def _parse_init_connect(self, line):
"""
Returns (host, port, httpversion) if line is a valid CONNECT line.
http://tools.ietf.org/html/draft-luotonen-web-proxy-tunneling-01 section 3.1
"""
v = self._parse_init(line)
if not v:
return None
method, url, httpversion = v
if method.upper() != 'CONNECT':
return None
try:
host, port = url.split(":")
except ValueError:
return None
try:
port = int(port)
except ValueError:
return None
if not utils.is_valid_port(port):
return None
if not utils.is_valid_host(host):
return None
return host, port, httpversion
def _parse_init_connect(self, line):
"""
Returns (host, port, httpversion) if line is a valid CONNECT line.
http://tools.ietf.org/html/draft-luotonen-web-proxy-tunneling-01 section 3.1
"""
v = self._parse_init(line)
if not v:
return None
method, url, httpversion = v
if method.upper() != 'CONNECT':
return None
try:
host, port = url.split(":")
except ValueError:
return None
try:
port = int(port)
except ValueError:
return None
if not utils.is_valid_port(port):
return None
if not utils.is_valid_host(host):
return None
return host, port, httpversion
def parse(url):
"""
URL-parsing function that checks that
- port is an integer 0-65535
- host is a valid IDNA-encoded hostname with no null-bytes
- path is valid ASCII
Args:
A URL (as bytes or as unicode)
Returns:
A (scheme, host, port, path) tuple
Raises:
ValueError, if the URL is not properly formatted.
"""
parsed = urllib.parse.urlparse(url)
if not parsed.hostname:
raise ValueError("No hostname given")
if isinstance(url, six.binary_type):
host = parsed.hostname
# this should not raise a ValueError,
# but we try to be very forgiving here and accept just everything.
# decode_parse_result(parsed, "ascii")
else:
host = parsed.hostname.encode("idna")
parsed = encode_parse_result(parsed, "ascii")
port = parsed.port
if not port:
port = 443 if parsed.scheme == b"https" else 80
full_path = urllib.parse.urlunparse(
(b"", b"", parsed.path, parsed.params, parsed.query, parsed.fragment)
)
if not full_path.startswith(b"/"):
full_path = b"/" + full_path
if not utils.is_valid_host(host):
raise ValueError("Invalid Host")
if not utils.is_valid_port(port):
raise ValueError("Invalid Port")
return parsed.scheme, host, port, full_path
def parse(url):
"""
URL-parsing function that checks that
- port is an integer 0-65535
- host is a valid IDNA-encoded hostname with no null-bytes
- path is valid ASCII
Args:
A URL (as bytes or as unicode)
Returns:
A (scheme, host, port, path) tuple
Raises:
ValueError, if the URL is not properly formatted.
"""
parsed = urllib.parse.urlparse(url)
if not parsed.hostname:
raise ValueError("No hostname given")
if isinstance(url, six.binary_type):
host = parsed.hostname
# this should not raise a ValueError,
# but we try to be very forgiving here and accept just everything.
# decode_parse_result(parsed, "ascii")
else:
host = parsed.hostname.encode("idna")
parsed = encode_parse_result(parsed, "ascii")
port = parsed.port
if not port:
port = 443 if parsed.scheme == b"https" else 80
full_path = urllib.parse.urlunparse(
(b"", b"", parsed.path, parsed.params, parsed.query, parsed.fragment)
)
if not full_path.startswith(b"/"):
full_path = b"/" + full_path
if not utils.is_valid_host(host):
raise ValueError("Invalid Host")
if not utils.is_valid_port(port):
raise ValueError("Invalid Port")
return parsed.scheme, host, port, full_path
def _parse_authority_form(hostport):
"""
Returns (host, port) if hostport is a valid authority-form host specification.
http://tools.ietf.org/html/draft-luotonen-web-proxy-tunneling-01 section 3.1
Raises:
ValueError, if the input is malformed
"""
try:
host, port = hostport.split(b":")
port = int(port)
if not utils.is_valid_host(host) or not utils.is_valid_port(port):
raise ValueError()
except ValueError:
raise exceptions.HttpSyntaxException("Invalid host specification: {}".format(hostport))
return host, port
