1
# Copyright (C) 2005-2010 Canonical Ltd
1
# Copyright (C) 2005 Robey Pointer <robey@lag.net>, Canonical Ltd
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
17
"""Implementation of Transport over SFTP, using paramiko."""
19
# TODO: Remove the transport-based lock_read and lock_write methods. They'll
20
# then raise TransportNotPossible, which will break remote access to any
21
# formats which rely on OS-level locks. That should be fine as those formats
22
# are pretty old, but these combinations may have to be removed from the test
23
# suite. Those formats all date back to 0.7; so we should be able to remove
24
# these methods when we officially drop support for those formats.
44
from bzrlib.errors import (FileExists,
45
NoSuchFile, PathNotChild,
32
from bzrlib.errors import (FileExists,
33
TransportNotPossible, NoSuchFile, PathNotChild,
51
from bzrlib.osutils import pathjoin, fancy_rename, getcwd
52
from bzrlib.symbol_versioning import (
55
from bzrlib.trace import mutter, warning
56
from bzrlib.transport import (
65
# Disable one particular warning that comes from paramiko in Python2.5; if
66
# this is emitted at the wrong time it tends to cause spurious test failures
67
# or at least noise in the test case::
69
# [1770/7639 in 86s, 1 known failures, 50 skipped, 2 missing features]
70
# test_permissions.TestSftpPermissions.test_new_files
71
# /var/lib/python-support/python2.5/paramiko/message.py:226: DeprecationWarning: integer argument expected, got float
72
# self.packet.write(struct.pack('>I', n))
73
warnings.filterwarnings('ignore',
74
'integer argument expected, got float',
75
category=DeprecationWarning,
76
module='paramiko.message')
36
from bzrlib.config import config_dir
37
from bzrlib.trace import mutter, warning, error
38
from bzrlib.transport import Transport, register_transport
80
except ImportError, e:
81
raise ParamikoNotPresent(e)
44
error('The SFTP transport requires paramiko.')
83
47
from paramiko.sftp import (SFTP_FLAG_WRITE, SFTP_FLAG_CREATE,
84
48
SFTP_FLAG_EXCL, SFTP_FLAG_TRUNC,
85
SFTP_OK, CMD_HANDLE, CMD_OPEN)
86
50
from paramiko.sftp_attr import SFTPAttributes
87
51
from paramiko.sftp_file import SFTPFile
90
_paramiko_version = getattr(paramiko, '__version_info__', (0, 0, 0))
91
# don't use prefetch unless paramiko version >= 1.5.5 (there were bugs earlier)
92
_default_do_prefetch = (_paramiko_version >= (1, 5, 5))
52
from paramiko.sftp_client import SFTPClient
54
if 'sftp' not in urlparse.uses_netloc: urlparse.uses_netloc.append('sftp')
58
if sys.platform == 'win32':
59
# close_fds not supported on win32
63
def _get_ssh_vendor():
64
"""Find out what version of SSH is on the system."""
66
if _ssh_vendor is not None:
72
p = subprocess.Popen(['ssh', '-V'],
74
stdin=subprocess.PIPE,
75
stdout=subprocess.PIPE,
76
stderr=subprocess.PIPE)
77
returncode = p.returncode
78
stdout, stderr = p.communicate()
82
if 'OpenSSH' in stderr:
83
mutter('ssh implementation is OpenSSH')
84
_ssh_vendor = 'openssh'
85
elif 'SSH Secure Shell' in stderr:
86
mutter('ssh implementation is SSH Corp.')
89
if _ssh_vendor != 'none':
92
# XXX: 20051123 jamesh
93
# A check for putty's plink or lsh would go here.
95
mutter('falling back to paramiko implementation')
100
"""A socket-like object that talks to an ssh subprocess via pipes."""
101
def __init__(self, hostname, port=None, user=None):
102
vendor = _get_ssh_vendor()
103
assert vendor in ['openssh', 'ssh']
104
if vendor == 'openssh':
106
'-oForwardX11=no', '-oForwardAgent=no',
107
'-oClearAllForwardings=yes', '-oProtocol=2',
108
'-oNoHostAuthenticationForLocalhost=yes']
110
args.extend(['-p', str(port)])
112
args.extend(['-l', user])
113
args.extend(['-s', hostname, 'sftp'])
114
elif vendor == 'ssh':
117
args.extend(['-p', str(port)])
119
args.extend(['-l', user])
120
args.extend(['-s', 'sftp', hostname])
122
self.proc = subprocess.Popen(args, close_fds=_close_fds,
123
stdin=subprocess.PIPE,
124
stdout=subprocess.PIPE)
126
def send(self, data):
127
return os.write(self.proc.stdin.fileno(), data)
129
def recv(self, count):
130
return os.read(self.proc.stdout.fileno(), count)
133
self.proc.stdin.close()
134
self.proc.stdout.close()
141
# This is a weakref dictionary, so that we can reuse connections
142
# that are still active. Long term, it might be nice to have some
143
# sort of expiration policy, such as disconnect if inactive for
144
# X seconds. But that requires a lot more fanciness.
145
_connected_hosts = weakref.WeakValueDictionary()
147
def load_host_keys():
149
Load system host keys (probably doesn't work on windows) and any
150
"discovered" keys from previous sessions.
152
global SYSTEM_HOSTKEYS, BZR_HOSTKEYS
154
SYSTEM_HOSTKEYS = paramiko.util.load_host_keys(os.path.expanduser('~/.ssh/known_hosts'))
156
mutter('failed to load system host keys: ' + str(e))
157
bzr_hostkey_path = os.path.join(config_dir(), 'ssh_host_keys')
159
BZR_HOSTKEYS = paramiko.util.load_host_keys(bzr_hostkey_path)
161
mutter('failed to load bzr host keys: ' + str(e))
164
def save_host_keys():
166
Save "discovered" host keys in $(config)/ssh_host_keys/.
168
global SYSTEM_HOSTKEYS, BZR_HOSTKEYS
169
bzr_hostkey_path = os.path.join(config_dir(), 'ssh_host_keys')
170
if not os.path.isdir(config_dir()):
171
os.mkdir(config_dir())
173
f = open(bzr_hostkey_path, 'w')
174
f.write('# SSH host keys collected by bzr\n')
175
for hostname, keys in BZR_HOSTKEYS.iteritems():
176
for keytype, key in keys.iteritems():
177
f.write('%s %s %s\n' % (hostname, keytype, key.get_base64()))
180
mutter('failed to save bzr host keys: ' + str(e))
95
183
class SFTPLock(object):
96
"""This fakes a lock in a remote location.
98
A present lock is indicated just by the existence of a file. This
99
doesn't work well on all transports and they are only used in
100
deprecated storage formats.
184
"""This fakes a lock in a remote location."""
103
185
__slots__ = ['path', 'lock_path', 'lock_file', 'transport']
105
186
def __init__(self, path, transport):
187
assert isinstance(transport, SFTPTransport)
106
189
self.lock_file = None
108
191
self.lock_path = path + '.write-lock'
109
192
self.transport = transport
111
# RBC 20060103 FIXME should we be using private methods here ?
112
abspath = transport._remote_path(self.lock_path)
113
self.lock_file = transport._sftp_open_exclusive(abspath)
194
self.lock_file = transport._sftp_open_exclusive(self.lock_path)
114
195
except FileExists:
115
196
raise LockError('File %r already locked' % (self.path,))
117
198
def __del__(self):
118
199
"""Should this warn, or actually try to cleanup?"""
119
200
if self.lock_file:
120
warning("SFTPLock %r not explicitly unlocked" % (self.path,))
201
warn("SFTPLock %r not explicitly unlocked" % (self.path,))
123
204
def unlock(self):
131
212
# What specific errors should we catch here?
135
class _SFTPReadvHelper(object):
136
"""A class to help with managing the state of a readv request."""
138
# See _get_requests for an explanation.
139
_max_request_size = 32768
141
def __init__(self, original_offsets, relpath, _report_activity):
142
"""Create a new readv helper.
144
:param original_offsets: The original requests given by the caller of
146
:param relpath: The name of the file (if known)
147
:param _report_activity: A Transport._report_activity bound method,
148
to be called as data arrives.
150
self.original_offsets = list(original_offsets)
151
self.relpath = relpath
152
self._report_activity = _report_activity
154
def _get_requests(self):
155
"""Break up the offsets into individual requests over sftp.
157
The SFTP spec only requires implementers to support 32kB requests. We
158
could try something larger (openssh supports 64kB), but then we have to
159
handle requests that fail.
160
So instead, we just break up our maximum chunks into 32kB chunks, and
161
asyncronously requests them.
162
Newer versions of paramiko would do the chunking for us, but we want to
163
start processing results right away, so we do it ourselves.
165
# TODO: Because we issue async requests, we don't 'fudge' any extra
166
# data. I'm not 100% sure that is the best choice.
168
# The first thing we do, is to collapse the individual requests as much
169
# as possible, so we don't issues requests <32kB
170
sorted_offsets = sorted(self.original_offsets)
171
coalesced = list(ConnectedTransport._coalesce_offsets(sorted_offsets,
172
limit=0, fudge_factor=0))
174
for c_offset in coalesced:
175
start = c_offset.start
176
size = c_offset.length
178
# Break this up into 32kB requests
180
next_size = min(size, self._max_request_size)
181
requests.append((start, next_size))
184
if 'sftp' in debug.debug_flags:
185
mutter('SFTP.readv(%s) %s offsets => %s coalesced => %s requests',
186
self.relpath, len(sorted_offsets), len(coalesced),
190
def request_and_yield_offsets(self, fp):
191
"""Request the data from the remote machine, yielding the results.
193
:param fp: A Paramiko SFTPFile object that supports readv.
194
:return: Yield the data requested by the original readv caller, one by
197
requests = self._get_requests()
198
offset_iter = iter(self.original_offsets)
199
cur_offset, cur_size = offset_iter.next()
200
# paramiko .readv() yields strings that are in the order of the requests
201
# So we track the current request to know where the next data is
202
# being returned from.
208
# This is used to buffer chunks which we couldn't process yet
209
# It is (start, end, data) tuples.
211
# Create an 'unlimited' data stream, so we stop based on requests,
212
# rather than just because the data stream ended. This lets us detect
214
data_stream = itertools.chain(fp.readv(requests),
215
itertools.repeat(None))
216
for (start, length), data in itertools.izip(requests, data_stream):
218
if cur_coalesced is not None:
219
raise errors.ShortReadvError(self.relpath,
220
start, length, len(data))
221
if len(data) != length:
222
raise errors.ShortReadvError(self.relpath,
223
start, length, len(data))
224
self._report_activity(length, 'read')
226
# This is the first request, just buffer it
227
buffered_data = [data]
228
buffered_len = length
230
elif start == last_end:
231
# The data we are reading fits neatly on the previous
232
# buffer, so this is all part of a larger coalesced range.
233
buffered_data.append(data)
234
buffered_len += length
215
class SFTPTransport (Transport):
217
Transport implementation for SFTP access.
219
_do_prefetch = False # Right now Paramiko's prefetch support causes things to hang
221
def __init__(self, base, clone_from=None):
222
assert base.startswith('sftp://')
223
self._parse_url(base)
224
base = self._unparse_url()
225
super(SFTPTransport, self).__init__(base)
226
if clone_from is None:
229
# use the same ssh connection, etc
230
self._sftp = clone_from._sftp
231
# super saves 'self.base'
233
def should_cache(self):
235
Return True if the data pulled across should be cached locally.
239
def clone(self, offset=None):
241
Return a new SFTPTransport with root at self.base + offset.
242
We share the same SFTP session between such transports, because it's
243
fairly expensive to set them up.
246
return SFTPTransport(self.base, self)
248
return SFTPTransport(self.abspath(offset), self)
250
def abspath(self, relpath):
252
Return the full url to the given relative path.
254
@param relpath: the relative path or path components
255
@type relpath: str or list
257
return self._unparse_url(self._abspath(relpath))
259
def _abspath(self, relpath):
260
"""Return the absolute path segment without the SFTP URL."""
261
# FIXME: share the common code across transports
262
assert isinstance(relpath, basestring)
263
relpath = [urllib.unquote(relpath)]
264
basepath = self._path.split('/')
265
if len(basepath) > 0 and basepath[-1] == '':
266
basepath = basepath[:-1]
270
if len(basepath) == 0:
271
# In most filesystems, a request for the parent
272
# of root, just returns root.
236
# We have an 'interrupt' in the data stream. So we know we are
237
# at a request boundary.
239
# We haven't consumed the buffer so far, so put it into
240
# data_chunks, and continue.
241
buffered = ''.join(buffered_data)
242
data_chunks.append((input_start, buffered))
244
buffered_data = [data]
245
buffered_len = length
246
last_end = start + length
247
if input_start == cur_offset and cur_size <= buffered_len:
248
# Simplify the next steps a bit by transforming buffered_data
249
# into a single string. We also have the nice property that
250
# when there is only one string ''.join([x]) == x, so there is
252
buffered = ''.join(buffered_data)
253
# Clean out buffered data so that we keep memory
257
# TODO: We *could* also consider the case where cur_offset is in
258
# in the buffered range, even though it doesn't *start*
259
# the buffered range. But for packs we pretty much always
260
# read in order, so you won't get any extra data in the
262
while (input_start == cur_offset
263
and (buffered_offset + cur_size) <= buffered_len):
264
# We've buffered enough data to process this request, spit it
266
cur_data = buffered[buffered_offset:buffered_offset + cur_size]
267
# move the direct pointer into our buffered data
268
buffered_offset += cur_size
269
# Move the start-of-buffer pointer
270
input_start += cur_size
271
# Yield the requested data
272
yield cur_offset, cur_data
273
cur_offset, cur_size = offset_iter.next()
274
# at this point, we've consumed as much of buffered as we can,
275
# so break off the portion that we consumed
276
if buffered_offset == len(buffered_data):
277
# No tail to leave behind
281
buffered = buffered[buffered_offset:]
282
buffered_data = [buffered]
283
buffered_len = len(buffered)
285
buffered = ''.join(buffered_data)
287
data_chunks.append((input_start, buffered))
289
if 'sftp' in debug.debug_flags:
290
mutter('SFTP readv left with %d out-of-order bytes',
291
sum(map(lambda x: len(x[1]), data_chunks)))
292
# We've processed all the readv data, at this point, anything we
293
# couldn't process is in data_chunks. This doesn't happen often, so
294
# this code path isn't optimized
295
# We use an interesting process for data_chunks
296
# Specifically if we have "bisect_left([(start, len, entries)],
298
# If start == qstart, then we get the specific node. Otherwise we
299
# get the previous node
301
idx = bisect.bisect_left(data_chunks, (cur_offset,))
302
if idx < len(data_chunks) and data_chunks[idx][0] == cur_offset:
303
# The data starts here
304
data = data_chunks[idx][1][:cur_size]
306
# The data is in a portion of a previous page
308
sub_offset = cur_offset - data_chunks[idx][0]
309
data = data_chunks[idx][1]
310
data = data[sub_offset:sub_offset + cur_size]
312
# We are missing the page where the data should be found,
315
if len(data) != cur_size:
316
raise AssertionError('We must have miscalulated.'
317
' We expected %d bytes, but only found %d'
318
% (cur_size, len(data)))
319
yield cur_offset, data
320
cur_offset, cur_size = offset_iter.next()
323
class SFTPTransport(ConnectedTransport):
324
"""Transport implementation for SFTP access."""
326
_do_prefetch = _default_do_prefetch
327
# TODO: jam 20060717 Conceivably these could be configurable, either
328
# by auto-tuning at run-time, or by a configuration (per host??)
329
# but the performance curve is pretty flat, so just going with
330
# reasonable defaults.
331
_max_readv_combine = 200
332
# Having to round trip to the server means waiting for a response,
333
# so it is better to download extra bytes.
334
# 8KiB had good performance for both local and remote network operations
335
_bytes_to_read_before_seek = 8192
337
# The sftp spec says that implementations SHOULD allow reads
338
# to be at least 32K. paramiko.readv() does an async request
339
# for the chunks. So we need to keep it within a single request
340
# size for paramiko <= 1.6.1. paramiko 1.6.2 will probably chop
341
# up the request itself, rather than us having to worry about it
342
_max_request_size = 32768
344
def __init__(self, base, _from_transport=None):
345
super(SFTPTransport, self).__init__(base,
346
_from_transport=_from_transport)
348
def _remote_path(self, relpath):
349
"""Return the path to be passed along the sftp protocol for relpath.
351
:param relpath: is a urlencoded string.
353
relative = urlutils.unescape(relpath).encode('utf-8')
354
remote_path = self._combine_paths(self._path, relative)
355
# the initial slash should be removed from the path, and treated as a
356
# homedir relative path (the path begins with a double slash if it is
357
# absolute). see draft-ietf-secsh-scp-sftp-ssh-uri-03.txt
358
# RBC 20060118 we are not using this as its too user hostile. instead
359
# we are following lftp and using /~/foo to mean '~/foo'
360
# vila--20070602 and leave absolute paths begin with a single slash.
361
if remote_path.startswith('/~/'):
362
remote_path = remote_path[3:]
363
elif remote_path == '/~':
367
def _create_connection(self, credentials=None):
368
"""Create a new connection with the provided credentials.
370
:param credentials: The credentials needed to establish the connection.
372
:return: The created connection and its associated credentials.
374
The credentials are only the password as it may have been entered
375
interactively by the user and may be different from the one provided
376
in base url at transport creation time.
378
if credentials is None:
379
password = self._password
381
password = credentials
383
vendor = ssh._get_ssh_vendor()
386
auth = config.AuthenticationConfig()
387
user = auth.get_user('ssh', self._host, self._port)
388
connection = vendor.connect_sftp(self._user, password,
389
self._host, self._port)
390
return connection, (user, password)
393
"""Ensures that a connection is established"""
394
connection = self._get_connection()
395
if connection is None:
396
# First connection ever
397
connection, credentials = self._create_connection()
398
self._set_connection(connection, credentials)
280
path = '/'.join(basepath)
281
# could still be a "relative" path here, but relative on the sftp server
284
def relpath(self, abspath):
285
username, password, host, port, path = self._split_url(abspath)
287
if (username != self._username):
288
error.append('username mismatch')
289
if (host != self._host):
290
error.append('host mismatch')
291
if (port != self._port):
292
error.append('port mismatch')
293
if (not path.startswith(self._path)):
294
error.append('path mismatch')
296
extra = ': ' + ', '.join(error)
297
raise PathNotChild(abspath, self.base, extra=extra)
299
return path[pl:].lstrip('/')
401
301
def has(self, relpath):
403
303
Does the target location exist?
406
self._get_sftp().stat(self._remote_path(relpath))
407
# stat result is about 20 bytes, let's say
408
self._report_activity(20, 'read')
306
self._sftp.stat(self._abspath(relpath))
413
def get(self, relpath):
414
"""Get the file at the given relative path.
311
def get(self, relpath, decode=False):
313
Get the file at the given relative path.
416
315
:param relpath: The relative path to the file
419
# FIXME: by returning the file directly, we don't pass this
420
# through to report_activity. We could try wrapping the object
421
# before it's returned. For readv and get_bytes it's handled in
422
# the higher-level function.
424
path = self._remote_path(relpath)
425
f = self._get_sftp().file(path, mode='rb')
426
if self._do_prefetch and (getattr(f, 'prefetch', None) is not None):
318
path = self._abspath(relpath)
319
f = self._sftp.file(path)
320
if self._do_prefetch and hasattr(f, 'prefetch'):
429
323
except (IOError, paramiko.SSHException), e:
430
self._translate_io_exception(e, path, ': error retrieving',
431
failure_exc=errors.ReadError)
433
def get_bytes(self, relpath):
434
# reimplement this here so that we can report how many bytes came back
324
self._translate_io_exception(e, path, ': error retrieving')
326
def get_partial(self, relpath, start, length=None):
328
Get just part of a file.
330
:param relpath: Path to the file, relative to base
331
:param start: The starting position to read from
332
:param length: The length to read. A length of None indicates
333
read to the end of the file.
334
:return: A file-like object containing at least the specified bytes.
335
Some implementations may return objects which can be read
336
past this length, but this is not guaranteed.
338
# TODO: implement get_partial_multi to help with knit support
435
339
f = self.get(relpath)
438
self._report_activity(len(bytes), 'read')
443
def _readv(self, relpath, offsets):
444
"""See Transport.readv()"""
445
# We overload the default readv() because we want to use a file
446
# that does not have prefetch enabled.
447
# Also, if we have a new paramiko, it implements an async readv()
452
path = self._remote_path(relpath)
453
fp = self._get_sftp().file(path, mode='rb')
454
readv = getattr(fp, 'readv', None)
456
return self._sftp_readv(fp, offsets, relpath)
457
if 'sftp' in debug.debug_flags:
458
mutter('seek and read %s offsets', len(offsets))
459
return self._seek_and_read(fp, offsets, relpath)
460
except (IOError, paramiko.SSHException), e:
461
self._translate_io_exception(e, path, ': error retrieving')
463
def recommended_page_size(self):
464
"""See Transport.recommended_page_size().
466
For SFTP we suggest a large page size to reduce the overhead
467
introduced by latency.
471
def _sftp_readv(self, fp, offsets, relpath):
472
"""Use the readv() member of fp to do async readv.
474
Then read them using paramiko.readv(). paramiko.readv()
475
does not support ranges > 64K, so it caps the request size, and
476
just reads until it gets all the stuff it wants.
478
helper = _SFTPReadvHelper(offsets, relpath, self._report_activity)
479
return helper.request_and_yield_offsets(fp)
481
def put_file(self, relpath, f, mode=None):
483
Copy the file-like object into the location.
341
if self._do_prefetch and hasattr(f, 'prefetch'):
345
def put(self, relpath, f):
347
Copy the file-like or string object into the location.
485
349
:param relpath: Location to put the contents, relative to base.
486
:param f: File-like object.
487
:param mode: The final mode for the file
350
:param f: File-like or string object.
489
final_path = self._remote_path(relpath)
490
return self._put(final_path, f, mode=mode)
492
def _put(self, abspath, f, mode=None):
493
"""Helper function so both put() and copy_abspaths can reuse the code"""
494
tmp_abspath = '%s.tmp.%.9f.%d.%d' % (abspath, time.time(),
352
final_path = self._abspath(relpath)
353
tmp_relpath = '%s.tmp.%.9f.%d.%d' % (relpath, time.time(),
495
354
os.getpid(), random.randint(0,0x7FFFFFFF))
496
fout = self._sftp_open_exclusive(tmp_abspath, mode=mode)
355
tmp_abspath = self._abspath(tmp_relpath)
356
fout = self._sftp_open_exclusive(tmp_relpath)
500
fout.set_pipelined(True)
501
length = self._pump(f, fout)
502
except (IOError, paramiko.SSHException), e:
503
self._translate_io_exception(e, tmp_abspath)
504
# XXX: This doesn't truly help like we would like it to.
505
# The problem is that openssh strips sticky bits. So while we
506
# can properly set group write permission, we lose the group
507
# sticky bit. So it is probably best to stop chmodding, and
508
# just tell users that they need to set the umask correctly.
509
# The attr.st_mode = mode, in _sftp_open_exclusive
510
# will handle when the user wants the final mode to be more
511
# restrictive. And then we avoid a round trip. Unless
512
# paramiko decides to expose an async chmod()
514
# This is designed to chmod() right before we close.
515
# Because we set_pipelined() earlier, theoretically we might
516
# avoid the round trip for fout.close()
518
self._get_sftp().chmod(tmp_abspath, mode)
521
self._rename_and_overwrite(tmp_abspath, abspath)
361
except (paramiko.SSHException, IOError), e:
362
self._translate_io_exception(e, relpath, ': unable to write')
523
363
except Exception, e:
524
364
# If we fail, try to clean up the temporary file
525
365
# before we throw the exception
526
366
# but don't let another exception mess things up
527
# Write out the traceback, because otherwise
528
# the catch and throw destroys it
530
mutter(traceback.format_exc())
534
self._get_sftp().remove(tmp_abspath)
536
# raise the saved except
538
# raise the original with its traceback if we can.
541
def _put_non_atomic_helper(self, relpath, writer, mode=None,
542
create_parent_dir=False,
544
abspath = self._remote_path(relpath)
546
# TODO: jam 20060816 paramiko doesn't publicly expose a way to
547
# set the file mode at create time. If it does, use it.
548
# But for now, we just chmod later anyway.
550
def _open_and_write_file():
551
"""Try to open the target file, raise error on failure"""
369
self._sftp.remove(tmp_abspath)
374
# sftp rename doesn't allow overwriting, so play tricks:
375
tmp_safety = 'bzr.tmp.%.9f.%d.%d' % (time.time(), os.getpid(), random.randint(0, 0x7FFFFFFF))
376
tmp_safety = self._abspath(tmp_safety)
378
self._sftp.rename(final_path, tmp_safety)
555
fout = self._get_sftp().file(abspath, mode='wb')
556
fout.set_pipelined(True)
385
self._sftp.rename(tmp_abspath, final_path)
558
386
except (paramiko.SSHException, IOError), e:
559
self._translate_io_exception(e, abspath,
562
# This is designed to chmod() right before we close.
563
# Because we set_pipelined() earlier, theoretically we might
564
# avoid the round trip for fout.close()
566
self._get_sftp().chmod(abspath, mode)
387
self._translate_io_exception(e, relpath, ': unable to rename')
571
if not create_parent_dir:
572
_open_and_write_file()
575
# Try error handling to create the parent directory if we need to
577
_open_and_write_file()
579
# Try to create the parent directory, and then go back to
581
parent_dir = os.path.dirname(abspath)
582
self._mkdir(parent_dir, dir_mode)
583
_open_and_write_file()
585
def put_file_non_atomic(self, relpath, f, mode=None,
586
create_parent_dir=False,
588
"""Copy the file-like object into the target location.
590
This function is not strictly safe to use. It is only meant to
591
be used when you already know that the target does not exist.
592
It is not safe, because it will open and truncate the remote
593
file. So there may be a time when the file has invalid contents.
595
:param relpath: The remote location to put the contents.
596
:param f: File-like object.
597
:param mode: Possible access permissions for new file.
598
None means do not set remote permissions.
599
:param create_parent_dir: If we cannot create the target file because
600
the parent directory does not exist, go ahead and
601
create it, and then try again.
605
self._put_non_atomic_helper(relpath, writer, mode=mode,
606
create_parent_dir=create_parent_dir,
609
def put_bytes_non_atomic(self, relpath, bytes, mode=None,
610
create_parent_dir=False,
614
self._put_non_atomic_helper(relpath, writer, mode=mode,
615
create_parent_dir=create_parent_dir,
393
self._sftp.unlink(tmp_safety)
395
self._sftp.rename(tmp_safety, final_path)
618
397
def iter_files_recursive(self):
619
398
"""Walk the relative paths of all files in this transport."""
620
# progress is handled by list_dir
621
399
queue = list(self.list_dir('.'))
623
relpath = queue.pop(0)
401
relpath = urllib.quote(queue.pop(0))
624
402
st = self.stat(relpath)
625
403
if stat.S_ISDIR(st.st_mode):
626
404
for i, basename in enumerate(self.list_dir(relpath)):
694
425
:param failure_exc: Paramiko has the super fun ability to raise completely
695
426
opaque errors that just set "e.args = ('Failure',)" with
696
427
no more information.
697
If this parameter is set, it defines the exception
698
to raise in these cases.
428
This sometimes means FileExists, but it also sometimes
700
431
# paramiko seems to generate detailless errors.
701
432
self._translate_error(e, path, raise_generic=False)
702
if getattr(e, 'args', None) is not None:
433
if hasattr(e, 'args'):
703
434
if (e.args == ('No such file or directory',) or
704
435
e.args == ('No such file',)):
705
436
raise NoSuchFile(path, str(e) + more_info)
706
if (e.args == ('mkdir failed',) or
707
e.args[0].startswith('syserr: File exists')):
437
if (e.args == ('mkdir failed',)):
708
438
raise FileExists(path, str(e) + more_info)
709
439
# strange but true, for the paramiko server.
710
440
if (e.args == ('Failure',)):
711
441
raise failure_exc(path, str(e) + more_info)
712
# Can be something like args = ('Directory not empty:
713
# '/srv/bazaar.launchpad.net/blah...: '
714
# [Errno 39] Directory not empty',)
715
if (e.args[0].startswith('Directory not empty: ')
716
or getattr(e, 'errno', None) == errno.ENOTEMPTY):
717
raise errors.DirectoryNotEmpty(path, str(e))
718
mutter('Raising exception with args %s', e.args)
719
if getattr(e, 'errno', None) is not None:
720
mutter('Raising exception with errno %s', e.errno)
723
def append_file(self, relpath, f, mode=None):
444
def append(self, relpath, f):
725
446
Append the text in the file-like object into the final
729
path = self._remote_path(relpath)
730
fout = self._get_sftp().file(path, 'ab')
732
self._get_sftp().chmod(path, mode)
450
path = self._abspath(relpath)
451
fout = self._sftp.file(path, 'ab')
734
452
self._pump(f, fout)
736
453
except (IOError, paramiko.SSHException), e:
737
454
self._translate_io_exception(e, relpath, ': unable to append')
739
def rename(self, rel_from, rel_to):
740
"""Rename without special overwriting"""
742
self._get_sftp().rename(self._remote_path(rel_from),
743
self._remote_path(rel_to))
744
except (IOError, paramiko.SSHException), e:
745
self._translate_io_exception(e, rel_from,
746
': unable to rename to %r' % (rel_to))
748
def _rename_and_overwrite(self, abs_from, abs_to):
749
"""Do a fancy rename on the remote server.
751
Using the implementation provided by osutils.
754
sftp = self._get_sftp()
755
fancy_rename(abs_from, abs_to,
756
rename_func=sftp.rename,
757
unlink_func=sftp.remove)
758
except (IOError, paramiko.SSHException), e:
759
self._translate_io_exception(e, abs_from,
760
': unable to rename to %r' % (abs_to))
456
def copy(self, rel_from, rel_to):
457
"""Copy the item at rel_from to the location at rel_to"""
458
path_from = self._abspath(rel_from)
459
path_to = self._abspath(rel_to)
460
self._copy_abspaths(path_from, path_to)
462
def _copy_abspaths(self, path_from, path_to):
463
"""Copy files given an absolute path
465
:param path_from: Path on remote server to read
466
:param path_to: Path on remote server to write
469
TODO: Should the destination location be atomically created?
470
This has not been specified
471
TODO: This should use some sort of remote copy, rather than
472
pulling the data locally, and then writing it remotely
475
fin = self._sftp.file(path_from, 'rb')
477
fout = self._sftp.file(path_to, 'wb')
479
fout.set_pipelined(True)
480
self._pump(fin, fout)
485
except (IOError, paramiko.SSHException), e:
486
self._translate_io_exception(e, path_from, ': unable copy to: %r' % path_to)
488
def copy_to(self, relpaths, other, pb=None):
489
"""Copy a set of entries from self into another Transport.
491
:param relpaths: A list/generator of entries to be copied.
493
if isinstance(other, SFTPTransport) and other._sftp is self._sftp:
494
# Both from & to are on the same remote filesystem
495
# We can use a remote copy, instead of pulling locally, and pushing
497
total = self._get_total(relpaths)
499
for path in relpaths:
500
path_from = self._abspath(relpath)
501
path_to = other._abspath(relpath)
502
self._update_pb(pb, 'copy-to', count, total)
503
self._copy_abspaths(path_from, path_to)
507
return super(SFTPTransport, self).copy_to(relpaths, other, pb=pb)
509
# The dummy implementation just does a simple get + put
510
def copy_entry(path):
511
other.put(path, self.get(path))
513
return self._iterate_over(relpaths, copy_entry, pb, 'copy_to', expand=False)
762
515
def move(self, rel_from, rel_to):
763
516
"""Move the item at rel_from to the location at rel_to"""
764
path_from = self._remote_path(rel_from)
765
path_to = self._remote_path(rel_to)
766
self._rename_and_overwrite(path_from, path_to)
517
path_from = self._abspath(rel_from)
518
path_to = self._abspath(rel_to)
520
self._sftp.rename(path_from, path_to)
521
except (IOError, paramiko.SSHException), e:
522
self._translate_io_exception(e, path_from, ': unable to move to: %r' % path_to)
768
524
def delete(self, relpath):
769
525
"""Delete the item at relpath"""
770
path = self._remote_path(relpath)
526
path = self._abspath(relpath)
772
self._get_sftp().remove(path)
528
self._sftp.remove(path)
773
529
except (IOError, paramiko.SSHException), e:
774
530
self._translate_io_exception(e, path, ': unable to delete')
776
def external_url(self):
777
"""See bzrlib.transport.Transport.external_url."""
778
# the external path for SFTP is the base
781
532
def listable(self):
782
533
"""Return True if this store supports listing."""
862
578
# that we have taken the lock.
863
579
return SFTPLock(relpath, self)
865
def _sftp_open_exclusive(self, abspath, mode=None):
582
def _unparse_url(self, path=None):
585
path = urllib.quote(path)
586
if path.startswith('/'):
587
path = '/%2F' + path[1:]
590
netloc = urllib.quote(self._host)
591
if self._username is not None:
592
netloc = '%s@%s' % (urllib.quote(self._username), netloc)
593
if self._port is not None:
594
netloc = '%s:%d' % (netloc, self._port)
596
return urlparse.urlunparse(('sftp', netloc, path, '', '', ''))
598
def _split_url(self, url):
599
if isinstance(url, unicode):
600
url = url.encode('utf-8')
601
(scheme, netloc, path, params,
602
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
603
assert scheme == 'sftp'
604
username = password = host = port = None
606
username, host = netloc.split('@', 1)
608
username, password = username.split(':', 1)
609
password = urllib.unquote(password)
610
username = urllib.unquote(username)
615
host, port = host.rsplit(':', 1)
619
# TODO: Should this be ConnectionError?
620
raise TransportError('%s: invalid port number' % port)
621
host = urllib.unquote(host)
623
path = urllib.unquote(path)
625
# the initial slash should be removed from the path, and treated
626
# as a homedir relative path (the path begins with a double slash
627
# if it is absolute).
628
# see draft-ietf-secsh-scp-sftp-ssh-uri-03.txt
629
if path.startswith('/'):
632
return (username, password, host, port, path)
634
def _parse_url(self, url):
635
(self._username, self._password,
636
self._host, self._port, self._path) = self._split_url(url)
638
def _sftp_connect(self):
639
"""Connect to the remote sftp server.
640
After this, self._sftp should have a valid connection (or
641
we raise an TransportError 'could not connect').
643
TODO: Raise a more reasonable ConnectionFailed exception
645
global _connected_hosts
647
idx = (self._host, self._port, self._username)
649
self._sftp = _connected_hosts[idx]
654
vendor = _get_ssh_vendor()
656
sock = SFTPSubprocess(self._host, self._port, self._username)
657
self._sftp = SFTPClient(sock)
659
self._paramiko_connect()
661
_connected_hosts[idx] = self._sftp
663
def _paramiko_connect(self):
664
global SYSTEM_HOSTKEYS, BZR_HOSTKEYS
669
t = paramiko.Transport((self._host, self._port or 22))
671
except paramiko.SSHException, e:
672
raise ConnectionError('Unable to reach SSH host %s:%d' %
673
(self._host, self._port), e)
675
server_key = t.get_remote_server_key()
676
server_key_hex = paramiko.util.hexify(server_key.get_fingerprint())
677
keytype = server_key.get_name()
678
if SYSTEM_HOSTKEYS.has_key(self._host) and SYSTEM_HOSTKEYS[self._host].has_key(keytype):
679
our_server_key = SYSTEM_HOSTKEYS[self._host][keytype]
680
our_server_key_hex = paramiko.util.hexify(our_server_key.get_fingerprint())
681
elif BZR_HOSTKEYS.has_key(self._host) and BZR_HOSTKEYS[self._host].has_key(keytype):
682
our_server_key = BZR_HOSTKEYS[self._host][keytype]
683
our_server_key_hex = paramiko.util.hexify(our_server_key.get_fingerprint())
685
warning('Adding %s host key for %s: %s' % (keytype, self._host, server_key_hex))
686
if not BZR_HOSTKEYS.has_key(self._host):
687
BZR_HOSTKEYS[self._host] = {}
688
BZR_HOSTKEYS[self._host][keytype] = server_key
689
our_server_key = server_key
690
our_server_key_hex = paramiko.util.hexify(our_server_key.get_fingerprint())
692
if server_key != our_server_key:
693
filename1 = os.path.expanduser('~/.ssh/known_hosts')
694
filename2 = os.path.join(config_dir(), 'ssh_host_keys')
695
raise TransportError('Host keys for %s do not match! %s != %s' % \
696
(self._host, our_server_key_hex, server_key_hex),
697
['Try editing %s or %s' % (filename1, filename2)])
702
self._sftp = t.open_sftp_client()
703
except paramiko.SSHException, e:
704
raise ConnectionError('Unable to start sftp client %s:%d' %
705
(self._host, self._port), e)
707
def _sftp_auth(self, transport):
708
# paramiko requires a username, but it might be none if nothing was supplied
709
# use the local username, just in case.
710
# We don't override self._username, because if we aren't using paramiko,
711
# the username might be specified in ~/.ssh/config and we don't want to
712
# force it to something else
713
# Also, it would mess up the self.relpath() functionality
714
username = self._username or getpass.getuser()
716
# Paramiko tries to open a socket.AF_UNIX in order to connect
717
# to ssh-agent. That attribute doesn't exist on win32 (it does in cygwin)
718
# so we get an AttributeError exception. For now, just don't try to
719
# connect to an agent if we are on win32
720
if sys.platform != 'win32':
721
agent = paramiko.Agent()
722
for key in agent.get_keys():
723
mutter('Trying SSH agent key %s' % paramiko.util.hexify(key.get_fingerprint()))
725
transport.auth_publickey(username, key)
727
except paramiko.SSHException, e:
730
# okay, try finding id_rsa or id_dss? (posix only)
731
if self._try_pkey_auth(transport, paramiko.RSAKey, username, 'id_rsa'):
733
if self._try_pkey_auth(transport, paramiko.DSSKey, username, 'id_dsa'):
739
transport.auth_password(username, self._password)
741
except paramiko.SSHException, e:
744
# FIXME: Don't keep a password held in memory if you can help it
745
#self._password = None
747
# give up and ask for a password
748
password = bzrlib.ui.ui_factory.get_password(
749
prompt='SSH %(user)s@%(host)s password',
750
user=username, host=self._host)
752
transport.auth_password(username, password)
753
except paramiko.SSHException, e:
754
raise ConnectionError('Unable to authenticate to SSH host as %s@%s' %
755
(username, self._host), e)
757
def _try_pkey_auth(self, transport, pkey_class, username, filename):
758
filename = os.path.expanduser('~/.ssh/' + filename)
760
key = pkey_class.from_private_key_file(filename)
761
transport.auth_publickey(username, key)
763
except paramiko.PasswordRequiredException:
764
password = bzrlib.ui.ui_factory.get_password(
765
prompt='SSH %(filename)s password',
768
key = pkey_class.from_private_key_file(filename, password)
769
transport.auth_publickey(username, key)
771
except paramiko.SSHException:
772
mutter('SSH authentication via %s key failed.' % (os.path.basename(filename),))
773
except paramiko.SSHException:
774
mutter('SSH authentication via %s key failed.' % (os.path.basename(filename),))
779
def _sftp_open_exclusive(self, relpath):
866
780
"""Open a remote path exclusively.
868
782
SFTP supports O_EXCL (SFTP_FLAG_EXCL), which fails if
added
removed
bzrlib/transport/sftp.py
