class Request(object):
"""An HTTP request.
This object represents the metadata of an HTTP request message;
that is, it contains attributes which describe the environment
in which the request URL, headers, and body were sent (if you
want tools to interpret the headers and body, those are elsewhere,
mostly in Tools). This 'metadata' consists of socket data,
transport characteristics, and the Request-Line. This object
also contains data regarding the configuration in effect for
the given URL, and the execution plan for generating a response.
"""
prev = None
"""
The previous Request object (if any). This should be None
unless we are processing an InternalRedirect."""
# Conversation/connection attributes
local = httputil.Host('127.0.0.1', 80)
'An httputil.Host(ip, port, hostname) object for the server socket.'
remote = httputil.Host('127.0.0.1', 1111)
'An httputil.Host(ip, port, hostname) object for the client socket.'
scheme = 'http'
"""
The protocol used between client and server. In most cases,
this will be either 'http' or 'https'."""
server_protocol = 'HTTP/1.1'
"""
The HTTP version for which the HTTP server is at least
conditionally compliant."""
base = ''
"""The (scheme://host) portion of the requested URL.
In some cases (e.g. when proxying via mod_rewrite), this may contain
path segments which cherrypy.url uses when constructing url's, but
which otherwise are ignored by CherryPy. Regardless, this value
MUST NOT end in a slash."""
# Request-Line attributes
request_line = ''
"""
The complete Request-Line received from the client. This is a
single string consisting of the request method, URI, and protocol
version (joined by spaces). Any final CRLF is removed."""
method = 'GET'
"""
Indicates the HTTP method to be performed on the resource identified
by the Request-URI. Common methods include GET, HEAD, POST, PUT, and
DELETE. CherryPy allows any extension method; however, various HTTP
servers and gateways may restrict the set of allowable methods.
CherryPy applications SHOULD restrict the set (on a per-URI basis)."""
query_string = ''
"""
The query component of the Request-URI, a string of information to be
interpreted by the resource. The query portion of a URI follows the
path component, and is separated by a '?'. For example, the URI
'http://www.cherrypy.org/wiki?a=3&b=4' has the query component,
'a=3&b=4'."""
query_string_encoding = 'utf8'
"""
The encoding expected for query string arguments after % HEX HEX decoding).
If a query string is provided that cannot be decoded with this encoding,
404 is raised (since technically it's a different URI). If you want
arbitrary encodings to not error, set this to 'Latin-1'; you can then
encode back to bytes and re-decode to whatever encoding you like later.
"""
protocol = (1, 1)
"""The HTTP protocol version corresponding to the set
of features which should be allowed in the response. If BOTH
the client's request message AND the server's level of HTTP
compliance is HTTP/1.1, this attribute will be the tuple (1, 1).
If either is 1.0, this attribute will be the tuple (1, 0).
Lower HTTP protocol versions are not explicitly supported."""
params = {}
"""
A dict which combines query string (GET) and request entity (POST)
variables. This is populated in two stages: GET params are added
before the 'on_start_resource' hook, and POST params are added
between the 'before_request_body' and 'before_handler' hooks."""
# Message attributes
header_list = []
"""
A list of the HTTP request headers as (name, value) tuples.
In general, you should use request.headers (a dict) instead."""
headers = httputil.HeaderMap()
"""
A dict-like object containing the request headers. Keys are header
names (in Title-Case format); however, you may get and set them in
a case-insensitive manner. That is, headers['Content-Type'] and
headers['content-type'] refer to the same value. Values are header
values (decoded according to :rfc:`2047` if necessary). See also:
httputil.HeaderMap, httputil.HeaderElement."""
cookie = SimpleCookie()
"""See help(Cookie)."""
rfile = None
"""
If the request included an entity (body), it will be available
as a stream in this attribute. However, the rfile will normally
be read for you between the 'before_request_body' hook and the
'before_handler' hook, and the resulting string is placed into
either request.params or the request.body attribute.
You may disable the automatic consumption of the rfile by setting
request.process_request_body to False, either in config for the desired
path, or in an 'on_start_resource' or 'before_request_body' hook.
WARNING: In almost every case, you should not attempt to read from the
rfile stream after CherryPy's automatic mechanism has read it. If you
turn off the automatic parsing of rfile, you should read exactly the
number of bytes specified in request.headers['Content-Length'].
Ignoring either of these warnings may result in a hung request thread
or in corruption of the next (pipelined) request.
"""
process_request_body = True
"""
If True, the rfile (if any) is automatically read and parsed,
and the result placed into request.params or request.body."""
methods_with_bodies = ('POST', 'PUT', 'PATCH')
"""
A sequence of HTTP methods for which CherryPy will automatically
attempt to read a body from the rfile. If you are going to change
this property, modify it on the configuration (recommended)
or on the "hook point" `on_start_resource`.
"""
body = None
"""
If the request Content-Type is 'application/x-www-form-urlencoded'
or multipart, this will be None. Otherwise, this will be an instance
of :class:`RequestBody<cherrypy._cpreqbody.RequestBody>` (which you
can .read()); this value is set between the 'before_request_body' and
'before_handler' hooks (assuming that process_request_body is True)."""
# Dispatch attributes
dispatch = cherrypy.dispatch.Dispatcher()
"""
The object which looks up the 'page handler' callable and collects
config for the current request based on the path_info, other
request attributes, and the application architecture. The core
calls the dispatcher as early as possible, passing it a 'path_info'
argument.
The default dispatcher discovers the page handler by matching path_info
to a hierarchical arrangement of objects, starting at request.app.root.
See help(cherrypy.dispatch) for more information."""
script_name = ''
"""
The 'mount point' of the application which is handling this request.
This attribute MUST NOT end in a slash. If the script_name refers to
the root of the URI, it MUST be an empty string (not "/").
"""
path_info = '/'
"""
The 'relative path' portion of the Request-URI. This is relative
to the script_name ('mount point') of the application which is
handling this request."""
login = None
"""
When authentication is used during the request processing this is
set to 'False' if it failed and to the 'username' value if it succeeded.
The default 'None' implies that no authentication happened."""
# Note that cherrypy.url uses "if request.app:" to determine whether
# the call is during a real HTTP request or not. So leave this None.
app = None
"""The cherrypy.Application object which is handling this request."""
handler = None
"""
The function, method, or other callable which CherryPy will call to
produce the response. The discovery of the handler and the arguments
it will receive are determined by the request.dispatch object.
By default, the handler is discovered by walking a tree of objects
starting at request.app.root, and is then passed all HTTP params
(from the query string and POST body) as keyword arguments."""
toolmaps = {}
"""
A nested dict of all Toolboxes and Tools in effect for this request,
of the form: {Toolbox.namespace: {Tool.name: config dict}}."""
config = None
"""
A flat dict of all configuration entries which apply to the
current request. These entries are collected from global config,
application config (based on request.path_info), and from handler
config (exactly how is governed by the request.dispatch object in
effect for this request; by default, handler config can be attached
anywhere in the tree between request.app.root and the final handler,
and inherits downward)."""
is_index = None
"""
This will be True if the current request is mapped to an 'index'
resource handler (also, a 'default' handler if path_info ends with
a slash). The value may be used to automatically redirect the
user-agent to a 'more canonical' URL which either adds or removes
the trailing slash. See cherrypy.tools.trailing_slash."""
hooks = HookMap(hookpoints)
"""
A HookMap (dict-like object) of the form: {hookpoint: [hook, ...]}.
Each key is a str naming the hook point, and each value is a list
of hooks which will be called at that hook point during this request.
The list of hooks is generally populated as early as possible (mostly
from Tools specified in config), but may be extended at any time.
See also: _cprequest.Hook, _cprequest.HookMap, and cherrypy.tools."""
error_response = cherrypy.HTTPError(500).set_response
"""
The no-arg callable which will handle unexpected, untrapped errors
during request processing. This is not used for expected exceptions
(like NotFound, HTTPError, or HTTPRedirect) which are raised in
response to expected conditions (those should be customized either
via request.error_page or by overriding HTTPError.set_response).
By default, error_response uses HTTPError(500) to return a generic
error response to the user-agent."""
error_page = {}
"""
A dict of {error code: response filename or callable} pairs.
The error code must be an int representing a given HTTP error code,
or the string 'default', which will be used if no matching entry
is found for a given numeric code.
If a filename is provided, the file should contain a Python string-
formatting template, and can expect by default to receive format
values with the mapping keys %(status)s, %(message)s, %(traceback)s,
and %(version)s. The set of format mappings can be extended by
overriding HTTPError.set_response.
If a callable is provided, it will be called by default with keyword
arguments 'status', 'message', 'traceback', and 'version', as for a
string-formatting template. The callable must return a string or
iterable of strings which will be set to response.body. It may also
override headers or perform any other processing.
If no entry is given for an error code, and no 'default' entry exists,
a default template will be used.
"""
show_tracebacks = True
"""
If True, unexpected errors encountered during request processing will
include a traceback in the response body."""
show_mismatched_params = True
"""
If True, mismatched parameters encountered during PageHandler invocation
processing will be included in the response body."""
throws = (KeyboardInterrupt, SystemExit, cherrypy.InternalRedirect)
"""The sequence of exceptions which Request.run does not trap."""
throw_errors = False
"""
If True, Request.run will not trap any errors (except HTTPRedirect and
HTTPError, which are more properly called 'exceptions', not errors)."""
closed = False
"""True once the close method has been called, False otherwise."""
stage = None
"""
A string containing the stage reached in the request-handling process.
This is useful when debugging a live server with hung requests."""
unique_id = None
"""A lazy object generating and memorizing UUID4 on ``str()`` render."""
namespaces = reprconf.NamespaceSet(
**{'hooks': hooks_namespace,
'request': request_namespace,
'response': response_namespace,
'error_page': error_page_namespace,
'tools': cherrypy.tools,
})
def __init__(self, local_host, remote_host, scheme='http',
server_protocol='HTTP/1.1'):
"""Populate a new Request object.
local_host should be an httputil.Host object with the server info.
remote_host should be an httputil.Host object with the client info.
scheme should be a string, either "http" or "https".
"""
self.local = local_host
self.remote = remote_host
self.scheme = scheme
self.server_protocol = server_protocol
self.closed = False
# Put a *copy* of the class error_page into self.
self.error_page = self.error_page.copy()
# Put a *copy* of the class namespaces into self.
self.namespaces = self.namespaces.copy()
self.stage = None
self.unique_id = LazyUUID4()
def close(self):
"""Run cleanup code. (Core)"""
if not self.closed:
self.closed = True
self.stage = 'on_end_request'
self.hooks.run('on_end_request')
self.stage = 'close'
def run(self, method, path, query_string, req_protocol, headers, rfile):
r"""Process the Request. (Core)
method, path, query_string, and req_protocol should be pulled directly
from the Request-Line (e.g. "GET /path?key=val HTTP/1.0").
path
This should be %XX-unquoted, but query_string should not be.
When using Python 2, they both MUST be byte strings,
not unicode strings.
When using Python 3, they both MUST be unicode strings,
not byte strings, and preferably not bytes \x00-\xFF
disguised as unicode.
headers
A list of (name, value) tuples.
rfile
A file-like object containing the HTTP request entity.
When run() is done, the returned object should have 3 attributes:
* status, e.g. "200 OK"
* header_list, a list of (name, value) tuples
* body, an iterable yielding strings
Consumer code (HTTP servers) should then access these response
attributes to build the outbound stream.
"""
response = cherrypy.serving.response
self.stage = 'run'
try:
self.error_response = cherrypy.HTTPError(500).set_response
self.method = method
path = path or '/'
self.query_string = query_string or ''
self.params = {}
# Compare request and server HTTP protocol versions, in case our
# server does not support the requested protocol. Limit our output
# to min(req, server). We want the following output:
# request server actual written supported response
# protocol protocol response protocol feature set
# a 1.0 1.0 1.0 1.0
# b 1.0 1.1 1.1 1.0
# c 1.1 1.0 1.0 1.0
# d 1.1 1.1 1.1 1.1
# Notice that, in (b), the response will be "HTTP/1.1" even though
# the client only understands 1.0. RFC 2616 10.5.6 says we should
# only return 505 if the _major_ version is different.
rp = int(req_protocol[5]), int(req_protocol[7])
sp = int(self.server_protocol[5]), int(self.server_protocol[7])
self.protocol = min(rp, sp)
response.headers.protocol = self.protocol
# Rebuild first line of the request (e.g. "GET /path HTTP/1.0").
url = path
if query_string:
url += '?' + query_string
self.request_line = '%s %s %s' % (method, url, req_protocol)
self.header_list = list(headers)
self.headers = httputil.HeaderMap()
self.rfile = rfile
self.body = None
self.cookie = SimpleCookie()
self.handler = None
# path_info should be the path from the
# app root (script_name) to the handler.
self.script_name = self.app.script_name
self.path_info = pi = path[len(self.script_name):]
self.stage = 'respond'
self.respond(pi)
except self.throws:
raise
except Exception:
if self.throw_errors:
raise
else:
# Failure in setup, error handler or finalize. Bypass them.
# Can't use handle_error because we may not have hooks yet.
cherrypy.log(traceback=True, severity=40)
if self.show_tracebacks:
body = format_exc()
else:
body = ''
r = bare_error(body)
response.output_status, response.header_list, response.body = r
if self.method == 'HEAD':
# HEAD requests MUST NOT return a message-body in the response.
response.body = []
try:
cherrypy.log.access()
except Exception:
cherrypy.log.error(traceback=True)
return response
def respond(self, path_info):
"""Generate a response for the resource at self.path_info. (Core)"""
try:
try:
try:
self._do_respond(path_info)
except (cherrypy.HTTPRedirect, cherrypy.HTTPError):
inst = sys.exc_info()[1]
inst.set_response()
self.stage = 'before_finalize (HTTPError)'
self.hooks.run('before_finalize')
cherrypy.serving.response.finalize()
finally:
self.stage = 'on_end_resource'
self.hooks.run('on_end_resource')
except self.throws:
raise
except Exception:
if self.throw_errors:
raise
self.handle_error()
def _do_respond(self, path_info):
response = cherrypy.serving.response
if self.app is None:
raise cherrypy.NotFound()
self.hooks = self.__class__.hooks.copy()
self.toolmaps = {}
# Get the 'Host' header, so we can HTTPRedirect properly.
self.stage = 'process_headers'
self.process_headers()
self.stage = 'get_resource'
self.get_resource(path_info)
self.body = _cpreqbody.RequestBody(
self.rfile, self.headers, request_params=self.params)
self.namespaces(self.config)
self.stage = 'on_start_resource'
self.hooks.run('on_start_resource')
# Parse the querystring
self.stage = 'process_query_string'
self.process_query_string()
# Process the body
if self.process_request_body:
if self.method not in self.methods_with_bodies:
self.process_request_body = False
self.stage = 'before_request_body'
self.hooks.run('before_request_body')
if self.process_request_body:
self.body.process()
# Run the handler
self.stage = 'before_handler'
self.hooks.run('before_handler')
if self.handler:
self.stage = 'handler'
response.body = self.handler()
# Finalize
self.stage = 'before_finalize'
self.hooks.run('before_finalize')
response.finalize()
def process_query_string(self):
"""Parse the query string into Python structures. (Core)"""
try:
p = httputil.parse_query_string(
self.query_string, encoding=self.query_string_encoding)
except UnicodeDecodeError:
raise cherrypy.HTTPError(
404, 'The given query string could not be processed. Query '
'strings for this resource must be encoded with %r.' %
self.query_string_encoding)
# Python 2 only: keyword arguments must be byte strings (type 'str').
if six.PY2:
for key, value in p.items():
if isinstance(key, six.text_type):
del p[key]
p[key.encode(self.query_string_encoding)] = value
self.params.update(p)
def process_headers(self):
"""Parse HTTP header data into Python structures. (Core)"""
# Process the headers into self.headers
headers = self.headers
for name, value in self.header_list:
# Call title() now (and use dict.__method__(headers))
# so title doesn't have to be called twice.
name = name.title()
value = value.strip()
headers[name] = httputil.decode_TEXT_maybe(value)
# Some clients, notably Konquoror, supply multiple
# cookies on different lines with the same key. To
# handle this case, store all cookies in self.cookie.
if name == 'Cookie':
try:
self.cookie.load(value)
except CookieError as exc:
raise cherrypy.HTTPError(400, str(exc))
if not dict.__contains__(headers, 'Host'):
# All Internet-based HTTP/1.1 servers MUST respond with a 400
# (Bad Request) status code to any HTTP/1.1 request message
# which lacks a Host header field.
if self.protocol >= (1, 1):
msg = "HTTP/1.1 requires a 'Host' request header."
raise cherrypy.HTTPError(400, msg)
host = dict.get(headers, 'Host')
if not host:
host = self.local.name or self.local.ip
self.base = '%s://%s' % (self.scheme, host)
def get_resource(self, path):
"""Call a dispatcher (which sets self.handler and .config). (Core)"""
# First, see if there is a custom dispatch at this URI. Custom
# dispatchers can only be specified in app.config, not in _cp_config
# (since custom dispatchers may not even have an app.root).
dispatch = self.app.find_config(
path, 'request.dispatch', self.dispatch)
# dispatch() should set self.handler and self.config
dispatch(path)
def handle_error(self):
"""Handle the last unanticipated exception. (Core)"""
try:
self.hooks.run('before_error_response')
if self.error_response:
self.error_response()
self.hooks.run('after_error_response')
cherrypy.serving.response.finalize()
except cherrypy.HTTPRedirect:
inst = sys.exc_info()[1]
inst.set_response()
cherrypy.serving.response.finalize()
class Application(object):
"""A CherryPy Application.
Servers and gateways should not instantiate Request objects directly.
Instead, they should ask an Application object for a request object.
An instance of this class may also be used as a WSGI callable
(WSGI application object) for itself.
"""
root = None
"""The top-most container of page handlers for this app. Handlers should
be arranged in a hierarchy of attributes, matching the expected URI
hierarchy; the default dispatcher then searches this hierarchy for a
matching handler. When using a dispatcher other than the default,
this value may be None."""
config = {}
"""A dict of {path: pathconf} pairs, where 'pathconf' is itself a dict
of {key: value} pairs."""
namespaces = reprconf.NamespaceSet()
toolboxes = {'tools': cherrypy.tools}
log = None
"""A LogManager instance. See _cplogging."""
wsgiapp = None
"""A CPWSGIApp instance. See _cpwsgi."""
request_class = _cprequest.Request
response_class = _cprequest.Response
relative_urls = False
def __init__(self, root, script_name='', config=None):
"""Initialize Application with given root."""
self.log = _cplogging.LogManager(id(self), cherrypy.log.logger_root)
self.root = root
self.script_name = script_name
self.wsgiapp = _cpwsgi.CPWSGIApp(self)
self.namespaces = self.namespaces.copy()
self.namespaces['log'] = lambda k, v: setattr(self.log, k, v)
self.namespaces['wsgi'] = self.wsgiapp.namespace_handler
self.config = self.__class__.config.copy()
if config:
self.merge(config)
def __repr__(self):
"""Generate a representation of the Application instance."""
return '%s.%s(%r, %r)' % (self.__module__, self.__class__.__name__,
self.root, self.script_name)
script_name_doc = """The URI "mount point" for this app. A mount point
is that portion of the URI which is constant for all URIs that are
serviced by this application; it does not include scheme, host, or proxy
("virtual host") portions of the URI.
For example, if script_name is "/my/cool/app", then the URL
"http://www.example.com/my/cool/app/page1" might be handled by a
"page1" method on the root object.
The value of script_name MUST NOT end in a slash. If the script_name
refers to the root of the URI, it MUST be an empty string (not "/").
If script_name is explicitly set to None, then the script_name will be
provided for each call from request.wsgi_environ['SCRIPT_NAME'].
"""
def _get_script_name(self):
if self._script_name is not None:
return self._script_name
# A `_script_name` with a value of None signals that the script name
# should be pulled from WSGI environ.
return cherrypy.serving.request.wsgi_environ['SCRIPT_NAME'].rstrip('/')
def _set_script_name(self, value):
if value:
value = value.rstrip('/')
self._script_name = value
script_name = property(fget=_get_script_name,
fset=_set_script_name,
doc=script_name_doc)
def merge(self, config):
"""Merge the given config into self.config."""
_cpconfig.merge(self.config, config)
# Handle namespaces specified in config.
self.namespaces(self.config.get('/', {}))
def find_config(self, path, key, default=None):
"""Return the most-specific value for key along path, or default."""
trail = path or '/'
while trail:
nodeconf = self.config.get(trail, {})
if key in nodeconf:
return nodeconf[key]
lastslash = trail.rfind('/')
if lastslash == -1:
break
elif lastslash == 0 and trail != '/':
trail = '/'
else:
trail = trail[:lastslash]
return default
def get_serving(self, local, remote, scheme, sproto):
"""Create and return a Request and Response object."""
req = self.request_class(local, remote, scheme, sproto)
req.app = self
for name, toolbox in self.toolboxes.items():
req.namespaces[name] = toolbox
resp = self.response_class()
cherrypy.serving.load(req, resp)
cherrypy.engine.publish('acquire_thread')
cherrypy.engine.publish('before_request')
return req, resp
def release_serving(self):
"""Release the current serving (request and response)."""
req = cherrypy.serving.request
cherrypy.engine.publish('after_request')
try:
req.close()
except:
cherrypy.log(traceback=True, severity=40)
cherrypy.serving.clear()
def __call__(self, environ, start_response):
"""Call a WSGI-callable."""
return self.wsgiapp(environ, start_response)