1
# Copyright (C) 2006 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
20
from SimpleHTTPServer import SimpleHTTPRequestHandler
31
from bzrlib.transport import Server
34
class WebserverNotAvailable(Exception):
38
class BadWebserverPath(ValueError):
40
return 'path %s is not in %s' % self.args
43
class TestingHTTPRequestHandler(SimpleHTTPRequestHandler):
45
def log_message(self, format, *args):
46
self.server.test_case.log('webserver - %s - - [%s] %s "%s" "%s"',
47
self.address_string(),
48
self.log_date_time_string(),
50
self.headers.get('referer', '-'),
51
self.headers.get('user-agent', '-'))
53
def handle_one_request(self):
54
"""Handle a single HTTP request.
56
You normally don't need to override this method; see the class
57
__doc__ string for information on how to handle specific HTTP
58
commands such as GET and POST.
61
for i in xrange(1,11): # Don't try more than 10 times
63
self.raw_requestline = self.rfile.readline()
64
except socket.error, e:
65
if e.args[0] in (errno.EAGAIN, errno.EWOULDBLOCK):
66
# omitted for now because some tests look at the log of
67
# the server and expect to see no errors. see recent
68
# email thread. -- mbp 20051021.
69
## self.log_message('EAGAIN (%d) while reading from raw_requestline' % i)
75
if not self.raw_requestline:
76
self.close_connection = 1
78
if not self.parse_request(): # An error code has been sent, just exit
80
mname = 'do_' + self.command
81
if getattr(self, mname, None) is None:
82
self.send_error(501, "Unsupported method (%r)" % self.command)
84
method = getattr(self, mname)
87
_range_regexp = re.compile(r'^(?P<start>\d+)-(?P<end>\d+)$')
88
_tail_regexp = re.compile(r'^-(?P<tail>\d+)$')
90
def parse_ranges(self, ranges_header):
91
"""Parse the range header value and returns ranges and tail.
93
RFC2616 14.35 says that syntactically invalid range
94
specifiers MUST be ignored. In that case, we return 0 for
95
tail and [] for ranges.
99
if not ranges_header.startswith('bytes='):
100
# Syntactically invalid header
103
ranges_header = ranges_header[len('bytes='):]
104
for range_str in ranges_header.split(','):
105
# FIXME: RFC2616 says end is optional and default to file_size
106
range_match = self._range_regexp.match(range_str)
107
if range_match is not None:
108
start = int(range_match.group('start'))
109
end = int(range_match.group('end'))
111
# Syntactically invalid range
113
ranges.append((start, end))
115
tail_match = self._tail_regexp.match(range_str)
116
if tail_match is not None:
117
tail = int(tail_match.group('tail'))
119
# Syntactically invalid range
123
def send_range_content(self, file, start, length):
125
self.wfile.write(file.read(length))
127
def get_single_range(self, file, file_size, start, end):
128
self.send_response(206)
129
length = end - start + 1
130
self.send_header('Accept-Ranges', 'bytes')
131
self.send_header("Content-Length", "%d" % length)
133
self.send_header("Content-Type", 'application/octet-stream')
134
self.send_header("Content-Range", "bytes %d-%d/%d" % (start,
138
self.send_range_content(file, start, length)
140
def get_multiple_ranges(self, file, file_size, ranges):
141
self.send_response(206)
142
self.send_header('Accept-Ranges', 'bytes')
143
boundary = "%d" % random.randint(0,0x7FFFFFFF)
144
self.send_header("Content-Type",
145
"multipart/byteranges; boundary=%s" % boundary)
147
for (start, end) in ranges:
148
self.wfile.write("--%s\r\n" % boundary)
149
self.send_header("Content-type", 'application/octet-stream')
150
self.send_header("Content-Range", "bytes %d-%d/%d" % (start,
154
self.send_range_content(file, start, end - start + 1)
155
self.wfile.write("--%s\r\n" % boundary)
159
"""Serve a GET request.
161
Handles the Range header.
164
path = self.translate_path(self.path)
165
ranges_header_value = self.headers.get('Range')
166
if ranges_header_value is None or os.path.isdir(path):
167
# Let the mother class handle most cases
168
return SimpleHTTPRequestHandler.do_GET(self)
171
# Always read in binary mode. Opening files in text
172
# mode may cause newline translations, making the
173
# actual size of the content transmitted *less* than
174
# the content-length!
175
file = open(path, 'rb')
177
self.send_error(404, "File not found")
180
file_size = os.fstat(file.fileno())[6]
181
tail, ranges = self.parse_ranges(ranges_header_value)
182
# Normalize tail into ranges
184
ranges.append((file_size - tail, file_size))
186
self._satisfiable_ranges = True
188
self._satisfiable_ranges = False
190
def check_range(range_specifier):
191
start, end = range_specifier
192
# RFC2616 14.35, ranges are invalid if start >= file_size
193
if start >= file_size:
194
self._satisfiable_ranges = False # Side-effect !
196
# RFC2616 14.35, end values should be truncated
197
# to file_size -1 if they exceed it
198
end = min(end, file_size - 1)
201
ranges = map(check_range, ranges)
203
if not self._satisfiable_ranges:
204
# RFC2616 14.16 and 14.35 says that when a server
205
# encounters unsatisfiable range specifiers, it
206
# SHOULD return a 416.
208
# FIXME: We SHOULD send a Content-Range header too,
209
# but the implementation of send_error does not
210
# allows that. So far.
211
self.send_error(416, "Requested range not satisfiable")
215
(start, end) = ranges[0]
216
self.get_single_range(file, file_size, start, end)
218
self.get_multiple_ranges(file, file_size, ranges)
221
if sys.platform == 'win32':
222
# On win32 you cannot access non-ascii filenames without
223
# decoding them into unicode first.
224
# However, under Linux, you can access bytestream paths
225
# without any problems. If this function was always active
226
# it would probably break tests when LANG=C was set
227
def translate_path(self, path):
228
"""Translate a /-separated PATH to the local filename syntax.
230
For bzr, all url paths are considered to be utf8 paths.
231
On Linux, you can access these paths directly over the bytestream
232
request, but on win32, you must decode them, and access them
235
# abandon query parameters
236
path = urlparse.urlparse(path)[2]
237
path = posixpath.normpath(urllib.unquote(path))
238
path = path.decode('utf-8')
239
words = path.split('/')
240
words = filter(None, words)
243
drive, word = os.path.splitdrive(word)
244
head, word = os.path.split(word)
245
if word in (os.curdir, os.pardir): continue
246
path = os.path.join(path, word)
250
class TestingHTTPServer(BaseHTTPServer.HTTPServer):
251
def __init__(self, server_address, RequestHandlerClass, test_case):
252
BaseHTTPServer.HTTPServer.__init__(self, server_address,
254
self.test_case = test_case
257
class HttpServer(Server):
258
"""A test server for http transports.
260
Subclasses can provide a specific request handler.
263
# used to form the url that connects to this server
264
_url_protocol = 'http'
266
# Subclasses can provide a specific request handler
267
def __init__(self, request_handler=TestingHTTPRequestHandler):
268
Server.__init__(self)
269
self.request_handler = request_handler
271
def _get_httpd(self):
272
return TestingHTTPServer(('localhost', 0),
273
self.request_handler,
276
def _http_start(self):
278
httpd = self._get_httpd()
279
host, self.port = httpd.socket.getsockname()
280
self._http_base_url = '%s://localhost:%s/' % (self._url_protocol,
282
self._http_starting.release()
283
httpd.socket.settimeout(0.1)
285
while self._http_running:
287
httpd.handle_request()
288
except socket.timeout:
291
def _get_remote_url(self, path):
292
path_parts = path.split(os.path.sep)
293
if os.path.isabs(path):
294
if path_parts[:len(self._local_path_parts)] != \
295
self._local_path_parts:
296
raise BadWebserverPath(path, self.test_dir)
297
remote_path = '/'.join(path_parts[len(self._local_path_parts):])
299
remote_path = '/'.join(path_parts)
301
return self._http_base_url + remote_path
303
def log(self, format, *args):
304
"""Capture Server log output."""
305
self.logs.append(format % args)
308
"""See bzrlib.transport.Server.setUp."""
309
self._home_dir = os.getcwdu()
310
self._local_path_parts = self._home_dir.split(os.path.sep)
311
self._http_starting = threading.Lock()
312
self._http_starting.acquire()
313
self._http_running = True
314
self._http_base_url = None
315
self._http_thread = threading.Thread(target=self._http_start)
316
self._http_thread.setDaemon(True)
317
self._http_thread.start()
318
# Wait for the server thread to start (i.e release the lock)
319
self._http_starting.acquire()
320
self._http_starting.release()
324
"""See bzrlib.transport.Server.tearDown."""
325
self._http_running = False
326
self._http_thread.join()
329
"""See bzrlib.transport.Server.get_url."""
330
return self._get_remote_url(self._home_dir)
332
def get_bogus_url(self):
333
"""See bzrlib.transport.Server.get_bogus_url."""
334
# this is chosen to try to prevent trouble with proxies, weird dns,
336
return 'http://127.0.0.1:1/'
339
class HttpServer_urllib(HttpServer):
340
"""Subclass of HttpServer that gives http+urllib urls.
342
This is for use in testing: connections to this server will always go
343
through urllib where possible.
346
# urls returned by this server should require the urllib client impl
347
_url_protocol = 'http+urllib'
350
class HttpServer_PyCurl(HttpServer):
351
"""Subclass of HttpServer that gives http+pycurl urls.
353
This is for use in testing: connections to this server will always go
354
through pycurl where possible.
357
# We don't care about checking the pycurl availability as
358
# this server will be required only when pycurl is present
360
# urls returned by this server should require the pycurl client impl
361
_url_protocol = 'http+pycurl'
added
removed
bzrlib/tests/HttpServer.py
