1
1
# Bazaar-NG -- distributed version control
3
3
# Copyright (C) 2005 by Canonical Ltd
5
5
# This program is free software; you can redistribute it and/or modify
6
6
# it under the terms of the GNU General Public License as published by
7
7
# the Free Software Foundation; either version 2 of the License, or
8
8
# (at your option) any later version.
10
10
# This program is distributed in the hope that it will be useful,
11
11
# but WITHOUT ANY WARRANTY; without even the implied warranty of
12
12
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13
13
# GNU General Public License for more details.
15
15
# You should have received a copy of the GNU General Public License
16
16
# along with this program; if not, write to the Free Software
17
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19
from cStringIO import StringIO
22
from os import listdir
26
from shutil import copyfile
28
from stat import (S_ISREG, S_ISDIR, S_ISLNK, ST_MODE, ST_SIZE,
29
S_ISCHR, S_ISBLK, S_ISFIFO, S_ISSOCK)
36
from ntpath import (abspath as _nt_abspath,
38
normpath as _nt_normpath,
39
realpath as _nt_realpath,
19
import os, types, re, time, errno
20
from stat import S_ISREG, S_ISDIR, S_ISLNK, ST_MODE, ST_SIZE
22
from errors import bailout, BzrError
23
from trace import mutter
43
from bzrlib.errors import (BzrError,
44
BzrBadParameterNotUnicode,
49
from bzrlib.symbol_versioning import *
50
from bzrlib.trace import mutter
51
import bzrlib.win32console
54
26
def make_readonly(filename):
55
27
"""Make a filename read-only."""
28
# TODO: probably needs to be fixed for windows
56
29
mod = os.stat(filename).st_mode
57
30
mod = mod & 0777555
58
31
os.chmod(filename, mod)
64
37
os.chmod(filename, mod)
40
_QUOTE_RE = re.compile(r'([^a-zA-Z0-9.,:/_~-])')
71
"""Return a quoted filename filename
73
This previously used backslash quoting, but that works poorly on
75
# TODO: I'm not really sure this is the best format either.x
78
_QUOTE_RE = re.compile(r'([^a-zA-Z0-9.,:/\\_~-])')
80
if _QUOTE_RE.search(f):
86
_directory_kind = 'directory'
89
stat.S_IFDIR:_directory_kind,
90
stat.S_IFCHR:'chardev',
94
stat.S_IFLNK:'symlink',
95
stat.S_IFSOCK:'socket',
99
def file_kind_from_stat_mode(stat_mode, _formats=_formats, _unknown='unknown'):
100
"""Generate a file kind from a stat mode. This is used in walkdirs.
102
Its performance is critical: Do not mutate without careful benchmarking.
105
return _formats[stat_mode & 0170000]
110
def file_kind(f, _lstat=os.lstat, _mapper=file_kind_from_stat_mode):
112
return _mapper(_lstat(f).st_mode)
114
if getattr(e, 'errno', None) == errno.ENOENT:
115
raise bzrlib.errors.NoSuchFile(f)
119
def kind_marker(kind):
122
elif kind == _directory_kind:
124
elif kind == 'symlink':
127
raise BzrError('invalid file kind %r' % kind)
129
lexists = getattr(os.path, 'lexists', None)
133
if hasattr(os, 'lstat'):
139
if e.errno == errno.ENOENT:
142
raise BzrError("lstat/stat of (%r): %r" % (f, e))
145
def fancy_rename(old, new, rename_func, unlink_func):
146
"""A fancy rename, when you don't have atomic rename.
148
:param old: The old path, to rename from
149
:param new: The new path, to rename to
150
:param rename_func: The potentially non-atomic rename function
151
:param unlink_func: A way to delete the target file if the full rename succeeds
154
# sftp rename doesn't allow overwriting, so play tricks:
156
base = os.path.basename(new)
157
dirname = os.path.dirname(new)
158
tmp_name = u'tmp.%s.%.9f.%d.%s' % (base, time.time(), os.getpid(), rand_chars(10))
159
tmp_name = pathjoin(dirname, tmp_name)
161
# Rename the file out of the way, but keep track if it didn't exist
162
# We don't want to grab just any exception
163
# something like EACCES should prevent us from continuing
164
# The downside is that the rename_func has to throw an exception
165
# with an errno = ENOENT, or NoSuchFile
168
rename_func(new, tmp_name)
169
except (NoSuchFile,), e:
172
# RBC 20060103 abstraction leakage: the paramiko SFTP clients rename
173
# function raises an IOError with errno == None when a rename fails.
174
# This then gets caught here.
175
if e.errno not in (None, errno.ENOENT, errno.ENOTDIR):
178
if (not hasattr(e, 'errno')
179
or e.errno not in (errno.ENOENT, errno.ENOTDIR)):
186
# This may throw an exception, in which case success will
188
rename_func(old, new)
192
# If the file used to exist, rename it back into place
193
# otherwise just delete it from the tmp location
195
unlink_func(tmp_name)
197
rename_func(tmp_name, new)
200
# In Python 2.4.2 and older, os.path.abspath and os.path.realpath
201
# choke on a Unicode string containing a relative path if
202
# os.getcwd() returns a non-sys.getdefaultencoding()-encoded
204
_fs_enc = sys.getfilesystemencoding()
205
def _posix_abspath(path):
206
return os.path.abspath(path.encode(_fs_enc)).decode(_fs_enc)
207
# jam 20060426 This is another possibility which mimics
208
# os.path.abspath, only uses unicode characters instead
209
# if not os.path.isabs(path):
210
# return os.path.join(os.getcwdu(), path)
214
def _posix_realpath(path):
215
return os.path.realpath(path.encode(_fs_enc)).decode(_fs_enc)
218
def _win32_abspath(path):
219
return _nt_abspath(path.encode(_fs_enc)).decode(_fs_enc).replace('\\', '/')
222
def _win32_realpath(path):
223
return _nt_realpath(path.encode(_fs_enc)).decode(_fs_enc).replace('\\', '/')
226
def _win32_pathjoin(*args):
227
return _nt_join(*args).replace('\\', '/')
230
def _win32_normpath(path):
231
return _nt_normpath(path).replace('\\', '/')
235
return os.getcwdu().replace('\\', '/')
238
def _win32_mkdtemp(*args, **kwargs):
239
return tempfile.mkdtemp(*args, **kwargs).replace('\\', '/')
242
def _win32_rename(old, new):
243
fancy_rename(old, new, rename_func=os.rename, unlink_func=os.unlink)
246
# Default is to just use the python builtins, but these can be rebound on
247
# particular platforms.
248
abspath = _posix_abspath
249
realpath = _posix_realpath
250
pathjoin = os.path.join
251
normpath = os.path.normpath
253
mkdtemp = tempfile.mkdtemp
255
dirname = os.path.dirname
256
basename = os.path.basename
257
rmtree = shutil.rmtree
259
MIN_ABS_PATHLENGTH = 1
262
if sys.platform == 'win32':
263
abspath = _win32_abspath
264
realpath = _win32_realpath
265
pathjoin = _win32_pathjoin
266
normpath = _win32_normpath
267
getcwd = _win32_getcwd
268
mkdtemp = _win32_mkdtemp
269
rename = _win32_rename
271
MIN_ABS_PATHLENGTH = 3
273
def _win32_delete_readonly(function, path, excinfo):
274
"""Error handler for shutil.rmtree function [for win32]
275
Helps to remove files and dirs marked as read-only.
277
type_, value = excinfo[:2]
278
if function in (os.remove, os.rmdir) \
279
and type_ == OSError \
280
and value.errno == errno.EACCES:
281
bzrlib.osutils.make_writable(path)
286
def rmtree(path, ignore_errors=False, onerror=_win32_delete_readonly):
287
"""Replacer for shutil.rmtree: could remove readonly dirs/files"""
288
return shutil.rmtree(path, ignore_errors, onerror)
291
def normalizepath(f):
292
if hasattr(os.path, 'realpath'):
296
[p,e] = os.path.split(f)
297
if e == "" or e == "." or e == "..":
300
return pathjoin(F(p), e)
304
"""Copy a file to a backup.
306
Backups are named in GNU-style, with a ~ suffix.
308
If the file is already a backup, it's not copied.
314
if has_symlinks() and os.path.islink(fn):
315
target = os.readlink(fn)
316
os.symlink(target, bfn)
324
outf = file(bfn, 'wb')
42
"""Return shell-quoted filename"""
43
## We could be a bit more terse by using double-quotes etc
44
f = _QUOTE_RE.sub(r'\\\1', f)
51
mode = os.lstat(f)[ST_MODE]
59
raise BzrError("can't handle file kind with mode %o of %r" % (mode, f))
347
"""True if f is a symlink."""
349
return S_ISLNK(os.lstat(f)[ST_MODE])
353
def is_inside(dir, fname):
354
"""True if fname is inside dir.
356
The parameters should typically be passed to osutils.normpath first, so
357
that . and .. and repeated slashes are eliminated, and the separators
358
are canonical for the platform.
360
The empty string as a dir name is taken as top-of-tree and matches
363
>>> is_inside('src', pathjoin('src', 'foo.c'))
365
>>> is_inside('src', 'srccontrol')
367
>>> is_inside('src', pathjoin('src', 'a', 'a', 'a', 'foo.c'))
369
>>> is_inside('foo.c', 'foo.c')
371
>>> is_inside('foo.c', '')
373
>>> is_inside('', 'foo.c')
376
# XXX: Most callers of this can actually do something smarter by
377
# looking at the inventory
387
return fname.startswith(dir)
390
def is_inside_any(dir_list, fname):
391
"""True if fname is inside any of given dirs."""
392
for dirname in dir_list:
393
if is_inside(dirname, fname):
399
def is_inside_or_parent_of_any(dir_list, fname):
400
"""True if fname is a child or a parent of any of the given files."""
401
for dirname in dir_list:
402
if is_inside(dirname, fname) or is_inside(fname, dirname):
408
80
def pumpfile(fromfile, tofile):
409
81
"""Copy contents of one file to another."""
412
b = fromfile.read(BUFSIZE)
418
def file_iterator(input_file, readsize=32768):
420
b = input_file.read(readsize)
82
tofile.write(fromfile.read())
86
"""Return a new UUID"""
88
return file('/proc/sys/kernel/random/uuid').readline().rstrip('\n')
90
return chomp(os.popen('uuidgen').readline())
427
95
if hasattr(f, 'tell'):
428
96
assert f.tell() == 0
459
122
'sha1': s.hexdigest()}
126
"""Return per-user configuration directory.
128
By default this is ~/.bzr.conf/
130
TODO: Global option --config-dir to override this.
132
return os.path.expanduser("~/.bzr.conf")
136
"""Calculate automatic user identification.
138
Returns (realname, email).
140
Only used when none is set in the environment or the id file.
142
This previously used the FQDN as the default domain, but that can
143
be very slow on machines where DNS is broken. So now we simply
148
# XXX: Any good way to get real user name on win32?
153
w = pwd.getpwuid(uid)
154
gecos = w.pw_gecos.decode(bzrlib.user_encoding)
155
username = w.pw_name.decode(bzrlib.user_encoding)
156
comma = gecos.find(',')
160
realname = gecos[:comma]
166
realname = username = getpass.getuser().decode(bzrlib.user_encoding)
168
return realname, (username + '@' + socket.gethostname())
172
"""Return the full user id from a file or environment variable.
174
TODO: Allow taking this from a file in the branch directory too
175
for per-branch ids."""
176
v = os.environ.get('BZREMAIL')
178
return v.decode(bzrlib.user_encoding)
181
return (open(os.path.join(config_dir(), "email"))
183
.decode(bzrlib.user_encoding)
186
if e.errno != errno.ENOENT:
189
v = os.environ.get('EMAIL')
191
return v.decode(bzrlib.user_encoding)
197
"""Return email-style username.
199
Something similar to 'Martin Pool <mbp@sourcefrog.net>'
201
TODO: Check it's reasonably well-formed.
207
name, email = _auto_user_id()
209
return '%s <%s>' % (name, email)
214
_EMAIL_RE = re.compile(r'[\w+.-]+@[\w+.-]+')
216
"""Return just the email component of a username."""
219
m = _EMAIL_RE.search(e)
221
bailout("%r doesn't seem to contain a reasonable email address" % e)
224
return _auto_user_id()[1]
462
228
def compare_files(a, b):
463
229
"""Returns true if equal in contents"""
521
283
return os.fstat(f.fileno())[ST_SIZE]
524
# Define rand_bytes based on platform.
526
# Python 2.4 and later have os.urandom,
527
# but it doesn't work on some arches
286
if hasattr(os, 'urandom'): # python 2.4 and later
529
287
rand_bytes = os.urandom
530
except (NotImplementedError, AttributeError):
531
# If python doesn't have os.urandom, or it doesn't work,
532
# then try to first pull random data from /dev/urandom
533
if os.path.exists("/dev/urandom"):
534
rand_bytes = file('/dev/urandom', 'rb').read
535
# Otherwise, use this hack as a last resort
537
# not well seeded, but better than nothing
542
s += chr(random.randint(0, 255))
547
ALNUM = '0123456789abcdefghijklmnopqrstuvwxyz'
549
"""Return a random string of num alphanumeric characters
551
The result only contains lowercase chars because it may be used on
552
case-insensitive filesystems.
555
for raw_byte in rand_bytes(num):
556
s += ALNUM[ord(raw_byte) % 36]
289
# FIXME: No good on non-Linux
290
_rand_file = file('/dev/urandom', 'rb')
291
rand_bytes = _rand_file.read
560
294
## TODO: We could later have path objects that remember their list
596
330
assert isinstance(p, list)
598
332
if (f == '..') or (f == None) or (f == ''):
599
raise BzrError("sorry, %r not allowed in path" % f)
603
@deprecated_function(zero_nine)
333
bailout("sorry, %r not allowed in path" % f)
334
return os.path.join(*p)
604
337
def appendpath(p1, p2):
608
return pathjoin(p1, p2)
612
"""Split s into lines, but without removing the newline characters."""
613
lines = s.split('\n')
614
result = [line + '\n' for line in lines[:-1]]
616
result.append(lines[-1])
620
def hardlinks_good():
621
return sys.platform not in ('win32', 'cygwin', 'darwin')
624
def link_or_copy(src, dest):
625
"""Hardlink a file, or copy it if it can't be hardlinked."""
626
if not hardlinks_good():
631
except (OSError, IOError), e:
632
if e.errno != errno.EXDEV:
636
def delete_any(full_path):
637
"""Delete a file or directory."""
641
# We may be renaming a dangling inventory id
642
if e.errno not in (errno.EISDIR, errno.EACCES, errno.EPERM):
648
if hasattr(os, 'symlink'):
654
def contains_whitespace(s):
655
"""True if there are any whitespace characters in s."""
656
for ch in string.whitespace:
663
def contains_linebreaks(s):
664
"""True if there is any vertical whitespace in s."""
672
def relpath(base, path):
673
"""Return path relative to base, or raise exception.
675
The path may be either an absolute path or a path relative to the
676
current working directory.
678
os.path.commonprefix (python2.4) has a bad bug that it works just
679
on string prefixes, assuming that '/u' is a prefix of '/u2'. This
683
assert len(base) >= MIN_ABS_PATHLENGTH, ('Length of base must be equal or'
684
' exceed the platform minimum length (which is %d)' %
691
while len(head) >= len(base):
694
head, tail = os.path.split(head)
698
raise PathNotChild(rp, base)
706
def safe_unicode(unicode_or_utf8_string):
707
"""Coerce unicode_or_utf8_string into unicode.
709
If it is unicode, it is returned.
710
Otherwise it is decoded from utf-8. If a decoding error
711
occurs, it is wrapped as a If the decoding fails, the exception is wrapped
712
as a BzrBadParameter exception.
714
if isinstance(unicode_or_utf8_string, unicode):
715
return unicode_or_utf8_string
717
return unicode_or_utf8_string.decode('utf8')
718
except UnicodeDecodeError:
719
raise BzrBadParameterNotUnicode(unicode_or_utf8_string)
722
_platform_normalizes_filenames = False
723
if sys.platform == 'darwin':
724
_platform_normalizes_filenames = True
727
def normalizes_filenames():
728
"""Return True if this platform normalizes unicode filenames.
730
Mac OSX does, Windows/Linux do not.
732
return _platform_normalizes_filenames
735
if _platform_normalizes_filenames:
736
def unicode_filename(path):
737
"""Make sure 'path' is a properly normalized filename.
739
On platforms where the system normalizes filenames (Mac OSX),
740
you can access a file by any path which will normalize
742
Internally, bzr only supports NFC/NFKC normalization, since
743
that is the standard for XML documents.
744
So we return an normalized path, and indicate this has been
747
:return: (path, is_normalized) Return a path which can
748
access the file, and whether or not this path is
751
return unicodedata.normalize('NFKC', path), True
753
def unicode_filename(path):
754
"""Make sure 'path' is a properly normalized filename.
756
On platforms where the system does not normalize filenames
757
(Windows, Linux), you have to access a file by its exact path.
758
Internally, bzr only supports NFC/NFKC normalization, since
759
that is the standard for XML documents.
760
So we return the original path, and indicate if this is
763
:return: (path, is_normalized) Return a path which can
764
access the file, and whether or not this path is
767
return path, unicodedata.normalize('NFKC', path) == path
770
def terminal_width():
771
"""Return estimated terminal width."""
772
if sys.platform == 'win32':
773
import bzrlib.win32console
774
return bzrlib.win32console.get_console_size()[0]
777
import struct, fcntl, termios
778
s = struct.pack('HHHH', 0, 0, 0, 0)
779
x = fcntl.ioctl(1, termios.TIOCGWINSZ, s)
780
width = struct.unpack('HHHH', x)[1]
785
width = int(os.environ['COLUMNS'])
793
def supports_executable():
794
return sys.platform != "win32"
797
_validWin32PathRE = re.compile(r'^([A-Za-z]:[/\\])?[^:<>*"?\|]*$')
800
def check_legal_path(path):
801
"""Check whether the supplied path is legal.
802
This is only required on Windows, so we don't test on other platforms
805
if sys.platform != "win32":
807
if _validWin32PathRE.match(path) is None:
808
raise IllegalPath(path)
811
def walkdirs(top, prefix=""):
812
"""Yield data about all the directories in a tree.
814
This yields all the data about the contents of a directory at a time.
815
After each directory has been yielded, if the caller has mutated the list
816
to exclude some directories, they are then not descended into.
818
The data yielded is of the form:
819
[(relpath, basename, kind, lstat, path_from_top), ...]
821
:param prefix: Prefix the relpaths that are yielded with 'prefix'. This
822
allows one to walk a subtree but get paths that are relative to a tree
824
:return: an iterator over the dirs.
828
_directory = _directory_kind
830
pending = [(prefix, "", _directory, None, top)]
833
currentdir = pending.pop()
834
# 0 - relpath, 1- basename, 2- kind, 3- stat, 4-toppath
837
relroot = currentdir[0] + '/'
840
for name in sorted(_listdir(top)):
841
abspath = top + '/' + name
842
statvalue = lstat(abspath)
843
dirblock.append ((relroot + name, name, file_kind_from_stat_mode(statvalue.st_mode), statvalue, abspath))
845
# push the user specified dirs from dirblock
846
for dir in reversed(dirblock):
847
if dir[2] == _directory:
341
return os.path.join(p1, p2)
344
def extern_command(cmd, ignore_errors = False):
345
mutter('external command: %s' % `cmd`)
347
if not ignore_errors:
348
bailout('command failed')
added
removed
bzrlib/osutils.py
