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"""
94
assert ranges_header.startswith('bytes=')
95
ranges_header = ranges_header[len('bytes='):]
96
for range_str in ranges_header.split(','):
97
range_match = self._range_regexp.match(range_str)
98
if range_match is not None:
99
ranges.append((int(range_match.group('start')),
100
int(range_match.group('end'))))
102
tail_match = self._tail_regexp.match(range_str)
103
if tail_match is not None:
104
tail = int(tail_match.group('tail'))
107
def send_range_content(self, file, start, length):
109
self.wfile.write(file.read(length))
111
def get_single_range(self, file, file_size, start, end):
112
self.send_response(206)
113
length = end - start + 1
114
self.send_header('Accept-Ranges', 'bytes')
115
self.send_header("Content-Length", "%d" % length)
117
self.send_header("Content-Type", 'application/octet-stream')
118
self.send_header("Content-Range", "bytes %d-%d/%d" % (start,
122
self.send_range_content(file, start, length)
124
def get_multiple_ranges(self, file, file_size, ranges):
125
self.send_response(206)
126
self.send_header('Accept-Ranges', 'bytes')
127
boundary = "%d" % random.randint(0,0x7FFFFFFF)
128
self.send_header("Content-Type",
129
"multipart/byteranges; boundary=%s" % boundary)
131
for (start, end) in ranges:
132
self.wfile.write("--%s\r\n" % boundary)
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, end - start + 1)
139
self.wfile.write("--%s\r\n" % boundary)
143
"""Serve a GET request.
145
Handles the Range header.
148
path = self.translate_path(self.path)
149
ranges_header_value = self.headers.get('Range')
150
if ranges_header_value is None or os.path.isdir(path):
151
# Let the mother class handle most cases
152
return SimpleHTTPRequestHandler.do_GET(self)
155
# Always read in binary mode. Opening files in text
156
# mode may cause newline translations, making the
157
# actual size of the content transmitted *less* than
158
# the content-length!
159
file = open(path, 'rb')
161
self.send_error(404, "File not found")
164
file_size = os.fstat(file.fileno())[6]
165
tail, ranges = self.parse_ranges(ranges_header_value)
166
# Normalize tail into ranges
168
ranges.append((file_size - tail, file_size))
174
for (start, end) in ranges:
175
if start >= file_size or end >= file_size:
179
# RFC2616 14-16 says that invalid Range headers
180
# should be ignored and in that case, the whole file
181
# should be returned as if no Range header was
183
file.close() # Will be reopened by the following call
184
return SimpleHTTPRequestHandler.do_GET(self)
187
(start, end) = ranges[0]
188
self.get_single_range(file, file_size, start, end)
190
self.get_multiple_ranges(file, file_size, ranges)
193
if sys.platform == 'win32':
194
# On win32 you cannot access non-ascii filenames without
195
# decoding them into unicode first.
196
# However, under Linux, you can access bytestream paths
197
# without any problems. If this function was always active
198
# it would probably break tests when LANG=C was set
199
def translate_path(self, path):
200
"""Translate a /-separated PATH to the local filename syntax.
202
For bzr, all url paths are considered to be utf8 paths.
203
On Linux, you can access these paths directly over the bytestream
204
request, but on win32, you must decode them, and access them
207
# abandon query parameters
208
path = urlparse.urlparse(path)[2]
209
path = posixpath.normpath(urllib.unquote(path))
210
path = path.decode('utf-8')
211
words = path.split('/')
212
words = filter(None, words)
215
drive, word = os.path.splitdrive(word)
216
head, word = os.path.split(word)
217
if word in (os.curdir, os.pardir): continue
218
path = os.path.join(path, word)
222
class TestingHTTPServer(BaseHTTPServer.HTTPServer):
223
def __init__(self, server_address, RequestHandlerClass, test_case):
224
BaseHTTPServer.HTTPServer.__init__(self, server_address,
226
self.test_case = test_case
229
class HttpServer(Server):
230
"""A test server for http transports.
232
Subclasses can provide a specific request handler.
235
# used to form the url that connects to this server
236
_url_protocol = 'http'
238
# Subclasses can provide a specific request handler
239
def __init__(self, request_handler=TestingHTTPRequestHandler):
240
Server.__init__(self)
241
self.request_handler = request_handler
243
def _get_httpd(self):
244
return TestingHTTPServer(('localhost', 0),
245
self.request_handler,
248
def _http_start(self):
250
httpd = self._get_httpd()
251
host, port = httpd.socket.getsockname()
252
self._http_base_url = '%s://localhost:%s/' % (self._url_protocol, port)
253
self._http_starting.release()
254
httpd.socket.settimeout(0.1)
256
while self._http_running:
258
httpd.handle_request()
259
except socket.timeout:
262
def _get_remote_url(self, path):
263
path_parts = path.split(os.path.sep)
264
if os.path.isabs(path):
265
if path_parts[:len(self._local_path_parts)] != \
266
self._local_path_parts:
267
raise BadWebserverPath(path, self.test_dir)
268
remote_path = '/'.join(path_parts[len(self._local_path_parts):])
270
remote_path = '/'.join(path_parts)
272
self._http_starting.acquire()
273
self._http_starting.release()
274
return self._http_base_url + remote_path
276
def log(self, format, *args):
277
"""Capture Server log output."""
278
self.logs.append(format % args)
281
"""See bzrlib.transport.Server.setUp."""
282
self._home_dir = os.getcwdu()
283
self._local_path_parts = self._home_dir.split(os.path.sep)
284
self._http_starting = threading.Lock()
285
self._http_starting.acquire()
286
self._http_running = True
287
self._http_base_url = None
288
self._http_thread = threading.Thread(target=self._http_start)
289
self._http_thread.setDaemon(True)
290
self._http_thread.start()
291
self._http_proxy = os.environ.get("http_proxy")
292
if self._http_proxy is not None:
293
del os.environ["http_proxy"]
297
"""See bzrlib.transport.Server.tearDown."""
298
self._http_running = False
299
self._http_thread.join()
300
if self._http_proxy is not None:
302
os.environ["http_proxy"] = self._http_proxy
305
"""See bzrlib.transport.Server.get_url."""
306
return self._get_remote_url(self._home_dir)
308
def get_bogus_url(self):
309
"""See bzrlib.transport.Server.get_bogus_url."""
310
# this is chosen to try to prevent trouble with proxies, weird dns,
312
return 'http://127.0.0.1:1/'
315
class HttpServer_urllib(HttpServer):
316
"""Subclass of HttpServer that gives http+urllib urls.
318
This is for use in testing: connections to this server will always go
319
through urllib where possible.
322
# urls returned by this server should require the urllib client impl
323
_url_protocol = 'http+urllib'
326
class HttpServer_PyCurl(HttpServer):
327
"""Subclass of HttpServer that gives http+pycurl urls.
329
This is for use in testing: connections to this server will always go
330
through pycurl where possible.
333
# We don't care about checking the pycurl availability as
334
# this server will be required only when pycurl is present
336
# urls returned by this server should require the pycurl client impl
337
_url_protocol = 'http+pycurl'