summaryrefslogtreecommitdiff
path: root/apps
diff options
context:
space:
mode:
authorAdam Kocoloski <adam@cloudant.com>2010-08-11 15:57:14 -0400
committerAdam Kocoloski <adam@cloudant.com>2010-08-11 17:39:38 -0400
commit67eff5876bca3108baf37b97cfc9f9e5920dcdb7 (patch)
tree31112ebfb5d38d088ab25bee5d1e78eef85da3cc /apps
parent498d5c85681ddfea577377b8e10051fb445ac0c2 (diff)
move ibrowse to rebar layout
Diffstat (limited to 'apps')
-rw-r--r--apps/ibrowse/src/ibrowse.app.src13
-rw-r--r--apps/ibrowse/src/ibrowse.erl760
-rw-r--r--apps/ibrowse/src/ibrowse.hrl12
-rw-r--r--apps/ibrowse/src/ibrowse_app.erl64
-rw-r--r--apps/ibrowse/src/ibrowse_http_client.erl1476
-rw-r--r--apps/ibrowse/src/ibrowse_lb.erl216
-rw-r--r--apps/ibrowse/src/ibrowse_lib.erl399
-rw-r--r--apps/ibrowse/src/ibrowse_sup.erl65
-rw-r--r--apps/ibrowse/src/ibrowse_test.erl377
9 files changed, 3382 insertions, 0 deletions
diff --git a/apps/ibrowse/src/ibrowse.app.src b/apps/ibrowse/src/ibrowse.app.src
new file mode 100644
index 00000000..4f43dd92
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse.app.src
@@ -0,0 +1,13 @@
+{application, ibrowse,
+ [{description, "HTTP client application"},
+ {vsn, "1.5.1"},
+ {modules, [ ibrowse,
+ ibrowse_http_client,
+ ibrowse_app,
+ ibrowse_sup,
+ ibrowse_lib,
+ ibrowse_lb ]},
+ {registered, []},
+ {applications, [kernel,stdlib,sasl]},
+ {env, []},
+ {mod, {ibrowse_app, []}}]}.
diff --git a/apps/ibrowse/src/ibrowse.erl b/apps/ibrowse/src/ibrowse.erl
new file mode 100644
index 00000000..1913ef59
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse.erl
@@ -0,0 +1,760 @@
+%%%-------------------------------------------------------------------
+%%% File : ibrowse.erl
+%%% Author : Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%% Description : Load balancer process for HTTP client connections.
+%%%
+%%% Created : 11 Oct 2003 by Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%%-------------------------------------------------------------------
+%% @author Chandrashekhar Mullaparthi <chandrashekhar dot mullaparthi at gmail dot com>
+%% @copyright 2005-2009 Chandrashekhar Mullaparthi
+%% @version 1.5.2
+%% @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
+%% one process to handle one TCP connection to a webserver
+%% (implemented in the module ibrowse_http_client). Multiple connections to a
+%% webserver are setup based on the settings for each webserver. The
+%% ibrowse process also determines which connection to pipeline a
+%% certain request on. The functions to call are send_req/3,
+%% send_req/4, send_req/5, send_req/6.
+%%
+%% <p>Here are a few sample invocations.</p>
+%%
+%% <code>
+%% ibrowse:send_req("http://intranet/messenger/", [], get).
+%% <br/><br/>
+%%
+%% ibrowse:send_req("http://www.google.com/", [], get, [],
+%% [{proxy_user, "XXXXX"},
+%% {proxy_password, "XXXXX"},
+%% {proxy_host, "proxy"},
+%% {proxy_port, 8080}], 1000).
+%% <br/><br/>
+%%
+%%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).
+%% <br/><br/>
+%%
+%% ibrowse:send_req("http://www.erlang.org", [], head).
+%%
+%% <br/><br/>
+%% ibrowse:send_req("http://www.sun.com", [], options).
+%%
+%% <br/><br/>
+%% ibrowse:send_req("http://www.bbc.co.uk", [], trace).
+%%
+%% <br/><br/>
+%% ibrowse:send_req("http://www.google.com", [], get, [],
+%% [{stream_to, self()}]).
+%% </code>
+%%
+%% <p>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.</p>
+
+-module(ibrowse).
+-vsn('$Id: ibrowse.erl,v 1.8 2009/07/01 22:43:19 chandrusf Exp $ ').
+
+-behaviour(gen_server).
+%%--------------------------------------------------------------------
+%% Include files
+%%--------------------------------------------------------------------
+
+%%--------------------------------------------------------------------
+%% External exports
+-export([start_link/0, start/0, stop/0]).
+
+%% gen_server callbacks
+-export([init/1, handle_call/3, handle_cast/2, handle_info/2,
+ 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
+ ]).
+
+-ifdef(debug).
+-compile(export_all).
+-endif.
+
+-import(ibrowse_lib, [
+ parse_url/1,
+ get_value/3,
+ do_trace/2
+ ]).
+
+-record(state, {trace = false}).
+
+-include("ibrowse.hrl").
+-include_lib("stdlib/include/ms_transform.hrl").
+
+-define(DEF_MAX_SESSIONS,10).
+-define(DEF_MAX_PIPELINE_SIZE,10).
+
+%%====================================================================
+%% External functions
+%%====================================================================
+%%--------------------------------------------------------------------
+%% Function: start_link/0
+%% Description: Starts the server
+%%--------------------------------------------------------------------
+%% @doc Starts the ibrowse process linked to the calling process. Usually invoked by the supervisor ibrowse_sup
+%% @spec start_link() -> {ok, pid()}
+start_link() ->
+ gen_server:start_link({local, ?MODULE}, ?MODULE, [], []).
+
+%% @doc Starts the ibrowse process without linking. Useful when testing using the shell
+start() ->
+ gen_server:start({local, ?MODULE}, ?MODULE, [], [{debug, []}]).
+
+%% @doc Stop the ibrowse process. Useful when testing using the shell.
+stop() ->
+ catch gen_server:call(ibrowse, stop).
+
+%% @doc This is the basic function to send a HTTP request.
+%% The Status return value indicates the HTTP status code returned by the webserver
+%% @spec send_req(Url::string(), Headers::headerList(), Method::method()) -> response()
+%% headerList() = [{header(), value()}]
+%% header() = atom() | string()
+%% value() = term()
+%% method() = get | post | head | options | put | delete | trace | mkcol | propfind | proppatch | lock | unlock | move | copy
+%% Status = string()
+%% ResponseHeaders = [respHeader()]
+%% respHeader() = {headerName(), headerValue()}
+%% headerName() = string()
+%% headerValue() = string()
+%% response() = {ok, Status, ResponseHeaders, ResponseBody} | {ibrowse_req_id, req_id() } | {error, Reason}
+%% req_id() = term()
+%% ResponseBody = string() | {file, Filename}
+%% Reason = term()
+send_req(Url, Headers, Method) ->
+ send_req(Url, Headers, Method, [], []).
+
+%% @doc Same as send_req/3.
+%% If a list is specified for the body it has to be a flat list. The body can also be a fun/0 or a fun/1. <br/>
+%% If fun/0, the connection handling process will repeatdely call the fun until it returns an error or eof. <pre>Fun() = {ok, Data} | eof</pre><br/>
+%% If fun/1, the connection handling process will repeatedly call the fun with the supplied state until it returns an error or eof. <pre>Fun(State) = {ok, Data} | {ok, Data, NewState} | eof</pre>
+%% @spec send_req(Url, Headers, Method::method(), Body::body()) -> response()
+%% body() = [] | string() | binary() | fun_arity_0() | {fun_arity_1(), initial_state()}
+%% initial_state() = term()
+send_req(Url, Headers, Method, Body) ->
+ send_req(Url, Headers, Method, Body, []).
+
+%% @doc Same as send_req/4.
+%% For a description of SSL Options, look in the <a href="http://www.erlang.org/doc/apps/ssl/index.html">ssl</a> manpage. If the
+%% HTTP Version to use is not specified, the default is 1.1.
+%% <br/>
+%% <p>The <code>host_header</code> option is useful in the case where ibrowse is
+%% connecting to a component such as <a
+%% href="http://www.stunnel.org">stunnel</a> 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 <code>Host</code> header in
+%% the request.</p>
+%% <ul>
+%% <li>The <code>stream_to</code> 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
+%% data is received from the server, the option <code>{stream_to,
+%% {process(), once}}</code> can be specified. The calling process
+%% will have to invoke <code>ibrowse:stream_next(Request_id)</code> to
+%% receive the next packet.</li>
+%%
+%% <li>When both the options <code>save_response_to_file</code> and <code>stream_to</code>
+%% are specified, the former takes precedence.</li>
+%%
+%% <li>For the <code>save_response_to_file</code> option, the response body is saved to
+%% file only if the status code is in the 200-299 range. If not, the response body is returned
+%% as a string.</li>
+%% <li>Whenever an error occurs in the processing of a request, ibrowse will return as much
+%% information as it has, such as HTTP Status Code and HTTP Headers. When this happens, the response
+%% is of the form <code>{error, {Reason, {stat_code, StatusCode}, HTTP_headers}}</code></li>
+%%
+%% <li>The <code>inactivity_timeout</code> option is useful when
+%% dealing with large response bodies and/or slow links. In these
+%% cases, it might be hard to estimate how long a request will take to
+%% complete. In such cases, the client might want to timeout if no
+%% data has been received on the link for a certain time interval.</li>
+%%
+%% <li>
+%% The <code>connect_timeout</code> option is to specify how long the
+%% client process should wait for connection establishment. This is
+%% useful in scenarios where connections to servers are usually setup
+%% very fast, but responses might take much longer compared to
+%% connection setup. In such cases, it is better for the calling
+%% process to timeout faster if there is a problem (DNS lookup
+%% delays/failures, network routing issues, etc). The total timeout
+%% value specified for the request will enforced. To illustrate using
+%% an example:
+%% <code>
+%% ibrowse:send_req("http://www.example.com/cgi-bin/request", [], get, [], [{connect_timeout, 100}], 1000).
+%% </code>
+%% In the above invocation, if the connection isn't established within
+%% 100 milliseconds, the request will fail with
+%% <code>{error, conn_failed}</code>.<br/>
+%% If connection setup succeeds, the total time allowed for the
+%% request to complete will be 1000 milliseconds minus the time taken
+%% for connection setup.
+%% </li>
+%% </ul>
+%%
+%% @spec send_req(Url::string(), Headers::headerList(), Method::method(), Body::body(), Options::optionList()) -> response()
+%% optionList() = [option()]
+%% option() = {max_sessions, integer()} |
+%% {response_format,response_format()}|
+%% {stream_chunk_size, integer()} |
+%% {max_pipeline_size, integer()} |
+%% {trace, boolean()} |
+%% {is_ssl, boolean()} |
+%% {ssl_options, [SSLOpt]} |
+%% {pool_name, atom()} |
+%% {proxy_host, string()} |
+%% {proxy_port, integer()} |
+%% {proxy_user, string()} |
+%% {proxy_password, string()} |
+%% {use_absolute_uri, boolean()} |
+%% {basic_auth, {username(), password()}} |
+%% {cookie, string()} |
+%% {content_length, integer()} |
+%% {content_type, string()} |
+%% {save_response_to_file, srtf()} |
+%% {stream_to, stream_to()} |
+%% {http_vsn, {MajorVsn, MinorVsn}} |
+%% {host_header, string()} |
+%% {inactivity_timeout, integer()} |
+%% {connect_timeout, integer()} |
+%% {transfer_encoding, {chunked, ChunkSize}}
+%%
+%% stream_to() = process() | {process(), once}
+%% process() = pid() | atom()
+%% username() = string()
+%% password() = string()
+%% SSLOpt = term()
+%% ChunkSize = integer()
+%% srtf() = boolean() | filename()
+%% filename() = string()
+%% response_format() = list | binary
+send_req(Url, Headers, Method, Body, Options) ->
+ send_req(Url, Headers, Method, Body, Options, 30000).
+
+%% @doc Same as send_req/5.
+%% All timeout values are in milliseconds.
+%% @spec send_req(Url, Headers::headerList(), Method::method(), Body::body(), Options::optionList(), Timeout) -> response()
+%% 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}}
+ 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
+ end, Options, Config_options).
+
+get_lb_pid(Url) ->
+ gen_server:call(?MODULE, {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_max_pipeline_size(Host, Port, Options) ->
+ get_value(max_pipeline_size, Options,
+ 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.
+set_dest(Host, Port, [{max_sessions, Max} | T]) ->
+ set_max_sessions(Host, Port, Max),
+ set_dest(Host, Port, T);
+set_dest(Host, Port, [{max_pipeline_size, Max} | T]) ->
+ set_max_pipeline_size(Host, Port, Max),
+ set_dest(Host, Port, T);
+set_dest(Host, Port, [{trace, Bool} | T]) when Bool == true; Bool == false ->
+ ibrowse ! {trace, true, Host, Port},
+ set_dest(Host, Port, T);
+set_dest(_Host, _Port, [H | _]) ->
+ exit({invalid_option, H});
+set_dest(_, _, []) ->
+ ok.
+
+%% @doc Set the maximum number of connections allowed to a specific Host:Port.
+%% @spec set_max_sessions(Host::string(), Port::integer(), Max::integer()) -> ok
+set_max_sessions(Host, Port, Max) when is_integer(Max), Max > 0 ->
+ gen_server:call(?MODULE, {set_config_value, {max_sessions, Host, Port}, Max}).
+
+%% @doc Set the maximum pipeline size for each connection to a specific Host:Port.
+%% @spec set_max_pipeline_size(Host::string(), Port::integer(), Max::integer()) -> ok
+set_max_pipeline_size(Host, Port, Max) when is_integer(Max), Max > 0 ->
+ gen_server:call(?MODULE, {set_config_value, {max_pipeline_size, Host, Port}, Max}).
+
+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
+ end.
+
+ensure_bin(L) when is_list(L) -> list_to_binary(L);
+ensure_bin(B) when is_binary(B) -> B;
+ensure_bin(Fun) when is_function(Fun) -> Fun;
+ensure_bin({Fun}) when is_function(Fun) -> Fun;
+ensure_bin({Fun, _} = Body) when is_function(Fun) -> Body.
+
+%% @doc 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
+%% some might take a very short time. To avoid getting these quick
+%% requests stuck in the pipeline behind time consuming requests, use
+%% this function to get a handle to a connection process. <br/>
+%% <b>Note:</b> Calling this function only creates a worker process. No connection
+%% is setup. The connection attempt is made only when the first
+%% request is sent via any of the send_req_direct/4,5,6,7 functions.<br/>
+%% <b>Note:</b> It is the responsibility of the calling process to control
+%% pipeline size on such connections.
+%%
+%% @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
+%% is linked to the worker process which is spawned.
+spawn_link_worker_process(Host, Port) ->
+ ibrowse_http_client:start_link({Host, Port}).
+
+%% @doc Terminate a worker process spawned using
+%% spawn_worker_process/2 or spawn_link_worker_process/2. Requests in
+%% progress will get the error response <pre>{error, closing_on_request}</pre>
+%% @spec stop_worker_process(Conn_pid::pid()) -> ok
+stop_worker_process(Conn_pid) ->
+ ibrowse_http_client:stop(Conn_pid).
+
+%% @doc Same as send_req/3 except that the first argument is the PID
+%% returned by spawn_worker_process/2 or spawn_link_worker_process/2
+send_req_direct(Conn_pid, Url, Headers, Method) ->
+ send_req_direct(Conn_pid, Url, Headers, Method, [], []).
+
+%% @doc Same as send_req/4 except that the first argument is the PID
+%% returned by spawn_worker_process/2 or spawn_link_worker_process/2
+send_req_direct(Conn_pid, Url, Headers, Method, Body) ->
+ send_req_direct(Conn_pid, Url, Headers, Method, Body, []).
+
+%% @doc Same as send_req/5 except that the first argument is the PID
+%% returned by spawn_worker_process/2 or spawn_link_worker_process/2
+send_req_direct(Conn_pid, Url, Headers, Method, Body, Options) ->
+ send_req_direct(Conn_pid, Url, Headers, Method, Body, Options, 30000).
+
+%% @doc Same as send_req/6 except that the first argument is the PID
+%% 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}}
+ end.
+
+%% @doc Tell ibrowse to stream the next chunk of data to the
+%% caller. Should be used in conjunction with the
+%% <code>stream_to</code> option
+%% @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
+ end.
+
+%% @doc Turn tracing on for the ibrowse process
+trace_on() ->
+ ibrowse ! {trace, true}.
+%% @doc Turn tracing off for the ibrowse process
+trace_off() ->
+ ibrowse ! {trace, false}.
+
+%% @doc Turn tracing on for all connections to the specified HTTP
+%% server. Host is whatever is specified as the domain name in the URL
+%% @spec trace_on(Host, Port) -> ok
+%% Host = string()
+%% Port = integer()
+trace_on(Host, Port) ->
+ ibrowse ! {trace, true, Host, Port},
+ ok.
+
+%% @doc Turn tracing OFF for all connections to the specified HTTP
+%% server.
+%% @spec trace_off(Host, Port) -> ok
+trace_off(Host, Port) ->
+ ibrowse ! {trace, false, Host, Port},
+ ok.
+
+%% @doc Turn Off ALL tracing
+%% @spec all_trace_off() -> ok
+all_trace_off() ->
+ ibrowse ! all_trace_off,
+ ok.
+
+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)),
+ All_ets = ets:all(),
+ io:format("~-40.40s | ~-5.5s | ~-10.10s | ~s~n",
+ ["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).
+
+%% @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
+ end.
+
+%% @doc Clear current configuration for ibrowse and load from the file
+%% ibrowse.conf in the IBROWSE_EBIN/../priv directory. Current
+%% configuration is cleared only if the ibrowse.conf file is readable
+%% using file:consult/1
+rescan_config() ->
+ gen_server:call(?MODULE, rescan_config).
+
+%% Clear current configuration for ibrowse and load from the specified
+%% file. Current configuration is cleared only if the specified
+%% file is readable using file:consult/1
+rescan_config(File) when is_list(File) ->
+ gen_server:call(?MODULE, {rescan_config, File}).
+
+%%====================================================================
+%% Server functions
+%%====================================================================
+
+%%--------------------------------------------------------------------
+%% Function: init/1
+%% Description: Initiates the server
+%% Returns: {ok, State} |
+%% {ok, State, Timeout} |
+%% ignore |
+%% {stop, Reason}
+%%--------------------------------------------------------------------
+init(_) ->
+ process_flag(trap_exit, true),
+ State = #state{},
+ put(my_trace_flag, State#state.trace),
+ put(ibrowse_trace_token, "ibrowse"),
+ ets:new(ibrowse_lb, [named_table, public, {keypos, 2}]),
+ ets:new(ibrowse_conf, [named_table, protected, {keypos, 2}]),
+ ets:new(ibrowse_stream, [named_table, public]),
+ import_config(),
+ {ok, #state{}}.
+
+import_config() ->
+ case code:priv_dir(ibrowse) of
+ {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
+ end.
+
+%% @doc Internal export
+get_config_value(Key) ->
+ [#ibrowse_conf{value = V}] = ets:lookup(ibrowse_conf, Key),
+ V.
+
+%% @doc Internal export
+get_config_value(Key, DefVal) ->
+ case ets:lookup(ibrowse_conf, Key) of
+ [] ->
+ DefVal;
+ [#ibrowse_conf{value = V}] ->
+ V
+ end.
+
+set_config_value(Key, Val) ->
+ ets:insert(ibrowse_conf, #ibrowse_conf{key = Key, value = Val}).
+%%--------------------------------------------------------------------
+%% Function: handle_call/3
+%% Description: Handling call messages
+%% Returns: {reply, Reply, State} |
+%% {reply, Reply, State, Timeout} |
+%% {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, Reply, State} | (terminate/2 is called)
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+handle_call({get_lb_pid, #url{host = Host, port = Port} = Url}, _From, State) ->
+ Pid = do_get_connection(Url, ets:lookup(ibrowse_lb, {Host, Port})),
+ {reply, Pid, State};
+
+handle_call(stop, _From, State) ->
+ do_trace("IBROWSE shutting down~n", []),
+ {stop, normal, ok, State};
+
+handle_call({set_config_value, Key, Val}, _From, State) ->
+ set_config_value(Key, Val),
+ {reply, ok, State};
+
+handle_call(rescan_config, _From, State) ->
+ Ret = (catch import_config()),
+ {reply, Ret, State};
+
+handle_call({rescan_config, File}, _From, State) ->
+ Ret = (catch import_config(File)),
+ {reply, Ret, State};
+
+handle_call(Request, _From, State) ->
+ Reply = {unknown_request, Request},
+ {reply, Reply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: handle_cast/2
+%% Description: Handling cast messages
+%% Returns: {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+
+handle_cast(_Msg, State) ->
+ {noreply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: handle_info/2
+%% Description: Handling all non call/cast messages
+%% Returns: {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+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,
+ 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,
+ ets:foldl(Fun, undefined, ibrowse_lb),
+ ets:insert(ibrowse_conf, #ibrowse_conf{key = {trace, Host, Port},
+ value = Bool}),
+ {noreply, State};
+
+handle_info(_Info, State) ->
+ {noreply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: terminate/2
+%% Description: Shutdown the server
+%% Returns: any (ignored by gen_server)
+%%--------------------------------------------------------------------
+terminate(_Reason, _State) ->
+ ok.
+
+%%--------------------------------------------------------------------
+%% Func: code_change/3
+%% Purpose: Convert process state when code is changed
+%% Returns: {ok, NewState}
+%%--------------------------------------------------------------------
+code_change(_OldVsn, State, _Extra) ->
+ {ok, State}.
+
+%%--------------------------------------------------------------------
+%%% Internal functions
+%%--------------------------------------------------------------------
+do_get_connection(#url{host = Host, port = Port}, []) ->
+ {ok, Pid} = ibrowse_lb:start_link([Host, Port]),
+ ets:insert(ibrowse_lb, #lb_pid{host_port = {Host, Port}, pid = Pid}),
+ Pid;
+do_get_connection(_Url, [#lb_pid{pid = Pid}]) ->
+ Pid.
diff --git a/apps/ibrowse/src/ibrowse.hrl b/apps/ibrowse/src/ibrowse.hrl
new file mode 100644
index 00000000..ebf3bb33
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse.hrl
@@ -0,0 +1,12 @@
+-ifndef(IBROWSE_HRL).
+-define(IBROWSE_HRL, "ibrowse.hrl").
+
+-record(url, {abspath, host, port, username, password, path, protocol}).
+
+-record(lb_pid, {host_port, pid}).
+
+-record(client_conn, {key, cur_pipeline_size = 0, reqs_served = 0}).
+
+-record(ibrowse_conf, {key, value}).
+
+-endif.
diff --git a/apps/ibrowse/src/ibrowse_app.erl b/apps/ibrowse/src/ibrowse_app.erl
new file mode 100644
index 00000000..8c83e8f1
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse_app.erl
@@ -0,0 +1,64 @@
+%%%-------------------------------------------------------------------
+%%% File : ibrowse_app.erl
+%%% Author : Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%% Description :
+%%%
+%%% Created : 15 Oct 2003 by Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%%-------------------------------------------------------------------
+-module(ibrowse_app).
+-vsn('$Id: ibrowse_app.erl,v 1.1 2005/05/05 22:28:28 chandrusf Exp $ ').
+
+-behaviour(application).
+%%--------------------------------------------------------------------
+%% Include files
+%%--------------------------------------------------------------------
+
+%%--------------------------------------------------------------------
+%% External exports
+%%--------------------------------------------------------------------
+-export([
+ start/2,
+ stop/1
+ ]).
+
+%%--------------------------------------------------------------------
+%% Internal exports
+%%--------------------------------------------------------------------
+-export([
+ ]).
+
+%%--------------------------------------------------------------------
+%% Macros
+%%--------------------------------------------------------------------
+
+%%--------------------------------------------------------------------
+%% Records
+%%--------------------------------------------------------------------
+
+%%====================================================================
+%% External functions
+%%====================================================================
+%%--------------------------------------------------------------------
+%% Func: start/2
+%% Returns: {ok, Pid} |
+%% {ok, Pid, State} |
+%% {error, Reason}
+%%--------------------------------------------------------------------
+start(_Type, _StartArgs) ->
+ case ibrowse_sup:start_link() of
+ {ok, Pid} ->
+ {ok, Pid};
+ Error ->
+ Error
+ end.
+
+%%--------------------------------------------------------------------
+%% Func: stop/1
+%% Returns: any
+%%--------------------------------------------------------------------
+stop(_State) ->
+ ok.
+
+%%====================================================================
+%% Internal functions
+%%====================================================================
diff --git a/apps/ibrowse/src/ibrowse_http_client.erl b/apps/ibrowse/src/ibrowse_http_client.erl
new file mode 100644
index 00000000..65d9cb9c
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse_http_client.erl
@@ -0,0 +1,1476 @@
+%%%-------------------------------------------------------------------
+%%% File : ibrowse_http_client.erl
+%%% Author : Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%% Description : The name says it all
+%%%
+%%% Created : 11 Oct 2003 by Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%%-------------------------------------------------------------------
+-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
+%%--------------------------------------------------------------------
+
+%%--------------------------------------------------------------------
+%% External exports
+-export([
+ start_link/1,
+ start/1,
+ stop/1,
+ send_req/7
+ ]).
+
+-ifdef(debug).
+-compile(export_all).
+-endif.
+
+%% gen_server callbacks
+-export([
+ 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(request, {url, method, options, from,
+ stream_to, caller_controls_socket = false,
+ 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
+ ]).
+
+-define(DEFAULT_STREAM_CHUNK_SIZE, 1024*1024).
+
+%%====================================================================
+%% External functions
+%%====================================================================
+%%--------------------------------------------------------------------
+%% Function: start_link/0
+%% Description: Starts the server
+%%--------------------------------------------------------------------
+start(Args) ->
+ gen_server:start(?MODULE, Args, []).
+
+start_link(Args) ->
+ gen_server:start_link(?MODULE, Args, []).
+
+stop(Conn_pid) ->
+ gen_server:call(Conn_pid, stop).
+
+send_req(Conn_Pid, Url, Headers, Method, Body, Options, Timeout) ->
+ gen_server:call(
+ Conn_Pid,
+ {send_req, {Url, Headers, Method, Body, Options, Timeout}}, Timeout).
+
+%%====================================================================
+%% Server functions
+%%====================================================================
+
+%%--------------------------------------------------------------------
+%% Function: init/1
+%% Description: Initiates the server
+%% Returns: {ok, State} |
+%% {ok, State, Timeout} |
+%% ignore |
+%% {stop, Reason}
+%%--------------------------------------------------------------------
+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},
+ put(ibrowse_trace_token, [Host, $:, integer_to_list(Port)]),
+ put(my_trace_flag, ibrowse_lib:get_trace_status(Host, Port)),
+ {ok, State};
+init({Host, Port}) ->
+ State = #state{host = Host,
+ port = Port},
+ put(ibrowse_trace_token, [Host, $:, integer_to_list(Port)]),
+ put(my_trace_flag, ibrowse_lib:get_trace_status(Host, Port)),
+ {ok, State};
+init(#url{host=Host, port=Port, protocol=Protocol}) ->
+ State = #state{
+ host = Host,
+ port = Port,
+ is_ssl = (Protocol == https),
+ ssl_options = [{ssl_imp, new}, {depth, 9}]
+ },
+ put(ibrowse_trace_token, [Host, $:, integer_to_list(Port)]),
+ put(my_trace_flag, ibrowse_lib:get_trace_status(Host, Port)),
+ {ok, State}.
+
+%%--------------------------------------------------------------------
+%% Function: handle_call/3
+%% Description: Handling call messages
+%% Returns: {reply, Reply, State} |
+%% {reply, Reply, State, Timeout} |
+%% {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, Reply, State} | (terminate/2 is called)
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+%% Received a request when the remote server has already sent us a
+%% Connection: Close header
+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) ->
+ send_req_1(From, Url, Headers, Method, Body, Options, Timeout, State);
+
+handle_call(stop, _From, State) ->
+ do_close(State),
+ do_error_reply(State, closing_on_request),
+ {stop, normal, ok, State#state{socket=undefined}};
+
+handle_call(Request, _From, State) ->
+ Reply = {unknown_request, Request},
+ {reply, Reply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: handle_cast/2
+%% Description: Handling cast messages
+%% Returns: {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+handle_cast(_Msg, State) ->
+ {noreply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: handle_info/2
+%% Description: Handling all non call/cast messages
+%% Returns: {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+handle_info({tcp, _Sock, Data}, #state{status = Status} = State) ->
+ do_trace("Data recvd in state: ~p. Size: ~p. ~p~n~n", [Status, size(Data), Data]),
+ handle_sock_data(Data, State);
+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),
+ {noreply, State};
+
+handle_info({stream_next, _Req_id}, State) ->
+ {noreply, State};
+
+handle_info({tcp_closed, _Sock}, State) ->
+ do_trace("TCP connection closed by peer!~n", []),
+ handle_sock_closed(State),
+ {stop, normal, State};
+handle_info({ssl_closed, _Sock}, State) ->
+ do_trace("SSL connection closed by peer!~n", []),
+ handle_sock_closed(State),
+ {stop, normal, State};
+
+handle_info({tcp_error, _Sock}, State) ->
+ io:format("Error on connection to ~1000.p:~1000.p~n", [State#state.host, State#state.port]),
+ handle_sock_closed(State),
+ {stop, normal, State};
+handle_info({ssl_error, _Sock}, State) ->
+ io:format("Error on SSL connection to ~1000.p:~1000.p~n", [State#state.host, State#state.port]),
+ handle_sock_closed(State),
+ {stop, normal, State};
+
+handle_info({req_timedout, From}, State) ->
+ case lists:keysearch(From, #request.from, queue:to_list(State#state.reqs)) of
+ false ->
+ {noreply, State};
+ {value, _} ->
+ shutting_down(State),
+ do_error_reply(State, req_timedout),
+ {stop, normal, State}
+ end;
+
+handle_info(timeout, State) ->
+ shutting_down(State),
+ do_error_reply(State, req_timedout),
+ {stop, normal, State};
+
+handle_info({trace, Bool}, State) ->
+ put(my_trace_flag, Bool),
+ {noreply, State};
+
+handle_info(Info, State) ->
+ io:format("Unknown message recvd for ~1000.p:~1000.p -> ~p~n",
+ [State#state.host, State#state.port, Info]),
+ io:format("Recvd unknown message ~p when in state: ~p~n", [Info, State]),
+ {noreply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: terminate/2
+%% Description: Shutdown the server
+%% Returns: any (ignored by gen_server)
+%%--------------------------------------------------------------------
+terminate(_Reason, State) ->
+ do_close(State).
+
+%%--------------------------------------------------------------------
+%% Func: code_change/3
+%% Purpose: Convert process state when code is changed
+%% Returns: {ok, NewState}
+%%--------------------------------------------------------------------
+code_change(_OldVsn, State, _Extra) ->
+ {ok, State}.
+
+%%--------------------------------------------------------------------
+%%% Internal functions
+%%--------------------------------------------------------------------
+
+%%--------------------------------------------------------------------
+%% Handles data recvd on the socket
+%%--------------------------------------------------------------------
+handle_sock_data(Data, #state{status=idle}=State) ->
+ do_trace("Data recvd on socket in state idle!. ~1000.p~n", [Data]),
+ shutting_down(State),
+ do_error_reply(State, data_in_status_idle),
+ do_close(State),
+ {stop, normal, State};
+
+handle_sock_data(Data, #state{status = get_header}=State) ->
+ case parse_response(Data, State) of
+ {error, _Reason} ->
+ shutting_down(State),
+ {stop, normal, State};
+ stop ->
+ shutting_down(State),
+ {stop, normal, State};
+ State_1 ->
+ active_once(State_1),
+ {noreply, State_1, get_inac_timeout(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) ->
+ 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};
+ stop ->
+ shutting_down(State),
+ {stop, normal, State};
+ State_1 ->
+ active_once(State_1),
+ {noreply, State_1, get_inac_timeout(State_1)}
+ end
+ end.
+
+accumulate_response(Data,
+ #state{
+ cur_req = #request{save_response_to_file = true,
+ tmp_file_fd = undefined} = CurReq,
+ http_status_code=[$2 | _]}=State) ->
+ TmpFilename = make_tmp_filename(),
+ 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}}
+ end;
+accumulate_response(Data, #state{cur_req = #request{save_response_to_file = true,
+ tmp_file_fd = Fd},
+ transfer_encoding=chunked,
+ reply_buffer = Reply_buf,
+ http_status_code=[$2 | _]
+ } = State) ->
+ case file:write(Fd, [Reply_buf, Data]) of
+ ok ->
+ State#state{reply_buffer = <<>>};
+ {error, Reason} ->
+ {error, {file_write_error, Reason}}
+ end;
+accumulate_response(Data, #state{cur_req = #request{save_response_to_file = true,
+ tmp_file_fd = Fd},
+ reply_buffer = RepBuf,
+ http_status_code=[$2 | _]
+ } = State) ->
+ case file:write(Fd, [RepBuf, Data]) of
+ 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) ->
+ #request{stream_to=StreamTo, req_id=ReqId,
+ stream_chunk_size = Stream_chunk_size,
+ response_format = Response_format,
+ caller_controls_socket = Caller_controls_socket} = CurReq,
+ RepBuf_1 = list_to_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}
+ end.
+
+make_tmp_filename() ->
+ 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)]).
+
+
+%%--------------------------------------------------------------------
+%% Handles the case when the server closes the socket
+%%--------------------------------------------------------------------
+handle_sock_closed(#state{status=get_header}=State) ->
+ shutting_down(State),
+ do_error_reply(State, connection_closed);
+
+handle_sock_closed(#state{cur_req=undefined} = State) ->
+ shutting_down(State);
+
+%% We check for IsClosing because this the server could have sent a
+%% 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) ->
+ #request{from=From, stream_to=StreamTo, req_id=ReqId,
+ response_format = Resp_format} = 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
+ end.
+
+do_connect(Host, Port, _Options, #state{is_ssl=true, ssl_options=SSLOptions}, Timeout) ->
+ ssl:connect(Host, Port,
+ [binary, {nodelay, true}, {active, false} | SSLOptions],
+ Timeout);
+do_connect(Host, Port, _Options, _State, Timeout) ->
+ gen_tcp:connect(Host, Port,
+ [binary, {nodelay, true}, {active, false}],
+ Timeout).
+
+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).
+
+%% @spec do_send_body(Sock::socket_descriptor(), Source::source_descriptor(), IsSSL::boolean()) -> ok | error()
+%% source_descriptor() = fun_arity_0 |
+%% {fun_arity_0} |
+%% {fun_arity_1, term()}
+%% error() = term()
+do_send_body(Source, State) when is_function(Source) ->
+ do_send_body({Source}, State);
+do_send_body({Source}, State) when is_function(Source) ->
+ do_send_body1(Source, Source(), State);
+do_send_body({Source, Source_state}, State) when is_function(Source) ->
+ do_send_body1(Source, Source(Source_state), State);
+do_send_body(Body, State) ->
+ do_send(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
+ end.
+
+do_close(#state{socket = undefined}) -> ok;
+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).
+
+do_setopts(Sock, Opts, true) -> ssl:setopts(Sock, Opts);
+do_setopts(Sock, Opts, false) -> 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)}
+ end.
+
+send_req_1(From,
+ #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,
+ 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}
+ end;
+send_req_1(From,
+ #url{abspath = AbsPath,
+ host = Host,
+ port = Port,
+ path = RelPath} = Url,
+ Headers, Method, Body, Options, Timeout,
+ #state{status = Status} = State) ->
+ ReqId = make_req_id(),
+ Resp_format = get_value(response_format, Options, list),
+ {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,
+ SaveResponseToFile = get_value(save_response_to_file, Options, false),
+ NewReq = #request{url = Url,
+ method = Method,
+ stream_to = StreamTo,
+ caller_controls_socket = Caller_controls_socket,
+ 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 = add_auth_headers(Url, Options, Headers, State),
+ HostHeaderValue = case lists:keysearch(host_header, 1, Options) of
+ false ->
+ case Port of
+ 80 -> Host;
+ _ -> [Host, ":", integer_to_list(Port)]
+ end;
+ {value, {_, Host_h_val}} ->
+ Host_h_val
+ end,
+ {Req, Body_1} = make_request(Method,
+ [{"Host", HostHeaderValue} | Headers_1],
+ AbsPath, RelPath, Body, Options, State#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,
+ case do_send(Req, State) of
+ ok ->
+ case do_send_body(Body_1, State) of
+ ok ->
+ State_2 = inc_pipeline_counter(State_1),
+ active_once(State_1),
+ 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}
+ end.
+
+add_auth_headers(#url{username = User,
+ password = UPw},
+ Options,
+ Headers,
+ #state{use_proxy = UseProxy,
+ proxy_auth_digest = ProxyAuthDigest}) ->
+ 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.
+
+http_auth_digest([], []) ->
+ [];
+http_auth_digest(Username, Password) ->
+ encode_base64(Username ++ [$: | Password]).
+
+encode_base64([]) ->
+ [];
+encode_base64([A]) ->
+ [e(A bsr 2), e((A band 3) bsl 4), $=, $=];
+encode_base64([A,B]) ->
+ [e(A bsr 2), e(((A band 3) bsl 4) bor (B bsr 4)), e((B band 15) bsl 2), $=];
+encode_base64([A,B,C|Ls]) ->
+ encode_base64_do(A,B,C, Ls).
+encode_base64_do(A,B,C, Rest) ->
+ BB = (A bsl 16) bor (B bsl 8) bor C,
+ [e(BB bsr 18), e((BB bsr 12) band 63),
+ e((BB bsr 6) band 63), e(BB band 63)|encode_base64(Rest)].
+
+e(X) when X >= 0, X < 26 -> X+65;
+e(X) when X>25, X<52 -> X+71;
+e(X) when X>51, X<62 -> X-4;
+e(62) -> $+;
+e(63) -> $/;
+e(X) -> exit({bad_encode_base64_token, X}).
+
+make_request(Method, Headers, AbsPath, RelPath, Body, Options, 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(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,
+ Headers_3 = cons_headers(Headers_2),
+ Uri = case get_value(use_absolute_uri, Options, false) or UseProxy of
+ true ->
+ AbsPath;
+ false ->
+ RelPath
+ end,
+ {[method(Method), " ", Uri, " ", HttpVsn, crnl(), Headers_3, crnl()], Body_1}.
+
+http_vsn_string({0,9}) -> "HTTP/0.9";
+http_vsn_string({1,0}) -> "HTTP/1.0";
+http_vsn_string({1,1}) -> "HTTP/1.1".
+
+cons_headers(Headers) ->
+ cons_headers(Headers, []).
+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]);
+cons_headers([{cookie, Cookie} | T], Acc) ->
+ cons_headers(T, [{"Cookie", Cookie} | Acc]);
+cons_headers([{content_length, L} | T], Acc) ->
+ cons_headers(T, [{"Content-Length", L} | Acc]);
+cons_headers([{content_type, L} | T], Acc) ->
+ cons_headers(T, [{"Content-Type", L} | Acc]);
+cons_headers([H | T], Acc) ->
+ cons_headers(T, [H | Acc]).
+
+encode_headers(L) ->
+ encode_headers(L, []).
+encode_headers([{http_vsn, _Val} | T], Acc) ->
+ encode_headers(T, Acc);
+encode_headers([{Name,Val} | T], Acc) when is_list(Name) ->
+ encode_headers(T, [[Name, ": ", fmt_val(Val), crnl()] | Acc]);
+encode_headers([{Name,Val} | T], Acc) when is_atom(Name) ->
+ encode_headers(T, [[atom_to_list(Name), ": ", fmt_val(Val), crnl()] | Acc]);
+encode_headers([], Acc) ->
+ lists:reverse(Acc).
+
+chunk_request_body(Body, ChunkSize) ->
+ 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 is_binary(Body),
+ size(Body) >= ChunkSize ->
+ <<ChunkBody:ChunkSize/binary, Rest/binary>> = Body,
+ Chunk = [ibrowse_lib:dec2hex(4, ChunkSize),"\r\n",
+ ChunkBody, "\r\n"],
+ chunk_request_body(Rest, ChunkSize, [Chunk | Acc]);
+chunk_request_body(Body, _ChunkSize, Acc) when is_binary(Body) ->
+ BodySize = size(Body),
+ Chunk = [ibrowse_lib:dec2hex(4, BodySize),"\r\n",
+ Body, "\r\n"],
+ LastChunk = "0\r\n",
+ lists:reverse(["\r\n", LastChunk, Chunk | Acc]);
+chunk_request_body(Body, ChunkSize, Acc) when is_list(Body),
+ length(Body) >= ChunkSize ->
+ {ChunkBody, Rest} = split_list_at(Body, ChunkSize),
+ Chunk = [ibrowse_lib:dec2hex(4, ChunkSize),"\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"],
+ LastChunk = "0\r\n",
+ lists:reverse(["\r\n", LastChunk, Chunk | Acc]).
+
+
+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) ->
+ #request{from=From, stream_to=StreamTo, req_id=ReqId,
+ method=Method, response_format = Resp_format} = 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 ->
+ 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}
+ end.
+
+is_connection_closing("HTTP/0.9", _) -> true;
+is_connection_closing(_, "close") -> true;
+is_connection_closing("HTTP/1.0", "false") -> true;
+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) ->
+ case scan_crlf(Chunk_sz_buf, DataRecvd) of
+ {yes, ChunkHeader, Data_1} ->
+ case parse_chunk_header(ChunkHeader) of
+ {error, Reason} ->
+ {error, Reason};
+ ChunkSize ->
+ %%
+ %% 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})
+ end;
+ {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) ->
+ 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}
+ 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) ->
+ do_trace("Detected end of chunked transfer...~n", []),
+ DataRecvd_1 = case DelCrlf of
+ 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}
+ 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) ->
+ 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})
+ 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) ->
+ 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)})
+ 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 ->
+ 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}),
+ 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) ->
+ Body = RepBuf,
+%% State_1 = set_cur_request(State),
+ 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
+ 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}.
+
+set_cur_request(#state{reqs = Reqs} = State) ->
+ case queue:to_list(Reqs) of
+ [] ->
+ 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, <<>>)
+ 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}
+ end.
+
+% From RFC 2616
+%
+% HTTP/1.1 header field values can be folded onto multiple lines if
+% the continuation line begins with a space or horizontal tab. All
+% linear white space, including folding, has the same semantics as
+% SP. A recipient MAY replace any linear white space with a single
+% SP before interpreting the field value or forwarding the message
+% downstream.
+ parse_headers_1(B) when is_binary(B) ->
+ parse_headers_1(binary_to_list(B));
+ parse_headers_1(String) ->
+ parse_headers_1(String, [], []).
+
+parse_headers_1([$\n, H |T], [$\r | L], Acc) when H == 32;
+ H == $\t ->
+ parse_headers_1(lists:dropwhile(fun(X) ->
+ is_whitespace(X)
+ end, T), [32 | L], Acc);
+parse_headers_1([$\n|T], [$\r | L], Acc) ->
+ case parse_header(lists:reverse(L)) of
+ invalid ->
+ parse_headers_1(T, [], Acc);
+ NewHeader ->
+ parse_headers_1(T, [], [NewHeader | Acc])
+ end;
+parse_headers_1([H|T], L, Acc) ->
+ parse_headers_1(T, [H|L], Acc);
+parse_headers_1([], [], Acc) ->
+ lists:reverse(Acc);
+parse_headers_1([], L, Acc) ->
+ Acc_1 = case parse_header(lists:reverse(L)) of
+ invalid ->
+ Acc;
+ NewHeader ->
+ [NewHeader | Acc]
+ end,
+ lists:reverse(Acc_1).
+
+parse_status_line(Line) when is_binary(Line) ->
+ parse_status_line(binary_to_list(Line));
+parse_status_line(Line) ->
+ parse_status_line(Line, get_prot_vsn, [], []).
+parse_status_line([32 | T], get_prot_vsn, ProtVsn, StatCode) ->
+ parse_status_line(T, get_status_code, ProtVsn, StatCode);
+parse_status_line([32 | T], get_status_code, ProtVsn, StatCode) ->
+ {ok, lists:reverse(ProtVsn), lists:reverse(StatCode), T};
+parse_status_line([H | T], get_prot_vsn, ProtVsn, StatCode) ->
+ parse_status_line(T, get_prot_vsn, [H|ProtVsn], StatCode);
+parse_status_line([H | T], get_status_code, ProtVsn, StatCode) ->
+ parse_status_line(T, get_status_code, ProtVsn, [H | StatCode]);
+parse_status_line([], _, _, _) ->
+ http_09.
+
+parse_header(B) when is_binary(B) ->
+ parse_header(binary_to_list(B));
+parse_header(L) ->
+ parse_header(L, []).
+parse_header([$: | V], Acc) ->
+ {lists:reverse(Acc), string:strip(V)};
+parse_header([H | T], Acc) ->
+ parse_header(T, [H | Acc]);
+parse_header([], _) ->
+ invalid.
+
+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}
+ end.
+
+scan_header(Bin1, Bin2) when size(Bin1) < 4 ->
+ scan_header(<<Bin1/binary, Bin2/binary>>);
+scan_header(Bin1, <<>>) ->
+ scan_header(Bin1);
+scan_header(Bin1, Bin2) ->
+ Bin1_already_scanned_size = size(Bin1) - 4,
+ <<Headers_prefix:Bin1_already_scanned_size/binary, Rest/binary>> = Bin1,
+ Bin_to_scan = <<Rest/binary, Bin2/binary>>,
+ 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, <<Headers_prefix/binary, Headers_suffix/binary>>, Body};
+ no ->
+ {no, <<Bin1/binary, Bin2/binary>>}
+ end.
+
+get_crlf_crlf_pos(<<$\r, $\n, $\r, $\n, _/binary>>, Pos) -> {yes, Pos};
+get_crlf_crlf_pos(<<_, Rest/binary>>, Pos) -> get_crlf_crlf_pos(Rest, Pos + 1);
+get_crlf_crlf_pos(<<>>, _) -> no.
+
+scan_crlf(Bin) ->
+ case get_crlf_pos(Bin) of
+ {yes, Pos} ->
+ {Prefix, <<_, _, Suffix/binary>>} = split_binary(Bin, Pos),
+ {yes, Prefix, Suffix};
+ no ->
+ {no, Bin}
+ end.
+
+scan_crlf(<<>>, Bin2) ->
+ scan_crlf(Bin2);
+scan_crlf(Bin1, Bin2) when size(Bin1) < 2 ->
+ scan_crlf(<<Bin1/binary, Bin2/binary>>);
+scan_crlf(Bin1, Bin2) ->
+ scan_crlf_1(size(Bin1) - 2, Bin1, Bin2).
+
+scan_crlf_1(Bin1_head_size, Bin1, Bin2) ->
+ <<Bin1_head:Bin1_head_size/binary, Bin1_tail/binary>> = Bin1,
+ Bin3 = <<Bin1_tail/binary, Bin2/binary>>,
+ case get_crlf_pos(Bin3) of
+ {yes, Pos} ->
+ {Prefix, <<_, _, Suffix/binary>>} = split_binary(Bin3, Pos),
+ {yes, list_to_binary([Bin1_head, Prefix]), Suffix};
+ no ->
+ {no, list_to_binary([Bin1, Bin2])}
+ end.
+
+get_crlf_pos(Bin) ->
+ get_crlf_pos(Bin, 0).
+
+get_crlf_pos(<<$\r, $\n, _/binary>>, Pos) -> {yes, Pos};
+get_crlf_pos(<<_, Rest/binary>>, Pos) -> get_crlf_pos(Rest, Pos + 1);
+get_crlf_pos(<<>>, _) -> no.
+
+%% scan_crlf(<<$\n, T/binary>>, [$\r | L]) -> {yes, lists:reverse(L), T};
+%% scan_crlf(<<H, T/binary>>, L) -> scan_crlf(T, [H|L]);
+%% scan_crlf(<<>>, L) -> {no, L};
+%% scan_crlf([$\n|T], [$\r | L]) -> {yes, lists:reverse(L), T};
+%% scan_crlf([H|T], L) -> scan_crlf(T, [H|L]);
+%% scan_crlf([], L) -> {no, L}.
+
+fmt_val(L) when is_list(L) -> L;
+fmt_val(I) when is_integer(I) -> integer_to_list(I);
+fmt_val(A) when is_atom(A) -> atom_to_list(A);
+fmt_val(Term) -> io_lib:format("~p", [Term]).
+
+crnl() -> "\r\n".
+
+method(get) -> "GET";
+method(post) -> "POST";
+method(head) -> "HEAD";
+method(options) -> "OPTIONS";
+method(put) -> "PUT";
+method(delete) -> "DELETE";
+method(trace) -> "TRACE";
+method(mkcol) -> "MKCOL";
+method(propfind) -> "PROPFIND";
+method(proppatch) -> "PROPPATCH";
+method(lock) -> "LOCK";
+method(unlock) -> "UNLOCK";
+method(move) -> "MOVE";
+method(copy) -> "COPY".
+
+%% From RFC 2616
+%%
+% The chunked encoding modifies the body of a message in order to
+% transfer it as a series of chunks, each with its own size indicator,
+% followed by an OPTIONAL trailer containing entity-header
+% 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)
+% 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
+% line.
+%%
+%% The parsing implemented here discards all chunk extensions. It also
+%% strips trailing spaces from the chunk size fields as Apache 1.3.27 was
+%% sending them.
+parse_chunk_header([]) ->
+ throw({error, invalid_chunk_size});
+parse_chunk_header(ChunkHeader) ->
+ parse_chunk_header(ChunkHeader, []).
+
+parse_chunk_header(<<$;, _/binary>>, Acc) ->
+ hexlist_to_integer(lists:reverse(Acc));
+parse_chunk_header(<<H, T/binary>>, Acc) ->
+ case is_whitespace(H) of
+ true ->
+ parse_chunk_header(T, Acc);
+ false ->
+ parse_chunk_header(T, [H | Acc])
+ end;
+parse_chunk_header(<<>>, Acc) ->
+ hexlist_to_integer(lists:reverse(Acc)).
+
+is_whitespace($\s) -> true;
+is_whitespace($\r) -> true;
+is_whitespace($\n) -> true;
+is_whitespace($\t) -> true;
+is_whitespace(_) -> false.
+
+
+send_async_headers(_ReqId, undefined, _StatCode, _Headers) ->
+ ok;
+send_async_headers(ReqId, StreamTo, StatCode, Headers) ->
+ catch StreamTo ! {ibrowse_async_headers, ReqId, StatCode, Headers}.
+
+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
+ end.
+
+do_reply(State, From, undefined, _, Resp_format, {ok, St_code, Headers, Body}) ->
+ Msg_1 = {ok, St_code, Headers, format_response_data(Resp_format, Body)},
+ gen_server:reply(From, Msg_1),
+ dec_pipeline_counter(State);
+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}) ->
+ 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}
+ end,
+ catch StreamTo ! {ibrowse_async_response_end, ReqId},
+ %% We don't want to delete the Req-id to Pid mapping straightaway
+ %% as the client may send a stream_next message just while we are
+ %% sending back this ibrowse_async_response_end message. If we
+ %% deleted this mapping straightaway, the caller will see a
+ %% {error, unknown_req_id} when it calls ibrowse:stream_next/1. To
+ %% get around this, we store the req id, and clear it after the
+ %% next request. If there are wierd combinations of stream,
+ %% stream_once and sync requests on the same connection, it will
+ %% take a while for the req_id-pid mapping to get cleared, but it
+ %% should do no harm.
+ ets:delete(ibrowse_stream, {req_id_pid, Prev_req_id}),
+ State_1#state{prev_req_id = ReqId};
+do_reply(State, _From, StreamTo, ReqId, Resp_format, Msg) ->
+ State_1 = dec_pipeline_counter(State),
+ Msg_1 = format_response_data(Resp_format, Msg),
+ catch StreamTo ! {ibrowse_async_response, ReqId, Msg_1},
+ State_1.
+
+do_interim_reply(undefined, _, _ReqId, _Msg) ->
+ ok;
+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) ->
+ 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}),
+ do_reply(State, From, StreamTo, ReqId, Resp_format, {error, Err})
+ end, ReqList).
+
+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,
+ do_reply(State, From, StreamTo, ReqId, Resp_format, Reply),
+ do_error_reply(State#state{reqs = Reqs_1}, previous_request_failed).
+
+split_list_at(List, N) ->
+ split_list_at(List, N, []).
+split_list_at([], _, Acc) ->
+ {lists:reverse(Acc), []};
+split_list_at(List2, 0, List1) ->
+ {lists:reverse(List1), List2};
+split_list_at([H | List2], N, List1) ->
+ split_list_at(List2, N-1, [H | List1]).
+
+hexlist_to_integer(List) ->
+ hexlist_to_integer(lists:reverse(List), 1, 0).
+hexlist_to_integer([H | T], Multiplier, Acc) ->
+ hexlist_to_integer(T, Multiplier*16, Multiplier*to_ascii(H) + Acc);
+hexlist_to_integer([], _, Acc) ->
+ Acc.
+
+to_ascii($A) -> 10;
+to_ascii($a) -> 10;
+to_ascii($B) -> 11;
+to_ascii($b) -> 11;
+to_ascii($C) -> 12;
+to_ascii($c) -> 12;
+to_ascii($D) -> 13;
+to_ascii($d) -> 13;
+to_ascii($E) -> 14;
+to_ascii($e) -> 14;
+to_ascii($F) -> 15;
+to_ascii($f) -> 15;
+to_ascii($1) -> 1;
+to_ascii($2) -> 2;
+to_ascii($3) -> 3;
+to_ascii($4) -> 4;
+to_ascii($5) -> 5;
+to_ascii($6) -> 6;
+to_ascii($7) -> 7;
+to_ascii($8) -> 8;
+to_ascii($9) -> 9;
+to_ascii($0) -> 0.
+
+cancel_timer(undefined) -> ok;
+cancel_timer(Ref) -> erlang:cancel_timer(Ref).
+
+cancel_timer(Ref, {eat_message, Msg}) ->
+ cancel_timer(Ref),
+ receive
+ Msg ->
+ ok
+ after 0 ->
+ ok
+ end.
+
+make_req_id() ->
+ now().
+
+to_lower(Str) ->
+ to_lower(Str, []).
+to_lower([H|T], Acc) when H >= $A, H =< $Z ->
+ to_lower(T, [H+32|Acc]);
+to_lower([H|T], Acc) ->
+ to_lower(T, [H|Acc]);
+to_lower([], Acc) ->
+ lists:reverse(Acc).
+
+shutting_down(#state{lb_ets_tid = undefined}) ->
+ ok;
+shutting_down(#state{lb_ets_tid = Tid,
+ cur_pipeline_size = Sz}) ->
+ catch ets:delete(Tid, {Sz, self()}).
+
+inc_pipeline_counter(#state{is_closing = true} = State) ->
+ State;
+inc_pipeline_counter(#state{cur_pipeline_size = Pipe_sz} = State) ->
+ State#state{cur_pipeline_size = Pipe_sz + 1}.
+
+dec_pipeline_counter(#state{is_closing = true} = State) ->
+ State;
+dec_pipeline_counter(#state{lb_ets_tid = undefined} = State) ->
+ State;
+dec_pipeline_counter(#state{cur_pipeline_size = Pipe_sz,
+ lb_ets_tid = Tid} = State) ->
+ ets:delete(Tid, {Pipe_sz, self()}),
+ ets:insert(Tid, {{Pipe_sz - 1, self()}, []}),
+ State#state{cur_pipeline_size = Pipe_sz - 1}.
+
+flatten([H | _] = L) when is_integer(H) ->
+ L;
+flatten([H | _] = L) when is_list(H) ->
+ lists:flatten(L);
+flatten([]) ->
+ [].
+
+get_stream_chunk_size(Options) ->
+ case lists:keysearch(stream_chunk_size, 1, Options) of
+ {value, {_, V}} when V > 0 ->
+ V;
+ _ ->
+ ?DEFAULT_STREAM_CHUNK_SIZE
+ end.
+
+get_inac_timeout(#state{cur_req = #request{options = Opts}}) ->
+ get_value(inactivity_timeout, Opts, infinity);
+get_inac_timeout(#state{cur_req = undefined}) ->
+ infinity.
diff --git a/apps/ibrowse/src/ibrowse_lb.erl b/apps/ibrowse/src/ibrowse_lb.erl
new file mode 100644
index 00000000..834054a7
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse_lb.erl
@@ -0,0 +1,216 @@
+%%%-------------------------------------------------------------------
+%%% File : ibrowse_lb.erl
+%%% Author : chandru <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%% Description :
+%%%
+%%% Created : 6 Mar 2008 by chandru <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%%-------------------------------------------------------------------
+-module(ibrowse_lb).
+
+-vsn('$Id: ibrowse_lb.erl,v 1.2 2009/07/01 22:43:19 chandrusf Exp $ ').
+-author(chandru).
+-behaviour(gen_server).
+%%--------------------------------------------------------------------
+%% Include files
+%%--------------------------------------------------------------------
+
+%%--------------------------------------------------------------------
+%% External exports
+-export([
+ start_link/1,
+ spawn_connection/5
+ ]).
+
+%% gen_server callbacks
+-export([
+ init/1,
+ handle_call/3,
+ handle_cast/2,
+ handle_info/2,
+ terminate/2,
+ code_change/3
+ ]).
+
+-record(state, {parent_pid,
+ ets_tid,
+ host,
+ port,
+ max_sessions,
+ max_pipeline_size,
+ num_cur_sessions = 0}).
+
+-include("ibrowse.hrl").
+
+%%====================================================================
+%% External functions
+%%====================================================================
+%%--------------------------------------------------------------------
+%% Function: start_link/0
+%% Description: Starts the server
+%%--------------------------------------------------------------------
+start_link(Args) ->
+ gen_server:start_link(?MODULE, Args, []).
+
+%%====================================================================
+%% Server functions
+%%====================================================================
+
+%%--------------------------------------------------------------------
+%% Function: init/1
+%% Description: Initiates the server
+%% Returns: {ok, State} |
+%% {ok, State, Timeout} |
+%% ignore |
+%% {stop, Reason}
+%%--------------------------------------------------------------------
+init([Host, Port]) ->
+ process_flag(trap_exit, true),
+ Max_sessions = ibrowse:get_config_value({max_sessions, Host, Port}, 10),
+ Max_pipe_sz = ibrowse:get_config_value({max_pipeline_size, Host, Port}, 10),
+ put(my_trace_flag, ibrowse_lib:get_trace_status(Host, Port)),
+ put(ibrowse_trace_token, ["LB: ", Host, $:, integer_to_list(Port)]),
+ Tid = ets:new(ibrowse_lb, [public, ordered_set]),
+ {ok, #state{parent_pid = whereis(ibrowse),
+ host = Host,
+ port = Port,
+ ets_tid = Tid,
+ max_pipeline_size = Max_pipe_sz,
+ max_sessions = Max_sessions}}.
+
+spawn_connection(Lb_pid, Url,
+ Max_sessions,
+ Max_pipeline_size,
+ SSL_options)
+ when is_pid(Lb_pid),
+ is_record(Url, url),
+ is_integer(Max_pipeline_size),
+ is_integer(Max_sessions) ->
+ gen_server:call(Lb_pid,
+ {spawn_connection, Url, Max_sessions, Max_pipeline_size, SSL_options}).
+%%--------------------------------------------------------------------
+%% Function: handle_call/3
+%% Description: Handling call messages
+%% Returns: {reply, Reply, State} |
+%% {reply, Reply, State, Timeout} |
+%% {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, Reply, State} | (terminate/2 is called)
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+% handle_call({spawn_connection, _Url, Max_sess, Max_pipe, _}, _From,
+% #state{max_sessions = Max_sess,
+% ets_tid = Tid,
+% max_pipeline_size = Max_pipe_sz,
+% num_cur_sessions = Num} = State)
+% when Num >= Max ->
+% Reply = find_best_connection(Tid),
+% {reply, sorry_dude_reuse, State};
+
+%% Update max_sessions in #state with supplied value
+handle_call({spawn_connection, _Url, Max_sess, Max_pipe, _}, _From,
+ #state{num_cur_sessions = Num} = State)
+ when Num >= Max_sess ->
+ State_1 = maybe_create_ets(State),
+ Reply = find_best_connection(State_1#state.ets_tid, Max_pipe),
+ {reply, Reply, State_1#state{max_sessions = Max_sess}};
+
+handle_call({spawn_connection, Url, _Max_sess, _Max_pipe, SSL_options}, _From,
+ #state{num_cur_sessions = Cur} = State) ->
+ State_1 = maybe_create_ets(State),
+ Tid = State_1#state.ets_tid,
+ {ok, Pid} = ibrowse_http_client:start_link({Tid, Url, SSL_options}),
+ ets:insert(Tid, {{1, Pid}, []}),
+ {reply, {ok, Pid}, State_1#state{num_cur_sessions = Cur + 1}};
+
+handle_call(Request, _From, State) ->
+ Reply = {unknown_request, Request},
+ {reply, Reply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: handle_cast/2
+%% Description: Handling cast messages
+%% Returns: {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+handle_cast(_Msg, State) ->
+ {noreply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: handle_info/2
+%% Description: Handling all non call/cast messages
+%% Returns: {noreply, State} |
+%% {noreply, State, Timeout} |
+%% {stop, Reason, State} (terminate/2 is called)
+%%--------------------------------------------------------------------
+handle_info({'EXIT', Parent, _Reason}, #state{parent_pid = Parent} = State) ->
+ {stop, normal, State};
+
+handle_info({'EXIT', _Pid, _Reason}, #state{ets_tid = undefined} = State) ->
+ {noreply, State};
+
+handle_info({'EXIT', Pid, _Reason},
+ #state{num_cur_sessions = Cur,
+ ets_tid = Tid} = State) ->
+ ets:match_delete(Tid, {{'_', Pid}, '_'}),
+ Cur_1 = Cur - 1,
+ State_1 = case Cur_1 of
+ 0 ->
+ ets:delete(Tid),
+ State#state{ets_tid = undefined};
+ _ ->
+ State
+ end,
+ {noreply, State_1#state{num_cur_sessions = Cur_1}};
+
+handle_info({trace, Bool}, #state{ets_tid = undefined} = State) ->
+ put(my_trace_flag, Bool),
+ {noreply, State};
+
+handle_info({trace, Bool}, #state{ets_tid = Tid} = State) ->
+ ets:foldl(fun({{_, Pid}, _}, Acc) when is_pid(Pid) ->
+ catch Pid ! {trace, Bool},
+ Acc;
+ (_, Acc) ->
+ Acc
+ end, undefined, Tid),
+ put(my_trace_flag, Bool),
+ {noreply, State};
+
+handle_info(_Info, State) ->
+ {noreply, State}.
+
+%%--------------------------------------------------------------------
+%% Function: terminate/2
+%% Description: Shutdown the server
+%% Returns: any (ignored by gen_server)
+%%--------------------------------------------------------------------
+terminate(_Reason, _State) ->
+ ok.
+
+%%--------------------------------------------------------------------
+%% Func: code_change/3
+%% Purpose: Convert process state when code is changed
+%% Returns: {ok, NewState}
+%%--------------------------------------------------------------------
+code_change(_OldVsn, State, _Extra) ->
+ {ok, State}.
+
+%%--------------------------------------------------------------------
+%%% Internal functions
+%%--------------------------------------------------------------------
+find_best_connection(Tid, Max_pipe) ->
+ case ets:first(Tid) of
+ {Cur_sz, Pid} when Cur_sz < Max_pipe ->
+ ets:delete(Tid, {Cur_sz, Pid}),
+ ets:insert(Tid, {{Cur_sz + 1, Pid}, []}),
+ {ok, Pid};
+ _ ->
+ {error, retry_later}
+ end.
+
+maybe_create_ets(#state{ets_tid = undefined} = State) ->
+ Tid = ets:new(ibrowse_lb, [public, ordered_set]),
+ State#state{ets_tid = Tid};
+maybe_create_ets(State) ->
+ State.
diff --git a/apps/ibrowse/src/ibrowse_lib.erl b/apps/ibrowse/src/ibrowse_lib.erl
new file mode 100644
index 00000000..6c7b1546
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse_lib.erl
@@ -0,0 +1,399 @@
+%%% File : ibrowse_lib.erl
+%%% Author : Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%% Description :
+%%% Created : 27 Feb 2004 by Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%% @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).
+-endif.
+
+-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(Host, Port) ->
+ ibrowse:get_config_value({trace, Host, Port}, false).
+
+drv_ue(Str) ->
+ [{port, Port}| _] = ets:lookup(ibrowse_table, port),
+ drv_ue(Str, Port).
+drv_ue(Str, Port) ->
+ case erlang:port_control(Port, 1, Str) of
+ [] ->
+ Str;
+ Res ->
+ Res
+ end.
+
+%% @doc URL-encodes a string based on RFC 1738. Returns a flat list.
+%% @spec url_encode(Str) -> UrlEncodedStr
+%% Str = string()
+%% UrlEncodedStr = string()
+url_encode(Str) when is_list(Str) ->
+ url_encode_char(lists:reverse(Str), []).
+
+url_encode_char([X | T], Acc) when X >= $0, X =< $9 ->
+ url_encode_char(T, [X | Acc]);
+url_encode_char([X | T], Acc) when X >= $a, X =< $z ->
+ url_encode_char(T, [X | Acc]);
+url_encode_char([X | T], Acc) when X >= $A, X =< $Z ->
+ url_encode_char(T, [X | Acc]);
+url_encode_char([X | T], Acc) when X == $-; X == $_; X == $. ->
+ url_encode_char(T, [X | Acc]);
+url_encode_char([32 | T], Acc) ->
+ url_encode_char(T, [$+ | Acc]);
+url_encode_char([X | T], Acc) ->
+ url_encode_char(T, [$%, d2h(X bsr 4), d2h(X band 16#0f) | Acc]);
+url_encode_char([], Acc) ->
+ Acc.
+
+d2h(N) when N<10 -> N+$0;
+d2h(N) -> N+$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
+ end.
+
+% TODO: Have to handle the Zone
+decode_rfc822_date_1([_,DayInt,Month,Year, Time,Zone]) ->
+ decode_rfc822_date_1([DayInt,Month,Year, Time,Zone]);
+decode_rfc822_date_1([Day,Month,Year, Time,_Zone]) ->
+ DayI = list_to_integer(Day),
+ 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,
+ {{YearI,MonthI,DayI}, TimeTup}.
+
+month_int("Jan") -> 1;
+month_int("Feb") -> 2;
+month_int("Mar") -> 3;
+month_int("Apr") -> 4;
+month_int("May") -> 5;
+month_int("Jun") -> 6;
+month_int("Jul") -> 7;
+month_int("Aug") -> 8;
+month_int("Sep") -> 9;
+month_int("Oct") -> 10;
+month_int("Nov") -> 11;
+month_int("Dec") -> 12.
+
+%% @doc Given a status code, returns an atom describing the status code.
+%% @spec status_code(StatusCode::status_code()) -> StatusDescription
+%% status_code() = string() | integer()
+%% StatusDescription = atom()
+status_code(100) -> continue;
+status_code(101) -> switching_protocols;
+status_code(102) -> processing;
+status_code(200) -> ok;
+status_code(201) -> created;
+status_code(202) -> accepted;
+status_code(203) -> non_authoritative_information;
+status_code(204) -> no_content;
+status_code(205) -> reset_content;
+status_code(206) -> partial_content;
+status_code(207) -> multi_status;
+status_code(300) -> multiple_choices;
+status_code(301) -> moved_permanently;
+status_code(302) -> found;
+status_code(303) -> see_other;
+status_code(304) -> not_modified;
+status_code(305) -> use_proxy;
+status_code(306) -> unused;
+status_code(307) -> temporary_redirect;
+status_code(400) -> bad_request;
+status_code(401) -> unauthorized;
+status_code(402) -> payment_required;
+status_code(403) -> forbidden;
+status_code(404) -> not_found;
+status_code(405) -> method_not_allowed;
+status_code(406) -> not_acceptable;
+status_code(407) -> proxy_authentication_required;
+status_code(408) -> request_timeout;
+status_code(409) -> conflict;
+status_code(410) -> gone;
+status_code(411) -> length_required;
+status_code(412) -> precondition_failed;
+status_code(413) -> request_entity_too_large;
+status_code(414) -> request_uri_too_long;
+status_code(415) -> unsupported_media_type;
+status_code(416) -> requested_range_not_satisfiable;
+status_code(417) -> expectation_failed;
+status_code(422) -> unprocessable_entity;
+status_code(423) -> locked;
+status_code(424) -> failed_dependency;
+status_code(500) -> internal_server_error;
+status_code(501) -> not_implemented;
+status_code(502) -> bad_gateway;
+status_code(503) -> service_unavailable;
+status_code(504) -> gateway_timeout;
+status_code(505) -> http_version_not_supported;
+status_code(507) -> insufficient_storage;
+status_code(X) when is_list(X) -> status_code(list_to_integer(X));
+status_code(_) -> unknown_status_code.
+
+%% @doc dec2hex taken from gtk.erl in std dist
+%% M = integer() -- number of hex digits required
+%% N = integer() -- the number to represent as hex
+%% @spec dec2hex(M::integer(), N::integer()) -> string()
+dec2hex(M,N) -> dec2hex(M,N,[]).
+
+dec2hex(0,_N,Ack) -> Ack;
+dec2hex(M,N,Ack) -> dec2hex(M-1,N bsr 4,[d2h(N band 15)|Ack]).
+
+%% @doc Implements the base64 encoding algorithm. The output data type matches in the input data type.
+%% @spec encode_base64(In) -> Out
+%% In = string() | binary()
+%% Out = string() | binary()
+encode_base64(List) when is_list(List) ->
+ encode_base64_1(list_to_binary(List));
+encode_base64(Bin) when is_binary(Bin) ->
+ List = encode_base64_1(Bin),
+ list_to_binary(List).
+
+encode_base64_1(<<A:6, B:6, C:6, D:6, Rest/binary>>) ->
+ [int_to_b64(A), int_to_b64(B),
+ int_to_b64(C), int_to_b64(D) | encode_base64_1(Rest)];
+encode_base64_1(<<A:6, B:6, C:4>>) ->
+ [int_to_b64(A), int_to_b64(B), int_to_b64(C bsl 2), $=];
+encode_base64_1(<<A:6, B:2>>) ->
+ [int_to_b64(A), int_to_b64(B bsl 4), $=, $=];
+encode_base64_1(<<>>) ->
+ [].
+
+%% @doc Implements the base64 decoding algorithm. The output data type matches in the input data type.
+%% @spec decode_base64(In) -> Out | exit({error, invalid_input})
+%% In = string() | binary()
+%% Out = string() | binary()
+decode_base64(List) when is_list(List) ->
+ decode_base64_1(List, []);
+decode_base64(Bin) when is_binary(Bin) ->
+ List = decode_base64_1(binary_to_list(Bin), []),
+ list_to_binary(List).
+
+decode_base64_1([H | T], Acc) when ((H == $\t) or
+ (H == 32) or
+ (H == $\r) or
+ (H == $\n)) ->
+ decode_base64_1(T, Acc);
+
+decode_base64_1([$=, $=], Acc) ->
+ lists:reverse(Acc);
+decode_base64_1([$=, _ | _], _Acc) ->
+ exit({error, invalid_input});
+
+decode_base64_1([A1, B1, $=, $=], Acc) ->
+ A = b64_to_int(A1),
+ B = b64_to_int(B1),
+ Oct1 = (A bsl 2) bor (B bsr 4),
+ decode_base64_1([], [Oct1 | Acc]);
+decode_base64_1([A1, B1, C1, $=], Acc) ->
+ A = b64_to_int(A1),
+ B = b64_to_int(B1),
+ C = b64_to_int(C1),
+ Oct1 = (A bsl 2) bor (B bsr 4),
+ Oct2 = ((B band 16#f) bsl 6) bor (C bsr 2),
+ decode_base64_1([], [Oct2, Oct1 | Acc]);
+decode_base64_1([A1, B1, C1, D1 | T], Acc) ->
+ A = b64_to_int(A1),
+ B = b64_to_int(B1),
+ C = b64_to_int(C1),
+ D = b64_to_int(D1),
+ Oct1 = (A bsl 2) bor (B bsr 4),
+ Oct2 = ((B band 16#f) bsl 4) bor (C bsr 2),
+ Oct3 = ((C band 2#11) bsl 6) bor D,
+ decode_base64_1(T, [Oct3, Oct2, Oct1 | Acc]);
+decode_base64_1([], Acc) ->
+ lists:reverse(Acc).
+
+%% Taken from httpd_util.erl
+int_to_b64(X) when X >= 0, X =< 25 -> X + $A;
+int_to_b64(X) when X >= 26, X =< 51 -> X - 26 + $a;
+int_to_b64(X) when X >= 52, X =< 61 -> X - 52 + $0;
+int_to_b64(62) -> $+;
+int_to_b64(63) -> $/.
+
+%% Taken from httpd_util.erl
+b64_to_int(X) when X >= $A, X =< $Z -> X - $A;
+b64_to_int(X) when X >= $a, X =< $z -> X - $a + 26;
+b64_to_int(X) when X >= $0, X =< $9 -> X - $0 + 52;
+b64_to_int($+) -> 62;
+b64_to_int($/) -> 63.
+
+get_value(Tag, TVL, DefVal) ->
+ case lists:keysearch(Tag, 1, TVL) of
+ false ->
+ DefVal;
+ {value, {_, Val}} ->
+ Val
+ end.
+
+get_value(Tag, TVL) ->
+ {value, {_, V}} = lists:keysearch(Tag,1,TVL),
+ V.
+
+parse_url(Url) ->
+ parse_url(Url, get_protocol, #url{abspath=Url}, []).
+
+parse_url([$:, $/, $/ | _], get_protocol, Url, []) ->
+ {invalid_uri_1, Url};
+parse_url([$:, $/, $/ | T], get_protocol, Url, TmpAcc) ->
+ Prot = list_to_atom(lists:reverse(TmpAcc)),
+ parse_url(T, get_username,
+ Url#url{protocol = Prot},
+ []);
+parse_url([$/ | T], get_username, Url, TmpAcc) ->
+ %% No username/password. No port number
+ Url#url{host = lists:reverse(TmpAcc),
+ port = default_port(Url#url.protocol),
+ path = [$/ | T]};
+parse_url([$: | 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
+ %% a username/password. If we encounter a '@' later on, there is a
+ %% username/password indeed. If we encounter a '/', it was
+ %% actually the hostname
+ parse_url(T, get_password,
+ Url#url{username = lists:reverse(TmpAcc)},
+ []);
+parse_url([$@ | T], get_username, Url, TmpAcc) ->
+ parse_url(T, get_host,
+ Url#url{username = lists:reverse(TmpAcc),
+ password = ""},
+ []);
+parse_url([$@ | T], get_password, Url, TmpAcc) ->
+ parse_url(T, get_host,
+ Url#url{password = lists:reverse(TmpAcc)},
+ []);
+parse_url([$/ | T], get_password, Url, TmpAcc) ->
+ %% Ok, what we thought was the username/password was the hostname
+ %% and portnumber
+ #url{username=User} = Url,
+ Port = list_to_integer(lists:reverse(TmpAcc)),
+ Url#url{host = User,
+ port = Port,
+ username = undefined,
+ password = undefined,
+ path = [$/ | T]};
+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),
+ port = default_port(Prot),
+ path = [$/ | T]};
+parse_url([$/ | T], get_port, #url{protocol=Prot} = Url, TmpAcc) ->
+ Port = case TmpAcc of
+ [] ->
+ default_port(Prot);
+ _ ->
+ list_to_integer(lists:reverse(TmpAcc))
+ end,
+ Url#url{port = Port, path = [$/ | T]};
+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 = "/"};
+parse_url([], get_username, Url, TmpAcc) when TmpAcc /= [] ->
+ Url#url{host = lists:reverse(TmpAcc),
+ 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,
+ Url#url{port = Port,
+ 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,
+ Url#url{host = User,
+ port = Port,
+ username = undefined,
+ password = undefined,
+ path = "/"};
+parse_url([], State, Url, TmpAcc) ->
+ {invalid_uri_2, State, Url, TmpAcc}.
+
+default_port(http) -> 80;
+default_port(https) -> 443;
+default_port(ftp) -> 21.
+
+printable_date() ->
+ {{Y,Mo,D},{H, M, S}} = calendar:local_time(),
+ {_,_,MicroSecs} = now(),
+ [integer_to_list(Y),
+ $-,
+ integer_to_list(Mo),
+ $-,
+ integer_to_list(D),
+ $_,
+ integer_to_list(H),
+ $:,
+ integer_to_list(M),
+ $:,
+ integer_to_list(S),
+ $:,
+ integer_to_list(MicroSecs div 1000)].
+
+do_trace(Fmt, Args) ->
+ do_trace(get(my_trace_flag), Fmt, Args).
+
+-ifdef(DEBUG).
+do_trace(_, Fmt, Args) ->
+ io:format("~s -- (~s) - "++Fmt,
+ [printable_date(),
+ get(ibrowse_trace_token) | Args]).
+-else.
+do_trace(true, Fmt, Args) ->
+ io:format("~s -- (~s) - "++Fmt,
+ [printable_date(),
+ get(ibrowse_trace_token) | Args]);
+do_trace(_, _, _) ->
+ ok.
+-endif.
diff --git a/apps/ibrowse/src/ibrowse_sup.erl b/apps/ibrowse/src/ibrowse_sup.erl
new file mode 100644
index 00000000..1b9b863a
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse_sup.erl
@@ -0,0 +1,65 @@
+%%%-------------------------------------------------------------------
+%%% File : ibrowse_sup.erl
+%%% Author : Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%% Description :
+%%%
+%%% Created : 15 Oct 2003 by Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%%-------------------------------------------------------------------
+-module(ibrowse_sup).
+-vsn('$Id: ibrowse_sup.erl,v 1.1 2005/05/05 22:28:28 chandrusf Exp $ ').
+
+-behaviour(supervisor).
+%%--------------------------------------------------------------------
+%% Include files
+%%--------------------------------------------------------------------
+
+%%--------------------------------------------------------------------
+%% External exports
+%%--------------------------------------------------------------------
+-export([
+ start_link/0
+ ]).
+
+%%--------------------------------------------------------------------
+%% Internal exports
+%%--------------------------------------------------------------------
+-export([
+ init/1
+ ]).
+
+%%--------------------------------------------------------------------
+%% Macros
+%%--------------------------------------------------------------------
+-define(SERVER, ?MODULE).
+
+%%--------------------------------------------------------------------
+%% Records
+%%--------------------------------------------------------------------
+
+%%====================================================================
+%% External functions
+%%====================================================================
+%%--------------------------------------------------------------------
+%% Function: start_link/0
+%% Description: Starts the supervisor
+%%--------------------------------------------------------------------
+start_link() ->
+ supervisor:start_link({local, ?SERVER}, ?MODULE, []).
+
+%%====================================================================
+%% Server functions
+%%====================================================================
+%%--------------------------------------------------------------------
+%% Func: init/1
+%% Returns: {ok, {SupFlags, [ChildSpec]}} |
+%% ignore |
+%% {error, Reason}
+%%--------------------------------------------------------------------
+init([]) ->
+ AChild = {ibrowse,{ibrowse,start_link,[]},
+ permanent,2000,worker,[ibrowse, ibrowse_http_client]},
+ {ok,{{one_for_all,10,1}, [AChild]}}.
+
+%%====================================================================
+%% Internal functions
+%%====================================================================
diff --git a/apps/ibrowse/src/ibrowse_test.erl b/apps/ibrowse/src/ibrowse_test.erl
new file mode 100644
index 00000000..3dc66ecf
--- /dev/null
+++ b/apps/ibrowse/src/ibrowse_test.erl
@@ -0,0 +1,377 @@
+%%% File : ibrowse_test.erl
+%%% Author : Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+%%% Description : Test ibrowse
+%%% Created : 14 Oct 2003 by Chandrashekhar Mullaparthi <chandrashekhar.mullaparthi@t-mobile.co.uk>
+
+-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,
+ do_send_req/2,
+ unit_tests/0,
+ unit_tests/1,
+ unit_tests_1/2,
+ drv_ue_test/0,
+ drv_ue_test/1,
+ ue_test/0,
+ ue_test/1,
+ verify_chunked_streaming/0,
+ verify_chunked_streaming/1,
+ i_do_async_req_list/4,
+ test_stream_once/3,
+ test_stream_once/4
+ ]).
+
+test_stream_once(Url, Method, Options) ->
+ test_stream_once(Url, Method, Options, 5000).
+
+test_stream_once(Url, Method, Options, Timeout) ->
+ case ibrowse:send_req(Url, [], Method, [], [{stream_to, {self(), once}} | Options], Timeout) of
+ {ibrowse_req_id, Req_id} ->
+ case ibrowse:stream_next(Req_id) of
+ ok ->
+ test_stream_once(Req_id);
+ Err ->
+ Err
+ end;
+ Err ->
+ Err
+ end.
+
+test_stream_once(Req_id) ->
+ receive
+ {ibrowse_async_headers, Req_id, StatCode, Headers} ->
+ io:format("Recvd headers~n~p~n", [{ibrowse_async_headers, Req_id, StatCode, Headers}]),
+ case ibrowse:stream_next(Req_id) of
+ ok ->
+ test_stream_once(Req_id);
+ Err ->
+ Err
+ end;
+ {ibrowse_async_response, Req_id, {error, Err}} ->
+ io:format("Recvd error: ~p~n", [Err]);
+ {ibrowse_async_response, Req_id, Body_1} ->
+ io:format("Recvd body part: ~n~p~n", [{ibrowse_async_response, Req_id, Body_1}]),
+ case ibrowse:stream_next(Req_id) of
+ ok ->
+ test_stream_once(Req_id);
+ Err ->
+ Err
+ end;
+ {ibrowse_async_response_end, Req_id} ->
+ ok
+ end.
+%% Use ibrowse:set_max_sessions/3 and ibrowse:set_max_pipeline_size/3 to
+%% tweak settings before running the load test. The defaults are 10 and 10.
+load_test(Url, NumWorkers, NumReqsPerWorker) when is_list(Url),
+ is_integer(NumWorkers),
+ is_integer(NumReqsPerWorker),
+ NumWorkers > 0,
+ NumReqsPerWorker > 0 ->
+ proc_lib:spawn(?MODULE, send_reqs_1, [Url, NumWorkers, NumReqsPerWorker]).
+
+send_reqs_1(Url, NumWorkers, NumReqsPerWorker) ->
+ Start_time = now(),
+ ets:new(pid_table, [named_table, public]),
+ ets:new(ibrowse_test_results, [named_table, public]),
+ ets:new(ibrowse_errors, [named_table, public, ordered_set]),
+ init_results(),
+ process_flag(trap_exit, true),
+ log_msg("Starting spawning of workers...~n", []),
+ spawn_workers(Url, NumWorkers, NumReqsPerWorker),
+ log_msg("Finished spawning workers...~n", []),
+ do_wait(),
+ End_time = now(),
+ log_msg("All workers are done...~n", []),
+ log_msg("ibrowse_test_results table: ~n~p~n", [ets:tab2list(ibrowse_test_results)]),
+ log_msg("Start time: ~1000.p~n", [calendar:now_to_local_time(Start_time)]),
+ log_msg("End time : ~1000.p~n", [calendar:now_to_local_time(End_time)]),
+ Elapsed_time_secs = trunc(timer:now_diff(End_time, Start_time) / 1000000),
+ log_msg("Elapsed : ~p~n", [Elapsed_time_secs]),
+ log_msg("Reqs/sec : ~p~n", [round(trunc((NumWorkers*NumReqsPerWorker) / Elapsed_time_secs))]),
+ dump_errors().
+
+init_results() ->
+ ets:insert(ibrowse_test_results, {crash, 0}),
+ ets:insert(ibrowse_test_results, {send_failed, 0}),
+ ets:insert(ibrowse_test_results, {other_error, 0}),
+ ets:insert(ibrowse_test_results, {success, 0}),
+ ets:insert(ibrowse_test_results, {retry_later, 0}),
+ ets:insert(ibrowse_test_results, {trid_mismatch, 0}),
+ ets:insert(ibrowse_test_results, {success_no_trid, 0}),
+ ets:insert(ibrowse_test_results, {failed, 0}),
+ ets:insert(ibrowse_test_results, {timeout, 0}),
+ ets:insert(ibrowse_test_results, {req_id, 0}).
+
+spawn_workers(_Url, 0, _) ->
+ ok;
+spawn_workers(Url, NumWorkers, NumReqsPerWorker) ->
+ Pid = proc_lib:spawn_link(?MODULE, do_send_req, [Url, NumReqsPerWorker]),
+ ets:insert(pid_table, {Pid, []}),
+ spawn_workers(Url, NumWorkers - 1, NumReqsPerWorker).
+
+do_wait() ->
+ receive
+ {'EXIT', _, normal} ->
+ do_wait();
+ {'EXIT', Pid, Reason} ->
+ ets:delete(pid_table, Pid),
+ ets:insert(ibrowse_errors, {Pid, Reason}),
+ ets:update_counter(ibrowse_test_results, crash, 1),
+ do_wait();
+ Msg ->
+ io:format("Recvd unknown message...~p~n", [Msg]),
+ do_wait()
+ after 1000 ->
+ case ets:info(pid_table, size) of
+ 0 ->
+ done;
+ _ ->
+ do_wait()
+ end
+ end.
+
+do_send_req(Url, NumReqs) ->
+ do_send_req_1(Url, NumReqs).
+
+do_send_req_1(_Url, 0) ->
+ ets:delete(pid_table, self());
+do_send_req_1(Url, NumReqs) ->
+ Counter = integer_to_list(ets:update_counter(ibrowse_test_results, req_id, 1)),
+ case ibrowse:send_req(Url, [{"ib_req_id", Counter}], get, [], [], 10000) of
+ {ok, _Status, Headers, _Body} ->
+ case lists:keysearch("ib_req_id", 1, Headers) of
+ {value, {_, Counter}} ->
+ ets:update_counter(ibrowse_test_results, success, 1);
+ {value, _} ->
+ ets:update_counter(ibrowse_test_results, trid_mismatch, 1);
+ false ->
+ ets:update_counter(ibrowse_test_results, success_no_trid, 1)
+ end;
+ {error, req_timedout} ->
+ ets:update_counter(ibrowse_test_results, timeout, 1);
+ {error, send_failed} ->
+ ets:update_counter(ibrowse_test_results, send_failed, 1);
+ {error, retry_later} ->
+ ets:update_counter(ibrowse_test_results, retry_later, 1);
+ Err ->
+ ets:insert(ibrowse_errors, {now(), Err}),
+ ets:update_counter(ibrowse_test_results, other_error, 1),
+ ok
+ end,
+ do_send_req_1(Url, NumReqs-1).
+
+dump_errors() ->
+ case ets:info(ibrowse_errors, size) of
+ 0 ->
+ ok;
+ _ ->
+ {A, B, C} = now(),
+ Filename = lists:flatten(
+ io_lib:format("ibrowse_errors_~p_~p_~p.txt" , [A, B, C])),
+ case file:open(Filename, [write, delayed_write, raw]) of
+ {ok, Iod} ->
+ dump_errors(ets:first(ibrowse_errors), Iod);
+ Err ->
+ io:format("failed to create file ~s. Reason: ~p~n", [Filename, Err]),
+ ok
+ end
+ end.
+
+dump_errors('$end_of_table', Iod) ->
+ file:close(Iod);
+dump_errors(Key, Iod) ->
+ [{_, Term}] = ets:lookup(ibrowse_errors, Key),
+ file:write(Iod, io_lib:format("~p~n", [Term])),
+ dump_errors(ets:next(ibrowse_errors, Key), Iod).
+
+%%------------------------------------------------------------------------------
+%% Unit Tests
+%%------------------------------------------------------------------------------
+-define(TEST_LIST, [{"http://intranet/messenger", get},
+ {"http://www.google.co.uk", get},
+ {"http://www.google.com", get},
+ {"http://www.google.com", options},
+ {"http://www.sun.com", get},
+ {"http://www.oracle.com", get},
+ {"http://www.bbc.co.uk", get},
+ {"http://www.bbc.co.uk", trace},
+ {"http://www.bbc.co.uk", options},
+ {"http://yaws.hyber.org", get},
+ {"http://jigsaw.w3.org/HTTP/ChunkedScript", get},
+ {"http://jigsaw.w3.org/HTTP/TE/foo.txt", get},
+ {"http://jigsaw.w3.org/HTTP/TE/bar.txt", get},
+ {"http://jigsaw.w3.org/HTTP/connection.html", get},
+ {"http://jigsaw.w3.org/HTTP/cc.html", get},
+ {"http://jigsaw.w3.org/HTTP/cc-private.html", get},
+ {"http://jigsaw.w3.org/HTTP/cc-proxy-revalidate.html", get},
+ {"http://jigsaw.w3.org/HTTP/cc-nocache.html", get},
+ {"http://jigsaw.w3.org/HTTP/h-content-md5.html", get},
+ {"http://jigsaw.w3.org/HTTP/h-retry-after.html", get},
+ {"http://jigsaw.w3.org/HTTP/h-retry-after-date.html", get},
+ {"http://jigsaw.w3.org/HTTP/neg", get},
+ {"http://jigsaw.w3.org/HTTP/negbad", get},
+ {"http://jigsaw.w3.org/HTTP/400/toolong/", get},
+ {"http://jigsaw.w3.org/HTTP/300/", get},
+ {"http://jigsaw.w3.org/HTTP/Basic/", get, [{basic_auth, {"guest", "guest"}}]},
+ {"http://jigsaw.w3.org/HTTP/CL/", get},
+ {"http://www.httpwatch.com/httpgallery/chunked/", get}
+ ]).
+
+unit_tests() ->
+ unit_tests([]).
+
+unit_tests(Options) ->
+ Options_1 = Options ++ [{connect_timeout, 5000}],
+ {Pid, Ref} = erlang:spawn_monitor(?MODULE, unit_tests_1, [self(), Options_1]),
+ receive
+ {done, Pid} ->
+ ok;
+ {'DOWN', Ref, _, _, Info} ->
+ io:format("Test process crashed: ~p~n", [Info])
+ after 60000 ->
+ exit(Pid, kill),
+ io:format("Timed out waiting for tests to complete~n", [])
+ end.
+
+unit_tests_1(Parent, Options) ->
+ lists:foreach(fun({Url, Method}) ->
+ execute_req(Url, Method, Options);
+ ({Url, Method, X_Opts}) ->
+ execute_req(Url, Method, X_Opts ++ Options)
+ end, ?TEST_LIST),
+ Parent ! {done, self()}.
+
+verify_chunked_streaming() ->
+ verify_chunked_streaming([]).
+
+verify_chunked_streaming(Options) ->
+ Url = "http://www.httpwatch.com/httpgallery/chunked/",
+ io:format("URL: ~s~n", [Url]),
+ io:format("Fetching data without streaming...~n", []),
+ Result_without_streaming = ibrowse:send_req(
+ Url, [], get, [],
+ [{response_format, binary} | Options]),
+ io:format("Fetching data with streaming as list...~n", []),
+ Async_response_list = do_async_req_list(
+ Url, get, [{response_format, list} | Options]),
+ io:format("Fetching data with streaming as binary...~n", []),
+ Async_response_bin = do_async_req_list(
+ Url, get, [{response_format, binary} | Options]),
+ compare_responses(Result_without_streaming, Async_response_list, Async_response_bin).
+
+compare_responses({ok, St_code, _, Body}, {ok, St_code, _, Body}, {ok, St_code, _, Body}) ->
+ success;
+compare_responses({ok, St_code, _, Body_1}, {ok, St_code, _, Body_2}, {ok, St_code, _, Body_3}) ->
+ case Body_1 of
+ Body_2 ->
+ io:format("Body_1 and Body_2 match~n", []);
+ Body_3 ->
+ io:format("Body_1 and Body_3 match~n", []);
+ _ when Body_2 == Body_3 ->
+ io:format("Body_2 and Body_3 match~n", []);
+ _ ->
+ io:format("All three bodies are different!~n", [])
+ end,
+ io:format("Body_1 -> ~p~n", [Body_1]),
+ io:format("Body_2 -> ~p~n", [Body_2]),
+ io:format("Body_3 -> ~p~n", [Body_3]),
+ fail_bodies_mismatch;
+compare_responses(R1, R2, R3) ->
+ io:format("R1 -> ~p~n", [R1]),
+ io:format("R2 -> ~p~n", [R2]),
+ io:format("R3 -> ~p~n", [R3]),
+ fail.
+
+%% do_async_req_list(Url) ->
+%% do_async_req_list(Url, get).
+
+%% do_async_req_list(Url, Method) ->
+%% do_async_req_list(Url, Method, [{stream_to, self()},
+%% {stream_chunk_size, 1000}]).
+
+do_async_req_list(Url, Method, Options) ->
+ {Pid,_} = erlang:spawn_monitor(?MODULE, i_do_async_req_list,
+ [self(), Url, Method,
+ Options ++ [{stream_chunk_size, 1000}]]),
+ io:format("Spawned process ~p~n", [Pid]),
+ wait_for_resp(Pid).
+
+wait_for_resp(Pid) ->
+ receive
+ {async_result, Pid, Res} ->
+ Res;
+ {async_result, Other_pid, _} ->
+ io:format("~p: Waiting for result from ~p: got from ~p~n", [self(), Pid, Other_pid]),
+ wait_for_resp(Pid);
+ {'DOWN', _, _, Pid, Reason} ->
+ {'EXIT', Reason};
+ {'DOWN', _, _, _, _} ->
+ wait_for_resp(Pid);
+ Msg ->
+ io:format("Recvd unknown message: ~p~n", [Msg]),
+ wait_for_resp(Pid)
+ after 10000 ->
+ {error, timeout}
+ end.
+
+i_do_async_req_list(Parent, Url, Method, Options) ->
+ Res = ibrowse:send_req(Url, [], Method, [], [{stream_to, self()} | Options]),
+ case Res of
+ {ibrowse_req_id, Req_id} ->
+ Result = wait_for_async_resp(Req_id, undefined, undefined, []),
+ Parent ! {async_result, self(), Result};
+ Err ->
+ Parent ! {async_result, self(), Err}
+ end.
+
+wait_for_async_resp(Req_id, Acc_Stat_code, Acc_Headers, Body) ->
+ receive
+ {ibrowse_async_headers, Req_id, StatCode, Headers} ->
+ wait_for_async_resp(Req_id, StatCode, Headers, Body);
+ {ibrowse_async_response_end, Req_id} ->
+ Body_1 = list_to_binary(lists:reverse(Body)),
+ {ok, Acc_Stat_code, Acc_Headers, Body_1};
+ {ibrowse_async_response, Req_id, Data} ->
+ wait_for_async_resp(Req_id, Acc_Stat_code, Acc_Headers, [Data | Body]);
+ Err ->
+ {ok, Acc_Stat_code, Acc_Headers, Err}
+ end.
+
+execute_req(Url, Method, Options) ->
+ io:format("~7.7w, ~50.50s: ", [Method, Url]),
+ Result = (catch ibrowse:send_req(Url, [], Method, [], Options)),
+ case Result of
+ {ok, SCode, _H, _B} ->
+ io:format("Status code: ~p~n", [SCode]);
+ Err ->
+ io:format("Err -> ~p~n", [Err])
+ end.
+
+drv_ue_test() ->
+ drv_ue_test(lists:duplicate(1024, 127)).
+drv_ue_test(Data) ->
+ [{port, Port}| _] = ets:lookup(ibrowse_table, port),
+% erl_ddll:unload_driver("ibrowse_drv"),
+% timer:sleep(1000),
+% erl_ddll:load_driver("../priv", "ibrowse_drv"),
+% Port = open_port({spawn, "ibrowse_drv"}, []),
+ {Time, Res} = timer:tc(ibrowse_lib, drv_ue, [Data, Port]),
+ io:format("Time -> ~p~n", [Time]),
+ io:format("Data Length -> ~p~n", [length(Data)]),
+ io:format("Res Length -> ~p~n", [length(Res)]).
+% io:format("Result -> ~s~n", [Res]).
+
+ue_test() ->
+ ue_test(lists:duplicate(1024, $?)).
+ue_test(Data) ->
+ {Time, Res} = timer:tc(ibrowse_lib, url_encode, [Data]),
+ io:format("Time -> ~p~n", [Time]),
+ io:format("Data Length -> ~p~n", [length(Data)]),
+ io:format("Res Length -> ~p~n", [length(Res)]).
+% io:format("Result -> ~s~n", [Res]).
+
+log_msg(Fmt, Args) ->
+ io:format("~s -- " ++ Fmt,
+ [ibrowse_lib:printable_date() | Args]).