diff options
author | Maximilian Hils <git@maximilianhils.com> | 2014-07-31 22:29:13 +0200 |
---|---|---|
committer | Maximilian Hils <git@maximilianhils.com> | 2014-07-31 22:29:13 +0200 |
commit | e9401a2123f416f8dd4c24bb4bbc74f297369133 (patch) | |
tree | afebf68d1091d487ef64083962992ea7b703a80e | |
parent | f5fb1138fdda77ec4a3bb5493479d2f56cdb0851 (diff) | |
parent | 3e3dbee936bb71e813d50937118eebff4ba23617 (diff) | |
download | mitmproxy-e9401a2123f416f8dd4c24bb4bbc74f297369133.tar.gz mitmproxy-e9401a2123f416f8dd4c24bb4bbc74f297369133.tar.bz2 mitmproxy-e9401a2123f416f8dd4c24bb4bbc74f297369133.zip |
Merge branch 'stream'
Conflicts:
libmproxy/protocol/http.py
-rw-r--r-- | doc-src/_nav.html | 1 | ||||
-rw-r--r-- | doc-src/features/index.py | 1 | ||||
-rw-r--r-- | doc-src/features/responsestreaming.html | 50 | ||||
-rw-r--r-- | doc-src/scripting/inlinescripts.html | 10 | ||||
-rw-r--r-- | examples/stream.py | 5 | ||||
-rw-r--r-- | examples/stub.py | 9 | ||||
-rw-r--r-- | libmproxy/cmdline.py | 102 | ||||
-rw-r--r-- | libmproxy/console/__init__.py | 5 | ||||
-rw-r--r-- | libmproxy/console/flowview.py | 2 | ||||
-rw-r--r-- | libmproxy/controller.py | 2 | ||||
-rw-r--r-- | libmproxy/dump.py | 14 | ||||
-rw-r--r-- | libmproxy/flow.py | 28 | ||||
-rw-r--r-- | libmproxy/protocol/http.py | 78 | ||||
-rw-r--r-- | libmproxy/proxy/server.py | 4 | ||||
-rw-r--r-- | test/scripts/all.py | 8 | ||||
-rw-r--r-- | test/test_server.py | 55 |
16 files changed, 292 insertions, 82 deletions
diff --git a/doc-src/_nav.html b/doc-src/_nav.html index ddc27dca..cd70cc25 100644 --- a/doc-src/_nav.html +++ b/doc-src/_nav.html @@ -19,6 +19,7 @@ $!nav("sticky.html", this, state)!$ $!nav("reverseproxy.html", this, state)!$ $!nav("upstreamcerts.html", this, state)!$ + $!nav("responsestreaming.html", this, state)!$ <li class="nav-header">Installing Certificates</li> $!nav("ssl.html", this, state)!$ diff --git a/doc-src/features/index.py b/doc-src/features/index.py index 0618681f..b4d441d1 100644 --- a/doc-src/features/index.py +++ b/doc-src/features/index.py @@ -12,4 +12,5 @@ pages = [ Page("replacements.html", "Replacements"), Page("reverseproxy.html", "Reverse proxy mode"), Page("upstreamcerts.html", "Upstream Certs"), + Page("responsestreaming.html", "Response Streaming"), ] diff --git a/doc-src/features/responsestreaming.html b/doc-src/features/responsestreaming.html new file mode 100644 index 00000000..d20af65c --- /dev/null +++ b/doc-src/features/responsestreaming.html @@ -0,0 +1,50 @@ +By using mitmproxy's streaming feature, response contents can be passed to the client incrementally before they have been fully received by the proxy. +This is especially useful for large binary files such as videos, where buffering the whole file slows down the client's browser. + +By default, mitmproxy will read the entire response, perform any indicated +manipulations on it and then send the (possibly modified) response to +the client. In some cases this is undesirable and you may wish to "stream" +the reponse back to the client. When streaming is enabled, the response is +not buffered on the proxy but directly sent back to the client instead. + +<h2>On the command-line</h2> + +Streaming can be enabled on the command line for all response bodies exceeding a certain size. The SIZE argument understands +k/m/g suffixes, e.g. 3m for 3 megabytes. + +<table class="table"> + <tbody> + <tr> + <th width="20%">command-line</th> + <td> + --stream SIZE + </td> + </tr> + </tbody> +</table> + + +<h2>Caveats</h2> + +When response streaming is enabled, <strong>streamed response contents will not be + recorded or preserved in any way.</strong> + +When response streaming is enabled, the response body cannot be modified. + +<h2>Customizing Response Streaming</h2> + +You can also use an <a href="@!urlTo("scripting/inlinescripts.html")!@">inline script</a> to customize exactly +which responses are streamed. + +Responses that should be tagged for streaming by setting their respective .stream attribute to True: + +$!example("examples/stream.py")!$ + + +<h2>Implementation Details</h2> + +When response streaming is enabled, portions of the code which would have otherwise performed changes +on the response body will see an empty response body instead (<code>libmproxy.protocol.http.CONTENT_MISSING</code>). Any modifications will be ignored. + +Streamed responses are usually sent in chunks of 4096 bytes. If the response is sent with a <code>Transfer-Encoding: + chunked</code> header, the response will be streamed one chunk at a time.
\ No newline at end of file diff --git a/doc-src/scripting/inlinescripts.html b/doc-src/scripting/inlinescripts.html index 32a98e99..65090cfb 100644 --- a/doc-src/scripting/inlinescripts.html +++ b/doc-src/scripting/inlinescripts.html @@ -46,12 +46,20 @@ a connection can correspond to multiple HTTP requests. Called when a client request has been received. The __Flow__ object is guaranteed to have a non-None __request__ attribute. +### responseheaders(ScriptContext, Flow) + +Called when the headers of a server response have been received. +This will always be called before the response hook. +The __Flow__ object is guaranteed to have non-None __request__ and +__response__ attributes. __response.content__ will not be valid, +as the response body has not been read yet. ### response(ScriptContext, Flow) Called when a server response has been received. The __Flow__ object is guaranteed to have non-None __request__ and __response__ attributes. - +Note that if response streaming is enabled for this response, +__response.content__ will not contain the response body. ### error(ScriptContext, Flow) diff --git a/examples/stream.py b/examples/stream.py new file mode 100644 index 00000000..f9f03c3e --- /dev/null +++ b/examples/stream.py @@ -0,0 +1,5 @@ +def responseheaders(ctx, flow): + """ + Enables streaming for all responses. + """ + flow.response.stream = True
\ No newline at end of file diff --git a/examples/stub.py b/examples/stub.py index 78cbfcf2..0cf67db7 100644 --- a/examples/stub.py +++ b/examples/stub.py @@ -27,6 +27,15 @@ def request(ctx, flow): """ ctx.log("request") + +def responseheaders(ctx, flow): + """ + Called when the response headers for a server response have been received, + but the response body has not been processed yet. Can be used to tell mitmproxy + to stream the response. + """ + ctx.log("responseheaders") + def response(ctx, flow): """ Called when a server response has been received. diff --git a/libmproxy/cmdline.py b/libmproxy/cmdline.py index e407cd10..c9c0c75e 100644 --- a/libmproxy/cmdline.py +++ b/libmproxy/cmdline.py @@ -3,7 +3,7 @@ import re import argparse from argparse import ArgumentTypeError from netlib import http -from . import proxy, filt +from . import filt, utils from .proxy import config APP_HOST = "mitm.it" @@ -23,13 +23,13 @@ def _parse_hook(s): elif len(parts) == 3: patt, a, b = parts else: - raise ParseException("Malformed hook specifier - too few clauses: %s"%s) + raise ParseException("Malformed hook specifier - too few clauses: %s" % s) if not a: - raise ParseException("Empty clause: %s"%str(patt)) + raise ParseException("Empty clause: %s" % str(patt)) if not filt.parse(patt): - raise ParseException("Malformed filter pattern: %s"%patt) + raise ParseException("Malformed filter pattern: %s" % patt) return patt, a, b @@ -64,7 +64,7 @@ def parse_replace_hook(s): try: re.compile(regex) except re.error, e: - raise ParseException("Malformed replacement regex: %s"%str(e.message)) + raise ParseException("Malformed replacement regex: %s" % str(e.message)) return patt, regex, replacement @@ -98,7 +98,6 @@ def parse_setheader(s): def parse_server_spec(url): - normalized_url = re.sub("^https?2", "", url) p = http.parse_url(normalized_url) @@ -125,6 +124,8 @@ def get_common_options(options): if options.stickyauth_filt: stickyauth = options.stickyauth_filt + stream_large_bodies = utils.parse_size(options.stream_large_bodies) + reps = [] for i in options.replace: try: @@ -140,10 +141,9 @@ def get_common_options(options): try: v = open(path, "rb").read() except IOError, e: - raise ArgumentTypeError("Could not read replace file: %s"%path) + raise ArgumentTypeError("Could not read replace file: %s" % path) reps.append((patt, rex, v)) - setheaders = [] for i in options.setheader: try: @@ -153,29 +153,30 @@ def get_common_options(options): setheaders.append(p) return dict( - app = options.app, - app_host = options.app_host, - app_port = options.app_port, - app_external = options.app_external, - - anticache = options.anticache, - anticomp = options.anticomp, - client_replay = options.client_replay, - kill = options.kill, - no_server = options.no_server, - refresh_server_playback = not options.norefresh, - rheaders = options.rheaders, - rfile = options.rfile, - replacements = reps, - setheaders = setheaders, - server_replay = options.server_replay, - scripts = options.scripts, - stickycookie = stickycookie, - stickyauth = stickyauth, - showhost = options.showhost, - wfile = options.wfile, - verbosity = options.verbose, - nopop = options.nopop, + app=options.app, + app_host=options.app_host, + app_port=options.app_port, + app_external=options.app_external, + + anticache=options.anticache, + anticomp=options.anticomp, + client_replay=options.client_replay, + kill=options.kill, + no_server=options.no_server, + refresh_server_playback=not options.norefresh, + rheaders=options.rheaders, + rfile=options.rfile, + replacements=reps, + setheaders=setheaders, + server_replay=options.server_replay, + scripts=options.scripts, + stickycookie=stickycookie, + stickyauth=stickyauth, + stream_large_bodies=stream_large_bodies, + showhost=options.showhost, + wfile=options.wfile, + verbosity=options.verbose, + nopop=options.nopop, ) @@ -187,8 +188,8 @@ def common_options(parser): ) parser.add_argument( "--confdir", - action="store", type = str, dest="confdir", default='~/.mitmproxy', - help = "Configuration directory. (~/.mitmproxy)" + action="store", type=str, dest="confdir", default='~/.mitmproxy', + help="Configuration directory. (~/.mitmproxy)" ) parser.add_argument( "--host", @@ -240,10 +241,16 @@ def common_options(parser): "-Z", action="store", dest="body_size_limit", default=None, metavar="SIZE", - help="Byte size limit of HTTP request and response bodies."\ + help="Byte size limit of HTTP request and response bodies." \ " Understands k/m/g suffixes, i.e. 3m for 3 megabytes." ) - + parser.add_argument( + "--stream", + action="store", dest="stream_large_bodies", default=None, + metavar="SIZE", + help="Stream data to the client if response body exceeds the given threshold. " + "If streamed, the body will not be stored in any way. Understands k/m/g suffixes, i.e. 3m for 3 megabytes." + ) group = parser.add_argument_group("Proxy Options") # We could make a mutually exclusive group out of -R, -U, -T, but we don't do that because @@ -251,8 +258,8 @@ def common_options(parser): # - our own error messages are more helpful group.add_argument( "-b", - action="store", type = str, dest="addr", default='', - help = "Address to bind proxy to (defaults to all interfaces)" + action="store", type=str, dest="addr", default='', + help="Address to bind proxy to (defaults to all interfaces)" ) group.add_argument( "-U", @@ -266,8 +273,8 @@ def common_options(parser): ) group.add_argument( "-p", - action="store", type = int, dest="port", default=8080, - help = "Proxy service port." + action="store", type=int, dest="port", default=8080, + help="Proxy service port." ) group.add_argument( "-R", @@ -280,7 +287,6 @@ def common_options(parser): help="Set transparent proxy mode." ) - group = parser.add_argument_group( "Advanced Proxy Options", """ @@ -304,7 +310,6 @@ def common_options(parser): help="Override the destination server all requests are sent to: http[s][2http[s]]://host[:port]" ) - group = parser.add_argument_group("Web App") group.add_argument( "-a", @@ -315,7 +320,7 @@ def common_options(parser): "--app-host", action="store", dest="app_host", default=APP_HOST, metavar="host", help="Domain to serve the app from. For transparent mode, use an IP when\ - a DNS entry for the app domain is not present. Default: %s"%APP_HOST + a DNS entry for the app domain is not present. Default: %s" % APP_HOST ) group.add_argument( @@ -329,7 +334,6 @@ def common_options(parser): help="Serve the app outside of the proxy." ) - group = parser.add_argument_group("Client Replay") group.add_argument( "-c", @@ -352,22 +356,21 @@ def common_options(parser): "--rheader", action="append", dest="rheaders", type=str, help="Request headers to be considered during replay. " - "Can be passed multiple times." + "Can be passed multiple times." ) group.add_argument( "--norefresh", action="store_true", dest="norefresh", default=False, - help= "Disable response refresh, " - "which updates times in cookies and headers for replayed responses." + help="Disable response refresh, " + "which updates times in cookies and headers for replayed responses." ) group.add_argument( "--no-pop", action="store_true", dest="nopop", default=False, help="Disable response pop from response flow. " - "This makes it possible to replay same response multiple times." + "This makes it possible to replay same response multiple times." ) - group = parser.add_argument_group( "Replacements", """ @@ -389,7 +392,6 @@ def common_options(parser): help="Replacement pattern, where the replacement clause is a path to a file." ) - group = parser.add_argument_group( "Set Headers", """ @@ -405,7 +407,6 @@ def common_options(parser): help="Header set pattern." ) - group = parser.add_argument_group( "Proxy Authentication", """ @@ -434,5 +435,4 @@ def common_options(parser): help="Allow access to users specified in an Apache htpasswd file." ) - config.ssl_option_group(parser) diff --git a/libmproxy/console/__init__.py b/libmproxy/console/__init__.py index e660f312..3d025739 100644 --- a/libmproxy/console/__init__.py +++ b/libmproxy/console/__init__.py @@ -165,6 +165,8 @@ class StatusBar(common.WWrap): opts.append("no-upstream-cert") if self.master.state.follow_focus: opts.append("following") + if self.master.stream_large_bodies: + opts.append("stream:%s" % utils.pretty_size(self.master.stream_large_bodies.max_size)) if opts: r.append("[%s]"%(":".join(opts))) @@ -343,6 +345,7 @@ class Options(object): "server_replay", "stickycookie", "stickyauth", + "stream_large_bodies", "verbosity", "wfile", "nopop", @@ -391,6 +394,8 @@ class ConsoleMaster(flow.FlowMaster): print >> sys.stderr, "Sticky auth error:", r sys.exit(1) + self.set_stream_large_bodies(options.stream_large_bodies) + self.refresh_server_playback = options.refresh_server_playback self.anticache = options.anticache self.anticomp = options.anticomp diff --git a/libmproxy/console/flowview.py b/libmproxy/console/flowview.py index ac10e809..4aaf8944 100644 --- a/libmproxy/console/flowview.py +++ b/libmproxy/console/flowview.py @@ -153,7 +153,7 @@ class FlowView(common.WWrap): def cont_view_handle_missing(self, conn, viewmode): if conn.content == CONTENT_MISSING: - msg, body = "", [urwid.Text([("error", "[content missing]")])], 0 + msg, body = "", [urwid.Text([("error", "[content missing]")])] else: msg, body = self.content_view(viewmode, conn) diff --git a/libmproxy/controller.py b/libmproxy/controller.py index 63e44241..4a72cf80 100644 --- a/libmproxy/controller.py +++ b/libmproxy/controller.py @@ -49,7 +49,7 @@ class Channel: try: # The timeout is here so we can handle a should_exit event. g = m.reply.q.get(timeout=0.5) - except Queue.Empty: # pragma: nocover + except Queue.Empty: # pragma: nocover continue return g diff --git a/libmproxy/dump.py b/libmproxy/dump.py index 1f107241..aeb34cc3 100644 --- a/libmproxy/dump.py +++ b/libmproxy/dump.py @@ -2,6 +2,7 @@ from __future__ import absolute_import import sys, os import netlib.utils from . import flow, filt, utils +from .protocol import http class DumpError(Exception): pass @@ -30,6 +31,7 @@ class Options(object): "showhost", "stickycookie", "stickyauth", + "stream_large_bodies", "verbosity", "wfile", ] @@ -69,6 +71,8 @@ class DumpMaster(flow.FlowMaster): self.showhost = options.showhost self.refresh_server_playback = options.refresh_server_playback + self.set_stream_large_bodies(options.stream_large_bodies) + if filtstr: self.filt = filt.parse(filtstr) else: @@ -80,6 +84,7 @@ class DumpMaster(flow.FlowMaster): if options.stickyauth: self.set_stickyauth(options.stickyauth) + if options.wfile: path = os.path.expanduser(options.wfile) try: @@ -157,12 +162,17 @@ class DumpMaster(flow.FlowMaster): if f.response: if self.o.flow_detail > 0: - sz = utils.pretty_size(len(f.response.content)) + if f.response.content == http.CONTENT_MISSING: + sz = "(content missing)" + else: + sz = utils.pretty_size(len(f.response.content)) result = " << %s %s"%(str_response(f.response), sz) if self.o.flow_detail > 1: result = result + "\n\n" + self.indent(4, f.response.headers) if self.o.flow_detail > 2: - if utils.isBin(f.response.content): + if f.response.content == http.CONTENT_MISSING: + cont = self.indent(4, "(content missing)") + elif utils.isBin(f.response.content): d = netlib.utils.hexdump(f.response.content) d = "\n".join("%s\t%s %s"%i for i in d) cont = self.indent(4, d) diff --git a/libmproxy/flow.py b/libmproxy/flow.py index 8ea51ce6..d4aa1805 100644 --- a/libmproxy/flow.py +++ b/libmproxy/flow.py @@ -145,6 +145,17 @@ class SetHeaders: f.request.headers.add(header, value) +class StreamLargeBodies(object): + def __init__(self, max_size): + self.max_size = max_size + + def run(self, flow, is_request): + r = flow.request if is_request else flow.response + code = flow.response.code if flow.response else None + expected_size = netlib.http.expected_http_body_size(r.headers, is_request, flow.request.method, code) + if not (0 <= expected_size <= self.max_size): + r.stream = True + class ClientPlaybackState: def __init__(self, flows, exit): self.flows, self.exit = flows, exit @@ -437,6 +448,7 @@ class FlowMaster(controller.Master): self.anticache = False self.anticomp = False + self.stream_large_bodies = False self.refresh_server_playback = False self.replacehooks = ReplaceHooks() self.setheaders = SetHeaders() @@ -522,6 +534,12 @@ class FlowMaster(controller.Master): self.stickycookie_state = None self.stickycookie_txt = None + def set_stream_large_bodies(self, max_size): + if max_size is not None: + self.stream_large_bodies = StreamLargeBodies(max_size) + else: + self.stream_large_bodies = False + def set_stickyauth(self, txt): if txt: flt = filt.parse(txt) @@ -701,6 +719,16 @@ class FlowMaster(controller.Master): self.process_new_request(f) return f + def handle_responseheaders(self, resp): + f = resp.flow + self.run_script_hook("responseheaders", f) + + if self.stream_large_bodies: + self.stream_large_bodies.run(f, False) + + resp.reply() + return f + def handle_response(self, r): f = self.state.add_response(r) if f: diff --git a/libmproxy/protocol/http.py b/libmproxy/protocol/http.py index 21fc2798..6837327b 100644 --- a/libmproxy/protocol/http.py +++ b/libmproxy/protocol/http.py @@ -293,7 +293,8 @@ class HTTPRequest(HTTPMessage): raise http.HttpError(400, "Invalid headers") if include_body: - content = http.read_http_body(rfile, headers, body_size_limit, True) + content = http.read_http_body(rfile, headers, body_size_limit, + method, None, True) timestamp_end = utils.timestamp() return HTTPRequest(form_in, method, scheme, host, port, path, httpversion, headers, @@ -305,7 +306,7 @@ class HTTPRequest(HTTPMessage): if form == "relative": path = self.path if self.method != "OPTIONS" else "*" request_line = '%s %s HTTP/%s.%s' % \ - (self.method, path, self.httpversion[0], self.httpversion[1]) + (self.method, path, self.httpversion[0], self.httpversion[1]) elif form == "authority": request_line = '%s %s:%s HTTP/%s.%s' % (self.method, self.host, self.port, self.httpversion[0], self.httpversion[1]) @@ -591,6 +592,7 @@ class HTTPResponse(HTTPMessage): # Is this request replayed? self.is_replay = False + self.stream = False _stateobject_attributes = HTTPMessage._stateobject_attributes.copy() _stateobject_attributes.update( @@ -632,24 +634,22 @@ class HTTPResponse(HTTPMessage): return 'HTTP/%s.%s %s %s' % \ (self.httpversion[0], self.httpversion[1], self.code, self.msg) - def _assemble_headers(self): + def _assemble_headers(self, preserve_transfer_encoding=False): headers = self.headers.copy() - utils.del_all( - headers, - [ - 'Proxy-Connection', - 'Transfer-Encoding' - ] - ) + utils.del_all(headers, ['Proxy-Connection']) + if not preserve_transfer_encoding: + utils.del_all(headers, ['Transfer-Encoding']) + if self.content: headers["Content-Length"] = [str(len(self.content))] - elif 'Transfer-Encoding' in self.headers: # add content-length for chuncked transfer-encoding with no content + elif not preserve_transfer_encoding and 'Transfer-Encoding' in self.headers: # add content-length for chuncked transfer-encoding with no content headers["Content-Length"] = ["0"] return str(headers) - def _assemble_head(self): - return '%s\r\n%s\r\n' % (self._assemble_first_line(), self._assemble_headers()) + def _assemble_head(self, preserve_transfer_encoding=False): + return '%s\r\n%s\r\n' % ( + self._assemble_first_line(), self._assemble_headers(preserve_transfer_encoding=preserve_transfer_encoding)) def _assemble(self): """ @@ -865,15 +865,16 @@ class HTTPHandler(ProtocolHandler, TemporaryServerChangeMixin): pass self.c.close = True - def get_response_from_server(self, request): + def get_response_from_server(self, request, include_body=True): self.c.establish_server_connection() request_raw = request._assemble() for i in range(2): try: self.c.server_conn.send(request_raw) - return HTTPResponse.from_stream(self.c.server_conn.rfile, request.method, - body_size_limit=self.c.config.body_size_limit) + res = HTTPResponse.from_stream(self.c.server_conn.rfile, request.method, + body_size_limit=self.c.config.body_size_limit, include_body=include_body) + return res except (tcp.NetLibDisconnect, http.HttpErrorConnClosed), v: self.c.log("error in server communication: %s" % str(v), level="debug") if i < 1: @@ -915,21 +916,53 @@ class HTTPHandler(ProtocolHandler, TemporaryServerChangeMixin): if isinstance(request_reply, HTTPResponse): flow.response = request_reply else: - flow.response = self.get_response_from_server(flow.request) - flow.server_conn = self.c.server_conn # no further manipulation of self.c.server_conn beyond this point + # read initially in "stream" mode, so we can get the headers separately + flow.response = self.get_response_from_server(flow.request, include_body=False) + + # call the appropriate script hook - this is an opportunity for an inline script to set flow.stream = True + self.c.channel.ask("responseheaders", flow.response) + + # now get the rest of the request body, if body still needs to be read but not streaming this response + if flow.response.stream: + flow.response.content = CONTENT_MISSING + else: + flow.response.content = http.read_http_body(self.c.server_conn.rfile, flow.response.headers, + self.c.config.body_size_limit, + flow.request.method, flow.response.code, False) + + # no further manipulation of self.c.server_conn beyond this point # we can safely set it as the final attribute value here. + flow.server_conn = self.c.server_conn self.c.log("response", "debug", [flow.response._assemble_first_line()]) response_reply = self.c.channel.ask("response", flow.response) if response_reply is None or response_reply == KILL: return False - self.c.client_conn.send(flow.response._assemble()) + if not flow.response.stream: + # no streaming: + # we already received the full response from the server and can send it to the client straight away. + self.c.client_conn.send(flow.response._assemble()) + else: + # streaming: + # First send the body and then transfer the response incrementally: + h = flow.response._assemble_head(preserve_transfer_encoding=True) + self.c.client_conn.send(h) + for chunk in http.read_http_body_chunked(self.c.server_conn.rfile, + flow.response.headers, + self.c.config.body_size_limit, flow.request.method, + flow.response.code, False, 4096): + for part in chunk: + self.c.client_conn.wfile.write(part) + self.c.client_conn.wfile.flush() + flow.timestamp_end = utils.timestamp() if (http.connection_close(flow.request.httpversion, flow.request.headers) or - http.connection_close(flow.response.httpversion, flow.response.headers)): + http.connection_close(flow.response.httpversion, flow.response.headers) or + http.expected_http_body_size(flow.response.headers, False, flow.request.method, + flow.response.code) == -1): if flow.request.form_in == "authority" and flow.response.code == 200: # Workaround for https://github.com/mitmproxy/mitmproxy/issues/313: # Some proxies (e.g. Charles) send a CONNECT response with HTTP/1.0 and no Content-Length header @@ -943,6 +976,7 @@ class HTTPHandler(ProtocolHandler, TemporaryServerChangeMixin): # If the user has changed the target server on this connection, # restore the original target server self.restore_server() + return True except (HttpAuthenticationError, http.HttpError, proxy.ProxyError, tcp.NetLibError), e: self.handle_error(e, flow) @@ -965,7 +999,7 @@ class HTTPHandler(ProtocolHandler, TemporaryServerChangeMixin): if flow.request and not flow.response: self.c.channel.ask("error", flow.error) else: - pass # FIXME: Do we want to persist errors without flows? + pass # FIXME: Do we want to persist errors without flows? try: self.send_error(code, message, headers) @@ -1068,7 +1102,7 @@ class HTTPHandler(ProtocolHandler, TemporaryServerChangeMixin): return True raise http.HttpError(400, "Invalid HTTP request form (expected: %s, got: %s)" % - (self.expected_form_in, request.form_in)) + (self.expected_form_in, request.form_in)) def authenticate(self, request): if self.c.config.authenticator: diff --git a/libmproxy/proxy/server.py b/libmproxy/proxy/server.py index 61369775..50f01bdc 100644 --- a/libmproxy/proxy/server.py +++ b/libmproxy/proxy/server.py @@ -95,14 +95,14 @@ class ConnectionHandler: except (ProxyError, tcp.NetLibError), e: handle_error(self.conntype, self, e) - except Exception, e: + except Exception: import traceback, sys self.log(traceback.format_exc(), "error") print >> sys.stderr, traceback.format_exc() print >> sys.stderr, "mitmproxy has crashed!" print >> sys.stderr, "Please lodge a bug report at: https://github.com/mitmproxy/mitmproxy" - raise e + raise self.del_server_connection() self.log("clientdisconnect", "info") diff --git a/test/scripts/all.py b/test/scripts/all.py index 7d30d757..3acaf694 100644 --- a/test/scripts/all.py +++ b/test/scripts/all.py @@ -7,14 +7,18 @@ def serverconnect(ctx, cc): ctx.log("XSERVERCONNECT") log.append("serverconnect") -def request(ctx, r): +def request(ctx, f): ctx.log("XREQUEST") log.append("request") -def response(ctx, r): +def response(ctx, f): ctx.log("XRESPONSE") log.append("response") +def responseheaders(ctx, f): + ctx.log("XRESPONSEHEADERS") + log.append("responseheaders") + def clientdisconnect(ctx, cc): ctx.log("XCLIENTDISCONNECT") log.append("clientdisconnect") diff --git a/test/test_server.py b/test/test_server.py index e1da3125..cf50f280 100644 --- a/test/test_server.py +++ b/test/test_server.py @@ -383,6 +383,61 @@ class TestRedirectRequest(tservers.HTTPProxTest): assert r3.content == r2.content == r1.content # Make sure that we actually use the same connection in this test case +class MasterStreamRequest(tservers.TestMaster): + """ + Enables the stream flag on the flow for all requests + """ + def handle_responseheaders(self, r): + r.stream = True + r.reply() + +class TestStreamRequest(tservers.HTTPProxTest): + masterclass = MasterStreamRequest + + def test_stream_simple(self): + p = self.pathoc() + + # a request with 100k of data but without content-length + self.server.clear_log() + r1 = p.request("get:'%s/p/200:r:b@100k:d102400'"%self.server.urlbase) + assert r1.status_code == 200 + assert len(r1.content) > 100000 + assert self.server.last_log() + + def test_stream_multiple(self): + p = self.pathoc() + + # simple request with streaming turned on + self.server.clear_log() + r1 = p.request("get:'%s/p/200'"%self.server.urlbase) + assert r1.status_code == 200 + assert self.server.last_log() + + # now send back 100k of data, streamed but not chunked + self.server.clear_log() + r1 = p.request("get:'%s/p/200:b@100k'"%self.server.urlbase) + assert r1.status_code == 200 + assert self.server.last_log() + + def test_stream_chunked(self): + + connection = socket.socket(socket.AF_INET, socket.SOCK_STREAM) + connection.connect(("127.0.0.1", self.proxy.port)) + fconn = connection.makefile() + spec = '200:h"Transfer-Encoding"="chunked":r:b"4\\r\\nthis\\r\\n7\\r\\nisatest\\r\\n0\\r\\n\\r\\n"' + connection.send("GET %s/p/%s HTTP/1.1\r\n"%(self.server.urlbase, spec)) + connection.send("\r\n") + + httpversion, code, msg, headers, content = http.read_response(fconn, "GET", None, include_body=False) + + assert headers["Transfer-Encoding"][0] == 'chunked' + assert code == 200 + + chunks = list(content for _, content, _ in http.read_http_body_chunked(fconn, headers, None, "GET", 200, False)) + assert chunks == ["this", "isatest", ""] + + connection.close() + class MasterFakeResponse(tservers.TestMaster): def handle_request(self, m): |