From 8a4653d4d4021241a4222bcd4eec16fd1b598a2c Mon Sep 17 00:00:00 2001 From: Chandrashekhar Mullaparthi Date: Mon, 17 May 2010 23:22:51 +0100 Subject: [PATCH] Added support for tunnelling through a proxy. Other minor changes as indicated in the README --- README | 19 +- doc/ibrowse.html | 76 +- ebin/ibrowse.app | 2 +- src/ibrowse.erl | 480 +++++------ src/ibrowse_app.erl | 1 - src/ibrowse_http_client.erl | 1577 ++++++++++++++++++++--------------- src/ibrowse_lb.erl | 2 - src/ibrowse_lib.erl | 213 ++--- src/ibrowse_sup.erl | 2 - src/ibrowse_test.erl | 1 - vsn.mk | 2 +- 11 files changed, 1325 insertions(+), 1050 deletions(-) diff --git a/README b/README index e534803..1a1dc42 100644 --- a/README +++ b/README @@ -18,12 +18,25 @@ ibrowse is available under two different licenses. LGPL and the BSD license. Comments to : Chandrashekhar.Mullaparthi@gmail.com -Version : 1.5.6 +Version : 1.5.7 Latest version : git://github.com/cmullaparthi/ibrowse.git CONTRIBUTIONS & CHANGE HISTORY ============================== +12-05-2010 - * Added support for the CONNECT method to tunnel HTTPS through + a proxy. When a https URL is requested through a proxy, + ibrowse will automatically use the CONNECT method to first + setup a tunnel through the proxy. Once this succeeds, the + actual request is dispatched. Successfully tested with the + new SSL implementation in R13B-03 + * Added SSL support for direct connections. + See ibrowse:spawn_worker_process/1 and + ibrowse:spawn_link_worker_process/1 + * Added option to return raw status line and raw unparsed headers + +23-04-2010 - * Fixes to URL parsing by Karol Skocik + 08-11-2009 - * Added option headers_as_is 04-10-2009 - * Patch from Kostis Sagonas to cleanup some code and suppress @@ -132,7 +145,7 @@ CONTRIBUTIONS & CHANGE HISTORY 12-01-2007 - Derek Upham sent in a bug fix. The reset_state function was not behaving correctly when the transfer encoding was not chunked. -13-11-2006 - Youns Hafri reported a bug where ibrowse was not returning the +13-11-2006 - Youns Hafri reported a bug where ibrowse was not returning the temporary filename when the server was closing the connection after sending the data (as in HTTP/1.0). Released ibrowse under the BSD license @@ -151,7 +164,7 @@ CONTRIBUTIONS & CHANGE HISTORY 22-Nov-2005 - Added ability to generate requests using the Chunked Transfer-Encoding. -08-May-2005 - Youns Hafri made a CRUX LINUX port of ibrowse. +08-May-2005 - Youns Hafri made a CRUX LINUX port of ibrowse. http://yhafri.club.fr/crux/index.html Here are some usage examples. Enjoy! diff --git a/doc/ibrowse.html b/doc/ibrowse.html index 79e4753..590ff9a 100644 --- a/doc/ibrowse.html +++ b/doc/ibrowse.html @@ -10,9 +10,9 @@

Module ibrowse

The ibrowse application implements an HTTP 1.1 client. -

Copyright 2005-2009 Chandrashekhar Mullaparthi

+

Copyright 2005-2010 Chandrashekhar Mullaparthi

-

Version: 1.5.2

+

Version: 1.6.0

Behaviours: gen_server.

Authors: Chandrashekhar Mullaparthi (chandrashekhar dot mullaparthi at gmail dot com).

@@ -28,23 +28,23 @@ send_req/4, send_req/5, send_req/6.

Here are a few sample invocations.

-

+ ibrowse:send_req("http://intranet/messenger/", [], get).

ibrowse:send_req("http://www.google.com/", [], get, [], - [{proxy_user, "XXXXX"}, - {proxy_password, "XXXXX"}, - {proxy_host, "proxy"}, - {proxy_port, 8080}], 1000). + [{proxy_user, "XXXXX"}, + {proxy_password, "XXXXX"}, + {proxy_host, "proxy"}, + {proxy_port, 8080}], 1000).

ibrowse:send_req("http://www.erlang.org/download/otp_src_R10B-3.tar.gz", [], get, [], - [{proxy_user, "XXXXX"}, - {proxy_password, "XXXXX"}, - {proxy_host, "proxy"}, - {proxy_port, 8080}, - {save_response_to_file, true}], 1000). + [{proxy_user, "XXXXX"}, + {proxy_password, "XXXXX"}, + {proxy_host, "proxy"}, + {proxy_port, 8080}, + {save_response_to_file, true}], 1000).

ibrowse:send_req("http://www.erlang.org", [], head). @@ -58,11 +58,8 @@ send_req/4, send_req/5, send_req/6.



ibrowse:send_req("http://www.google.com", [], get, [], [{stream_to, self()}]). -

+ -

A driver exists which implements URL encoding in C, but the - speed achieved using only erlang has been good enough, so the - driver isn't actually used.

Function Index

@@ -93,10 +90,12 @@ send_req/4, send_req/5, send_req/6.

- - + + @@ -203,7 +202,7 @@ send_req/4, send_req/5, send_req/6.

send_req(Url::string(), Headers::headerList(), Method::method(), Body::body(), Options::optionList()) -> response()

Same as spawn_worker_process/1 except the the calling process is linked to the worker process which is spawned.

-

spawn_worker_process/2

+

spawn_link_worker_process/2

-

spawn_worker_process(Host::string(), Port::integer()) -> {ok, pid()}

+

spawn_link_worker_process(Host::string(), Port::integer()) -> {ok, pid()}

+
+ +

spawn_worker_process/1

+
+

spawn_worker_process(Url::string()) -> {ok, pid()}

Creates a HTTP client process to the specified Host:Port which is not part of the load balancing pool. This is useful in cases where some requests to a webserver might take a long time whereas @@ -362,6 +375,11 @@ send_req/4, send_req/5, send_req/6.

pipeline size on such connections.

+

spawn_worker_process/2

+
+

spawn_worker_process(Host::string(), Port::integer()) -> {ok, pid()}

+
+

start/0

start() -> any()

@@ -423,6 +441,6 @@ send_req/4, send_req/5, send_req/6.

all_trace_off/0Turn Off ALL tracing.
code_change/3
show_dest_status/0
show_dest_status/2Shows some internal information about load balancing to a specified Host:Port.
spawn_link_worker_process/2Same as spawn_worker_process/2 except the the calling process +
spawn_link_worker_process/1Same as spawn_worker_process/1 except the the calling process is linked to the worker process which is spawned.
spawn_worker_process/2Creates a HTTP client process to the specified Host:Port which +
spawn_link_worker_process/2
spawn_worker_process/1Creates a HTTP client process to the specified Host:Port which is not part of the load balancing pool.
spawn_worker_process/2
start/0Starts the ibrowse process without linking.
start_link/0Starts the ibrowse process linked to the calling process.
stop/0Stop the ibrowse process.
Overviewerlang logo
-

Generated by EDoc, Sep 5 2009, 23:59:48.

+

Generated by EDoc, May 17 2010, 23:21:42.

diff --git a/ebin/ibrowse.app b/ebin/ibrowse.app index ccab68c..dd0e8bc 100644 --- a/ebin/ibrowse.app +++ b/ebin/ibrowse.app @@ -1,6 +1,6 @@ {application, ibrowse, [{description, "HTTP client application"}, - {vsn, "1.5.3"}, + {vsn, "1.6.0"}, {modules, [ ibrowse, ibrowse_http_client, ibrowse_app, diff --git a/src/ibrowse.erl b/src/ibrowse.erl index eaf9aae..c1043ef 100644 --- a/src/ibrowse.erl +++ b/src/ibrowse.erl @@ -6,8 +6,8 @@ %%% Created : 11 Oct 2003 by Chandrashekhar Mullaparthi %%%------------------------------------------------------------------- %% @author Chandrashekhar Mullaparthi -%% @copyright 2005-2009 Chandrashekhar Mullaparthi -%% @version 1.5.4 +%% @copyright 2005-2010 Chandrashekhar Mullaparthi +%% @version 1.6.0 %% @doc The ibrowse application implements an HTTP 1.1 client. This %% module implements the API of the HTTP client. There is one named %% process called 'ibrowse' which assists in load balancing and maintaining configuration. There is one load balancing process per unique webserver. There is @@ -25,18 +25,18 @@ %%

%% %% ibrowse:send_req("http://www.google.com/", [], get, [], -%% [{proxy_user, "XXXXX"}, -%% {proxy_password, "XXXXX"}, -%% {proxy_host, "proxy"}, -%% {proxy_port, 8080}], 1000). +%% [{proxy_user, "XXXXX"}, +%% {proxy_password, "XXXXX"}, +%% {proxy_host, "proxy"}, +%% {proxy_port, 8080}], 1000). %%

%% %%ibrowse:send_req("http://www.erlang.org/download/otp_src_R10B-3.tar.gz", [], get, [], -%% [{proxy_user, "XXXXX"}, -%% {proxy_password, "XXXXX"}, -%% {proxy_host, "proxy"}, -%% {proxy_port, 8080}, -%% {save_response_to_file, true}], 1000). +%% [{proxy_user, "XXXXX"}, +%% {proxy_password, "XXXXX"}, +%% {proxy_host, "proxy"}, +%% {proxy_port, 8080}, +%% {save_response_to_file, true}], 1000). %%

%% %% ibrowse:send_req("http://www.erlang.org", [], head). @@ -52,13 +52,8 @@ %% [{stream_to, self()}]). %% %% -%%

A driver exists which implements URL encoding in C, but the -%% speed achieved using only erlang has been good enough, so the -%% driver isn't actually used.

-module(ibrowse). --vsn('$Id: ibrowse.erl,v 1.8 2009/07/01 22:43:19 chandrusf Exp $ '). - -behaviour(gen_server). %%-------------------------------------------------------------------- %% Include files @@ -70,48 +65,50 @@ %% gen_server callbacks -export([init/1, handle_call/3, handle_cast/2, handle_info/2, - terminate/2, code_change/3]). + terminate/2, code_change/3]). %% API interface -export([ - rescan_config/0, - rescan_config/1, - get_config_value/1, - get_config_value/2, - spawn_worker_process/2, - spawn_link_worker_process/2, - stop_worker_process/1, - send_req/3, - send_req/4, - send_req/5, - send_req/6, - send_req_direct/4, - send_req_direct/5, - send_req_direct/6, - send_req_direct/7, - stream_next/1, - set_max_sessions/3, - set_max_pipeline_size/3, - set_dest/3, - trace_on/0, - trace_off/0, - trace_on/2, - trace_off/2, - all_trace_off/0, - show_dest_status/0, - show_dest_status/2 - ]). + rescan_config/0, + rescan_config/1, + get_config_value/1, + get_config_value/2, + spawn_worker_process/1, + spawn_worker_process/2, + spawn_link_worker_process/1, + spawn_link_worker_process/2, + stop_worker_process/1, + send_req/3, + send_req/4, + send_req/5, + send_req/6, + send_req_direct/4, + send_req_direct/5, + send_req_direct/6, + send_req_direct/7, + stream_next/1, + set_max_sessions/3, + set_max_pipeline_size/3, + set_dest/3, + trace_on/0, + trace_off/0, + trace_on/2, + trace_off/2, + all_trace_off/0, + show_dest_status/0, + show_dest_status/2 + ]). -ifdef(debug). -compile(export_all). -endif. -import(ibrowse_lib, [ - parse_url/1, - get_value/3, - do_trace/2 - ]). - + parse_url/1, + get_value/3, + do_trace/2 + ]). + -record(state, {trace = false}). -include("ibrowse.hrl"). @@ -173,15 +170,15 @@ send_req(Url, Headers, Method, Body) -> %% For a description of SSL Options, look in the ssl manpage. If the %% HTTP Version to use is not specified, the default is 1.1. %%
-%%

The host_header option is useful in the case where ibrowse is +%%

    +%%
  • The host_header option is useful in the case where ibrowse is %% connecting to a component such as stunnel which then sets up a %% secure connection to a webserver. In this case, the URL supplied to %% ibrowse must have the stunnel host/port details, but that won't %% make sense to the destination webserver. This option can then be %% used to specify what should go in the Host header in -%% the request.

    -%%
      +%% the request. %%
    • The stream_to option can be used to have the HTTP %% response streamed to a process as messages as data arrives on the %% socket. If the calling process wishes to control the rate at which @@ -234,6 +231,11 @@ send_req(Url, Headers, Method, Body) -> %%
    • The headers_as_is option is to enable the caller %% to send headers exactly as specified in the request without ibrowse %% adding some of its own. Required for some picky servers apparently.
      • +%% +%%
    • The give_raw_headers option is to enable the +%% caller to get access to the raw status line and raw unparsed +%% headers. Not quite sure why someone would want this, but one of my +%% users asked for it, so here it is.
      • %%
    %% %% @spec send_req(Url::string(), Headers::headerList(), Method::method(), Body::body(), Options::optionList()) -> response() @@ -263,7 +265,8 @@ send_req(Url, Headers, Method, Body) -> %% {connect_timeout, integer()} | %% {socket_options, Sock_opts} | %% {transfer_encoding, {chunked, ChunkSize}} | -%% {headers_as_is, boolean()} +%% {headers_as_is, boolean()} | +%% {give_raw_headers, boolean()} %% %% stream_to() = process() | {process(), once} %% process() = pid() | atom() @@ -285,48 +288,48 @@ send_req(Url, Headers, Method, Body, Options) -> %% Timeout = integer() | infinity send_req(Url, Headers, Method, Body, Options, Timeout) -> case catch parse_url(Url) of - #url{host = Host, - port = Port, - protocol = Protocol} = Parsed_url -> - Lb_pid = case ets:lookup(ibrowse_lb, {Host, Port}) of - [] -> - get_lb_pid(Parsed_url); - [#lb_pid{pid = Lb_pid_1}] -> - Lb_pid_1 - end, - Max_sessions = get_max_sessions(Host, Port, Options), - Max_pipeline_size = get_max_pipeline_size(Host, Port, Options), - Options_1 = merge_options(Host, Port, Options), - {SSLOptions, IsSSL} = - case (Protocol == https) orelse - get_value(is_ssl, Options_1, false) of - false -> {[], false}; - true -> {get_value(ssl_options, Options_1, []), true} - end, - case ibrowse_lb:spawn_connection(Lb_pid, Parsed_url, - Max_sessions, - Max_pipeline_size, - {SSLOptions, IsSSL}) of - {ok, Conn_Pid} -> - do_send_req(Conn_Pid, Parsed_url, Headers, - Method, Body, Options_1, Timeout); - Err -> - Err - end; - Err -> - {error, {url_parsing_failed, Err}} + #url{host = Host, + port = Port, + protocol = Protocol} = Parsed_url -> + Lb_pid = case ets:lookup(ibrowse_lb, {Host, Port}) of + [] -> + get_lb_pid(Parsed_url); + [#lb_pid{pid = Lb_pid_1}] -> + Lb_pid_1 + end, + Max_sessions = get_max_sessions(Host, Port, Options), + Max_pipeline_size = get_max_pipeline_size(Host, Port, Options), + Options_1 = merge_options(Host, Port, Options), + {SSLOptions, IsSSL} = + case (Protocol == https) orelse + get_value(is_ssl, Options_1, false) of + false -> {[], false}; + true -> {get_value(ssl_options, Options_1, []), true} + end, + case ibrowse_lb:spawn_connection(Lb_pid, Parsed_url, + Max_sessions, + Max_pipeline_size, + {SSLOptions, IsSSL}) of + {ok, Conn_Pid} -> + do_send_req(Conn_Pid, Parsed_url, Headers, + Method, Body, Options_1, Timeout); + Err -> + Err + end; + Err -> + {error, {url_parsing_failed, Err}} end. merge_options(Host, Port, Options) -> Config_options = get_config_value({options, Host, Port}, []), lists:foldl( fun({Key, Val}, Acc) -> - case lists:keysearch(Key, 1, Options) of - false -> - [{Key, Val} | Acc]; - _ -> - Acc - end + case lists:keysearch(Key, 1, Options) of + false -> + [{Key, Val} | Acc]; + _ -> + Acc + end end, Options, Config_options). get_lb_pid(Url) -> @@ -334,11 +337,11 @@ get_lb_pid(Url) -> get_max_sessions(Host, Port, Options) -> get_value(max_sessions, Options, - get_config_value({max_sessions, Host, Port}, ?DEF_MAX_SESSIONS)). + get_config_value({max_sessions, Host, Port}, ?DEF_MAX_SESSIONS)). get_max_pipeline_size(Host, Port, Options) -> get_value(max_pipeline_size, Options, - get_config_value({max_pipeline_size, Host, Port}, ?DEF_MAX_PIPELINE_SIZE)). + get_config_value({max_pipeline_size, Host, Port}, ?DEF_MAX_PIPELINE_SIZE)). %% @doc Deprecated. Use set_max_sessions/3 and set_max_pipeline_size/3 %% for achieving the same effect. @@ -368,21 +371,21 @@ set_max_pipeline_size(Host, Port, Max) when is_integer(Max), Max > 0 -> do_send_req(Conn_Pid, Parsed_url, Headers, Method, Body, Options, Timeout) -> case catch ibrowse_http_client:send_req(Conn_Pid, Parsed_url, - Headers, Method, ensure_bin(Body), - Options, Timeout) of - {'EXIT', {timeout, _}} -> - {error, req_timedout}; - {'EXIT', Reason} -> - {error, {'EXIT', Reason}}; - {ok, St_code, Headers, Body} = Ret when is_binary(Body) -> - case get_value(response_format, Options, list) of - list -> - {ok, St_code, Headers, binary_to_list(Body)}; - binary -> - Ret - end; - Ret -> - Ret + Headers, Method, ensure_bin(Body), + Options, Timeout) of + {'EXIT', {timeout, _}} -> + {error, req_timedout}; + {'EXIT', Reason} -> + {error, {'EXIT', Reason}}; + {ok, St_code, Headers, Body} = Ret when is_binary(Body) -> + case get_value(response_format, Options, list) of + list -> + {ok, St_code, Headers, binary_to_list(Body)}; + binary -> + Ret + end; + Ret -> + Ret end. ensure_bin(L) when is_list(L) -> list_to_binary(L); @@ -403,12 +406,21 @@ ensure_bin({Fun, _} = Body) when is_function(Fun) -> Body. %% Note: It is the responsibility of the calling process to control %% pipeline size on such connections. %% +%% @spec spawn_worker_process(Url::string()) -> {ok, pid()} +spawn_worker_process(Url) -> + ibrowse_http_client:start(Url). + %% @spec spawn_worker_process(Host::string(), Port::integer()) -> {ok, pid()} spawn_worker_process(Host, Port) -> ibrowse_http_client:start({Host, Port}). -%% @doc Same as spawn_worker_process/2 except the the calling process +%% @doc Same as spawn_worker_process/1 except the the calling process %% is linked to the worker process which is spawned. +%% @spec spawn_link_worker_process(Url::string()) -> {ok, pid()} +spawn_link_worker_process(Url) -> + ibrowse_http_client:start_link(Url). + +%% @spec spawn_link_worker_process(Host::string(), Port::integer()) -> {ok, pid()} spawn_link_worker_process(Host, Port) -> ibrowse_http_client:start_link({Host, Port}). @@ -438,17 +450,17 @@ send_req_direct(Conn_pid, Url, Headers, Method, Body, Options) -> %% returned by spawn_worker_process/2 or spawn_link_worker_process/2 send_req_direct(Conn_pid, Url, Headers, Method, Body, Options, Timeout) -> case catch parse_url(Url) of - #url{host = Host, - port = Port} = Parsed_url -> - Options_1 = merge_options(Host, Port, Options), - case do_send_req(Conn_pid, Parsed_url, Headers, Method, Body, Options_1, Timeout) of - {error, {'EXIT', {noproc, _}}} -> - {error, worker_is_dead}; - Ret -> - Ret - end; - Err -> - {error, {url_parsing_failed, Err}} + #url{host = Host, + port = Port} = Parsed_url -> + Options_1 = merge_options(Host, Port, Options), + case do_send_req(Conn_pid, Parsed_url, Headers, Method, Body, Options_1, Timeout) of + {error, {'EXIT', {noproc, _}}} -> + {error, worker_is_dead}; + Ret -> + Ret + end; + Err -> + {error, {url_parsing_failed, Err}} end. %% @doc Tell ibrowse to stream the next chunk of data to the @@ -457,11 +469,11 @@ send_req_direct(Conn_pid, Url, Headers, Method, Body, Options, Timeout) -> %% @spec stream_next(Req_id :: req_id()) -> ok | {error, unknown_req_id} stream_next(Req_id) -> case ets:lookup(ibrowse_stream, {req_id_pid, Req_id}) of - [] -> - {error, unknown_req_id}; - [{_, Pid}] -> - catch Pid ! {stream_next, Req_id}, - ok + [] -> + {error, unknown_req_id}; + [{_, Pid}] -> + catch Pid ! {stream_next, Req_id}, + ok end. %% @doc Turn tracing on for the ibrowse process @@ -495,75 +507,75 @@ all_trace_off() -> show_dest_status() -> Dests = lists:filter(fun({lb_pid, {Host, Port}, _}) when is_list(Host), - is_integer(Port) -> - true; - (_) -> - false - end, ets:tab2list(ibrowse_lb)), + is_integer(Port) -> + true; + (_) -> + false + end, ets:tab2list(ibrowse_lb)), All_ets = ets:all(), io:format("~-40.40s | ~-5.5s | ~-10.10s | ~s~n", - ["Server:port", "ETS", "Num conns", "LB Pid"]), + ["Server:port", "ETS", "Num conns", "LB Pid"]), io:format("~80.80.=s~n", [""]), lists:foreach(fun({lb_pid, {Host, Port}, Lb_pid}) -> - case lists:dropwhile( - fun(Tid) -> - ets:info(Tid, owner) /= Lb_pid - end, All_ets) of - [] -> - io:format("~40.40s | ~-5.5s | ~-5.5s | ~s~n", - [Host ++ ":" ++ integer_to_list(Port), - "", - "", - io_lib:format("~p", [Lb_pid])] - ); - [Tid | _] -> - catch ( - begin - Size = ets:info(Tid, size), - io:format("~40.40s | ~-5.5s | ~-5.5s | ~s~n", - [Host ++ ":" ++ integer_to_list(Port), - integer_to_list(Tid), - integer_to_list(Size), - io_lib:format("~p", [Lb_pid])] - ) - end - ) - end - end, Dests). - + case lists:dropwhile( + fun(Tid) -> + ets:info(Tid, owner) /= Lb_pid + end, All_ets) of + [] -> + io:format("~40.40s | ~-5.5s | ~-5.5s | ~s~n", + [Host ++ ":" ++ integer_to_list(Port), + "", + "", + io_lib:format("~p", [Lb_pid])] + ); + [Tid | _] -> + catch ( + begin + Size = ets:info(Tid, size), + io:format("~40.40s | ~-5.5s | ~-5.5s | ~s~n", + [Host ++ ":" ++ integer_to_list(Port), + integer_to_list(Tid), + integer_to_list(Size), + io_lib:format("~p", [Lb_pid])] + ) + end + ) + end + end, Dests). + %% @doc Shows some internal information about load balancing to a %% specified Host:Port. Info about workers spawned using %% spawn_worker_process/2 or spawn_link_worker_process/2 is not %% included. show_dest_status(Host, Port) -> case ets:lookup(ibrowse_lb, {Host, Port}) of - [] -> - no_active_processes; - [#lb_pid{pid = Lb_pid}] -> - io:format("Load Balancer Pid : ~p~n", [Lb_pid]), - io:format("LB process msg q size : ~p~n", [(catch process_info(Lb_pid, message_queue_len))]), - case lists:dropwhile( - fun(Tid) -> - ets:info(Tid, owner) /= Lb_pid - end, ets:all()) of - [] -> - io:format("Couldn't locate ETS table for ~p~n", [Lb_pid]); - [Tid | _] -> - First = ets:first(Tid), - Last = ets:last(Tid), - Size = ets:info(Tid, size), - io:format("LB ETS table id : ~p~n", [Tid]), - io:format("Num Connections : ~p~n", [Size]), - case Size of - 0 -> - ok; - _ -> - {First_p_sz, _} = First, - {Last_p_sz, _} = Last, - io:format("Smallest pipeline : ~1000.p~n", [First_p_sz]), - io:format("Largest pipeline : ~1000.p~n", [Last_p_sz]) - end - end + [] -> + no_active_processes; + [#lb_pid{pid = Lb_pid}] -> + io:format("Load Balancer Pid : ~p~n", [Lb_pid]), + io:format("LB process msg q size : ~p~n", [(catch process_info(Lb_pid, message_queue_len))]), + case lists:dropwhile( + fun(Tid) -> + ets:info(Tid, owner) /= Lb_pid + end, ets:all()) of + [] -> + io:format("Couldn't locate ETS table for ~p~n", [Lb_pid]); + [Tid | _] -> + First = ets:first(Tid), + Last = ets:last(Tid), + Size = ets:info(Tid, size), + io:format("LB ETS table id : ~p~n", [Tid]), + io:format("Num Connections : ~p~n", [Size]), + case Size of + 0 -> + ok; + _ -> + {First_p_sz, _} = First, + {Last_p_sz, _} = Last, + io:format("Smallest pipeline : ~1000.p~n", [First_p_sz]), + io:format("Largest pipeline : ~1000.p~n", [Last_p_sz]) + end + end end. %% @doc Clear current configuration for ibrowse and load from the file @@ -604,40 +616,40 @@ init(_) -> import_config() -> case code:priv_dir(ibrowse) of - {error, _} = Err -> - Err; - PrivDir -> - Filename = filename:join(PrivDir, "ibrowse.conf"), - import_config(Filename) + {error, _} = Err -> + Err; + PrivDir -> + Filename = filename:join(PrivDir, "ibrowse.conf"), + import_config(Filename) end. import_config(Filename) -> case file:consult(Filename) of - {ok, Terms} -> - ets:delete_all_objects(ibrowse_conf), - Fun = fun({dest, Host, Port, MaxSess, MaxPipe, Options}) - when is_list(Host), is_integer(Port), - is_integer(MaxSess), MaxSess > 0, - is_integer(MaxPipe), MaxPipe > 0, is_list(Options) -> - I = [{{max_sessions, Host, Port}, MaxSess}, - {{max_pipeline_size, Host, Port}, MaxPipe}, - {{options, Host, Port}, Options}], - lists:foreach( - fun({X, Y}) -> - ets:insert(ibrowse_conf, - #ibrowse_conf{key = X, - value = Y}) - end, I); - ({K, V}) -> - ets:insert(ibrowse_conf, - #ibrowse_conf{key = K, - value = V}); - (X) -> - io:format("Skipping unrecognised term: ~p~n", [X]) - end, - lists:foreach(Fun, Terms); - Err -> - Err + {ok, Terms} -> + ets:delete_all_objects(ibrowse_conf), + Fun = fun({dest, Host, Port, MaxSess, MaxPipe, Options}) + when is_list(Host), is_integer(Port), + is_integer(MaxSess), MaxSess > 0, + is_integer(MaxPipe), MaxPipe > 0, is_list(Options) -> + I = [{{max_sessions, Host, Port}, MaxSess}, + {{max_pipeline_size, Host, Port}, MaxPipe}, + {{options, Host, Port}, Options}], + lists:foreach( + fun({X, Y}) -> + ets:insert(ibrowse_conf, + #ibrowse_conf{key = X, + value = Y}) + end, I); + ({K, V}) -> + ets:insert(ibrowse_conf, + #ibrowse_conf{key = K, + value = V}); + (X) -> + io:format("Skipping unrecognised term: ~p~n", [X]) + end, + lists:foreach(Fun, Terms); + Err -> + Err end. %% @doc Internal export @@ -648,10 +660,10 @@ get_config_value(Key) -> %% @doc Internal export get_config_value(Key, DefVal) -> case ets:lookup(ibrowse_conf, Key) of - [] -> - DefVal; - [#ibrowse_conf{value = V}] -> - V + [] -> + DefVal; + [#ibrowse_conf{value = V}] -> + V end. set_config_value(Key, Val) -> @@ -712,36 +724,36 @@ handle_info(all_trace_off, State) -> Mspec = [{{ibrowse_conf,{trace,'$1','$2'},true},[],[{{'$1','$2'}}]}], Trace_on_dests = ets:select(ibrowse_conf, Mspec), Fun = fun(#lb_pid{host_port = {H, P}, pid = Pid}, _) -> - case lists:member({H, P}, Trace_on_dests) of - false -> - ok; - true -> - catch Pid ! {trace, false} - end; - (_, Acc) -> - Acc - end, + case lists:member({H, P}, Trace_on_dests) of + false -> + ok; + true -> + catch Pid ! {trace, false} + end; + (_, Acc) -> + Acc + end, ets:foldl(Fun, undefined, ibrowse_lb), ets:select_delete(ibrowse_conf, [{{ibrowse_conf,{trace,'$1','$2'},true},[],['true']}]), {noreply, State}; - + handle_info({trace, Bool}, State) -> put(my_trace_flag, Bool), {noreply, State}; handle_info({trace, Bool, Host, Port}, State) -> Fun = fun(#lb_pid{host_port = {H, P}, pid = Pid}, _) - when H == Host, - P == Port -> - catch Pid ! {trace, Bool}; - (_, Acc) -> - Acc - end, + when H == Host, + P == Port -> + catch Pid ! {trace, Bool}; + (_, Acc) -> + Acc + end, ets:foldl(Fun, undefined, ibrowse_lb), ets:insert(ibrowse_conf, #ibrowse_conf{key = {trace, Host, Port}, - value = Bool}), + value = Bool}), {noreply, State}; - + handle_info(_Info, State) -> {noreply, State}. diff --git a/src/ibrowse_app.erl b/src/ibrowse_app.erl index f5e523c..d3a0f7b 100644 --- a/src/ibrowse_app.erl +++ b/src/ibrowse_app.erl @@ -6,7 +6,6 @@ %%% Created : 15 Oct 2003 by Chandrashekhar Mullaparthi %%%------------------------------------------------------------------- -module(ibrowse_app). --vsn('$Id: ibrowse_app.erl,v 1.1 2005/05/05 22:28:28 chandrusf Exp $ '). -behaviour(application). %%-------------------------------------------------------------------- diff --git a/src/ibrowse_http_client.erl b/src/ibrowse_http_client.erl index 4fdb334..619bf57 100644 --- a/src/ibrowse_http_client.erl +++ b/src/ibrowse_http_client.erl @@ -6,8 +6,6 @@ %%% Created : 11 Oct 2003 by Chandrashekhar Mullaparthi %%%------------------------------------------------------------------- -module(ibrowse_http_client). --vsn('$Id: ibrowse_http_client.erl,v 1.19 2009/07/01 22:43:19 chandrusf Exp $ '). - -behaviour(gen_server). %%-------------------------------------------------------------------- %% Include files @@ -16,11 +14,11 @@ %%-------------------------------------------------------------------- %% External exports -export([ - start_link/1, - start/1, - stop/1, - send_req/7 - ]). + start_link/1, + start/1, + stop/1, + send_req/7 + ]). -ifdef(debug). -compile(export_all). @@ -28,42 +26,45 @@ %% gen_server callbacks -export([ - init/1, - handle_call/3, - handle_cast/2, - handle_info/2, - terminate/2, - code_change/3 - ]). + init/1, + handle_call/3, + handle_cast/2, + handle_info/2, + terminate/2, + code_change/3 + ]). -include("ibrowse.hrl"). --record(state, {host, port, - use_proxy = false, proxy_auth_digest, - ssl_options = [], is_ssl = false, socket, - reqs=queue:new(), cur_req, status=idle, http_status_code, - reply_buffer = <<>>, rep_buf_size=0, streamed_size = 0, - recvd_headers=[], - is_closing, send_timer, content_length, - deleted_crlf = false, transfer_encoding, - chunk_size, chunk_size_buffer = <<>>, recvd_chunk_size, - lb_ets_tid, cur_pipeline_size = 0, prev_req_id - }). +-record(state, {host, port, connect_timeout, + use_proxy = false, proxy_auth_digest, + ssl_options = [], is_ssl = false, socket, + proxy_tunnel_setup = false, + tunnel_setup_queue = [], + reqs=queue:new(), cur_req, status=idle, http_status_code, + reply_buffer = <<>>, rep_buf_size=0, streamed_size = 0, + recvd_headers=[], + status_line, raw_headers, + is_closing, send_timer, content_length, + deleted_crlf = false, transfer_encoding, + chunk_size, chunk_size_buffer = <<>>, recvd_chunk_size, + lb_ets_tid, cur_pipeline_size = 0, prev_req_id + }). -record(request, {url, method, options, from, - stream_to, caller_controls_socket = false, - caller_socket_options = [], - req_id, - stream_chunk_size, - save_response_to_file = false, - tmp_file_name, tmp_file_fd, - response_format}). + stream_to, caller_controls_socket = false, + caller_socket_options = [], + req_id, + stream_chunk_size, + save_response_to_file = false, + tmp_file_name, tmp_file_fd, + response_format}). -import(ibrowse_lib, [ - get_value/2, - get_value/3, - do_trace/2 - ]). + get_value/2, + get_value/3, + do_trace/2 + ]). -define(DEFAULT_STREAM_CHUNK_SIZE, 1024*1024). @@ -103,16 +104,23 @@ send_req(Conn_Pid, Url, Headers, Method, Body, Options, Timeout) -> %%-------------------------------------------------------------------- init({Lb_Tid, #url{host = Host, port = Port}, {SSLOptions, Is_ssl}}) -> State = #state{host = Host, - port = Port, - ssl_options = SSLOptions, - is_ssl = Is_ssl, - lb_ets_tid = Lb_Tid}, + port = Port, + ssl_options = SSLOptions, + is_ssl = Is_ssl, + lb_ets_tid = Lb_Tid}, put(ibrowse_trace_token, [Host, $:, integer_to_list(Port)]), put(my_trace_flag, ibrowse_lib:get_trace_status(Host, Port)), {ok, State}; +init(Url) when is_list(Url) -> + case catch ibrowse_lib:parse_url(Url) of + #url{protocol = Protocol} = Url_rec -> + init({undefined, Url_rec, {[], Protocol == https}}); + {'EXIT', _} -> + {error, invalid_url} + end; init({Host, Port}) -> State = #state{host = Host, - port = Port}, + port = Port}, put(ibrowse_trace_token, [Host, $:, integer_to_list(Port)]), put(my_trace_flag, ibrowse_lib:get_trace_status(Host, Port)), {ok, State}. @@ -133,7 +141,7 @@ handle_call({send_req, _}, _From, #state{is_closing = true} = State) -> {reply, {error, connection_closing}, State}; handle_call({send_req, {Url, Headers, Method, Body, Options, Timeout}}, - From, State) -> + From, State) -> send_req_1(From, Url, Headers, Method, Body, Options, Timeout, State); handle_call(stop, _From, State) -> @@ -169,9 +177,8 @@ handle_info({ssl, _Sock, Data}, State) -> handle_sock_data(Data, State); handle_info({stream_next, Req_id}, #state{socket = Socket, - is_ssl = Is_ssl, - cur_req = #request{req_id = Req_id}} = State) -> - do_setopts(Socket, [{active, once}], Is_ssl), + cur_req = #request{req_id = Req_id}} = State) -> + do_setopts(Socket, [{active, once}], State), {noreply, State}; handle_info({stream_next, _Req_id}, State) -> @@ -197,12 +204,12 @@ handle_info({ssl_error, _Sock}, State) -> handle_info({req_timedout, From}, State) -> case lists:keymember(From, #request.from, queue:to_list(State#state.reqs)) of - false -> - {noreply, State}; - true -> - shutting_down(State), - do_error_reply(State, req_timedout), - {stop, normal, State} + false -> + {noreply, State}; + true -> + shutting_down(State), + do_error_reply(State, req_timedout), + {stop, normal, State} end; handle_info(timeout, State) -> @@ -216,7 +223,7 @@ handle_info({trace, Bool}, State) -> handle_info(Info, State) -> io:format("Unknown message recvd for ~1000.p:~1000.p -> ~p~n", - [State#state.host, State#state.port, Info]), + [State#state.host, State#state.port, Info]), io:format("Recvd unknown message ~p when in state: ~p~n", [Info, State]), {noreply, State}. @@ -252,121 +259,124 @@ handle_sock_data(Data, #state{status=idle}=State) -> handle_sock_data(Data, #state{status = get_header}=State) -> case parse_response(Data, State) of - {error, _Reason} -> - shutting_down(State), - {stop, normal, State}; - State_1 -> - active_once(State_1), - {noreply, State_1, get_inac_timeout(State_1)} + {error, _Reason} -> + shutting_down(State), + {stop, normal, State}; + State_1 -> + active_once(State_1), + set_inac_timer(State_1), + {noreply, State_1} end; handle_sock_data(Data, #state{status = get_body, - content_length = CL, - http_status_code = StatCode, - recvd_headers = Headers, - chunk_size = CSz} = State) -> + content_length = CL, + http_status_code = StatCode, + recvd_headers = Headers, + chunk_size = CSz} = State) -> case (CL == undefined) and (CSz == undefined) of - true -> - case accumulate_response(Data, State) of - {error, Reason} -> - shutting_down(State), - fail_pipelined_requests(State, - {error, {Reason, {stat_code, StatCode}, Headers}}), - {stop, normal, State}; - State_1 -> - active_once(State_1), - {noreply, State_1, get_inac_timeout(State_1)} - end; - _ -> - case parse_11_response(Data, State) of - {error, Reason} -> - shutting_down(State), - fail_pipelined_requests(State, - {error, {Reason, {stat_code, StatCode}, Headers}}), - {stop, normal, State}; - State_1 -> - active_once(State_1), - {noreply, State_1, get_inac_timeout(State_1)} - end + true -> + case accumulate_response(Data, State) of + {error, Reason} -> + shutting_down(State), + fail_pipelined_requests(State, + {error, {Reason, {stat_code, StatCode}, Headers}}), + {stop, normal, State}; + State_1 -> + active_once(State_1), + set_inac_timer(State_1), + {noreply, State_1} + end; + _ -> + case parse_11_response(Data, State) of + {error, Reason} -> + shutting_down(State), + fail_pipelined_requests(State, + {error, {Reason, {stat_code, StatCode}, Headers}}), + {stop, normal, State}; + State_1 -> + active_once(State_1), + set_inac_timer(State_1), + {noreply, State_1} + end end. accumulate_response(Data, - #state{ - cur_req = #request{save_response_to_file = Srtf, - tmp_file_fd = undefined} = CurReq, - http_status_code=[$2 | _]}=State) when Srtf /= false -> + #state{ + cur_req = #request{save_response_to_file = Srtf, + tmp_file_fd = undefined} = CurReq, + http_status_code=[$2 | _]}=State) when Srtf /= false -> TmpFilename = make_tmp_filename(Srtf), case file:open(TmpFilename, [write, delayed_write, raw]) of - {ok, Fd} -> - accumulate_response(Data, State#state{ - cur_req = CurReq#request{ - tmp_file_fd = Fd, - tmp_file_name = TmpFilename}}); - {error, Reason} -> - {error, {file_open_error, Reason}} + {ok, Fd} -> + accumulate_response(Data, State#state{ + cur_req = CurReq#request{ + tmp_file_fd = Fd, + tmp_file_name = TmpFilename}}); + {error, Reason} -> + {error, {file_open_error, Reason}} end; accumulate_response(Data, #state{cur_req = #request{save_response_to_file = Srtf, - tmp_file_fd = Fd}, - transfer_encoding=chunked, - reply_buffer = Reply_buf, - http_status_code=[$2 | _] - } = State) when Srtf /= false -> + tmp_file_fd = Fd}, + transfer_encoding=chunked, + reply_buffer = Reply_buf, + http_status_code=[$2 | _] + } = State) when Srtf /= false -> case file:write(Fd, [Reply_buf, Data]) of - ok -> - State#state{reply_buffer = <<>>}; - {error, Reason} -> - {error, {file_write_error, Reason}} + ok -> + State#state{reply_buffer = <<>>}; + {error, Reason} -> + {error, {file_write_error, Reason}} end; accumulate_response(Data, #state{cur_req = #request{save_response_to_file = Srtf, - tmp_file_fd = Fd}, - reply_buffer = RepBuf, - http_status_code=[$2 | _] - } = State) when Srtf /= false -> + tmp_file_fd = Fd}, + reply_buffer = RepBuf, + http_status_code=[$2 | _] + } = State) when Srtf /= false -> case file:write(Fd, [RepBuf, Data]) of - ok -> - State#state{reply_buffer = <<>>}; - {error, Reason} -> - {error, {file_write_error, Reason}} + ok -> + State#state{reply_buffer = <<>>}; + {error, Reason} -> + {error, {file_write_error, Reason}} end; accumulate_response(<<>>, State) -> State; accumulate_response(Data, #state{reply_buffer = RepBuf, - rep_buf_size = RepBufSize, - streamed_size = Streamed_size, - cur_req = CurReq}=State) -> + rep_buf_size = RepBufSize, + streamed_size = Streamed_size, + cur_req = CurReq}=State) -> #request{stream_to=StreamTo, req_id=ReqId, - stream_chunk_size = Stream_chunk_size, - response_format = Response_format, - caller_controls_socket = Caller_controls_socket} = CurReq, + stream_chunk_size = Stream_chunk_size, + response_format = Response_format, + caller_controls_socket = Caller_controls_socket} = CurReq, RepBuf_1 = concat_binary([RepBuf, Data]), New_data_size = RepBufSize - Streamed_size, case StreamTo of - undefined -> - State#state{reply_buffer = RepBuf_1}; - _ when Caller_controls_socket == true -> - do_interim_reply(StreamTo, Response_format, ReqId, RepBuf_1), - State#state{reply_buffer = <<>>, - streamed_size = Streamed_size + size(RepBuf_1)}; - _ when New_data_size >= Stream_chunk_size -> - {Stream_chunk, Rem_data} = split_binary(RepBuf_1, Stream_chunk_size), - do_interim_reply(StreamTo, Response_format, ReqId, Stream_chunk), - accumulate_response( - Rem_data, - State#state{ - reply_buffer = <<>>, - streamed_size = Streamed_size + Stream_chunk_size}); - _ -> - State#state{reply_buffer = RepBuf_1} + undefined -> + State#state{reply_buffer = RepBuf_1}; + _ when Caller_controls_socket == true -> + do_interim_reply(StreamTo, Response_format, ReqId, RepBuf_1), + State#state{reply_buffer = <<>>, + streamed_size = Streamed_size + size(RepBuf_1)}; + _ when New_data_size >= Stream_chunk_size -> + {Stream_chunk, Rem_data} = split_binary(RepBuf_1, Stream_chunk_size), + do_interim_reply(StreamTo, Response_format, ReqId, Stream_chunk), + accumulate_response( + Rem_data, + State#state{ + reply_buffer = <<>>, + streamed_size = Streamed_size + Stream_chunk_size}); + _ -> + State#state{reply_buffer = RepBuf_1} end. make_tmp_filename(true) -> DownloadDir = ibrowse:get_config_value(download_dir, filename:absname("./")), {A,B,C} = now(), filename:join([DownloadDir, - "ibrowse_tmp_file_"++ - integer_to_list(A) ++ - integer_to_list(B) ++ - integer_to_list(C)]); + "ibrowse_tmp_file_"++ + integer_to_list(A) ++ + integer_to_list(B) ++ + integer_to_list(C)]); make_tmp_filename(File) when is_list(File) -> File. @@ -385,56 +395,72 @@ handle_sock_closed(#state{cur_req=undefined} = State) -> %% Connection-Close header and has closed the socket to indicate end %% of response. There maybe requests pipelined which need a response. handle_sock_closed(#state{reply_buffer = Buf, reqs = Reqs, http_status_code = SC, - is_closing = IsClosing, - cur_req = #request{tmp_file_name=TmpFilename, - tmp_file_fd=Fd} = CurReq, - status = get_body, recvd_headers = Headers}=State) -> + is_closing = IsClosing, + cur_req = #request{tmp_file_name=TmpFilename, + tmp_file_fd=Fd} = CurReq, + status = get_body, + recvd_headers = Headers, + status_line = Status_line, + raw_headers = Raw_headers + }=State) -> #request{from=From, stream_to=StreamTo, req_id=ReqId, - response_format = Resp_format} = CurReq, + response_format = Resp_format, + options = Options} = CurReq, case IsClosing of - true -> - {_, Reqs_1} = queue:out(Reqs), - case TmpFilename of - undefined -> - do_reply(State, From, StreamTo, ReqId, Resp_format, - {ok, SC, Headers, Buf}); - _ -> - file:close(Fd), - do_reply(State, From, StreamTo, ReqId, Resp_format, - {ok, SC, Headers, {file, TmpFilename}}) - end, - do_error_reply(State#state{reqs = Reqs_1}, connection_closed), - State; - _ -> - do_error_reply(State, connection_closed), - State + true -> + {_, Reqs_1} = queue:out(Reqs), + Body = case TmpFilename of + undefined -> + Buf; + _ -> + file:close(Fd), + {file, TmpFilename} + end, + Reply = case get_value(give_raw_headers, Options, false) of + true -> + {ok, Status_line, Raw_headers, Body}; + false -> + {ok, SC, Headers, Buf} + end, + do_reply(State, From, StreamTo, ReqId, Resp_format, Reply), + do_error_reply(State#state{reqs = Reqs_1}, connection_closed), + State; + _ -> + do_error_reply(State, connection_closed), + State end. -do_connect(Host, Port, Options, #state{is_ssl=true, ssl_options=SSLOptions}, Timeout) -> +do_connect(Host, Port, Options, #state{is_ssl = true, + use_proxy = false, + ssl_options = SSLOptions}, + Timeout) -> Caller_socket_options = get_value(socket_options, Options, []), Other_sock_options = filter_sock_options(SSLOptions ++ Caller_socket_options), ssl:connect(Host, Port, - [binary, {nodelay, true}, {active, false} | Other_sock_options], - Timeout); + [binary, {nodelay, true}, {active, false} | Other_sock_options], + Timeout); do_connect(Host, Port, Options, _State, Timeout) -> Caller_socket_options = get_value(socket_options, Options, []), Other_sock_options = filter_sock_options(Caller_socket_options), - gen_tcp:connect(Host, Port, - [binary, {nodelay, true}, {active, false} | Other_sock_options], - Timeout). + gen_tcp:connect(Host, to_integer(Port), + [binary, {nodelay, true}, {active, false} | Other_sock_options], + Timeout). %% We don't want the caller to specify certain options filter_sock_options(Opts) -> lists:filter(fun({active, _}) -> - false; - ({packet, _}) -> - false; - (list) -> - false; - (_) -> - true - end, Opts). - + false; + ({packet, _}) -> + false; + (list) -> + false; + (_) -> + true + end, Opts). + +do_send(Req, #state{socket = Sock, + is_ssl = true, + proxy_tunnel_setup = Pts}) when Pts /= done -> gen_tcp:send(Sock, Req); do_send(Req, #state{socket = Sock, is_ssl = true}) -> ssl:send(Sock, Req); do_send(Req, #state{socket = Sock, is_ssl = false}) -> gen_tcp:send(Sock, Req). @@ -454,171 +480,244 @@ do_send_body(Body, State) -> do_send_body1(Source, Resp, State) -> case Resp of - {ok, Data} -> - do_send(Data, State), - do_send_body({Source}, State); - {ok, Data, New_source_state} -> - do_send(Data, State), - do_send_body({Source, New_source_state}, State); - eof -> - ok; - Err -> - Err + {ok, Data} -> + do_send(Data, State), + do_send_body({Source}, State); + {ok, Data, New_source_state} -> + do_send(Data, State), + do_send_body({Source, New_source_state}, State); + eof -> + ok; + Err -> + Err end. do_close(#state{socket = undefined}) -> ok; +do_close(#state{socket = Sock, + is_ssl = true, + use_proxy = true, + proxy_tunnel_setup = Pts + }) when Pts /= done -> gen_tcp:close(Sock); do_close(#state{socket = Sock, is_ssl = true}) -> ssl:close(Sock); do_close(#state{socket = Sock, is_ssl = false}) -> gen_tcp:close(Sock). active_once(#state{cur_req = #request{caller_controls_socket = true}}) -> ok; -active_once(#state{socket = Socket, is_ssl = Is_ssl}) -> - do_setopts(Socket, [{active, once}], Is_ssl). +active_once(#state{socket = Socket} = State) -> + do_setopts(Socket, [{active, once}], State). do_setopts(_Sock, [], _) -> ok; -do_setopts(Sock, Opts, true) -> ssl:setopts(Sock, Opts); -do_setopts(Sock, Opts, false) -> inet:setopts(Sock, Opts). +do_setopts(Sock, Opts, #state{is_ssl = true, + use_proxy = true, + proxy_tunnel_setup = Pts} + ) when Pts /= done -> inet:setopts(Sock, Opts); +do_setopts(Sock, Opts, #state{is_ssl = true}) -> ssl:setopts(Sock, Opts); +do_setopts(Sock, Opts, _) -> inet:setopts(Sock, Opts). check_ssl_options(Options, State) -> case get_value(is_ssl, Options, false) of - false -> - State; - true -> - State#state{is_ssl=true, ssl_options=get_value(ssl_options, Options)} + false -> + State; + true -> + State#state{is_ssl=true, ssl_options=get_value(ssl_options, Options)} end. send_req_1(From, - #url{host = Host, - port = Port} = Url, - Headers, Method, Body, Options, Timeout, - #state{socket = undefined} = State) -> + #url{host = Host, + port = Port} = Url, + Headers, Method, Body, Options, Timeout, + #state{socket = undefined} = State) -> {Host_1, Port_1, State_1} = - case get_value(proxy_host, Options, false) of - false -> - {Host, Port, State}; - PHost -> - ProxyUser = get_value(proxy_user, Options, []), - ProxyPassword = get_value(proxy_password, Options, []), - Digest = http_auth_digest(ProxyUser, ProxyPassword), - {PHost, get_value(proxy_port, Options, 80), - State#state{use_proxy = true, - proxy_auth_digest = Digest}} - end, + case get_value(proxy_host, Options, false) of + false -> + {Host, Port, State}; + PHost -> + ProxyUser = get_value(proxy_user, Options, []), + ProxyPassword = get_value(proxy_password, Options, []), + Digest = http_auth_digest(ProxyUser, ProxyPassword), + {PHost, get_value(proxy_port, Options, 80), + State#state{use_proxy = true, + proxy_auth_digest = Digest}} + end, State_2 = check_ssl_options(Options, State_1), do_trace("Connecting...~n", []), Start_ts = now(), Conn_timeout = get_value(connect_timeout, Options, Timeout), case do_connect(Host_1, Port_1, Options, State_2, Conn_timeout) of - {ok, Sock} -> - do_trace("Connected!~n", []), - End_ts = now(), - Timeout_1 = case Timeout of - infinity -> - infinity; - _ -> - Timeout - trunc(round(timer:now_diff(End_ts, Start_ts) / 1000)) - end, - State_3 = State_2#state{socket = Sock}, - send_req_1(From, Url, Headers, Method, Body, Options, Timeout_1, State_3); - Err -> - shutting_down(State_2), - do_trace("Error connecting. Reason: ~1000.p~n", [Err]), - gen_server:reply(From, {error, conn_failed}), - {stop, normal, State_2} + {ok, Sock} -> + do_trace("Connected!~n", []), + End_ts = now(), + Timeout_1 = case Timeout of + infinity -> + infinity; + _ -> + Timeout - trunc(round(timer:now_diff(End_ts, Start_ts) / 1000)) + end, + State_3 = State_2#state{socket = Sock, + connect_timeout = Conn_timeout}, + send_req_1(From, Url, Headers, Method, Body, Options, Timeout_1, State_3); + Err -> + shutting_down(State_2), + do_trace("Error connecting. Reason: ~1000.p~n", [Err]), + gen_server:reply(From, {error, conn_failed}), + {stop, normal, State_2} + end; + +%% Send a CONNECT request. +%% Wait for 200 OK +%% Upgrade to SSL connection +%% Then send request + +send_req_1(From, + #url{ + host = Server_host, + port = Server_port + } = Url, + Headers, Method, Body, Options, Timeout, + #state{ + proxy_tunnel_setup = false, + use_proxy = true, + is_ssl = true} = State) -> + NewReq = #request{ + method = connect, + options = Options + }, + State_1 = State#state{reqs=queue:in(NewReq, State#state.reqs)}, + Pxy_auth_headers = maybe_modify_headers(Url, Method, Options, [], State_1), + Path = [Server_host, $:, integer_to_list(Server_port)], + {Req, Body_1} = make_request(connect, Pxy_auth_headers, + Path, Path, + [], Options, State_1), + trace_request(Req), + case do_send(Req, State) of + ok -> + case do_send_body(Body_1, State_1) of + ok -> + active_once(State_1), + Ref = case Timeout of + infinity -> + undefined; + _ -> + erlang:send_after(Timeout, self(), {req_timedout, From}) + end, + State_2 = State_1#state{status = get_header, + cur_req = NewReq, + send_timer = Ref, + proxy_tunnel_setup = in_progress, + tunnel_setup_queue = [{From, Url, Headers, Method, Body, Options, Timeout}]}, + set_inac_timer(State_1), + {noreply, State_2}; + Err -> + shutting_down(State_1), + do_trace("Send failed... Reason: ~p~n", [Err]), + gen_server:reply(From, {error, send_failed}), + {stop, normal, State_1} + end; + Err -> + shutting_down(State_1), + do_trace("Send failed... Reason: ~p~n", [Err]), + gen_server:reply(From, {error, send_failed}), + {stop, normal, State_1} end; + +send_req_1(From, Url, Headers, Method, Body, Options, Timeout, + #state{proxy_tunnel_setup = in_progress, + tunnel_setup_queue = Q} = State) -> + do_trace("Queued SSL request awaiting tunnel setup: ~n" + "URL : ~s~n" + "Method : ~p~n" + "Headers : ~p~n", [Url, Method, Headers]), + {noreply, State#state{tunnel_setup_queue = [{From, Url, Headers, Method, Body, Options, Timeout} | Q]}}; + send_req_1(From, - #url{abspath = AbsPath, - path = RelPath} = Url, - Headers, Method, Body, Options, Timeout, - #state{status = Status, - socket = Socket, - is_ssl = Is_ssl} = State) -> + #url{abspath = AbsPath, + path = RelPath} = Url, + Headers, Method, Body, Options, Timeout, + #state{status = Status, + socket = Socket, + is_ssl = Is_ssl} = State) -> ReqId = make_req_id(), Resp_format = get_value(response_format, Options, list), Caller_socket_options = get_value(socket_options, Options, []), {StreamTo, Caller_controls_socket} = - case get_value(stream_to, Options, undefined) of - {Caller, once} when is_pid(Caller) or - is_atom(Caller) -> - Async_pid_rec = {{req_id_pid, ReqId}, self()}, - true = ets:insert(ibrowse_stream, Async_pid_rec), - {Caller, true}; - undefined -> - {undefined, false}; - Caller when is_pid(Caller) or - is_atom(Caller) -> - {Caller, false}; - Stream_to_inv -> - exit({invalid_option, {stream_to, Stream_to_inv}}) - end, + case get_value(stream_to, Options, undefined) of + {Caller, once} when is_pid(Caller) or + is_atom(Caller) -> + Async_pid_rec = {{req_id_pid, ReqId}, self()}, + true = ets:insert(ibrowse_stream, Async_pid_rec), + {Caller, true}; + undefined -> + {undefined, false}; + Caller when is_pid(Caller) or + is_atom(Caller) -> + {Caller, false}; + Stream_to_inv -> + exit({invalid_option, {stream_to, Stream_to_inv}}) + end, SaveResponseToFile = get_value(save_response_to_file, Options, false), NewReq = #request{url = Url, - method = Method, - stream_to = StreamTo, - caller_controls_socket = Caller_controls_socket, - caller_socket_options = Caller_socket_options, - options = Options, - req_id = ReqId, - save_response_to_file = SaveResponseToFile, - stream_chunk_size = get_stream_chunk_size(Options), - response_format = Resp_format, - from = From}, + method = Method, + stream_to = StreamTo, + caller_controls_socket = Caller_controls_socket, + caller_socket_options = Caller_socket_options, + options = Options, + req_id = ReqId, + save_response_to_file = SaveResponseToFile, + stream_chunk_size = get_stream_chunk_size(Options), + response_format = Resp_format, + from = From}, State_1 = State#state{reqs=queue:in(NewReq, State#state.reqs)}, - Headers_1 = maybe_modify_headers(Url, Options, Headers, State_1), + Headers_1 = maybe_modify_headers(Url, Method, Options, Headers, State_1), {Req, Body_1} = make_request(Method, - Headers_1, - AbsPath, RelPath, Body, Options, State_1#state.use_proxy), - case get(my_trace_flag) of - true -> - %%Avoid the binary operations if trace is not on... - NReq = binary_to_list(list_to_binary(Req)), - do_trace("Sending request: ~n" - "--- Request Begin ---~n~s~n" - "--- Request End ---~n", [NReq]); - _ -> ok - end, + Headers_1, + AbsPath, RelPath, Body, Options, State_1), + trace_request(Req), do_setopts(Socket, Caller_socket_options, Is_ssl), case do_send(Req, State_1) of - ok -> - case do_send_body(Body_1, State_1) of - ok -> - State_2 = inc_pipeline_counter(State_1), - active_once(State_2), - Ref = case Timeout of - infinity -> - undefined; - _ -> - erlang:send_after(Timeout, self(), {req_timedout, From}) - end, - State_3 = case Status of - idle -> - State_2#state{status = get_header, - cur_req = NewReq, - send_timer = Ref}; - _ -> - State_2#state{send_timer = Ref} - end, - case StreamTo of - undefined -> - ok; - _ -> - gen_server:reply(From, {ibrowse_req_id, ReqId}) - end, - {noreply, State_3, get_inac_timeout(State_3)}; - Err -> - shutting_down(State_1), - do_trace("Send failed... Reason: ~p~n", [Err]), - gen_server:reply(From, {error, send_failed}), - {stop, normal, State_1} - end; - Err -> - shutting_down(State_1), - do_trace("Send failed... Reason: ~p~n", [Err]), - gen_server:reply(From, {error, send_failed}), - {stop, normal, State_1} + ok -> + case do_send_body(Body_1, State_1) of + ok -> + State_2 = inc_pipeline_counter(State_1), + active_once(State_2), + Ref = case Timeout of + infinity -> + undefined; + _ -> + erlang:send_after(Timeout, self(), {req_timedout, From}) + end, + State_3 = case Status of + idle -> + State_2#state{status = get_header, + cur_req = NewReq, + send_timer = Ref}; + _ -> + State_2#state{send_timer = Ref} + end, + case StreamTo of + undefined -> + ok; + _ -> + gen_server:reply(From, {ibrowse_req_id, ReqId}) + end, + set_inac_timer(State_1), + {noreply, State_3}; + Err -> + shutting_down(State_1), + do_trace("Send failed... Reason: ~p~n", [Err]), + gen_server:reply(From, {error, send_failed}), + {stop, normal, State_1} + end; + Err -> + shutting_down(State_1), + do_trace("Send failed... Reason: ~p~n", [Err]), + gen_server:reply(From, {error, send_failed}), + {stop, normal, State_1} end. +maybe_modify_headers(#url{}, connect, _, Headers, State) -> + add_proxy_auth_headers(State, Headers); maybe_modify_headers(#url{host = Host, port = Port} = Url, + _Method, Options, Headers, State) -> case get_value(headers_as_is, Options, false) of false -> @@ -638,30 +737,29 @@ maybe_modify_headers(#url{host = Host, port = Port} = Url, end. add_auth_headers(#url{username = User, - password = UPw}, - Options, - Headers, - #state{use_proxy = UseProxy, - proxy_auth_digest = ProxyAuthDigest}) -> + password = UPw}, + Options, + Headers, + State) -> Headers_1 = case User of - undefined -> - case get_value(basic_auth, Options, undefined) of - undefined -> - Headers; - {U,P} -> - [{"Authorization", ["Basic ", http_auth_digest(U, P)]} | Headers] - end; - _ -> - [{"Authorization", ["Basic ", http_auth_digest(User, UPw)]} | Headers] - end, - case UseProxy of - false -> - Headers_1; - true when ProxyAuthDigest == [] -> - Headers_1; - true -> - [{"Proxy-Authorization", ["Basic ", ProxyAuthDigest]} | Headers_1] - end. + undefined -> + case get_value(basic_auth, Options, undefined) of + undefined -> + Headers; + {U,P} -> + [{"Authorization", ["Basic ", http_auth_digest(U, P)]} | Headers] + end; + _ -> + [{"Authorization", ["Basic ", http_auth_digest(User, UPw)]} | Headers] + end, + add_proxy_auth_headers(State, Headers_1). + +add_proxy_auth_headers(#state{use_proxy = false}, Headers) -> + Headers; +add_proxy_auth_headers(#state{proxy_auth_digest = []}, Headers) -> + Headers; +add_proxy_auth_headers(#state{proxy_auth_digest = Auth_digest}, Headers) -> + [{"Proxy-Authorization", ["Basic ", Auth_digest]} | Headers]. http_auth_digest([], []) -> []; @@ -688,41 +786,42 @@ e(62) -> $+; e(63) -> $/; e(X) -> exit({bad_encode_base64_token, X}). -make_request(Method, Headers, AbsPath, RelPath, Body, Options, UseProxy) -> +make_request(Method, Headers, AbsPath, RelPath, Body, Options, + #state{use_proxy = UseProxy}) -> HttpVsn = http_vsn_string(get_value(http_vsn, Options, {1,1})), Headers_1 = - case get_value(content_length, Headers, false) of - false when (Body == []) or - (Body == <<>>) or - is_tuple(Body) or - is_function(Body) -> - Headers; - false when is_binary(Body) -> - [{"content-length", integer_to_list(byte_size(Body))} | Headers]; - false -> - [{"content-length", integer_to_list(length(Body))} | Headers]; - _ -> - Headers - end, + case get_value(content_length, Headers, false) of + false when (Body == []) or + (Body == <<>>) or + is_tuple(Body) or + is_function(Body) -> + Headers; + false when is_binary(Body) -> + [{"content-length", integer_to_list(size(Body))} | Headers]; + false -> + [{"content-length", integer_to_list(length(Body))} | Headers]; + _ -> + Headers + end, {Headers_2, Body_1} = - case get_value(transfer_encoding, Options, false) of - false -> - {Headers_1, Body}; - {chunked, ChunkSize} -> - {[{X, Y} || {X, Y} <- Headers_1, - X /= "Content-Length", - X /= "content-length", - X /= content_length] ++ - [{"Transfer-Encoding", "chunked"}], - chunk_request_body(Body, ChunkSize)} - end, + case get_value(transfer_encoding, Options, false) of + false -> + {Headers_1, Body}; + {chunked, ChunkSize} -> + {[{X, Y} || {X, Y} <- Headers_1, + X /= "Content-Length", + X /= "content-length", + X /= content_length] ++ + [{"Transfer-Encoding", "chunked"}], + chunk_request_body(Body, ChunkSize)} + end, Headers_3 = cons_headers(Headers_2), Uri = case get_value(use_absolute_uri, Options, false) or UseProxy of - true -> - AbsPath; - false -> - RelPath - end, + true -> + AbsPath; + false -> + RelPath + end, {[method(Method), " ", Uri, " ", HttpVsn, crnl(), Headers_3, crnl()], Body_1}. http_vsn_string({0,9}) -> "HTTP/0.9"; @@ -735,7 +834,7 @@ cons_headers([], Acc) -> encode_headers(Acc); cons_headers([{basic_auth, {U,P}} | T], Acc) -> cons_headers(T, [{"Authorization", - ["Basic ", ibrowse_lib:encode_base64(U++":"++P)]} | Acc]); + ["Basic ", ibrowse_lib:encode_base64(U++":"++P)]} | Acc]); cons_headers([{cookie, Cookie} | T], Acc) -> cons_headers(T, [{"Cookie", Cookie} | Acc]); cons_headers([{content_length, L} | T], Acc) -> @@ -762,26 +861,27 @@ chunk_request_body(Body, ChunkSize) -> chunk_request_body(Body, _ChunkSize, Acc) when Body == <<>>; Body == [] -> LastChunk = "0\r\n", lists:reverse(["\r\n", LastChunk | Acc]); -chunk_request_body(Body, ChunkSize, Acc) when byte_size(Body) >= ChunkSize -> +chunk_request_body(Body, ChunkSize, Acc) when is_binary(Body), + size(Body) >= ChunkSize -> <> = Body, Chunk = [ibrowse_lib:dec2hex(4, ChunkSize),"\r\n", - ChunkBody, "\r\n"], + ChunkBody, "\r\n"], chunk_request_body(Rest, ChunkSize, [Chunk | Acc]); chunk_request_body(Body, _ChunkSize, Acc) when is_binary(Body) -> - BodySize = byte_size(Body), + BodySize = size(Body), Chunk = [ibrowse_lib:dec2hex(4, BodySize),"\r\n", - Body, "\r\n"], + Body, "\r\n"], LastChunk = "0\r\n", lists:reverse(["\r\n", LastChunk, Chunk | Acc]); chunk_request_body(Body, ChunkSize, Acc) when length(Body) >= ChunkSize -> {ChunkBody, Rest} = split_list_at(Body, ChunkSize), Chunk = [ibrowse_lib:dec2hex(4, ChunkSize),"\r\n", - ChunkBody, "\r\n"], + ChunkBody, "\r\n"], chunk_request_body(Rest, ChunkSize, [Chunk | Acc]); chunk_request_body(Body, _ChunkSize, Acc) when is_list(Body) -> BodySize = length(Body), Chunk = [ibrowse_lib:dec2hex(4, BodySize),"\r\n", - Body, "\r\n"], + Body, "\r\n"], LastChunk = "0\r\n", lists:reverse(["\r\n", LastChunk, Chunk | Acc]). @@ -789,114 +889,172 @@ chunk_request_body(Body, _ChunkSize, Acc) when is_list(Body) -> parse_response(_Data, #state{cur_req = undefined}=State) -> State#state{status = idle}; parse_response(Data, #state{reply_buffer = Acc, reqs = Reqs, - cur_req = CurReq} = State) -> + cur_req = CurReq} = State) -> #request{from=From, stream_to=StreamTo, req_id=ReqId, - method=Method, response_format = Resp_format} = CurReq, + method=Method, response_format = Resp_format, + options = Options + } = CurReq, MaxHeaderSize = ibrowse:get_config_value(max_headers_size, infinity), case scan_header(Acc, Data) of - {yes, Headers, Data_1} -> - do_trace("Recvd Header Data -> ~s~n----~n", [Headers]), - do_trace("Recvd headers~n--- Headers Begin ---~n~s~n--- Headers End ---~n~n", [Headers]), - {HttpVsn, StatCode, Headers_1} = parse_headers(Headers), - do_trace("HttpVsn: ~p StatusCode: ~p Headers_1 -> ~1000.p~n", [HttpVsn, StatCode, Headers_1]), - LCHeaders = [{to_lower(X), Y} || {X,Y} <- Headers_1], - ConnClose = to_lower(get_value("connection", LCHeaders, "false")), - IsClosing = is_connection_closing(HttpVsn, ConnClose), - case IsClosing of - true -> + {yes, Headers, Data_1} -> + do_trace("Recvd Header Data -> ~s~n----~n", [Headers]), + do_trace("Recvd headers~n--- Headers Begin ---~n~s~n--- Headers End ---~n~n", [Headers]), + {HttpVsn, StatCode, Headers_1, Status_line, Raw_headers} = parse_headers(Headers), + do_trace("HttpVsn: ~p StatusCode: ~p Headers_1 -> ~1000.p~n", [HttpVsn, StatCode, Headers_1]), + LCHeaders = [{to_lower(X), Y} || {X,Y} <- Headers_1], + ConnClose = to_lower(get_value("connection", LCHeaders, "false")), + IsClosing = is_connection_closing(HttpVsn, ConnClose), + case IsClosing of + true -> shutting_down(State); - false -> - ok - end, - State_1 = State#state{recvd_headers=Headers_1, status=get_body, - reply_buffer = <<>>, - http_status_code=StatCode, is_closing=IsClosing}, - put(conn_close, ConnClose), - TransferEncoding = to_lower(get_value("transfer-encoding", LCHeaders, "false")), - case get_value("content-length", LCHeaders, undefined) of - _ when Method == head -> - {_, Reqs_1} = queue:out(Reqs), - send_async_headers(ReqId, StreamTo, StatCode, Headers_1), - State_1_1 = do_reply(State_1, From, StreamTo, ReqId, Resp_format, - {ok, StatCode, Headers_1, []}), - cancel_timer(State_1_1#state.send_timer, {eat_message, {req_timedout, From}}), - State_2 = reset_state(State_1_1), - State_3 = set_cur_request(State_2#state{reqs = Reqs_1}), - parse_response(Data_1, State_3); - _ when hd(StatCode) =:= $1 -> - %% No message body is expected. Server may send - %% one or more 1XX responses before a proper - %% response. - send_async_headers(ReqId, StreamTo, StatCode, Headers_1), - do_trace("Recvd a status code of ~p. Ignoring and waiting for a proper response~n", [StatCode]), - parse_response(Data_1, State_1#state{recvd_headers = [], - status = get_header}); - _ when StatCode =:= "204"; - StatCode =:= "304" -> - %% No message body is expected for these Status Codes. - %% RFC2616 - Sec 4.4 - {_, Reqs_1} = queue:out(Reqs), - send_async_headers(ReqId, StreamTo, StatCode, Headers_1), - State_1_1 = do_reply(State_1, From, StreamTo, ReqId, Resp_format, - {ok, StatCode, Headers_1, []}), - cancel_timer(State_1_1#state.send_timer, {eat_message, {req_timedout, From}}), - State_2 = reset_state(State_1_1), - State_3 = set_cur_request(State_2#state{reqs = Reqs_1}), - parse_response(Data_1, State_3); - _ when TransferEncoding =:= "chunked" -> - do_trace("Chunked encoding detected...~n",[]), - send_async_headers(ReqId, StreamTo, StatCode, Headers_1), - case parse_11_response(Data_1, State_1#state{transfer_encoding=chunked, - chunk_size=chunk_start, - reply_buffer = <<>>}) of - {error, Reason} -> - fail_pipelined_requests(State_1, - {error, {Reason, - {stat_code, StatCode}, Headers_1}}), - {error, Reason}; - State_2 -> - State_2 - end; - undefined when HttpVsn =:= "HTTP/1.0"; - ConnClose =:= "close" -> - send_async_headers(ReqId, StreamTo, StatCode, Headers_1), - State_1#state{reply_buffer = Data_1}; - undefined -> - fail_pipelined_requests(State_1, - {error, {content_length_undefined, - {stat_code, StatCode}, Headers}}), - {error, content_length_undefined}; - V -> - case catch list_to_integer(V) of - V_1 when is_integer(V_1), V_1 >= 0 -> - send_async_headers(ReqId, StreamTo, StatCode, Headers_1), - do_trace("Recvd Content-Length of ~p~n", [V_1]), - State_2 = State_1#state{rep_buf_size=0, - reply_buffer = <<>>, - content_length=V_1}, - case parse_11_response(Data_1, State_2) of - {error, Reason} -> - fail_pipelined_requests(State_1, - {error, {Reason, - {stat_code, StatCode}, Headers_1}}), - {error, Reason}; - State_3 -> - State_3 - end; - _ -> - fail_pipelined_requests(State_1, - {error, {content_length_undefined, - {stat_code, StatCode}, Headers}}), - {error, content_length_undefined} - end - end; - {no, Acc_1} when MaxHeaderSize == infinity -> - State#state{reply_buffer = Acc_1}; - {no, Acc_1} when size(Acc_1) < MaxHeaderSize -> - State#state{reply_buffer = Acc_1}; - {no, _Acc_1} -> - fail_pipelined_requests(State, {error, max_headers_size_exceeded}), - {error, max_headers_size_exceeded} + false -> + ok + end, + Give_raw_headers = get_value(give_raw_headers, Options, false), + State_1 = case Give_raw_headers of + true -> + State#state{recvd_headers=Headers_1, status=get_body, + reply_buffer = <<>>, + status_line = Status_line, + raw_headers = Raw_headers, + http_status_code=StatCode, is_closing=IsClosing}; + false -> + State#state{recvd_headers=Headers_1, status=get_body, + reply_buffer = <<>>, + http_status_code=StatCode, is_closing=IsClosing} + end, + put(conn_close, ConnClose), + TransferEncoding = to_lower(get_value("transfer-encoding", LCHeaders, "false")), + case get_value("content-length", LCHeaders, undefined) of + _ when Method == connect, + hd(StatCode) == $2 -> + cancel_timer(State#state.send_timer), + {_, Reqs_1} = queue:out(Reqs), + upgrade_to_ssl(set_cur_request(State#state{reqs = Reqs_1, + recvd_headers = [], + status = idle + })); + _ when Method == connect -> + {_, Reqs_1} = queue:out(Reqs), + do_error_reply(State#state{reqs = Reqs_1}, + {error, proxy_tunnel_failed}), + {error, proxy_tunnel_failed}; + _ when Method == head -> + {_, Reqs_1} = queue:out(Reqs), + send_async_headers(ReqId, StreamTo, Give_raw_headers, State_1), + State_1_1 = do_reply(State_1, From, StreamTo, ReqId, Resp_format, + {ok, StatCode, Headers_1, []}), + cancel_timer(State_1_1#state.send_timer, {eat_message, {req_timedout, From}}), + State_2 = reset_state(State_1_1), + State_3 = set_cur_request(State_2#state{reqs = Reqs_1}), + parse_response(Data_1, State_3); + _ when hd(StatCode) =:= $1 -> + %% No message body is expected. Server may send + %% one or more 1XX responses before a proper + %% response. + send_async_headers(ReqId, StreamTo, Give_raw_headers, State_1), + do_trace("Recvd a status code of ~p. Ignoring and waiting for a proper response~n", [StatCode]), + parse_response(Data_1, State_1#state{recvd_headers = [], + status = get_header}); + _ when StatCode =:= "204"; + StatCode =:= "304" -> + %% No message body is expected for these Status Codes. + %% RFC2616 - Sec 4.4 + {_, Reqs_1} = queue:out(Reqs), + send_async_headers(ReqId, StreamTo, Give_raw_headers, State_1), + State_1_1 = do_reply(State_1, From, StreamTo, ReqId, Resp_format, + {ok, StatCode, Headers_1, []}), + cancel_timer(State_1_1#state.send_timer, {eat_message, {req_timedout, From}}), + State_2 = reset_state(State_1_1), + State_3 = set_cur_request(State_2#state{reqs = Reqs_1}), + parse_response(Data_1, State_3); + _ when TransferEncoding =:= "chunked" -> + do_trace("Chunked encoding detected...~n",[]), + send_async_headers(ReqId, StreamTo, Give_raw_headers, State_1), + case parse_11_response(Data_1, State_1#state{transfer_encoding=chunked, + chunk_size=chunk_start, + reply_buffer = <<>>}) of + {error, Reason} -> + fail_pipelined_requests(State_1, + {error, {Reason, + {stat_code, StatCode}, Headers_1}}), + {error, Reason}; + State_2 -> + State_2 + end; + undefined when HttpVsn =:= "HTTP/1.0"; + ConnClose =:= "close" -> + send_async_headers(ReqId, StreamTo, Give_raw_headers, State_1), + State_1#state{reply_buffer = Data_1}; + undefined -> + fail_pipelined_requests(State_1, + {error, {content_length_undefined, + {stat_code, StatCode}, Headers}}), + {error, content_length_undefined}; + V -> + case catch list_to_integer(V) of + V_1 when is_integer(V_1), V_1 >= 0 -> + send_async_headers(ReqId, StreamTo, Give_raw_headers, State_1), + do_trace("Recvd Content-Length of ~p~n", [V_1]), + State_2 = State_1#state{rep_buf_size=0, + reply_buffer = <<>>, + content_length=V_1}, + case parse_11_response(Data_1, State_2) of + {error, Reason} -> + fail_pipelined_requests(State_1, + {error, {Reason, + {stat_code, StatCode}, Headers_1}}), + {error, Reason}; + State_3 -> + State_3 + end; + _ -> + fail_pipelined_requests(State_1, + {error, {content_length_undefined, + {stat_code, StatCode}, Headers}}), + {error, content_length_undefined} + end + end; + {no, Acc_1} when MaxHeaderSize == infinity -> + State#state{reply_buffer = Acc_1}; + {no, Acc_1} when size(Acc_1) < MaxHeaderSize -> + State#state{reply_buffer = Acc_1}; + {no, _Acc_1} -> + fail_pipelined_requests(State, {error, max_headers_size_exceeded}), + {error, max_headers_size_exceeded} + end. + +upgrade_to_ssl(#state{socket = Socket, + connect_timeout = Conn_timeout, + ssl_options = Ssl_options, + tunnel_setup_queue = Q} = State) -> + case ssl:connect(Socket, Ssl_options, Conn_timeout) of + {ok, Ssl_socket} -> + do_trace("Upgraded to SSL socket!!~n", []), + State_1 = State#state{socket = Ssl_socket, + proxy_tunnel_setup = done}, + send_queued_requests(lists:reverse(Q), State_1); + Err -> + do_trace("Upgrade to SSL socket failed. Reson: ~p~n", [Err]), + do_error_reply(State, {error, send_failed}), + {error, send_failed} + end. + +send_queued_requests([], State) -> + do_trace("Sent all queued requests via SSL connection~n", []), + State#state{tunnel_setup_queue = done}; +send_queued_requests([{From, Url, Headers, Method, Body, Options, Timeout} | Q], + State) -> + case send_req_1(From, Url, Headers, Method, Body, Options, Timeout, State) of + {noreply, State_1} -> + send_queued_requests(Q, State_1); + _ -> + do_trace("Error sending queued SSL request: ~n" + "URL : ~s~n" + "Method : ~p~n" + "Headers : ~p~n", [Url, Method, Headers]), + do_error_reply(State, {error, send_failed}), + {error, send_failed} end. is_connection_closing("HTTP/0.9", _) -> true; @@ -906,197 +1064,215 @@ is_connection_closing(_, _) -> false. %% This clause determines the chunk size when given data from the beginning of the chunk parse_11_response(DataRecvd, - #state{transfer_encoding = chunked, - chunk_size = chunk_start, - chunk_size_buffer = Chunk_sz_buf - } = State) -> + #state{transfer_encoding = chunked, + chunk_size = chunk_start, + chunk_size_buffer = Chunk_sz_buf + } = State) -> case scan_crlf(Chunk_sz_buf, DataRecvd) of - {yes, ChunkHeader, Data_1} -> - ChunkSize = parse_chunk_header(ChunkHeader), - %% - %% Do we have to preserve the chunk encoding when - %% streaming? NO. This should be transparent to the client - %% process. Chunked encoding was only introduced to make - %% it efficient for the server. - %% - RemLen = size(Data_1), - do_trace("Determined chunk size: ~p. Already recvd: ~p~n", - [ChunkSize, RemLen]), - parse_11_response(Data_1, State#state{chunk_size_buffer = <<>>, - deleted_crlf = true, - recvd_chunk_size = 0, - chunk_size = ChunkSize}); - {no, Data_1} -> - State#state{chunk_size_buffer = Data_1} + {yes, ChunkHeader, Data_1} -> + ChunkSize = parse_chunk_header(ChunkHeader), + %% + %% Do we have to preserve the chunk encoding when + %% streaming? NO. This should be transparent to the client + %% process. Chunked encoding was only introduced to make + %% it efficient for the server. + %% + RemLen = size(Data_1), + do_trace("Determined chunk size: ~p. Already recvd: ~p~n", + [ChunkSize, RemLen]), + parse_11_response(Data_1, State#state{chunk_size_buffer = <<>>, + deleted_crlf = true, + recvd_chunk_size = 0, + chunk_size = ChunkSize}); + {no, Data_1} -> + State#state{chunk_size_buffer = Data_1} end; %% This clause is to remove the CRLF between two chunks %% parse_11_response(DataRecvd, - #state{transfer_encoding = chunked, - chunk_size = tbd, - chunk_size_buffer = Buf}=State) -> + #state{transfer_encoding = chunked, + chunk_size = tbd, + chunk_size_buffer = Buf}=State) -> case scan_crlf(Buf, DataRecvd) of - {yes, _, NextChunk} -> - State_1 = State#state{chunk_size = chunk_start, - chunk_size_buffer = <<>>, - deleted_crlf = true}, - parse_11_response(NextChunk, State_1); - {no, Data_1} -> - State#state{chunk_size_buffer = Data_1} + {yes, _, NextChunk} -> + State_1 = State#state{chunk_size = chunk_start, + chunk_size_buffer = <<>>, + deleted_crlf = true}, + parse_11_response(NextChunk, State_1); + {no, Data_1} -> + State#state{chunk_size_buffer = Data_1} end; %% This clause deals with the end of a chunked transfer. ibrowse does %% not support Trailers in the Chunked Transfer encoding. Any trailer %% received is silently discarded. parse_11_response(DataRecvd, - #state{transfer_encoding = chunked, chunk_size = 0, - cur_req = CurReq, - deleted_crlf = DelCrlf, - chunk_size_buffer = Trailer, reqs = Reqs}=State) -> + #state{transfer_encoding = chunked, chunk_size = 0, + cur_req = CurReq, + deleted_crlf = DelCrlf, + chunk_size_buffer = Trailer, reqs = Reqs}=State) -> do_trace("Detected end of chunked transfer...~n", []), DataRecvd_1 = case DelCrlf of - false -> - DataRecvd; - true -> - <<$\r, $\n, DataRecvd/binary>> + false -> + DataRecvd; + true -> + <<$\r, $\n, DataRecvd/binary>> end, case scan_header(Trailer, DataRecvd_1) of - {yes, _TEHeaders, Rem} -> - {_, Reqs_1} = queue:out(Reqs), - State_1 = handle_response(CurReq, State#state{reqs = Reqs_1}), - parse_response(Rem, reset_state(State_1)); - {no, Rem} -> - State#state{chunk_size_buffer = Rem, deleted_crlf = false} + {yes, _TEHeaders, Rem} -> + {_, Reqs_1} = queue:out(Reqs), + State_1 = handle_response(CurReq, State#state{reqs = Reqs_1}), + parse_response(Rem, reset_state(State_1)); + {no, Rem} -> + State#state{chunk_size_buffer = Rem, deleted_crlf = false} end; %% This clause extracts a chunk, given the size. parse_11_response(DataRecvd, - #state{transfer_encoding = chunked, - chunk_size = CSz, - recvd_chunk_size = Recvd_csz, - rep_buf_size = RepBufSz} = State) -> + #state{transfer_encoding = chunked, + chunk_size = CSz, + recvd_chunk_size = Recvd_csz, + rep_buf_size = RepBufSz} = State) -> NeedBytes = CSz - Recvd_csz, DataLen = size(DataRecvd), do_trace("Recvd more data: size: ~p. NeedBytes: ~p~n", [DataLen, NeedBytes]), case DataLen >= NeedBytes of - true -> - {RemChunk, RemData} = split_binary(DataRecvd, NeedBytes), - do_trace("Recvd another chunk...~n", []), - do_trace("RemData -> ~p~n", [RemData]), - case accumulate_response(RemChunk, State) of - {error, Reason} -> - do_trace("Error accumulating response --> ~p~n", [Reason]), - {error, Reason}; - #state{} = State_1 -> - State_2 = State_1#state{chunk_size=tbd}, - parse_11_response(RemData, State_2) - end; - false -> - accumulate_response(DataRecvd, - State#state{rep_buf_size = RepBufSz + DataLen, - recvd_chunk_size = Recvd_csz + DataLen}) + true -> + {RemChunk, RemData} = split_binary(DataRecvd, NeedBytes), + do_trace("Recvd another chunk...~n", []), + do_trace("RemData -> ~p~n", [RemData]), + case accumulate_response(RemChunk, State) of + {error, Reason} -> + do_trace("Error accumulating response --> ~p~n", [Reason]), + {error, Reason}; + #state{} = State_1 -> + State_2 = State_1#state{chunk_size=tbd}, + parse_11_response(RemData, State_2) + end; + false -> + accumulate_response(DataRecvd, + State#state{rep_buf_size = RepBufSz + DataLen, + recvd_chunk_size = Recvd_csz + DataLen}) end; %% This clause to extract the body when Content-Length is specified parse_11_response(DataRecvd, - #state{content_length=CL, rep_buf_size=RepBufSz, - reqs=Reqs}=State) -> + #state{content_length=CL, rep_buf_size=RepBufSz, + reqs=Reqs}=State) -> NeedBytes = CL - RepBufSz, DataLen = size(DataRecvd), case DataLen >= NeedBytes of - true -> - {RemBody, Rem} = split_binary(DataRecvd, NeedBytes), - {_, Reqs_1} = queue:out(Reqs), - State_1 = accumulate_response(RemBody, State), - State_2 = handle_response(State_1#state.cur_req, State_1#state{reqs=Reqs_1}), - State_3 = reset_state(State_2), - parse_response(Rem, State_3); - false -> - accumulate_response(DataRecvd, State#state{rep_buf_size = (RepBufSz+DataLen)}) + true -> + {RemBody, Rem} = split_binary(DataRecvd, NeedBytes), + {_, Reqs_1} = queue:out(Reqs), + State_1 = accumulate_response(RemBody, State), + State_2 = handle_response(State_1#state.cur_req, State_1#state{reqs=Reqs_1}), + State_3 = reset_state(State_2), + parse_response(Rem, State_3); + false -> + accumulate_response(DataRecvd, State#state{rep_buf_size = (RepBufSz+DataLen)}) end. handle_response(#request{from=From, stream_to=StreamTo, req_id=ReqId, - response_format = Resp_format, - save_response_to_file = SaveResponseToFile, - tmp_file_name = TmpFilename, - tmp_file_fd = Fd - }, - #state{http_status_code = SCode, - send_timer = ReqTimer, - reply_buffer = RepBuf, - recvd_headers = RespHeaders}=State) when SaveResponseToFile /= false -> + response_format = Resp_format, + save_response_to_file = SaveResponseToFile, + tmp_file_name = TmpFilename, + tmp_file_fd = Fd, + options = Options + }, + #state{http_status_code = SCode, + status_line = Status_line, + raw_headers = Raw_headers, + send_timer = ReqTimer, + reply_buffer = RepBuf, + recvd_headers = RespHeaders}=State) when SaveResponseToFile /= false -> Body = RepBuf, State_1 = set_cur_request(State), file:close(Fd), ResponseBody = case TmpFilename of - undefined -> - Body; - _ -> - {file, TmpFilename} - end, - State_2 = do_reply(State_1, From, StreamTo, ReqId, Resp_format, - {ok, SCode, RespHeaders, ResponseBody}), + undefined -> + Body; + _ -> + {file, TmpFilename} + end, + Reply = case get_value(give_raw_headers, Options, false) of + true -> + {ok, Status_line, Raw_headers, ResponseBody}; + false -> + {ok, SCode, RespHeaders, ResponseBody} + end, + State_2 = do_reply(State_1, From, StreamTo, ReqId, Resp_format, Reply), cancel_timer(ReqTimer, {eat_message, {req_timedout, From}}), State_2; handle_response(#request{from=From, stream_to=StreamTo, req_id=ReqId, - response_format = Resp_format}, - #state{http_status_code=SCode, recvd_headers=RespHeaders, - reply_buffer = RepBuf, - send_timer=ReqTimer}=State) -> + response_format = Resp_format, + options = Options}, + #state{http_status_code = SCode, + status_line = Status_line, + raw_headers = Raw_headers, + recvd_headers = RespHeaders, + reply_buffer = RepBuf, + send_timer = ReqTimer} = State) -> Body = RepBuf, %% State_1 = set_cur_request(State), + Reply = case get_value(give_raw_headers, Options, false) of + true -> + {ok, Status_line, Raw_headers, Body}; + false -> + {ok, SCode, RespHeaders, Body} + end, State_1 = case get(conn_close) of - "close" -> - do_reply(State, From, StreamTo, ReqId, Resp_format, - {ok, SCode, RespHeaders, Body}), - exit(normal); - _ -> - State_1_1 = do_reply(State, From, StreamTo, ReqId, Resp_format, - {ok, SCode, RespHeaders, Body}), - cancel_timer(ReqTimer, {eat_message, {req_timedout, From}}), - State_1_1 + "close" -> + do_reply(State, From, StreamTo, ReqId, Resp_format, Reply), + exit(normal); + _ -> + State_1_1 = do_reply(State, From, StreamTo, ReqId, Resp_format, Reply), + cancel_timer(ReqTimer, {eat_message, {req_timedout, From}}), + State_1_1 end, set_cur_request(State_1). reset_state(State) -> State#state{status = get_header, - rep_buf_size = 0, - streamed_size = 0, - content_length = undefined, - reply_buffer = <<>>, - chunk_size_buffer = <<>>, - recvd_headers = [], - deleted_crlf = false, - http_status_code = undefined, - chunk_size = undefined, - transfer_encoding = undefined}. + rep_buf_size = 0, + streamed_size = 0, + content_length = undefined, + reply_buffer = <<>>, + chunk_size_buffer = <<>>, + recvd_headers = [], + status_line = undefined, + raw_headers = undefined, + deleted_crlf = false, + http_status_code = undefined, + chunk_size = undefined, + transfer_encoding = undefined}. set_cur_request(#state{reqs = Reqs} = State) -> case queue:to_list(Reqs) of - [] -> - State#state{cur_req = undefined}; - [NextReq | _] -> - State#state{cur_req = NextReq} + [] -> + State#state{cur_req = undefined}; + [NextReq | _] -> + State#state{cur_req = NextReq} end. parse_headers(Headers) -> case scan_crlf(Headers) of - {yes, StatusLine, T} -> - parse_headers(StatusLine, T); - {no, StatusLine} -> - parse_headers(StatusLine, <<>>) + {yes, StatusLine, T} -> + parse_headers(StatusLine, T); + {no, StatusLine} -> + parse_headers(StatusLine, <<>>) end. parse_headers(StatusLine, Headers) -> Headers_1 = parse_headers_1(Headers), case parse_status_line(StatusLine) of - {ok, HttpVsn, StatCode, _Msg} -> - put(http_prot_vsn, HttpVsn), - {HttpVsn, StatCode, Headers_1}; - _ -> %% A HTTP 0.9 response? - put(http_prot_vsn, "HTTP/0.9"), - {"HTTP/0.9", undefined, Headers} + {ok, HttpVsn, StatCode, _Msg} -> + put(http_prot_vsn, HttpVsn), + {HttpVsn, StatCode, Headers_1, StatusLine, Headers}; + _ -> %% A HTTP 0.9 response? + put(http_prot_vsn, "HTTP/0.9"), + {"HTTP/0.9", undefined, Headers, StatusLine, Headers} end. % From RFC 2616 @@ -1113,16 +1289,16 @@ parse_headers_1(String) -> parse_headers_1(String, [], []). parse_headers_1([$\n, H |T], [$\r | L], Acc) when H =:= 32; - H =:= $\t -> + H =:= $\t -> parse_headers_1(lists:dropwhile(fun(X) -> - is_whitespace(X) - end, T), [32 | L], Acc); + is_whitespace(X) + end, T), [32 | L], Acc); parse_headers_1([$\n|T], [$\r | L], Acc) -> case parse_header(lists:reverse(L)) of - invalid -> - parse_headers_1(T, [], Acc); - NewHeader -> - parse_headers_1(T, [], [NewHeader | Acc]) + invalid -> + parse_headers_1(T, [], Acc); + NewHeader -> + parse_headers_1(T, [], [NewHeader | Acc]) end; parse_headers_1([H|T], L, Acc) -> parse_headers_1(T, [H|L], Acc); @@ -1130,11 +1306,11 @@ parse_headers_1([], [], Acc) -> lists:reverse(Acc); parse_headers_1([], L, Acc) -> Acc_1 = case parse_header(lists:reverse(L)) of - invalid -> - Acc; - NewHeader -> - [NewHeader | Acc] - end, + invalid -> + Acc; + NewHeader -> + [NewHeader | Acc] + end, lists:reverse(Acc_1). parse_status_line(Line) when is_binary(Line) -> @@ -1164,27 +1340,27 @@ parse_header([], _) -> scan_header(Bin) -> case get_crlf_crlf_pos(Bin, 0) of - {yes, Pos} -> - {Headers, <<_:4/binary, Body/binary>>} = split_binary(Bin, Pos), - {yes, Headers, Body}; - no -> - {no, Bin} + {yes, Pos} -> + {Headers, <<_:4/binary, Body/binary>>} = split_binary(Bin, Pos), + {yes, Headers, Body}; + no -> + {no, Bin} end. -scan_header(Bin1, Bin2) when byte_size(Bin1) < 4 -> +scan_header(Bin1, Bin2) when size(Bin1) < 4 -> scan_header(<>); scan_header(Bin1, <<>>) -> scan_header(Bin1); scan_header(Bin1, Bin2) -> - Bin1_already_scanned_size = byte_size(Bin1) - 4, + Bin1_already_scanned_size = size(Bin1) - 4, <> = Bin1, Bin_to_scan = <>, case get_crlf_crlf_pos(Bin_to_scan, 0) of - {yes, Pos} -> - {Headers_suffix, <<_:4/binary, Body/binary>>} = split_binary(Bin_to_scan, Pos), - {yes, <>, Body}; - no -> - {no, <>} + {yes, Pos} -> + {Headers_suffix, <<_:4/binary, Body/binary>>} = split_binary(Bin_to_scan, Pos), + {yes, <>, Body}; + no -> + {no, <>} end. get_crlf_crlf_pos(<<$\r, $\n, $\r, $\n, _/binary>>, Pos) -> {yes, Pos}; @@ -1193,29 +1369,29 @@ get_crlf_crlf_pos(<<>>, _) -> no. scan_crlf(Bin) -> case get_crlf_pos(Bin) of - {yes, Pos} -> - {Prefix, <<_, _, Suffix/binary>>} = split_binary(Bin, Pos), - {yes, Prefix, Suffix}; - no -> - {no, Bin} + {yes, Pos} -> + {Prefix, <<_, _, Suffix/binary>>} = split_binary(Bin, Pos), + {yes, Prefix, Suffix}; + no -> + {no, Bin} end. scan_crlf(<<>>, Bin2) -> scan_crlf(Bin2); -scan_crlf(Bin1, Bin2) when byte_size(Bin1) < 2 -> +scan_crlf(Bin1, Bin2) when size(Bin1) < 2 -> scan_crlf(<>); scan_crlf(Bin1, Bin2) -> - scan_crlf_1(byte_size(Bin1) - 2, Bin1, Bin2). + scan_crlf_1(size(Bin1) - 2, Bin1, Bin2). scan_crlf_1(Bin1_head_size, Bin1, Bin2) -> <> = Bin1, Bin3 = <>, case get_crlf_pos(Bin3) of - {yes, Pos} -> - {Prefix, <<_, _, Suffix/binary>>} = split_binary(Bin3, Pos), - {yes, concat_binary([Bin1_head, Prefix]), Suffix}; - no -> - {no, concat_binary([Bin1, Bin2])} + {yes, Pos} -> + {Prefix, <<_, _, Suffix/binary>>} = split_binary(Bin3, Pos), + {yes, concat_binary([Bin1_head, Prefix]), Suffix}; + no -> + {no, concat_binary([Bin1, Bin2])} end. get_crlf_pos(Bin) -> @@ -1245,7 +1421,8 @@ method(proppatch) -> "PROPPATCH"; method(lock) -> "LOCK"; method(unlock) -> "UNLOCK"; method(move) -> "MOVE"; -method(copy) -> "COPY". +method(copy) -> "COPY"; +method(connect) -> "CONNECT". %% From RFC 2616 %% @@ -1255,19 +1432,19 @@ method(copy) -> "COPY". % fields. This allows dynamically produced content to be transferred % along with the information necessary for the recipient to verify % that it has received the full message. -% Chunked-Body = *chunk -% last-chunk -% trailer -% CRLF -% chunk = chunk-size [ chunk-extension ] CRLF -% chunk-data CRLF -% chunk-size = 1*HEX -% last-chunk = 1*("0") [ chunk-extension ] CRLF -% chunk-extension= *( ";" chunk-ext-name [ "=" chunk-ext-val ] ) -% chunk-ext-name = token -% chunk-ext-val = token | quoted-string -% chunk-data = chunk-size(OCTET) -% trailer = *(entity-header CRLF) +% Chunked-Body = *chunk +% last-chunk +% trailer +% CRLF +% chunk = chunk-size [ chunk-extension ] CRLF +% chunk-data CRLF +% chunk-size = 1*HEX +% last-chunk = 1*("0") [ chunk-extension ] CRLF +% chunk-extension= *( ";" chunk-ext-name [ "=" chunk-ext-val ] ) +% chunk-ext-name = token +% chunk-ext-val = token | quoted-string +% chunk-data = chunk-size(OCTET) +% trailer = *(entity-header CRLF) % The chunk-size field is a string of hex digits indicating the size % of the chunk. The chunked encoding is ended by any chunk whose size % is zero, followed by the trailer, which is terminated by an empty @@ -1283,10 +1460,10 @@ parse_chunk_header(<<$;, _/binary>>, Acc) -> hexlist_to_integer(lists:reverse(Acc)); parse_chunk_header(<>, Acc) -> case is_whitespace(H) of - true -> - parse_chunk_header(T, Acc); - false -> - parse_chunk_header(T, [H | Acc]) + true -> + parse_chunk_header(T, Acc); + false -> + parse_chunk_header(T, [H | Acc]) end; parse_chunk_header(<<>>, Acc) -> hexlist_to_integer(lists:reverse(Acc)). @@ -1297,23 +1474,31 @@ is_whitespace($\n) -> true; is_whitespace($\t) -> true; is_whitespace(_) -> false. -send_async_headers(_ReqId, undefined, _StatCode, _Headers) -> +send_async_headers(_ReqId, undefined, _, _State) -> ok; -send_async_headers(ReqId, StreamTo, StatCode, Headers) -> - catch StreamTo ! {ibrowse_async_headers, ReqId, StatCode, Headers}. +send_async_headers(ReqId, StreamTo, Give_raw_headers, + #state{status_line = Status_line, raw_headers = Raw_headers, + recvd_headers = Headers, http_status_code = StatCode + }) -> + case Give_raw_headers of + false -> + catch StreamTo ! {ibrowse_async_headers, ReqId, StatCode, Headers}; + true -> + catch StreamTo ! {ibrowse_async_headers, ReqId, Status_line, Raw_headers} + end. format_response_data(Resp_format, Body) -> case Resp_format of - list when is_list(Body) -> - flatten(Body); - list when is_binary(Body) -> - binary_to_list(Body); - binary when is_list(Body) -> - list_to_binary(Body); - _ -> - %% This is to cater for sending messages such as - %% {chunk_start, _}, chunk_end etc - Body + list when is_list(Body) -> + flatten(Body); + list when is_binary(Body) -> + binary_to_list(Body); + binary when is_list(Body) -> + list_to_binary(Body); + _ -> + %% This is to cater for sending messages such as + %% {chunk_start, _}, chunk_end etc + Body end. do_reply(State, From, undefined, _, Resp_format, {ok, St_code, Headers, Body}) -> @@ -1324,14 +1509,14 @@ do_reply(State, From, undefined, _, _, Msg) -> gen_server:reply(From, Msg), dec_pipeline_counter(State); do_reply(#state{prev_req_id = Prev_req_id} = State, - _From, StreamTo, ReqId, Resp_format, {ok, _, _, Body}) -> + _From, StreamTo, ReqId, Resp_format, {ok, _, _, Body}) -> State_1 = dec_pipeline_counter(State), case Body of - [] -> - ok; - _ -> - Body_1 = format_response_data(Resp_format, Body), - catch StreamTo ! {ibrowse_async_response, ReqId, Body_1} + [] -> + ok; + _ -> + Body_1 = format_response_data(Resp_format, Body), + catch StreamTo ! {ibrowse_async_response, ReqId, Body_1} end, catch StreamTo ! {ibrowse_async_response_end, ReqId}, %% We don't want to delete the Req-id to Pid mapping straightaway @@ -1358,18 +1543,22 @@ do_interim_reply(StreamTo, Response_format, ReqId, Msg) -> Msg_1 = format_response_data(Response_format, Msg), catch StreamTo ! {ibrowse_async_response, ReqId, Msg_1}. -do_error_reply(#state{reqs = Reqs} = State, Err) -> +do_error_reply(#state{reqs = Reqs, tunnel_setup_queue = Tun_q} = State, Err) -> ReqList = queue:to_list(Reqs), lists:foreach(fun(#request{from=From, stream_to=StreamTo, req_id=ReqId, - response_format = Resp_format}) -> - ets:delete(ibrowse_stream, {req_id_pid, ReqId}), + response_format = Resp_format}) -> + ets:delete(ibrowse_stream, {req_id_pid, ReqId}), do_reply(State, From, StreamTo, ReqId, Resp_format, {error, Err}) - end, ReqList). + end, ReqList), + lists:foreach( + fun({From, _Url, _Headers, _Method, _Body, _Options, _Timeout}) -> + do_reply(State, From, undefined, undefined, undefined, Err) + end, Tun_q). fail_pipelined_requests(#state{reqs = Reqs, cur_req = CurReq} = State, Reply) -> {_, Reqs_1} = queue:out(Reqs), #request{from=From, stream_to=StreamTo, req_id=ReqId, - response_format = Resp_format} = CurReq, + response_format = Resp_format} = CurReq, do_reply(State, From, StreamTo, ReqId, Resp_format, Reply), do_error_reply(State#state{reqs = Reqs_1}, previous_request_failed). @@ -1420,10 +1609,10 @@ cancel_timer(Ref) -> erlang:cancel_timer(Ref). cancel_timer(Ref, {eat_message, Msg}) -> cancel_timer(Ref), receive - Msg -> - ok + Msg -> + ok after 0 -> - ok + ok end. make_req_id() -> @@ -1441,7 +1630,7 @@ to_lower([], Acc) -> shutting_down(#state{lb_ets_tid = undefined}) -> ok; shutting_down(#state{lb_ets_tid = Tid, - cur_pipeline_size = Sz}) -> + cur_pipeline_size = Sz}) -> catch ets:delete(Tid, {Sz, self()}). inc_pipeline_counter(#state{is_closing = true} = State) -> @@ -1454,7 +1643,7 @@ dec_pipeline_counter(#state{is_closing = true} = State) -> dec_pipeline_counter(#state{lb_ets_tid = undefined} = State) -> State; dec_pipeline_counter(#state{cur_pipeline_size = Pipe_sz, - lb_ets_tid = Tid} = State) -> + lb_ets_tid = Tid} = State) -> ets:delete(Tid, {Pipe_sz, self()}), ets:insert(Tid, {{Pipe_sz - 1, self()}, []}), State#state{cur_pipeline_size = Pipe_sz - 1}. @@ -1468,13 +1657,35 @@ flatten([]) -> get_stream_chunk_size(Options) -> case lists:keysearch(stream_chunk_size, 1, Options) of - {value, {_, V}} when V > 0 -> - V; - _ -> - ?DEFAULT_STREAM_CHUNK_SIZE + {value, {_, V}} when V > 0 -> + V; + _ -> + ?DEFAULT_STREAM_CHUNK_SIZE end. -get_inac_timeout(#state{cur_req = #request{options = Opts}}) -> +set_inac_timer(State) -> + set_inac_timer(State, get_inac_timeout(State)). + +set_inac_timer(_State, Timeout) when is_integer(Timeout) -> + erlang:send_after(Timeout, self(), timeout); +set_inac_timer(_, _) -> + undefined. + +get_inac_timeout(#state{cur_req = #request{options = Opts}}) -> get_value(inactivity_timeout, Opts, infinity); get_inac_timeout(#state{cur_req = undefined}) -> infinity. + +trace_request(Req) -> + case get(my_trace_flag) of + true -> + %%Avoid the binary operations if trace is not on... + NReq = binary_to_list(list_to_binary(Req)), + do_trace("Sending request: ~n" + "--- Request Begin ---~n~s~n" + "--- Request End ---~n", [NReq]); + _ -> ok + end. + +to_integer(X) when is_list(X) -> list_to_integer(X); +to_integer(X) when is_integer(X) -> X. diff --git a/src/ibrowse_lb.erl b/src/ibrowse_lb.erl index 5c3e30e..6bc600b 100644 --- a/src/ibrowse_lb.erl +++ b/src/ibrowse_lb.erl @@ -6,8 +6,6 @@ %%% Created : 6 Mar 2008 by chandru %%%------------------------------------------------------------------- -module(ibrowse_lb). - --vsn('$Id: ibrowse_lb.erl,v 1.2 2009/07/01 22:43:19 chandrusf Exp $ '). -author(chandru). -behaviour(gen_server). %%-------------------------------------------------------------------- diff --git a/src/ibrowse_lib.erl b/src/ibrowse_lib.erl index 7e1861a..9dc71cd 100644 --- a/src/ibrowse_lib.erl +++ b/src/ibrowse_lib.erl @@ -5,7 +5,6 @@ %% @doc Module with a few useful functions -module(ibrowse_lib). --vsn('$Id: ibrowse_lib.erl,v 1.6 2008/03/27 01:35:50 chandrusf Exp $ '). -author('chandru'). -ifdef(debug). -compile(export_all). @@ -14,22 +13,22 @@ -include("ibrowse.hrl"). -export([ - get_trace_status/2, - do_trace/2, - do_trace/3, - url_encode/1, - decode_rfc822_date/1, - status_code/1, - dec2hex/2, - drv_ue/1, - drv_ue/2, - encode_base64/1, - decode_base64/1, - get_value/2, - get_value/3, - parse_url/1, - printable_date/0 - ]). + get_trace_status/2, + do_trace/2, + do_trace/3, + url_encode/1, + decode_rfc822_date/1, + status_code/1, + dec2hex/2, + drv_ue/1, + drv_ue/2, + encode_base64/1, + decode_base64/1, + get_value/2, + get_value/3, + parse_url/1, + printable_date/0 + ]). get_trace_status(Host, Port) -> ibrowse:get_config_value({trace, Host, Port}, false). @@ -39,10 +38,10 @@ drv_ue(Str) -> drv_ue(Str, Port). drv_ue(Str, Port) -> case erlang:port_control(Port, 1, Str) of - [] -> - Str; - Res -> - Res + [] -> + Str; + Res -> + Res end. %% @doc URL-encodes a string based on RFC 1738. Returns a flat list. @@ -72,10 +71,10 @@ d2h(N) -> N+$a-10. decode_rfc822_date(String) when is_list(String) -> case catch decode_rfc822_date_1(string:tokens(String, ", \t\r\n")) of - {'EXIT', _} -> - {error, invalid_date}; - Res -> - Res + {'EXIT', _} -> + {error, invalid_date}; + Res -> + Res end. % TODO: Have to handle the Zone @@ -86,15 +85,15 @@ decode_rfc822_date_1([Day,Month,Year, Time,_Zone]) -> MonthI = month_int(Month), YearI = list_to_integer(Year), TimeTup = case string:tokens(Time, ":") of - [H,M] -> - {list_to_integer(H), - list_to_integer(M), - 0}; - [H,M,S] -> - {list_to_integer(H), - list_to_integer(M), - list_to_integer(S)} - end, + [H,M] -> + {list_to_integer(H), + list_to_integer(M), + 0}; + [H,M,S] -> + {list_to_integer(H), + list_to_integer(M), + list_to_integer(S)} + end, {{YearI,MonthI,DayI}, TimeTup}. month_int("Jan") -> 1; @@ -204,9 +203,9 @@ decode_base64(Bin) when is_binary(Bin) -> list_to_binary(List). decode_base64_1([H | T], Acc) when ((H == $\t) or - (H == 32) or - (H == $\r) or - (H == $\n)) -> + (H == 32) or + (H == $\r) or + (H == $\n)) -> decode_base64_1(T, Acc); decode_base64_1([$=, $=], Acc) -> @@ -254,10 +253,10 @@ b64_to_int($/) -> 63. get_value(Tag, TVL, DefVal) -> case lists:keysearch(Tag, 1, TVL) of - false -> - DefVal; - {value, {_, Val}} -> - Val + false -> + DefVal; + {value, {_, Val}} -> + Val end. get_value(Tag, TVL) -> @@ -272,13 +271,20 @@ parse_url([$:, $/, $/ | _], get_protocol, Url, []) -> parse_url([$:, $/, $/ | T], get_protocol, Url, TmpAcc) -> Prot = list_to_atom(lists:reverse(TmpAcc)), parse_url(T, get_username, - Url#url{protocol = Prot}, - []); -parse_url([$/ | T], get_username, Url, TmpAcc) -> + Url#url{protocol = Prot}, + []); +parse_url([H | T], get_username, Url, TmpAcc) when H == $/; + H == $? -> + Path = case H of + $/ -> + [$/ | T]; + $? -> + [$/, $? | T] + end, %% No username/password. No port number Url#url{host = lists:reverse(TmpAcc), - port = default_port(Url#url.protocol), - path = [$/ | T]}; + port = default_port(Url#url.protocol), + path = Path}; parse_url([$: | T], get_username, Url, TmpAcc) -> %% It is possible that no username/password has been %% specified. But we'll continue with the assumption that there is @@ -286,77 +292,98 @@ parse_url([$: | T], get_username, Url, TmpAcc) -> %% username/password indeed. If we encounter a '/', it was %% actually the hostname parse_url(T, get_password, - Url#url{username = lists:reverse(TmpAcc)}, - []); + Url#url{username = lists:reverse(TmpAcc)}, + []); parse_url([$@ | T], get_username, Url, TmpAcc) -> parse_url(T, get_host, - Url#url{username = lists:reverse(TmpAcc), - password = ""}, - []); + Url#url{username = lists:reverse(TmpAcc), + password = ""}, + []); parse_url([$@ | T], get_password, Url, TmpAcc) -> parse_url(T, get_host, - Url#url{password = lists:reverse(TmpAcc)}, - []); -parse_url([$/ | T], get_password, Url, TmpAcc) -> + Url#url{password = lists:reverse(TmpAcc)}, + []); +parse_url([H | T], get_password, Url, TmpAcc) when H == $/; + H == $? -> %% Ok, what we thought was the username/password was the hostname %% and portnumber #url{username=User} = Url, Port = list_to_integer(lists:reverse(TmpAcc)), + Path = case H of + $/ -> + [$/ | T]; + $? -> + [$/, $? | T] + end, Url#url{host = User, - port = Port, - username = undefined, - password = undefined, - path = [$/ | T]}; + port = Port, + username = undefined, + password = undefined, + path = Path}; parse_url([$: | T], get_host, #url{} = Url, TmpAcc) -> parse_url(T, get_port, - Url#url{host = lists:reverse(TmpAcc)}, - []); -parse_url([$/ | T], get_host, #url{protocol=Prot} = Url, TmpAcc) -> + Url#url{host = lists:reverse(TmpAcc)}, + []); +parse_url([H | T], get_host, #url{protocol=Prot} = Url, TmpAcc) when H == $/; + H == $? -> + Path = case H of + $/ -> + [$/ | T]; + $? -> + [$/, $? | T] + end, Url#url{host = lists:reverse(TmpAcc), - port = default_port(Prot), - path = [$/ | T]}; -parse_url([$/ | T], get_port, #url{protocol=Prot} = Url, TmpAcc) -> + port = default_port(Prot), + path = Path}; +parse_url([H | T], get_port, #url{protocol=Prot} = Url, TmpAcc) when H == $/; + H == $? -> + Path = case H of + $/ -> + [$/ | T]; + $? -> + [$/, $? | T] + end, Port = case TmpAcc of - [] -> - default_port(Prot); - _ -> - list_to_integer(lists:reverse(TmpAcc)) - end, - Url#url{port = Port, path = [$/ | T]}; + [] -> + default_port(Prot); + _ -> + list_to_integer(lists:reverse(TmpAcc)) + end, + Url#url{port = Port, path = Path}; parse_url([H | T], State, Url, TmpAcc) -> parse_url(T, State, Url, [H | TmpAcc]); parse_url([], get_host, Url, TmpAcc) when TmpAcc /= [] -> Url#url{host = lists:reverse(TmpAcc), - port = default_port(Url#url.protocol), - path = "/"}; + port = default_port(Url#url.protocol), + path = "/"}; parse_url([], get_username, Url, TmpAcc) when TmpAcc /= [] -> Url#url{host = lists:reverse(TmpAcc), - port = default_port(Url#url.protocol), - path = "/"}; + port = default_port(Url#url.protocol), + path = "/"}; parse_url([], get_port, #url{protocol=Prot} = Url, TmpAcc) -> Port = case TmpAcc of - [] -> - default_port(Prot); - _ -> - list_to_integer(lists:reverse(TmpAcc)) - end, + [] -> + default_port(Prot); + _ -> + list_to_integer(lists:reverse(TmpAcc)) + end, Url#url{port = Port, - path = "/"}; + path = "/"}; parse_url([], get_password, Url, TmpAcc) -> %% Ok, what we thought was the username/password was the hostname %% and portnumber #url{username=User} = Url, Port = case TmpAcc of - [] -> - default_port(Url#url.protocol); - _ -> - list_to_integer(lists:reverse(TmpAcc)) - end, + [] -> + default_port(Url#url.protocol); + _ -> + list_to_integer(lists:reverse(TmpAcc)) + end, Url#url{host = User, - port = Port, - username = undefined, - password = undefined, - path = "/"}; + port = Port, + username = undefined, + password = undefined, + path = "/"}; parse_url([], State, Url, TmpAcc) -> {invalid_uri_2, State, Url, TmpAcc}. @@ -387,13 +414,13 @@ do_trace(Fmt, Args) -> -ifdef(DEBUG). do_trace(_, Fmt, Args) -> io:format("~s -- (~s) - "++Fmt, - [printable_date(), - get(ibrowse_trace_token) | Args]). + [printable_date(), + get(ibrowse_trace_token) | Args]). -else. do_trace(true, Fmt, Args) -> io:format("~s -- (~s) - "++Fmt, - [printable_date(), - get(ibrowse_trace_token) | Args]); + [printable_date(), + get(ibrowse_trace_token) | Args]); do_trace(_, _, _) -> ok. -endif. diff --git a/src/ibrowse_sup.erl b/src/ibrowse_sup.erl index 300435d..ace33d1 100644 --- a/src/ibrowse_sup.erl +++ b/src/ibrowse_sup.erl @@ -6,8 +6,6 @@ %%% Created : 15 Oct 2003 by Chandrashekhar Mullaparthi %%%------------------------------------------------------------------- -module(ibrowse_sup). --vsn('$Id: ibrowse_sup.erl,v 1.1 2005/05/05 22:28:28 chandrusf Exp $ '). - -behaviour(supervisor). %%-------------------------------------------------------------------- %% Include files diff --git a/src/ibrowse_test.erl b/src/ibrowse_test.erl index 5f7f5c9..84cdfc0 100644 --- a/src/ibrowse_test.erl +++ b/src/ibrowse_test.erl @@ -4,7 +4,6 @@ %%% Created : 14 Oct 2003 by Chandrashekhar Mullaparthi -module(ibrowse_test). --vsn('$Id: ibrowse_test.erl,v 1.4 2009/07/01 22:43:19 chandrusf Exp $ '). -export([ load_test/3, send_reqs_1/3, diff --git a/vsn.mk b/vsn.mk index 4b3c4df..1e91755 100644 --- a/vsn.mk +++ b/vsn.mk @@ -1,2 +1,2 @@ -IBROWSE_VSN = 1.5.6 +IBROWSE_VSN = 1.6.0