improve connection to remote

NOTICE

@@ -21,3 +21,6 @@ Copyright (c) 2009, Erlang Training and Consulting Ltd.
 Copyright (C) 1998 - 2014, Daniel Stenberg, <daniel@haxx.se>, et al.
 
 *) hackney_trace (C) 2015 under the Erlang Public LicensE
+
+*) hackney_cidr is based on inet_cidr 1.2.1. vendored for customer purpose.
+Copyright (c) 2024, Enki Multimedia , MIT License

include/hackney.hrl

@@ -59,3 +59,4 @@
 
 -define(HTTP_PROXY_ENV_VARS, ["http_proxy", "HTTP_PROXY", "all_proxy", "ALL_PROXY"]).
 -define(HTTPS_PROXY_ENV_VARS, ["https_proxy", "HTTPS_PROXY", "all_proxy", "ALL_PROXY"]).
+-define(HTTP_NO_PROXY_ENV_VARS, ["no_proxy", "NO_PROXY"]).

rebar.lock

@@ -1,5 +1,5 @@
 {"1.2.0",
-[{<<"certifi">>,{pkg,<<"certifi">>,<<"2.12.0">>},0},
+[{<<"certifi">>,{pkg,<<"certifi">>,<<"2.13.0">>},0},
  {<<"idna">>,{pkg,<<"idna">>,<<"6.1.1">>},0},
  {<<"metrics">>,{pkg,<<"metrics">>,<<"1.0.1">>},0},
  {<<"mimerl">>,{pkg,<<"mimerl">>,<<"1.2.0">>},0},
@@ -8,15 +8,15 @@
  {<<"unicode_util_compat">>,{pkg,<<"unicode_util_compat">>,<<"0.7.0">>},0}]}.
 [
 {pkg_hash,[
- {<<"certifi">>, <<"2D1CCA2EC95F59643862AF91F001478C9863C2AC9CB6E2F89780BFD8DE987329">>},
+ {<<"certifi">>, <<"E52BE248590050B2DD33B0BB274B56678F9068E67805DCA8AA8B1CCDB016BBF6">>},
  {<<"idna">>, <<"8A63070E9F7D0C62EB9D9FCB360A7DE382448200FBBD1B106CC96D3D8099DF8D">>},
  {<<"metrics">>, <<"25F094DEA2CDA98213CECC3AEFF09E940299D950904393B2A29D191C346A8486">>},
  {<<"mimerl">>, <<"67E2D3F571088D5CFD3E550C383094B47159F3EEE8FFA08E64106CDF5E981BE3">>},
  {<<"parse_trans">>, <<"6E6AA8167CB44CC8F39441D05193BE6E6F4E7C2946CB2759F015F8C56B76E5FF">>},
  {<<"ssl_verify_fun">>, <<"354C321CF377240C7B8716899E182CE4890C5938111A1296ADD3EC74CF1715DF">>},
  {<<"unicode_util_compat">>, <<"BC84380C9AB48177092F43AC89E4DFA2C6D62B40B8BD132B1059ECC7232F9A78">>}]},
 {pkg_hash_ext,[
- {<<"certifi">>, <<"EE68D85DF22E554040CDB4BE100F33873AC6051387BAF6A8F6CE82272340FF1C">>},
+ {<<"certifi">>, <<"8F3D9533A0F06070AFDFD5D596B32E21C6580667A492891851B0E2737BC507A1">>},
  {<<"idna">>, <<"92376EB7894412ED19AC475E4A86F7B413C1B9FBB5BD16DCCD57934157944CEA">>},
  {<<"metrics">>, <<"69B09ADDDC4F74A40716AE54D140F93BEB0FB8978D8636EADED0C31B6F099F16">>},
  {<<"mimerl">>, <<"F278585650AA581986264638EBF698F8BB19DF297F66AD91B18910DFC6E19323">>},

src/hackney.erl

@@ -676,14 +676,22 @@ maybe_proxy(Transport, Scheme, Host, Port, Options)
   end.
 
 maybe_proxy_from_env(Transport, _Scheme, Host, Port, Options, true) ->
-  ?report_debug("request without proxy", []),
+  ?report_debug("no proxy env is forced, request without proxy", []),
   hackney_connect:connect(Transport, Host, Port, Options, true);
 maybe_proxy_from_env(Transport, Scheme, Host, Port, Options, _) ->
   case get_proxy_env(Scheme) of
     {ok, Url} ->
-      proxy_from_url(Url, Transport, Host, Port, Options);
+      NoProxyEnv = get_no_proxy_env(),
+      case match_no_proxy_env(NoProxyEnv, Host) of
+        false ->
+          ?report_debug("request with proxy", [{proxy, Url}, {host, Host}]),
+          proxy_from_url(Url, Transport, Host, Port, Options);
+        true ->
+         ?report_debug("request without proxy", []),
+         hackney_connect:connect(Transport, Host, Port, Options, true)
+      end;
     false ->
-      ?report_debug("request without proxy", []),
+      ?report_debug("no proxy env setup, request without proxy", []),
       hackney_connect:connect(Transport, Host, Port, Options, true)
   end.
 
@@ -705,17 +713,121 @@ proxy_from_url(Url, Transport, Host, Port, Options) ->
       end
   end.
 
+get_no_proxy_env() ->
+  case application:get_env(hackney, no_proxy) of
+    undefined ->
+      case get_no_proxy_env(?HTTP_NO_PROXY_ENV_VARS) of
+        false ->
+          application:set_env(hackney, no_proxy, false),
+          false;
+        NoProxyEnv ->
+          parse_no_proxy_env(NoProxyEnv, [])
+      end;
+    {ok, NoProxyEnv} ->
+      NoProxyEnv
+  end.
+
+get_no_proxy_env([Key | Rest]) ->
+  case os:getenv(Key) of
+    false -> get_no_proxy_env(Rest);
+    NoProxyStr ->
+      lists:usort(string:split(NoProxyStr, ","))
+  end;
+get_no_proxy_env([]) ->
+  false.
+
+parse_no_proxy_env(["*" | _], _Acc) ->
+  application:set_env(hackney, no_proxy, '*'),
+  '*';
+parse_no_proxy_env([S | Rest], Acc) ->
+  try
+    CIDR = hackney_cidr:parse(S),
+    parse_no_proxy_env(Rest, [{cidr, CIDR} | Acc])
+  catch
+    _:_ ->
+      Labels = string:tokens(S, "."),
+      parse_no_proxy_env(Rest, [{host, lists:reverse(Labels)}])
+  end;
+parse_no_proxy_env([], Acc) ->
+  NoProxy = lists:reverse(Acc),
+  application:set_env(hackney, no_proxy, NoProxy),
+  NoProxy.
+
+match_no_proxy_env(false, _Host) -> false;
+match_no_proxy_env('*', _Host) -> true;
+match_no_proxy_env(Patterns, Host) ->
+  do_match_no_proxy_env(Patterns, undefined, undefined, Host).
+
+do_match_no_proxy_env([{cidr, _CIDR} | _]=Patterns, undefined, Labels, Host) ->
+  Addrs = case inet:parse_address(Host) of
+            {ok, Addr} -> [Addr];
+            _ -> getaddrs(Host)
+          end,
+  do_match_no_proxy_env(Patterns, Addrs, Labels, Host);
+do_match_no_proxy_env([{cidr, CIDR} | Rest], Addrs, Labels, Host) ->
+  case test_host_cidr(Addrs, CIDR) of
+    true -> true;
+    false -> do_match_no_proxy_env(Rest, Addrs, Labels, Host)
+  end;
+do_match_no_proxy_env([{host, _Labels} | _] = Patterns, Addrs, undefined, Host) ->
+  HostLabels = string:tokens(Host, "."),
+  do_match_no_proxy_env(Patterns, Addrs, lists:reverse(HostLabels), Host);
+do_match_no_proxy_env([{host, Labels} | Rest], Addrs, HostLabels, Host) ->
+  case test_host_labels(Labels, HostLabels) of
+    true -> true;
+    false -> do_match_no_proxy_env(Rest, Addrs, Labels, Host)
+  end;
+do_match_no_proxy_env([], _, _, _) ->
+  false.
+
+test_host_labels(["*" | R1], [_ | R2]) -> test_host_labels(R1, R2);
+test_host_labels([ A | R1], [A | R2]) -> test_host_labels(R1, R2);
+test_host_labels([], _) -> true;
+test_host_labels(_, _) -> false.
+
+test_host_cidr([Addr, Rest], CIDR) ->
+  case hackney_cidr:contains(CIDR, Addr) of
+    true -> true;
+    false -> test_host_cidr(Rest, CIDR)
+  end;
+test_host_cidr([], _) ->
+  false.
+
+getaddrs(Host) ->
+  IP4Addrs = case inet:getaddrs(Host, inet) of
+               {ok, Addrs} -> Addrs;
+               {error, nxdomain} -> []
+             end,
+  case inet:getaddrs(Host, inet6) of
+    {ok, IP6Addrs} -> [IP6Addrs | IP4Addrs];
+    {error, nxdomain} -> IP4Addrs
+  end.
+
 get_proxy_env(https) ->
-  get_proxy_env(?HTTPS_PROXY_ENV_VARS);
+  case application:get_env(hackney, https_proxy) of
+    undefined ->
+      ProxyEnv = do_get_proxy_env(?HTTPS_PROXY_ENV_VARS),
+      application:set_env(hackney, https_proxy, ProxyEnv),
+      ProxyEnv;
+    {ok, Cached} ->
+      Cached
+  end;
 get_proxy_env(S) when S =:= http; S =:= http_unix ->
-  get_proxy_env(?HTTP_PROXY_ENV_VARS);
+  case application:get_env(hackney, http_proxy) of
+    undefined ->
+      ProxyEnv = do_get_proxy_env(?HTTP_PROXY_ENV_VARS),
+      application:set_env(hackney, http_proxy, ProxyEnv),
+      ProxyEnv;
+    {ok, Cached} ->
+      Cached
+  end.
 
-get_proxy_env([Var | Rest]) ->
+do_get_proxy_env([Var | Rest]) ->
   case os:getenv(Var) of
-    false -> get_proxy_env(Rest);
+    false -> do_get_proxy_env(Rest);
     Url -> {ok, Url}
   end;
-get_proxy_env([]) ->
+do_get_proxy_env([]) ->
   false.
 
 do_connect(ProxyHost, ProxyPort, undefined, Transport, Host, Port, Options) ->

src/hackney_connection.erl

@@ -102,27 +102,13 @@ connect_options(hackney_local_tcp, _Host, ClientOptions) ->
   proplists:get_value(connect_options, ClientOptions, []);
 
 connect_options(Transport, Host, ClientOptions) ->
-  ConnectOpts0 = proplists:get_value(connect_options, ClientOptions, []),
-
-  %% handle ipv6
-  ConnectOpts1 = case lists:member(inet, ConnectOpts0) orelse
-                      lists:member(inet6, ConnectOpts0) of
-                   true ->
-                     ConnectOpts0;
-                   false ->
-                     case hackney_util:is_ipv6(Host) of
-                       true ->
-                         [inet6 | ConnectOpts0];
-                       false ->
-                         ConnectOpts0
-                     end
-                 end,
+  ConnectOpts = proplists:get_value(connect_options, ClientOptions, []),
 
   case Transport of
     hackney_ssl ->
-      ConnectOpts1 ++ ssl_opts(Host, ClientOptions);
+      [{ssl_options, ssl_opts(Host, ClientOptions)} | ConnectOpts];
     _ ->
-      ConnectOpts1
+      ConnectOpts
   end.
 
 

src/hackney_happy.erl

@@ -0,0 +1,72 @@
+-module(hackney_happy).
+
+-export([connect/3, connect/4]).
+
+-include("hackney_internal.hrl").
+-include_lib("kernel/include/inet.hrl").
+
+-define(TIMEOUT, 250).
+-define(CONNECT_TIMEOUT, 5000).
+
+connect(Hostname, Port, Opts) ->
+  connect(Hostname, Port, Opts, ?CONNECT_TIMEOUT).
+
+connect(Hostname, Port, Opts, Timeout) ->
+  case hackney_cidr:is_ipv6(Hostname) of
+    true ->
+      ?report_debug("connect using IPv6", [{hostname, Hostname}, {port, Port}]),
+      gen_tcp:connect(Hostname, Port, [inet6 | Opts], Timeout);
+    false ->
+      case hackney_cidr:is_ipv4(Hostname) of
+        true ->
+          ?report_debug("connect using IPv4", [{hostname, Hostname}, {port, Port}]),
+          gen_tcp:connect(Hostname, Port, [inet | Opts], Timeout);
+        false ->
+          ?report_debug("happy eyeballs, try to connect using IPv6",  [{hostname, Hostname}, {port, Port}]),
+          Self = self(),
+          Addrs = getaddrs(Hostname),
+          Pid = spawn_link( fun() -> try_connect(Addrs, Port, Opts, Self, {error, nxdomain}) end),
+          MRef = erlang:monitor(process, Pid),
+          receive
+            {happy_connect, OK} ->
+              erlang:demonitor(MRef, [flush]),
+              OK;
+            {'DOWN', MRef, _Type, _Pid, Info} ->
+              {'error', {'connect_error', Info}}
+          after Timeout -> 
+                  erlang:demonitor(MRef, [flush]),
+                  {error, connect_timeout}
+          end
+      end
+  end.
+
+getaddrs(Hostname) ->
+  IP6Addrs = [{Addr, 'inet6'} || Addr <- getbyname(Hostname, 'aaaa')],
+  IP4Addrs = [{Addr, 'inet'} || Addr <- getbyname(Hostname, 'a')],
+  IP6Addrs ++ IP4Addrs.
+
+getbyname(Hostname, Type) ->
+  case (catch inet_res:getbyname(Hostname, Type)) of
+    {'ok', #hostent{h_addr_list=AddrList}} -> lists:usort(AddrList);
+    {error, _Reason} -> [];
+    Else ->
+      %% ERLANG 22 has an issue when g matching somee DNS server messages
+      ?report_debug("DNS error", [{hostname, Hostname}
+                                 ,{type, Type}
+                                 ,{error, Else}]),
+      []
+  end.
+
+try_connect([], _Port, _Opts, ServerPid, LastError) ->
+  ?report_trace("happy eyeball: failed to connect", [{error, LastError}]),
+  ServerPid ! {hackney_happy, LastError};
+try_connect([{IP, Type} | Rest], Port, Opts, ServerPid, _LastError) ->
+  ?report_trace("try to connect", [{ip, IP}, {type, Type}]),
+  case gen_tcp:connect(IP, Port, [Type | Opts], ?TIMEOUT) of
+    {ok, Socket} = OK ->
+      ?report_trace("success to connect", [{ip, IP}, {type, Type}]),
+      ok = gen_tcp:controlling_process(Socket, ServerPid),
+      ServerPid ! {happy_connect, OK};
+    Error ->
+      try_connect(Rest, Port, Opts, ServerPid, Error)
+  end.

src/hackney_http_connect.erl

@@ -63,7 +63,7 @@ connect(ProxyHost, ProxyPort, Opts, Timeout)
   ConnectOpts = hackney_util:filter_options(Opts, AcceptedOpts, BaseOpts),
 
   %% connect to the proxy, and upgrade the socket if needed.
-  case gen_tcp:connect(ProxyHost, ProxyPort, ConnectOpts) of
+  case hackney_happy:connect(ProxyHost, ProxyPort, ConnectOpts) of
     {ok, Socket} ->
       case do_handshake(Socket, Host, Port, Opts) of
         ok ->

src/hackney_pool.erl

@@ -63,12 +63,13 @@ checkout(Host, Port, Transport, Client) ->
   Requester = self(),
   try
     do_checkout(Requester, Host, Port, Transport, Client)
-  catch _:_ ->
+  catch _:Error ->
+    ?report_trace("pool: checkout failure", [{error, Error}]),
     {error, checkout_failure}
   end.
 
 do_checkout(Requester, Host, _Port, Transport, #client{options=Opts,
-  mod_metrics=Metrics}=Client) ->
+                                                       mod_metrics=Metrics}=Client) ->
   ConnectTimeout = proplists:get_value(connect_timeout, Opts, 8000),
   %% Fall back to using connect_timeout if checkout_timeout is not set
   CheckoutTimeout = proplists:get_value(checkout_timeout, Opts, ConnectTimeout),
@@ -78,7 +79,6 @@ do_checkout(Requester, Host, _Port, Transport, #client{options=Opts,
   Pool = find_pool(PoolName, Opts),
   case catch gen_server:call(Pool, {checkout, Connection, Requester, RequestRef}, CheckoutTimeout) of
     {ok, Socket, Owner} ->
-
       %% stats
       ?report_debug("reuse a connection", [{pool, PoolName}]),
       _ = metrics:update_meter(Metrics, [hackney_pool, PoolName, take_rate], 1),
@@ -105,7 +105,7 @@ do_checkout(Requester, Host, _Port, Transport, #client{options=Opts,
           _ = metrics:increment_counter(Metrics, [hackney, Host, connect_timeout]),
           {error, timeout};
         Error ->
-          ?report_trace("connect error", []),
+          ?report_trace("connect error", [{pool, PoolName}, {error, Error}]),
           _ = metrics:increment_counter(Metrics, [hackney, Host, connect_error]),
           Error
       end;

src/hackney_socks5.erl

@@ -62,7 +62,7 @@ connect(Host, Port, Opts, Timeout) when is_list(Host), is_integer(Port),
   ConnectOpts = hackney_util:filter_options(Opts, AcceptedOpts, BaseOpts),
 
   %% connect to the socks 5 proxy
-  case gen_tcp:connect(ProxyHost, ProxyPort, ConnectOpts, Timeout) of
+  case hackney_happy:connect(ProxyHost, ProxyPort, ConnectOpts, Timeout) of
     {ok, Socket} ->
       case do_handshake(Socket, Host, Port, Opts) of
         ok ->

src/hackney_ssl.erl

@@ -25,16 +25,7 @@
 %% @doc Atoms used to identify messages in {active, once | true} mode.
 messages(_) -> {ssl, ssl_closed, ssl_error}.
 
-%% @doc The ssl:connect/4 (and related) doesn't work with textual representation
-%% of IP addresses. It accepts either a string with a DNS-resolvable name or a
-%% tuple with parts of an IP as numbers. This function attempts to parse given
-%% string and either returns such tuple, or the string if it's impossible to
-%% parse.
-parse_address(Host) when is_list(Host) ->
-  case inet:parse_address(Host) of
-    {ok, Address} -> Address;
-    {error, _} -> Host
-  end.
+
 
 
 check_hostname_opts(Host0) ->
@@ -134,14 +125,17 @@ find(_Fun, []) ->
 connect(Host, Port, Opts) ->
   connect(Host, Port, Opts, infinity).
 
-connect(Host, Port, Opts, Timeout) when is_list(Host), is_integer(Port),
-                                        (Timeout =:= infinity orelse is_integer(Timeout)) ->
-  BaseOpts = [binary, {active, false}, {packet, raw},
-              {secure_renegotiate, true},
-              {reuse_sessions, true}],
-  Opts1 = hackney_util:merge_opts(BaseOpts, Opts),
-  %% connect
-  ssl:connect(parse_address(Host), Port, Opts1, Timeout).
+connect(Host, Port, Opts0, Timeout) when is_list(Host), is_integer(Port),
+                                        (Timeout =:= 5000 orelse is_integer(Timeout)) ->
+  SSLOpts = [{secure_renegotiate, true}, proplists:get_value(ssl_options, Opts0)],
+  BaseOpts = [binary, {active, false}, {packet, raw}],
+  Opts1 = hackney_util:merge_opts(BaseOpts, proplists:delete(ssl_options, Opts0)),
+  case hackney_happy:connect(Host, Port, Opts1, Timeout) of
+    {ok, Sock} ->
+      ssl:connect(Sock, SSLOpts);
+    Error ->
+      Error
+  end.
 
 
 

src/hackney_tcp.erl

@@ -28,7 +28,7 @@ connect(Host, Port, Opts, Timeout) when is_list(Host), is_integer(Port),
   BaseOpts = [binary, {active, false}, {packet, raw}],
   Opts1 = hackney_util:merge_opts(BaseOpts, Opts),
   %% connect
-  gen_tcp:connect(Host, Port, Opts1, Timeout).
+ hackney_happy:connect(Host, Port, Opts1, Timeout).
 
 recv(Socket, Length) ->
   recv(Socket, Length, infinity).