1
# Copyright (C) 2006 Michael Ellerman
2
# modified by John Arbash Meinel (Canonical Ltd)
4
# This program is free software; you can redistribute it and/or modify
5
# it under the terms of the GNU General Public License as published by
6
# the Free Software Foundation; either version 2 of the License, or
7
# (at your option) any later version.
9
# This program is distributed in the hope that it will be useful,
10
# but WITHOUT ANY WARRANTY; without even the implied warranty of
11
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12
# GNU General Public License for more details.
14
# You should have received a copy of the GNU General Public License
15
# along with this program; if not, write to the Free Software
16
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
18
"""Handlers for HTTP Responses.
20
The purpose of these classes is to provide a uniform interface for clients
21
to standard HTTP responses, single range responses and multipart range
26
from bisect import bisect
27
from cStringIO import StringIO
30
from bzrlib import errors
31
from bzrlib.trace import mutter
34
class ResponseRange(object):
35
"""A range in a RangeFile-object."""
37
__slots__ = ['_ent_start', '_ent_end', '_data_start']
39
def __init__(self, ent_start, ent_end, data_start):
40
self._ent_start = ent_start
41
self._ent_end = ent_end
42
self._data_start = data_start
44
def __cmp__(self, other):
45
"""Compare this to other.
47
We need this both for sorting, and so that we can
48
bisect the list of ranges.
50
if isinstance(other, int):
51
# Later on we bisect for a starting point
52
# so we allow comparing against a single integer
53
return cmp(self._ent_start, other)
55
return cmp((self._ent_start, self._ent_end, self._data_start),
56
(other._ent_start, other._ent_end, other._data_start))
59
return "%s(%s-%s,%s)" % (self.__class__.__name__,
60
self._ent_start, self._ent_end,
64
class RangeFile(object):
65
"""File-like object that allow access to partial available data.
67
Specified by a set of ranges.
70
def __init__(self, path, input_file):
75
self._data = input_file.read()
77
def _add_range(self, ent_start, ent_end, data_start):
78
"""Add an entity range.
80
:param ent_start: Start offset of entity
81
:param ent_end: End offset of entity (inclusive)
82
:param data_start: Start offset of data in data stream.
84
self._ranges.append(ResponseRange(ent_start, ent_end, data_start))
85
self._len = max(self._len, ent_end)
87
def _finish_ranges(self):
91
"""Read size bytes from the current position in the file.
93
Reading across ranges is not supported.
95
# find the last range which has a start <= pos
96
i = bisect(self._ranges, self._pos) - 1
98
if i < 0 or self._pos > self._ranges[i]._ent_end:
99
raise errors.InvalidRange(self._path, self._pos)
103
# mutter('found range %s %s for pos %s', i, self._ranges[i], self._pos)
105
if (self._pos + size - 1) > r._ent_end:
106
raise errors.InvalidRange(self._path, self._pos)
108
start = r._data_start + (self._pos - r._ent_start)
110
# mutter("range read %d bytes at %d == %d-%d", size, self._pos,
112
self._pos += (end-start)
113
return self._data[start:end]
115
def seek(self, offset, whence=0):
121
self._pos = self._len + offset
123
raise ValueError("Invalid value %s for whence." % whence)
132
class HttpRangeResponse(RangeFile):
133
"""A single-range HTTP response."""
135
# TODO: jam 20060706 Consider compiling these regexes on demand
136
_CONTENT_RANGE_RE = re.compile(
137
'\s*([^\s]+)\s+([0-9]+)-([0-9]+)/([0-9]+)\s*$')
139
def __init__(self, path, content_range, input_file):
140
# mutter("parsing 206 non-multipart response for %s", path)
141
RangeFile.__init__(self, path, input_file)
142
start, end = self._parse_range(content_range, path)
143
self._add_range(start, end, 0)
144
self._finish_ranges()
147
def _parse_range(range, path='<unknown>'):
148
"""Parse an http Content-range header and return start + end
150
:param range: The value for Content-range
151
:param path: Provide to give better error messages.
152
:return: (start, end) A tuple of integers
154
match = HttpRangeResponse._CONTENT_RANGE_RE.match(range)
156
raise errors.InvalidHttpRange(path, range,
157
"Invalid Content-range")
159
rtype, start, end, total = match.groups()
162
raise errors.InvalidHttpRange(path, range,
163
"Unsupported range type '%s'" % (rtype,))
168
except ValueError, e:
169
raise errors.InvalidHttpRange(path, range, str(e))
174
class HttpMultipartRangeResponse(RangeFile):
175
"""A multi-range HTTP response."""
177
_CONTENT_TYPE_RE = re.compile(
178
'^\s*multipart/byteranges\s*;\s*boundary\s*=\s*(.*?)\s*$')
180
# Start with --<boundary>\r\n
181
# and ignore all headers ending in \r\n
182
# except for content-range:
183
# and find the two trailing \r\n separators
184
# indicating the start of the text
185
# TODO: jam 20060706 This requires exact conformance
186
# to the spec, we probably could relax the requirement
187
# of \r\n, and use something more like (\r?\n)
189
"^--%s(?:\r\n(?:(?:content-range:([^\r]+))|[^\r]+))+\r\n\r\n")
191
def __init__(self, path, content_type, input_file):
192
# mutter("parsing 206 multipart response for %s", path)
193
# TODO: jam 20060706 Is it valid to initialize a
194
# grandparent without initializing parent?
195
RangeFile.__init__(self, path, input_file)
197
self.boundary_regex = self._parse_boundary(content_type, path)
199
for match in self.boundary_regex.finditer(self._data):
200
ent_start, ent_end = HttpRangeResponse._parse_range(match.group(1),
202
self._add_range(ent_start, ent_end, match.end())
204
self._finish_ranges()
207
def _parse_boundary(ctype, path='<unknown>'):
208
"""Parse the Content-type field.
210
This expects a multipart Content-type, and returns a
211
regex which is capable of finding the boundaries
212
in the multipart data.
214
match = HttpMultipartRangeResponse._CONTENT_TYPE_RE.match(ctype)
216
raise errors.InvalidHttpContentType(path, ctype,
217
"Expected multipart/byteranges with boundary")
219
boundary = match.group(1)
220
# mutter('multipart boundary is %s', boundary)
221
pattern = HttpMultipartRangeResponse._BOUNDARY_PATT
222
return re.compile(pattern % re.escape(boundary),
223
re.IGNORECASE | re.MULTILINE)
226
def _is_multipart(content_type):
227
return content_type.startswith('multipart/byteranges;')
230
def handle_response(url, code, headers, data):
231
"""Interpret the code & headers and return a HTTP response.
233
This is a factory method which returns an appropriate HTTP response
234
based on the code & headers it's given.
236
:param url: The url being processed. Mostly for error reporting
237
:param code: The integer HTTP response code
238
:param headers: A dict-like object that contains the HTTP response headers
239
:param data: A file-like object that can be read() to get the
241
:return: A file-like object that can seek()+read() the
242
ranges indicated by the headers.
247
content_type = headers['Content-Type']
249
raise errors.InvalidHttpContentType(url, '',
250
msg='Missing Content-Type')
252
if _is_multipart(content_type):
253
# Full fledged multipart response
254
return HttpMultipartRangeResponse(url, content_type, data)
256
# A response to a range request, but not multipart
258
content_range = headers['Content-Range']
260
raise errors.InvalidHttpResponse(url,
261
'Missing the Content-Range header in a 206 range response')
262
return HttpRangeResponse(url, content_range, data)
264
# A regular non-range response, unfortunately the result from
265
# urllib doesn't support seek, so we wrap it in a StringIO
266
tell = getattr(data, 'tell', None)
268
return StringIO(data.read())
271
raise errors.NoSuchFile(url)
273
# TODO: jam 20060713 Properly handle redirects (302 Found, etc)
274
# The '_get' code says to follow redirects, we probably
275
# should actually handle the return values
277
raise errors.InvalidHttpResponse(url, "Unknown response code %s"
added
removed
bzrlib/transport/http/response.py
