1
# Copyright (C) 2005 Canonical Ltd
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.
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.
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
"""Implementation of Transport over http.
19
from bzrlib.transport import Transport, register_transport
20
from bzrlib.errors import (TransportNotPossible, NoSuchFile,
21
NonRelativePath, TransportError)
23
from cStringIO import StringIO
27
from bzrlib.errors import BzrError, BzrCheckError
28
from bzrlib.branch import Branch
29
from bzrlib.trace import mutter
31
# velocitynet.com.au transparently proxies connections and thereby
32
# breaks keep-alive -- sucks!
37
mutter("get_url %s" % url)
38
url_f = urllib2.urlopen(url)
41
class HttpTransportError(TransportError):
44
class HttpTransport(Transport):
45
"""This is the transport agent for http:// access.
47
TODO: Implement pipelined versions of all of the *_multi() functions.
50
def __init__(self, base):
51
"""Set the base path where files will be stored."""
52
assert base.startswith('http://') or base.startswith('https://')
53
super(HttpTransport, self).__init__(base)
54
# In the future we might actually connect to the remote host
55
# rather than using get_url
56
# self._connection = None
57
(self._proto, self._host,
58
self._path, self._parameters,
59
self._query, self._fragment) = urlparse.urlparse(self.base)
61
def should_cache(self):
62
"""Return True if the data pulled across should be cached locally.
66
def clone(self, offset=None):
67
"""Return a new HttpTransport with root at self.base + offset
68
For now HttpTransport does not actually connect, so just return
69
a new HttpTransport object.
72
return HttpTransport(self.base)
74
return HttpTransport(self.abspath(offset))
76
def abspath(self, relpath):
77
"""Return the full url to the given relative path.
78
This can be supplied with a string or a list
80
if isinstance(relpath, basestring):
82
basepath = self._path.split('/')
83
if len(basepath) > 0 and basepath[-1] == '':
84
basepath = basepath[:-1]
89
# In most filesystems, a request for the parent
90
# of root, just returns root.
99
# Possibly, we could use urlparse.urljoin() here, but
100
# I'm concerned about when it chooses to strip the last
101
# portion of the path, and when it doesn't.
102
path = '/'.join(basepath)
103
return urlparse.urlunparse((self._proto,
104
self._host, path, '', '', ''))
106
def relpath(self, abspath):
107
if not abspath.startswith(self.base):
108
raise NonRelativePath('path %r is not under base URL %r'
109
% (abspath, self.base))
111
return abspath[pl:].lstrip('/')
113
def has(self, relpath):
114
"""Does the target location exist?
116
TODO: HttpTransport.has() should use a HEAD request,
117
not a full GET request.
119
TODO: This should be changed so that we don't use
120
urllib2 and get an exception, the code path would be
121
cleaner if we just do an http HEAD request, and parse
125
f = get_url(self.abspath(relpath))
126
# Without the read and then close()
127
# we tend to have busy sockets.
133
except urllib2.URLError:
136
if e.errno == errno.ENOENT:
138
raise HttpTransportError(orig_error=e)
140
def get(self, relpath, decode=False):
141
"""Get the file at the given relative path.
143
:param relpath: The relative path to the file
146
return get_url(self.abspath(relpath))
147
except (BzrError, urllib2.URLError, IOError), e:
148
raise NoSuchFile(msg = "Error retrieving %s"
149
% self.abspath(relpath),
152
raise HttpTransportError(orig_error=e)
154
def get_partial(self, relpath, start, length=None):
155
"""Get just part of a file.
157
:param relpath: Path to the file, relative to base
158
:param start: The starting position to read from
159
:param length: The length to read. A length of None indicates
160
read to the end of the file.
161
:return: A file-like object containing at least the specified bytes.
162
Some implementations may return objects which can be read
163
past this length, but this is not guaranteed.
165
# TODO: You can make specialized http requests for just
166
# a portion of the file. Figure out how to do that.
167
# For now, urllib2 returns files that cannot seek() so
168
# we just read bytes off the beginning, until we
169
# get to the point that we care about.
170
f = self.get(relpath)
171
# TODO: read in smaller chunks, in case things are
172
# buffered internally.
176
def put(self, relpath, f):
177
"""Copy the file-like or string object into the location.
179
:param relpath: Location to put the contents, relative to base.
180
:param f: File-like or string object.
182
raise TransportNotPossible('http PUT not supported')
184
def mkdir(self, relpath):
185
"""Create a directory at the given path."""
186
raise TransportNotPossible('http does not support mkdir()')
188
def append(self, relpath, f):
189
"""Append the text in the file-like object into the final
192
raise TransportNotPossible('http does not support append()')
194
def copy(self, rel_from, rel_to):
195
"""Copy the item at rel_from to the location at rel_to"""
196
raise TransportNotPossible('http does not support copy()')
198
def copy_to(self, relpaths, other, pb=None):
199
"""Copy a set of entries from self into another Transport.
201
:param relpaths: A list/generator of entries to be copied.
203
TODO: if other is LocalTransport, is it possible to
204
do better than put(get())?
206
# At this point HttpTransport might be able to check and see if
207
# the remote location is the same, and rather than download, and
208
# then upload, it could just issue a remote copy_this command.
209
if isinstance(other, HttpTransport):
210
raise TransportNotPossible('http cannot be the target of copy_to()')
212
return super(HttpTransport, self).copy_to(relpaths, other, pb=pb)
214
def move(self, rel_from, rel_to):
215
"""Move the item at rel_from to the location at rel_to"""
216
raise TransportNotPossible('http does not support move()')
218
def delete(self, relpath):
219
"""Delete the item at relpath"""
220
raise TransportNotPossible('http does not support delete()')
223
"""See Transport.listable."""
226
def stat(self, relpath):
227
"""Return the stat information for a file.
229
raise TransportNotPossible('http does not support stat()')
231
def lock_read(self, relpath):
232
"""Lock the given file for shared (read) access.
233
:return: A lock object, which should be passed to Transport.unlock()
235
# The old RemoteBranch ignore lock for reading, so we will
236
# continue that tradition and return a bogus lock object.
237
class BogusLock(object):
238
def __init__(self, path):
242
return BogusLock(relpath)
244
def lock_write(self, relpath):
245
"""Lock the given file for exclusive (write) access.
246
WARNING: many transports do not support this, so trying avoid using it
248
:return: A lock object, which should be passed to Transport.unlock()
250
raise TransportNotPossible('http does not support lock_write()')
252
register_transport('http://', HttpTransport)
253
register_transport('https://', HttpTransport)