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
TransportNotPossible, NoSuchFile, NonRelativePath, \
23
from cStringIO import StringIO
27
from bzrlib.errors import BzrError, BzrCheckError
28
from bzrlib.branch import Branch, BZR_BRANCH_FORMAT
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.
98
# Possibly, we could use urlparse.urljoin() here, but
99
# I'm concerned about when it chooses to strip the last
100
# portion of the path, and when it doesn't.
101
path = '/'.join(basepath)
102
return urlparse.urlunparse((self._proto,
103
self._host, path, '', '', ''))
105
def relpath(self, abspath):
106
if not abspath.startswith(self.base):
107
raise NonRelativePath('path %r is not under base URL %r'
108
% (abspath, self.base))
110
return abspath[pl:].lstrip('/')
112
def has(self, relpath):
113
"""Does the target location exist?
115
TODO: HttpTransport.has() should use a HEAD request,
116
not a full GET request.
118
TODO: This should be changed so that we don't use
119
urllib2 and get an exception, the code path would be
120
cleaner if we just do an http HEAD request, and parse
124
f = get_url(self.abspath(relpath))
125
# Without the read and then close()
126
# we tend to have busy sockets.
132
except urllib2.URLError:
135
if e.errno == errno.ENOENT:
137
raise HttpTransportError(orig_error=e)
139
def get(self, relpath, decode=False):
140
"""Get the file at the given relative path.
142
:param relpath: The relative path to the file
145
return get_url(self.abspath(relpath))
146
except (BzrError, urllib2.URLError, IOError), e:
147
raise NoSuchFile(orig_error=e)
149
raise HttpTransportError(orig_error=e)
151
def get_partial(self, relpath, start, length=None):
152
"""Get just part of a file.
154
:param relpath: Path to the file, relative to base
155
:param start: The starting position to read from
156
:param length: The length to read. A length of None indicates
157
read to the end of the file.
158
:return: A file-like object containing at least the specified bytes.
159
Some implementations may return objects which can be read
160
past this length, but this is not guaranteed.
162
# TODO: You can make specialized http requests for just
163
# a portion of the file. Figure out how to do that.
164
# For now, urllib2 returns files that cannot seek() so
165
# we just read bytes off the beginning, until we
166
# get to the point that we care about.
167
f = self.get(relpath)
168
# TODO: read in smaller chunks, in case things are
169
# buffered internally.
173
def put(self, relpath, f):
174
"""Copy the file-like or string object into the location.
176
:param relpath: Location to put the contents, relative to base.
177
:param f: File-like or string object.
179
raise TransportNotPossible('http PUT not supported')
181
def mkdir(self, relpath):
182
"""Create a directory at the given path."""
183
raise TransportNotPossible('http does not support mkdir()')
185
def append(self, relpath, f):
186
"""Append the text in the file-like object into the final
189
raise TransportNotPossible('http does not support append()')
191
def copy(self, rel_from, rel_to):
192
"""Copy the item at rel_from to the location at rel_to"""
193
raise TransportNotPossible('http does not support copy()')
195
def copy_to(self, relpaths, other, pb=None):
196
"""Copy a set of entries from self into another Transport.
198
:param relpaths: A list/generator of entries to be copied.
200
TODO: if other is LocalTransport, is it possible to
201
do better than put(get())?
203
# At this point HttpTransport might be able to check and see if
204
# the remote location is the same, and rather than download, and
205
# then upload, it could just issue a remote copy_this command.
206
if isinstance(other, HttpTransport):
207
raise TransportNotPossible('http cannot be the target of copy_to()')
209
return super(HttpTransport, self).copy_to(relpaths, other, pb=pb)
211
def move(self, rel_from, rel_to):
212
"""Move the item at rel_from to the location at rel_to"""
213
raise TransportNotPossible('http does not support move()')
215
def delete(self, relpath):
216
"""Delete the item at relpath"""
217
raise TransportNotPossible('http does not support delete()')
219
def list_dir(self, relpath):
220
"""Return a list of all files at the given location.
221
WARNING: many transports do not support this, so trying avoid using
222
it if at all possible.
224
raise TransportNotPossible('http does not support list_dir()')
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)