = os.path.splitext
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
346
# These were already imported into local scope
347
# mkdtemp = tempfile.mkdtemp
348
# rmtree = shutil.rmtree
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
349
1551.2.53 by abentley
Strip trailing slashes in a platform-sensible way
350
MIN_ABS_PATHLENGTH = 1
351
1685.1.9 by John Arbash Meinel
Updated LocalTransport so that it's base is now a URL rather than a local path. This helps consistency with all other functions. To do so, I added local_abspath() which returns the local path, and local_path_to/from_url
352
1185.31.47 by John Arbash Meinel
Added a fancy footwork rename to osutils, made SftpTransport use it.
353
if sys.platform == 'win32':
1685.1.20 by John Arbash Meinel
More changes to get 'bzr branch' and 'bzr pull' to work
354
    abspath = _win32_abspath
355
    realpath = _win32_realpath
356
    pathjoin = _win32_pathjoin
357
    normpath = _win32_normpath
358
    getcwd = _win32_getcwd
359
    mkdtemp = _win32_mkdtemp
360
    rename = _win32_rename
361
1551.2.53 by abentley
Strip trailing slashes in a platform-sensible way
362
    MIN_ABS_PATHLENGTH = 3
1532 by Robert Collins
Merge in John Meinels integration branch.
363
1692.7.6 by Martin Pool
[patch] force deletion of trees containing readonly files (alexander)
364
    def _win32_delete_readonly(function, path, excinfo):
365
        """Error handler for shutil.rmtree function [for win32]
366
        Helps to remove files and dirs marked as read-only.
367
        """
2116.5.1 by Henri Wiechers
Fixes osutils.rmtree on Windows with Python 2.5
368
        exception = excinfo[1]
1692.7.6 by Martin Pool
[patch] force deletion of trees containing readonly files (alexander)
369
        if function in (os.remove, os.rmdir) \
2116.5.1 by Henri Wiechers
Fixes osutils.rmtree on Windows with Python 2.5
370
            and isinstance(exception, OSError) \
371
            and exception.errno == errno.EACCES:
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
372
            make_writable(path)
1692.7.6 by Martin Pool
[patch] force deletion of trees containing readonly files (alexander)
373
            function(path)
374
        else:
375
            raise
376
377
    def rmtree(path, ignore_errors=False, onerror=_win32_delete_readonly):
378
        """Replacer for shutil.rmtree: could remove readonly dirs/files"""
379
        return shutil.rmtree(path, ignore_errors, onerror)
1830.3.11 by John Arbash Meinel
Create a mac version of 'getcwd()' which normalizes the path.
380
elif sys.platform == 'darwin':
381
    getcwd = _mac_getcwd
1692.7.6 by Martin Pool
[patch] force deletion of trees containing readonly files (alexander)
382
1685.1.31 by John Arbash Meinel
Adding tests for the rest of the _win32 functions.
383
1711.4.10 by John Arbash Meinel
Pull out sys.stdout.encoding handling into a separate function so it can be tested, and used elsewhere.
384
def get_terminal_encoding():
385
    """Find the best encoding for printing to the screen.
386
387
    This attempts to check both sys.stdout and sys.stdin to see
388
    what encoding they are in, and if that fails it falls back to
389
    bzrlib.user_encoding.
390
    The problem is that on Windows, locale.getpreferredencoding()
391
    is not the same encoding as that used by the console:
392
    http://mail.python.org/pipermail/python-list/2003-May/162357.html
393
394
    On my standard US Windows XP, the preferred encoding is
395
    cp1252, but the console is cp437
396
    """
397
    output_encoding = getattr(sys.stdout, 'encoding', None)
398
    if not output_encoding:
399
        input_encoding = getattr(sys.stdin, 'encoding', None)
400
        if not input_encoding:
401
            output_encoding = bzrlib.user_encoding
402
            mutter('encoding stdout as bzrlib.user_encoding %r', output_encoding)
403
        else:
404
            output_encoding = input_encoding
405
            mutter('encoding stdout as sys.stdin encoding %r', output_encoding)
406
    else:
407
        mutter('encoding stdout as sys.stdout encoding %r', output_encoding)
2127.4.1 by Alexander Belchenko
(jam, bialix) Workaround for cp0 console encoding on Windows
408
    if output_encoding == 'cp0':
409
        # invalid encoding (cp0 means 'no codepage' on Windows)
410
        output_encoding = bzrlib.user_encoding
411
        mutter('cp0 is invalid encoding.'
412
               ' encoding stdout as bzrlib.user_encoding %r', output_encoding)
2192.1.1 by Alexander Belchenko
Before actually using encoding need to check that Python has corresponding codec
413
    # check encoding
414
    try:
415
        codecs.lookup(output_encoding)
416
    except LookupError:
417
        sys.stderr.write('bzr: warning:'
2192.1.9 by Alexander Belchenko
final fix suggested by John Meinel
418
                         ' unknown terminal encoding %s.\n'
2192.1.1 by Alexander Belchenko
Before actually using encoding need to check that Python has corresponding codec
419
                         '  Using encoding %s instead.\n'
420
                         % (output_encoding, bzrlib.user_encoding)
421
                        )
422
        output_encoding = bzrlib.user_encoding
423
1711.4.10 by John Arbash Meinel
Pull out sys.stdout.encoding handling into a separate function so it can be tested, and used elsewhere.
424
    return output_encoding
425
426
1185.31.32 by John Arbash Meinel
Updated the bzr sourcecode to use bzrlib.osutils.pathjoin rather than os.path.join to enforce internal use of / instead of \
427
def normalizepath(f):
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
428
    if getattr(os.path, 'realpath', None) is not None:
1185.31.32 by John Arbash Meinel
Updated the bzr sourcecode to use bzrlib.osutils.pathjoin rather than os.path.join to enforce internal use of / instead of \
429
        F = realpath
430
    else:
431
        F = abspath
432
    [p,e] = os.path.split(f)
433
    if e == "" or e == "." or e == "..":
434
        return F(f)
435
    else:
436
        return pathjoin(F(p), e)
437
1 by mbp at sourcefrog
import from baz patch-364
438
779 by Martin Pool
- better quotefn for windows: use doublequotes for strings with
439
def backup_file(fn):
440
    """Copy a file to a backup.
441
442
    Backups are named in GNU-style, with a ~ suffix.
443
444
    If the file is already a backup, it's not copied.
445
    """
446
    if fn[-1] == '~':
447
        return
448
    bfn = fn + '~'
449
1448 by Robert Collins
revert symlinks correctly
450
    if has_symlinks() and os.path.islink(fn):
451
        target = os.readlink(fn)
452
        os.symlink(target, bfn)
453
        return
779 by Martin Pool
- better quotefn for windows: use doublequotes for strings with
454
    inf = file(fn, 'rb')
455
    try:
456
        content = inf.read()
457
    finally:
458
        inf.close()
459
    
460
    outf = file(bfn, 'wb')
461
    try:
462
        outf.write(content)
463
    finally:
464
        outf.close()
465
466
1 by mbp at sourcefrog
import from baz patch-364
467
def isdir(f):
468
    """True if f is an accessible directory."""
469
    try:
470
        return S_ISDIR(os.lstat(f)[ST_MODE])
471
    except OSError:
472
        return False
473
474
475
def isfile(f):
476
    """True if f is a regular file."""
477
    try:
478
        return S_ISREG(os.lstat(f)[ST_MODE])
479
    except OSError:
480
        return False
481
1092.2.6 by Robert Collins
symlink support updated to work
482
def islink(f):
483
    """True if f is a symlink."""
484
    try:
485
        return S_ISLNK(os.lstat(f)[ST_MODE])
486
    except OSError:
487
        return False
1 by mbp at sourcefrog
import from baz patch-364
488
485 by Martin Pool
- move commit code into its own module
489
def is_inside(dir, fname):
490
    """True if fname is inside dir.
969 by Martin Pool
- Add less-sucky is_within_any
491
    
1185.31.38 by John Arbash Meinel
Changing os.path.normpath to osutils.normpath
492
    The parameters should typically be passed to osutils.normpath first, so
969 by Martin Pool
- Add less-sucky is_within_any
493
    that . and .. and repeated slashes are eliminated, and the separators
494
    are canonical for the platform.
495
    
974.1.26 by aaron.bentley at utoronto
merged mbp@sourcefrog.net-20050817233101-0939da1cf91f2472
496
    The empty string as a dir name is taken as top-of-tree and matches 
497
    everything.
485 by Martin Pool
- move commit code into its own module
498
    """
969 by Martin Pool
- Add less-sucky is_within_any
499
    # XXX: Most callers of this can actually do something smarter by 
500
    # looking at the inventory
972 by Martin Pool
- less dodgy is_inside function
501
    if dir == fname:
502
        return True
503
    
974.1.26 by aaron.bentley at utoronto
merged mbp@sourcefrog.net-20050817233101-0939da1cf91f2472
504
    if dir == '':
505
        return True
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
506
1185.31.34 by John Arbash Meinel
Removing instances of os.sep
507
    if dir[-1] != '/':
508
        dir += '/'
1185.1.41 by Robert Collins
massive patch from Alexander Belchenko - many PEP8 fixes, removes unused function uuid
509
972 by Martin Pool
- less dodgy is_inside function
510
    return fname.startswith(dir)
511
485 by Martin Pool
- move commit code into its own module
512
513
def is_inside_any(dir_list, fname):
514
    """True if fname is inside any of given dirs."""
515
    for dirname in dir_list:
516
        if is_inside(dirname, fname):
517
            return True
518
    else:
519
        return False
520
521
1740.3.4 by Jelmer Vernooij
Move inventory to commit builder.
522
def is_inside_or_parent_of_any(dir_list, fname):
523
    """True if fname is a child or a parent of any of the given files."""
524
    for dirname in dir_list:
525
        if is_inside(dirname, fname) or is_inside(fname, dirname):
526
            return True
527
    else:
528
        return False
529
530
1 by mbp at sourcefrog
import from baz patch-364
531
def pumpfile(fromfile, tofile):
532
    """Copy contents of one file to another."""
1185.49.12 by John Arbash Meinel
Changed pumpfile to work on blocks, rather than reading the entire file at once.
533
    BUFSIZE = 32768
534
    while True:
535
        b = fromfile.read(BUFSIZE)
536
        if not b:
537
            break
1185.49.13 by John Arbash Meinel
Removed delayed setup, since it broke some tests. Fixed other small bugs. All tests pass.
538
        tofile.write(b)
1 by mbp at sourcefrog
import from baz patch-364
539
540
1185.67.7 by Aaron Bentley
Refactored a bit
541
def file_iterator(input_file, readsize=32768):
542
    while True:
543
        b = input_file.read(readsize)
544
        if len(b) == 0:
545
            break
546
        yield b
547
548
1 by mbp at sourcefrog
import from baz patch-364
549
def sha_file(f):
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
550
    if getattr(f, 'tell', None) is not None:
1 by mbp at sourcefrog
import from baz patch-364
551
        assert f.tell() == 0
552
    s = sha.new()
320 by Martin Pool
- Compute SHA-1 of files in chunks
553
    BUFSIZE = 128<<10
554
    while True:
555
        b = f.read(BUFSIZE)
556
        if not b:
557
            break
558
        s.update(b)
1 by mbp at sourcefrog
import from baz patch-364
559
    return s.hexdigest()
560
561
1235 by Martin Pool
- split sha_strings into osutils
562
563
def sha_strings(strings):
564
    """Return the sha-1 of concatenation of strings"""
565
    s = sha.new()
566
    map(s.update, strings)
567
    return s.hexdigest()
568
569
1 by mbp at sourcefrog
import from baz patch-364
570
def sha_string(f):
571
    s = sha.new()
572
    s.update(f)
573
    return s.hexdigest()
574
575
124 by mbp at sourcefrog
- check file text for past revisions is correct
576
def fingerprint_file(f):
577
    s = sha.new()
126 by mbp at sourcefrog
Use just one big read to fingerprint files
578
    b = f.read()
579
    s.update(b)
580
    size = len(b)
124 by mbp at sourcefrog
- check file text for past revisions is correct
581
    return {'size': size,
582
            'sha1': s.hexdigest()}
583
584
1 by mbp at sourcefrog
import from baz patch-364
585
def compare_files(a, b):
586
    """Returns true if equal in contents"""
74 by mbp at sourcefrog
compare_files: read in one page at a time rather than
587
    BUFSIZE = 4096
588
    while True:
589
        ai = a.read(BUFSIZE)
590
        bi = b.read(BUFSIZE)
591
        if ai != bi:
592
            return False
593
        if ai == '':
594
            return True
1 by mbp at sourcefrog
import from baz patch-364
595
596
49 by mbp at sourcefrog
fix local-time-offset calculation
597
def local_time_offset(t=None):
598
    """Return offset of local zone from GMT, either at present or at time t."""
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
599
    if t is None:
73 by mbp at sourcefrog
fix time.localtime call for python 2.3
600
        t = time.time()
2215.6.1 by James Henstridge
Don't rely on time.timezone and time.altzone in local_time_offset(),
601
    offset = datetime.fromtimestamp(t) - datetime.utcfromtimestamp(t)
602
    return offset.days * 86400 + offset.seconds
8 by mbp at sourcefrog
store committer's timezone in revision and show
603
604
    
1185.12.24 by Aaron Bentley
Made format_date more flexible
605
def format_date(t, offset=0, timezone='original', date_fmt=None, 
606
                show_offset=True):
1 by mbp at sourcefrog
import from baz patch-364
607
    ## TODO: Perhaps a global option to use either universal or local time?
608
    ## Or perhaps just let people set $TZ?
609
    assert isinstance(t, float)
610
    
8 by mbp at sourcefrog
store committer's timezone in revision and show
611
    if timezone == 'utc':
1 by mbp at sourcefrog
import from baz patch-364
612
        tt = time.gmtime(t)
613
        offset = 0
8 by mbp at sourcefrog
store committer's timezone in revision and show
614
    elif timezone == 'original':
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
615
        if offset is None:
23 by mbp at sourcefrog
format_date: handle revisions with no timezone offset
616
            offset = 0
16 by mbp at sourcefrog
fix inverted calculation for original timezone -> utc
617
        tt = time.gmtime(t + offset)
12 by mbp at sourcefrog
new --timezone option for bzr log
618
    elif timezone == 'local':
1 by mbp at sourcefrog
import from baz patch-364
619
        tt = time.localtime(t)
49 by mbp at sourcefrog
fix local-time-offset calculation
620
        offset = local_time_offset(t)
12 by mbp at sourcefrog
new --timezone option for bzr log
621
    else:
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
622
        raise errors.BzrError("unsupported timezone format %r" % timezone,
623
                              ['options are "utc", "original", "local"'])
1185.12.24 by Aaron Bentley
Made format_date more flexible
624
    if date_fmt is None:
625
        date_fmt = "%a %Y-%m-%d %H:%M:%S"
626
    if show_offset:
627
        offset_str = ' %+03d%02d' % (offset / 3600, (offset / 60) % 60)
628
    else:
629
        offset_str = ''
630
    return (time.strftime(date_fmt, tt) +  offset_str)
1 by mbp at sourcefrog
import from baz patch-364
631
632
633
def compact_date(when):
634
    return time.strftime('%Y%m%d%H%M%S', time.gmtime(when))
635
    
636
1957.1.4 by John Arbash Meinel
create a helper for formatting a time delta
637
def format_delta(delta):
638
    """Get a nice looking string for a time delta.
639
640
    :param delta: The time difference in seconds, can be positive or negative.
641
        positive indicates time in the past, negative indicates time in the
642
        future. (usually time.time() - stored_time)
643
    :return: String formatted to show approximate resolution
644
    """
645
    delta = int(delta)
646
    if delta >= 0:
647
        direction = 'ago'
648
    else:
649
        direction = 'in the future'
650
        delta = -delta
651
652
    seconds = delta
653
    if seconds < 90: # print seconds up to 90 seconds
654
        if seconds == 1:
655
            return '%d second %s' % (seconds, direction,)
656
        else:
657
            return '%d seconds %s' % (seconds, direction)
658
659
    minutes = int(seconds / 60)
660
    seconds -= 60 * minutes
661
    if seconds == 1:
662
        plural_seconds = ''
663
    else:
664
        plural_seconds = 's'
665
    if minutes < 90: # print minutes, seconds up to 90 minutes
666
        if minutes == 1:
667
            return '%d minute, %d second%s %s' % (
668
                    minutes, seconds, plural_seconds, direction)
669
        else:
670
            return '%d minutes, %d second%s %s' % (
671
                    minutes, seconds, plural_seconds, direction)
672
673
    hours = int(minutes / 60)
674
    minutes -= 60 * hours
675
    if minutes == 1:
676
        plural_minutes = ''
677
    else:
678
        plural_minutes = 's'
679
680
    if hours == 1:
681
        return '%d hour, %d minute%s %s' % (hours, minutes,
682
                                            plural_minutes, direction)
683
    return '%d hours, %d minute%s %s' % (hours, minutes,
684
                                         plural_minutes, direction)
1 by mbp at sourcefrog
import from baz patch-364
685
686
def filesize(f):
687
    """Return size of given open file."""
688
    return os.fstat(f.fileno())[ST_SIZE]
689
1553.5.5 by Martin Pool
New utility routine rand_chars
690
1185.1.7 by Robert Collins
Nathaniel McCallums patch for urandom friendliness on aix.
691
# Define rand_bytes based on platform.
692
try:
693
    # Python 2.4 and later have os.urandom,
694
    # but it doesn't work on some arches
695
    os.urandom(1)
1 by mbp at sourcefrog
import from baz patch-364
696
    rand_bytes = os.urandom
1185.1.7 by Robert Collins
Nathaniel McCallums patch for urandom friendliness on aix.
697
except (NotImplementedError, AttributeError):
698
    # If python doesn't have os.urandom, or it doesn't work,
699
    # then try to first pull random data from /dev/urandom
2067.1.1 by John Arbash Meinel
Catch an exception while opening /dev/urandom rather than using os.path.exists()
700
    try:
1185.1.7 by Robert Collins
Nathaniel McCallums patch for urandom friendliness on aix.
701
        rand_bytes = file('/dev/urandom', 'rb').read
702
    # Otherwise, use this hack as a last resort
2067.1.1 by John Arbash Meinel
Catch an exception while opening /dev/urandom rather than using os.path.exists()
703
    except (IOError, OSError):
1185.1.7 by Robert Collins
Nathaniel McCallums patch for urandom friendliness on aix.
704
        # not well seeded, but better than nothing
705
        def rand_bytes(n):
706
            import random
707
            s = ''
708
            while n:
709
                s += chr(random.randint(0, 255))
710
                n -= 1
711
            return s
1 by mbp at sourcefrog
import from baz patch-364
712
1553.5.5 by Martin Pool
New utility routine rand_chars
713
714
ALNUM = '0123456789abcdefghijklmnopqrstuvwxyz'
715
def rand_chars(num):
716
    """Return a random string of num alphanumeric characters
717
    
718
    The result only contains lowercase chars because it may be used on 
719
    case-insensitive filesystems.
720
    """
721
    s = ''
722
    for raw_byte in rand_bytes(num):
723
        s += ALNUM[ord(raw_byte) % 36]
724
    return s
725
726
1 by mbp at sourcefrog
import from baz patch-364
727
## TODO: We could later have path objects that remember their list
1759.2.2 by Jelmer Vernooij
Revert some of my spelling fixes and fix some typos after review by Aaron.
728
## decomposition (might be too tricksy though.)
1 by mbp at sourcefrog
import from baz patch-364
729
730
def splitpath(p):
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
731
    """Turn string into list of parts."""
732
    assert isinstance(p, basestring)
271 by Martin Pool
- Windows path fixes
733
734
    # split on either delimiter because people might use either on
735
    # Windows
736
    ps = re.split(r'[\\/]', p)
737
738
    rps = []
1 by mbp at sourcefrog
import from baz patch-364
739
    for f in ps:
740
        if f == '..':
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
741
            raise errors.BzrError("sorry, %r not allowed in path" % f)
271 by Martin Pool
- Windows path fixes
742
        elif (f == '.') or (f == ''):
743
            pass
744
        else:
745
            rps.append(f)
746
    return rps
1 by mbp at sourcefrog
import from baz patch-364
747
748
def joinpath(p):
749
    assert isinstance(p, list)
750
    for f in p:
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
751
        if (f == '..') or (f is None) or (f == ''):
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
752
            raise errors.BzrError("sorry, %r not allowed in path" % f)
1185.31.32 by John Arbash Meinel
Updated the bzr sourcecode to use bzrlib.osutils.pathjoin rather than os.path.join to enforce internal use of / instead of \
753
    return pathjoin(*p)
1 by mbp at sourcefrog
import from baz patch-364
754
755
1732.1.1 by John Arbash Meinel
deprecating appendpath, it does exactly what pathjoin does
756
@deprecated_function(zero_nine)
1 by mbp at sourcefrog
import from baz patch-364
757
def appendpath(p1, p2):
758
    if p1 == '':
759
        return p2
760
    else:
1185.31.32 by John Arbash Meinel
Updated the bzr sourcecode to use bzrlib.osutils.pathjoin rather than os.path.join to enforce internal use of / instead of \
761
        return pathjoin(p1, p2)
1 by mbp at sourcefrog
import from baz patch-364
762
    
763
1231 by Martin Pool
- more progress on fetch on top of weaves
764
def split_lines(s):
765
    """Split s into lines, but without removing the newline characters."""
1666.1.6 by Robert Collins
Make knit the default format.
766
    lines = s.split('\n')
767
    result = [line + '\n' for line in lines[:-1]]
768
    if lines[-1]:
769
        result.append(lines[-1])
770
    return result
1391 by Robert Collins
merge from integration
771
772
1185.10.4 by Aaron Bentley
Disabled hardlinks on cygwin, mac OS
773
def hardlinks_good():
1185.10.5 by Aaron Bentley
Fixed hardlinks_good test
774
    return sys.platform not in ('win32', 'cygwin', 'darwin')
1185.10.4 by Aaron Bentley
Disabled hardlinks on cygwin, mac OS
775
1185.1.46 by Robert Collins
Aarons branch --basis patch
776
1185.10.3 by Aaron Bentley
Made copy_multi_immutable create hardlinks opportunistically
777
def link_or_copy(src, dest):
778
    """Hardlink a file, or copy it if it can't be hardlinked."""
1185.10.4 by Aaron Bentley
Disabled hardlinks on cygwin, mac OS
779
    if not hardlinks_good():
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
780
        shutil.copyfile(src, dest)
1185.10.3 by Aaron Bentley
Made copy_multi_immutable create hardlinks opportunistically
781
        return
782
    try:
783
        os.link(src, dest)
784
    except (OSError, IOError), e:
785
        if e.errno != errno.EXDEV:
786
            raise
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
787
        shutil.copyfile(src, dest)
1399.1.4 by Robert Collins
move diff and symlink conditionals into inventory.py from diff.py
788
1558.12.9 by Aaron Bentley
Handle resolving conflicts with directories properly
789
def delete_any(full_path):
790
    """Delete a file or directory."""
791
    try:
792
        os.unlink(full_path)
793
    except OSError, e:
794
    # We may be renaming a dangling inventory id
795
        if e.errno not in (errno.EISDIR, errno.EACCES, errno.EPERM):
796
            raise
797
        os.rmdir(full_path)
798
1399.1.4 by Robert Collins
move diff and symlink conditionals into inventory.py from diff.py
799
800
def has_symlinks():
1963.2.6 by Robey Pointer
pychecker is on crack; go back to using 'is None'.
801
    if getattr(os, 'symlink', None) is not None:
1399.1.4 by Robert Collins
move diff and symlink conditionals into inventory.py from diff.py
802
        return True
803
    else:
804
        return False
1185.16.38 by Martin Pool
- move contains_whitespace and contains_linebreaks to osutils
805
        
806
807
def contains_whitespace(s):
808
    """True if there are any whitespace characters in s."""
2249.2.1 by John Arbash Meinel
(John Arbash Meinel) hard-code the whitespace chars to avoid problems in some locales.
809
    # string.whitespace can include '\xa0' in certain locales, because it is
810
    # considered "non-breaking-space" as part of ISO-8859-1. But it
811
    # 1) Isn't a breaking whitespace
812
    # 2) Isn't one of ' \t\r\n' which are characters we sometimes use as
813
    #    separators
814
    # 3) '\xa0' isn't unicode safe since it is >128.
815
    # So we are following textwrap's example and hard-coding our own.
816
    # We probably could ignore \v and \f, too.
817
    for ch in u' \t\n\r\v\f':
1185.16.38 by Martin Pool
- move contains_whitespace and contains_linebreaks to osutils
818
        if ch in s:
819
            return True
820
    else:
821
        return False
822
823
824
def contains_linebreaks(s):
825
    """True if there is any vertical whitespace in s."""
826
    for ch in '\f\n\r':
827
        if ch in s:
828
            return True
829
    else:
830
        return False
1457.1.2 by Robert Collins
move branch._relpath into osutils as relpath
831
832
833
def relpath(base, path):
834
    """Return path relative to base, or raise exception.
835
836
    The path may be either an absolute path or a path relative to the
837
    current working directory.
838
839
    os.path.commonprefix (python2.4) has a bad bug that it works just
840
    on string prefixes, assuming that '/u' is a prefix of '/u2'.  This
1636.1.1 by Robert Collins
Fix calling relpath() and abspath() on transports at their root.
841
    avoids that problem.
842
    """
1685.1.12 by John Arbash Meinel
Some more work to get LocalTransport to only support URLs
843
1551.2.53 by abentley
Strip trailing slashes in a platform-sensible way
844
    assert len(base) >= MIN_ABS_PATHLENGTH, ('Length of base must be equal or'
845
        ' exceed the platform minimum length (which is %d)' % 
846
        MIN_ABS_PATHLENGTH)
1685.1.9 by John Arbash Meinel
Updated LocalTransport so that it's base is now a URL rather than a local path. This helps consistency with all other functions. To do so, I added local_abspath() which returns the local path, and local_path_to/from_url
847
1685.1.12 by John Arbash Meinel
Some more work to get LocalTransport to only support URLs
848
    rp = abspath(path)
1457.1.2 by Robert Collins
move branch._relpath into osutils as relpath
849
850
    s = []
1685.1.12 by John Arbash Meinel
Some more work to get LocalTransport to only support URLs
851
    head = rp
1457.1.2 by Robert Collins
move branch._relpath into osutils as relpath
852
    while len(head) >= len(base):
853
        if head == base:
854
            break
855
        head, tail = os.path.split(head)
856
        if tail:
857
            s.insert(0, tail)
858
    else:
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
859
        raise errors.PathNotChild(rp, base)
1457.1.2 by Robert Collins
move branch._relpath into osutils as relpath
860
1185.31.35 by John Arbash Meinel
Couple small fixes, all tests pass on cygwin.
861
    if s:
862
        return pathjoin(*s)
863
    else:
864
        return ''
1185.33.60 by Martin Pool
Use full terminal width for verbose test output.
865
866
1534.3.1 by Robert Collins
* bzrlib.osutils.safe_unicode now exists to provide parameter coercion
867
def safe_unicode(unicode_or_utf8_string):
868
    """Coerce unicode_or_utf8_string into unicode.
869
870
    If it is unicode, it is returned.
871
    Otherwise it is decoded from utf-8. If a decoding error
872
    occurs, it is wrapped as a If the decoding fails, the exception is wrapped 
873
    as a BzrBadParameter exception.
874
    """
875
    if isinstance(unicode_or_utf8_string, unicode):
876
        return unicode_or_utf8_string
877
    try:
878
        return unicode_or_utf8_string.decode('utf8')
879
    except UnicodeDecodeError:
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
880
        raise errors.BzrBadParameterNotUnicode(unicode_or_utf8_string)
1534.3.1 by Robert Collins
* bzrlib.osutils.safe_unicode now exists to provide parameter coercion
881
882
1185.85.75 by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths.
883
_platform_normalizes_filenames = False
884
if sys.platform == 'darwin':
885
    _platform_normalizes_filenames = True
886
887
888
def normalizes_filenames():
889
    """Return True if this platform normalizes unicode filenames.
890
891
    Mac OSX does, Windows/Linux do not.
892
    """
893
    return _platform_normalizes_filenames
894
895
1830.3.2 by John Arbash Meinel
normalized_filename is a much better name
896
def _accessible_normalized_filename(path):
1830.3.1 by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms
897
    """Get the unicode normalized path, and if you can access the file.
898
899
    On platforms where the system normalizes filenames (Mac OSX),
900
    you can access a file by any path which will normalize correctly.
901
    On platforms where the system does not normalize filenames 
902
    (Windows, Linux), you have to access a file by its exact path.
903
904
    Internally, bzr only supports NFC/NFKC normalization, since that is 
905
    the standard for XML documents.
906
907
    So return the normalized path, and a flag indicating if the file
908
    can be accessed by that path.
909
    """
910
1830.3.8 by John Arbash Meinel
unicodedata.normalize requires unicode strings
911
    return unicodedata.normalize('NFKC', unicode(path)), True
1830.3.1 by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms
912
913
1830.3.2 by John Arbash Meinel
normalized_filename is a much better name
914
def _inaccessible_normalized_filename(path):
915
    __doc__ = _accessible_normalized_filename.__doc__
1830.3.1 by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms
916
1830.3.8 by John Arbash Meinel
unicodedata.normalize requires unicode strings
917
    normalized = unicodedata.normalize('NFKC', unicode(path))
1830.3.1 by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms
918
    return normalized, normalized == path
919
920
1185.85.75 by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths.
921
if _platform_normalizes_filenames:
1830.3.2 by John Arbash Meinel
normalized_filename is a much better name
922
    normalized_filename = _accessible_normalized_filename
1185.85.75 by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths.
923
else:
1830.3.2 by John Arbash Meinel
normalized_filename is a much better name
924
    normalized_filename = _inaccessible_normalized_filename
1185.85.75 by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths.
925
926
1185.33.60 by Martin Pool
Use full terminal width for verbose test output.
927
def terminal_width():
928
    """Return estimated terminal width."""
1704.2.3 by Martin Pool
(win32) Detect terminal width using GetConsoleScreenBufferInfo (Alexander)
929
    if sys.platform == 'win32':
2245.4.6 by Alexander Belchenko
osutils.py: terminal_width() now use win32utils.get_console_size()
930
        return win32utils.get_console_size()[0]
1704.2.2 by Martin Pool
Detect terminal width using ioctl
931
    width = 0
1185.33.60 by Martin Pool
Use full terminal width for verbose test output.
932
    try:
1704.2.2 by Martin Pool
Detect terminal width using ioctl
933
        import struct, fcntl, termios
934
        s = struct.pack('HHHH', 0, 0, 0, 0)
935
        x = fcntl.ioctl(1, termios.TIOCGWINSZ, s)
936
        width = struct.unpack('HHHH', x)[1]
937
    except IOError:
938
        pass
939
    if width <= 0:
940
        try:
941
            width = int(os.environ['COLUMNS'])
942
        except:
943
            pass
944
    if width <= 0:
945
        width = 80
946
947
    return width
1534.7.25 by Aaron Bentley
Added set_executability
948
1963.1.5 by John Arbash Meinel
Create an osutils helper function for modifying the environment
949
1534.7.25 by Aaron Bentley
Added set_executability
950
def supports_executable():
1534.7.160 by Aaron Bentley
Changed implementation of supports_executable
951
    return sys.platform != "win32"
1551.2.53 by abentley
Strip trailing slashes in a platform-sensible way
952
953
1551.10.4 by Aaron Bentley
Update to skip on win32
954
def supports_posix_readonly():
955
    """Return True if 'readonly' has POSIX semantics, False otherwise.
956
957
    Notably, a win32 readonly file cannot be deleted, unlike POSIX where the
958
    directory controls creation/deletion, etc.
959
960
    And under win32, readonly means that the directory itself cannot be
961
    deleted.  The contents of a readonly directory can be changed, unlike POSIX
962
    where files in readonly directories cannot be added, deleted or renamed.
963
    """
964
    return sys.platform != "win32"
965
966
1963.1.5 by John Arbash Meinel
Create an osutils helper function for modifying the environment
967
def set_or_unset_env(env_variable, value):
968
    """Modify the environment, setting or removing the env_variable.
969
970
    :param env_variable: The environment variable in question
971
    :param value: The value to set the environment to. If None, then
972
        the variable will be removed.
973
    :return: The original value of the environment variable.
974
    """
975
    orig_val = os.environ.get(env_variable)
976
    if value is None:
977
        if orig_val is not None:
978
            del os.environ[env_variable]
979
    else:
980
        if isinstance(value, unicode):
981
            value = value.encode(bzrlib.user_encoding)
982
        os.environ[env_variable] = value
983
    return orig_val
984
985
1551.2.56 by Aaron Bentley
Better illegal pathname check for Windows
986
_validWin32PathRE = re.compile(r'^([A-Za-z]:[/\\])?[^:<>*"?\|]*$')
987
988
989
def check_legal_path(path):
990
    """Check whether the supplied path is legal.  
991
    This is only required on Windows, so we don't test on other platforms
992
    right now.
993
    """
994
    if sys.platform != "win32":
995
        return
996
    if _validWin32PathRE.match(path) is None:
1996.3.25 by John Arbash Meinel
Make importing errors lazy for osutils
997
        raise errors.IllegalPath(path)
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
998
999
1757.2.8 by Robert Collins
Teach walkdirs to walk a subdir of a tree.
1000
def walkdirs(top, prefix=""):
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
1001
    """Yield data about all the directories in a tree.
1002
    
1003
    This yields all the data about the contents of a directory at a time.
1004
    After each directory has been yielded, if the caller has mutated the list
1005
    to exclude some directories, they are then not descended into.
1006
    
1007
    The data yielded is of the form:
1897.1.2 by Robert Collins
cleanup osutils.walkdirs changes after review.
1008
    ((directory-relpath, directory-path-from-top),
1897.1.1 by Robert Collins
Add some useful summary data to osutils.walkdirs output.
1009
    [(relpath, basename, kind, lstat), ...]),
1897.1.2 by Robert Collins
cleanup osutils.walkdirs changes after review.
1010
     - directory-relpath is the relative path of the directory being returned
1011
       with respect to top. prefix is prepended to this.
1012
     - directory-path-from-root is the path including top for this directory. 
1013
       It is suitable for use with os functions.
1897.1.1 by Robert Collins
Add some useful summary data to osutils.walkdirs output.
1014
     - relpath is the relative path within the subtree being walked.
1015
     - basename is the basename of the path
1897.1.2 by Robert Collins
cleanup osutils.walkdirs changes after review.
1016
     - kind is the kind of the file now. If unknown then the file is not
1897.1.1 by Robert Collins
Add some useful summary data to osutils.walkdirs output.
1017
       present within the tree - but it may be recorded as versioned. See
1018
       versioned_kind.
1019
     - lstat is the stat data *if* the file was statted.
1020
     - planned, not implemented: 
1021
       path_from_tree_root is the path from the root of the tree.
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
1022
1757.2.16 by Robert Collins
Review comments.
1023
    :param prefix: Prefix the relpaths that are yielded with 'prefix'. This 
1024
        allows one to walk a subtree but get paths that are relative to a tree
1025
        rooted higher up.
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
1026
    :return: an iterator over the dirs.
1027
    """
1897.1.1 by Robert Collins
Add some useful summary data to osutils.walkdirs output.
1028
    #TODO there is a bit of a smell where the results of the directory-
1029
    # summary in this, and the path from the root, may not agree 
1030
    # depending on top and prefix - i.e. ./foo and foo as a pair leads to
1031
    # potentially confusing output. We should make this more robust - but
1897.1.2 by Robert Collins
cleanup osutils.walkdirs changes after review.
1032
    # not at a speed cost. RBC 20060731
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
1033
    lstat = os.lstat
1034
    pending = []
1035
    _directory = _directory_kind
1996.3.14 by John Arbash Meinel
lazy_import osutils and sign_my_commits
1036
    _listdir = os.listdir
1757.2.8 by Robert Collins
Teach walkdirs to walk a subdir of a tree.
1037
    pending = [(prefix, "", _directory, None, top)]
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
1038
    while pending:
1039
        dirblock = []
1040
        currentdir = pending.pop()
1041
        # 0 - relpath, 1- basename, 2- kind, 3- stat, 4-toppath
1042
        top = currentdir[4]
1043
        if currentdir[0]:
1044
            relroot = currentdir[0] + '/'
1045
        else:
1046
            relroot = ""
1047
        for name in sorted(_listdir(top)):
1048
            abspath = top + '/' + name
1049
            statvalue = lstat(abspath)
1897.1.2 by Robert Collins
cleanup osutils.walkdirs changes after review.
1050
            dirblock.append((relroot + name, name,
1051
                file_kind_from_stat_mode(statvalue.st_mode),
1052
                statvalue, abspath))
1897.1.1 by Robert Collins
Add some useful summary data to osutils.walkdirs output.
1053
        yield (currentdir[0], top), dirblock
1753.1.1 by Robert Collins
(rbc, jam, mbp)Add bzrlib.osutils.walkdirs, an optimised walk-and-stat routine.
1054
        # push the user specified dirs from dirblock
1055
        for dir in reversed(dirblock):
1056
            if dir[2] == _directory:
1057
                pending.append(dir)
1773.3.1 by Robert Collins
Add path_prefix_key and compare_paths_prefix_order utility functions.
1058
1059
1907.3.2 by John Arbash Meinel
Updated the copy_tree function to allow overriding functionality.
1060
def copy_tree(from_path, to_path, handlers={}):
1907.3.1 by John Arbash Meinel
create a copy_tree wrapper around walkdirs()
1061
    """Copy all of the entries in from_path into to_path.
1062
1063
    :param from_path: The base directory to copy. 
1064
    :param to_path: The target directory. If it does not exist, it will
1065
        be created.
1907.3.2 by John Arbash Meinel
Updated the copy_tree function to allow overriding functionality.
1066
    :param handlers: A dictionary of functions, which takes a source and
1067
        destinations for files, directories, etc.
1068
        It is keyed on the file kind, such as 'directory', 'symlink', or 'file'
1069
        'file', 'directory', and 'symlink' should always exist.
1070
        If they are missing, they will be replaced with 'os.mkdir()',
1071
        'os.readlink() + os.symlink()', and 'shutil.copy2()', respectively.
1907.3.1 by John Arbash Meinel
create a copy_tree wrapper around walkdirs()
1072
    """
1073
    # Now, just copy the existing cached tree to the new location
1074
    # We use a cheap trick here.
1075
    # Absolute paths are prefixed with the first parameter
1076
    # relative paths are prefixed with the second.
1077
    # So we can get both the source and target returned
1078
    # without any extra work.
1079
1907.3.2 by John Arbash Meinel
Updated the copy_tree function to allow overriding functionality.
1080
    def copy_dir(source, dest):
1081
        os.mkdir(dest)
1082
1083
    def copy_link(source, dest):
1084
        """Copy the contents of a symlink"""
1085
        link_to = os.readlink(source)
1086
        os.symlink(link_to, dest)
1087
1088
    real_handlers = {'file':shutil.copy2,
1089
                     'symlink':copy_link,
1090
                     'directory':copy_dir,
1091
                    }
1092
    real_handlers.update(handlers)
1093
1907.3.1 by John Arbash Meinel
create a copy_tree wrapper around walkdirs()
1094
    if not os.path.exists(to_path):
1907.3.2 by John Arbash Meinel
Updated the copy_tree function to allow overriding functionality.
1095
        real_handlers['directory'](from_path, to_path)
1907.3.1 by John Arbash Meinel
create a copy_tree wrapper around walkdirs()
1096
1097
    for dir_info, entries in walkdirs(from_path, prefix=to_path):
1098
        for relpath, name, kind, st, abspath in entries:
1907.3.2 by John Arbash Meinel
Updated the copy_tree function to allow overriding functionality.
1099
            real_handlers[kind](abspath, relpath)
1907.3.1 by John Arbash Meinel
create a copy_tree wrapper around walkdirs()
1100
1101
1773.3.1 by Robert Collins
Add path_prefix_key and compare_paths_prefix_order utility functions.
1102
def path_prefix_key(path):
1103
    """Generate a prefix-order path key for path.
1104
1105
    This can be used to sort paths in the same way that walkdirs does.
1106
    """
1773.3.2 by Robert Collins
New corner case from John Meinel, showing up the need to check the directory lexographically outside of a single tree's root. Fixed.
1107
    return (dirname(path) , path)
1773.3.1 by Robert Collins
Add path_prefix_key and compare_paths_prefix_order utility functions.
1108
1109
1110
def compare_paths_prefix_order(path_a, path_b):
1111
    """Compare path_a and path_b to generate the same order walkdirs uses."""
1112
    key_a = path_prefix_key(path_a)
1113
    key_b = path_prefix_key(path_b)
1114
    return cmp(key_a, key_b)
1955.2.2 by John Arbash Meinel
Change the name of the test classes (test_lang => test_locale), move the function into osutils.py
1115
1116
1117
_cached_user_encoding = None
1118
1119
2192.1.3 by Alexander Belchenko
Tests for osutils.get_user_encoding
1120
def get_user_encoding(use_cache=True):
1955.2.2 by John Arbash Meinel
Change the name of the test classes (test_lang => test_locale), move the function into osutils.py
1121
    """Find out what the preferred user encoding is.
1122
1123
    This is generally the encoding that is used for command line parameters
1124
    and file contents. This may be different from the terminal encoding
1125
    or the filesystem encoding.
1126
2192.1.3 by Alexander Belchenko
Tests for osutils.get_user_encoding
1127
    :param  use_cache:  Enable cache for detected encoding.
1128
                        (This parameter is turned on by default,
1129
                        and required only for selftesting)
1130
1955.2.2 by John Arbash Meinel
Change the name of the test classes (test_lang => test_locale), move the function into osutils.py
1131
    :return: A string defining the preferred user encoding
1132
    """
1133
    global _cached_user_encoding
2192.1.3 by Alexander Belchenko
Tests for osutils.get_user_encoding
1134
    if _cached_user_encoding is not None and use_cache:
1955.2.2 by John Arbash Meinel
Change the name of the test classes (test_lang => test_locale), move the function into osutils.py
1135
        return _cached_user_encoding
1136
1137
    if sys.platform == 'darwin':
1138
        # work around egregious python 2.4 bug
1139
        sys.platform = 'posix'
1140
        try:
1141
            import locale
1142
        finally:
1143
            sys.platform = 'darwin'
1144
    else:
1145
        import locale
1146
1147
    try:
2192.1.3 by Alexander Belchenko
Tests for osutils.get_user_encoding
1148
        user_encoding = locale.getpreferredencoding()
1955.2.2 by John Arbash Meinel
Change the name of the test classes (test_lang => test_locale), move the function into osutils.py
1149
    except locale.Error, e:
1955.2.3 by John Arbash Meinel
Change error message text
1150
        sys.stderr.write('bzr: warning: %s\n'
2001.2.1 by Jelmer Vernooij
Fix typo in encoding warning.
1151
                         '  Could not determine what text encoding to use.\n'
1955.2.3 by John Arbash Meinel
Change error message text
1152
                         '  This error usually means your Python interpreter\n'
1153
                         '  doesn\'t support the locale set by $LANG (%s)\n'
1154
                         "  Continuing with ascii encoding.\n"
1955.2.2 by John Arbash Meinel
Change the name of the test classes (test_lang => test_locale), move the function into osutils.py
1155
                         % (e, os.environ.get('LANG')))
2192.1.7 by Alexander Belchenko
get_user_encoding: if locale.Error raised we need to set user_encoding to 'ascii' as warning says
1156
        user_encoding = 'ascii'
1955.2.2 by John Arbash Meinel
Change the name of the test classes (test_lang => test_locale), move the function into osutils.py
1157
2127.4.1 by Alexander Belchenko
(jam, bialix) Workaround for cp0 console encoding on Windows
1158
    # Windows returns 'cp0' to indicate there is no code page. So we'll just
1159
    # treat that as ASCII, and not support printing unicode characters to the
1160
    # console.
2192.1.3 by Alexander Belchenko
Tests for osutils.get_user_encoding
1161
    if user_encoding in (None, 'cp0'):
1162
        user_encoding = 'ascii'
2192.1.1 by Alexander Belchenko
Before actually using encoding need to check that Python has corresponding codec
1163
    else:
1164
        # check encoding
1165
        try:
2192.1.3 by Alexander Belchenko
Tests for osutils.get_user_encoding
1166
            codecs.lookup(user_encoding)
2192.1.1 by Alexander Belchenko
Before actually using encoding need to check that Python has corresponding codec
1167
        except LookupError:
1168
            sys.stderr.write('bzr: warning:'
1169
                             ' unknown encoding %s.'
1170
                             ' Continuing with ascii encoding.\n'
2192.1.3 by Alexander Belchenko
Tests for osutils.get_user_encoding
1171
                             % user_encoding
2192.1.1 by Alexander Belchenko
Before actually using encoding need to check that Python has corresponding codec
1172
                            )
2192.1.3 by Alexander Belchenko
Tests for osutils.get_user_encoding
1173
            user_encoding = 'ascii'
1174
1175
    if use_cache:
1176
        _cached_user_encoding = user_encoding
1177
1178
    return user_encoding
2091.1.1 by Martin Pool
Avoid MSG_WAITALL as it doesn't work on Windows
1179
1180
1181
def recv_all(socket, bytes):
1182
    """Receive an exact number of bytes.
1183
1184
    Regular Socket.recv() may return less than the requested number of bytes,
1185
    dependning on what's in the OS buffer.  MSG_WAITALL is not available
1186
    on all platforms, but this should work everywhere.  This will return
1187
    less than the requested amount if the remote end closes.
1188
1189
    This isn't optimized and is intended mostly for use in testing.
1190
    """
1191
    b = ''
1192
    while len(b) < bytes:
1193
        new = socket.recv(bytes - len(b))
1194
        if new == '':
1195
            break # eof
1196
        b += new
1197
    return b
1198
2091.3.7 by Aaron Bentley
Rename real_parent to dereferenced_path
1199
def dereference_path(path):
1200
    """Determine the real path to a file.
1201
1202
    All parent elements are dereferenced.  But the file itself is not
1203
    dereferenced.
1204
    :param path: The original path.  May be absolute or relative.
1205
    :return: the real path *to* the file
1206
    """
2091.3.5 by Aaron Bentley
Move realpath functionality into osutils
1207
    parent, base = os.path.split(path)
1208
    # The pathjoin for '.' is a workaround for Python bug #1213894.
1209
    # (initial path components aren't dereferenced)
1210
    return pathjoin(realpath(pathjoin('.', parent)), base)