From 7b1a9e6086588afa7840198bd01b4510581e7cdd Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?P=C3=A9ter=20Dimitrov?= Date: Fri, 10 Nov 2017 15:47:35 +0100 Subject: inets: Add support for URI-references in Location RFC 2616 requires an absolute URI in 'Location' header field for redirects. RFC 7231 obsoleted RFC 2616 and allows URI-references. Updated httpc_response to support URI-references, based on the URI resolution algorithm defined by RFC 3986 (5.2.2. Transform References). Change-Id: I42227d32f458b6e7a60d55b40407c4092e69b222 --- lib/inets/src/http_client/httpc_response.erl | 224 +++++++++++++++++++-------- lib/inets/src/http_lib/http_util.erl | 32 +++- lib/inets/test/httpc_SUITE.erl | 37 +++++ 3 files changed, 230 insertions(+), 63 deletions(-) diff --git a/lib/inets/src/http_client/httpc_response.erl b/lib/inets/src/http_client/httpc_response.erl index 91638f5d2e..58ab9144df 100644 --- a/lib/inets/src/http_client/httpc_response.erl +++ b/lib/inets/src/http_client/httpc_response.erl @@ -190,7 +190,7 @@ parse_status_code(<>, StatusCodeStr, MaxHeaderSize, Result, true) -> parse_headers(Rest, [], [], MaxHeaderSize, [" ", list_to_integer(lists:reverse( - string:strip(StatusCodeStr))) + string:trim(StatusCodeStr))) | Result], true); parse_status_code(<>, StatusCodeStr, @@ -377,58 +377,173 @@ status_server_error_50x(Response, Request) -> {stop, {Request#request.id, Msg}}. -redirect(Response = {StatusLine, Headers, Body}, Request) -> +redirect(Response = {_, Headers, _}, Request) -> {_, Data} = format_response(Response), case Headers#http_response_h.location of - undefined -> - transparent(Response, Request); - RedirUrl -> - UrlParseOpts = [{ipv6_host_with_brackets, - Request#request.ipv6_host_with_brackets}], - case uri_parse(RedirUrl, UrlParseOpts) of - {error, no_scheme} when - (Request#request.settings)#http_options.relaxed -> - NewLocation = fix_relative_uri(Request, RedirUrl), - redirect({StatusLine, Headers#http_response_h{ - location = NewLocation}, - Body}, Request); - {error, Reason} -> - {ok, error(Request, Reason), Data}; - %% Automatic redirection - {ok, {Scheme, _, Host, Port, Path, Query}} -> - HostPort = http_request:normalize_host(Scheme, Host, Port), - NewHeaders = - (Request#request.headers)#http_request_h{host = HostPort}, - NewRequest = - Request#request{redircount = - Request#request.redircount+1, - scheme = Scheme, - headers = NewHeaders, - address = {Host,Port}, - path = Path, - pquery = Query, - abs_uri = - atom_to_list(Scheme) ++ "://" ++ - Host ++ ":" ++ - integer_to_list(Port) ++ - Path ++ Query}, - {redirect, NewRequest, Data} - end + undefined -> + transparent(Response, Request); + RedirUrl -> + Brackets = Request#request.ipv6_host_with_brackets, + case uri_string:parse(RedirUrl) of + {error, Reason, _} -> + {ok, error(Request, Reason), Data}; + %% Automatic redirection + URI -> + {Host, Port0} = Request#request.address, + Port = maybe_to_integer(Port0), + Path = Request#request.path, + Scheme = atom_to_list(Request#request.scheme), + Query = Request#request.pquery, + URIMap = resolve_uri(Scheme, Host, Port, Path, Query, URI), + TScheme = list_to_atom(maps:get(scheme, URIMap)), + THost = http_util:maybe_add_brackets(maps:get(host, URIMap), Brackets), + TPort = maps:get(port, URIMap), + TPath = maps:get(path, URIMap), + TQuery = maps:get(query, URIMap, ""), + NewURI = uri_string:normalize( + uri_string:recompose(URIMap)), + HostPort = http_request:normalize_host(TScheme, THost, TPort), + NewHeaders = + (Request#request.headers)#http_request_h{host = HostPort}, + NewRequest = + Request#request{redircount = + Request#request.redircount+1, + scheme = TScheme, + headers = NewHeaders, + address = {THost,TPort}, + path = TPath, + pquery = TQuery, + abs_uri = NewURI}, + {redirect, NewRequest, Data} + end + end. + + +%% RFC3986 - 5.2.2. Transform References +resolve_uri(Scheme, Host, Port, Path, Query, URI) -> + resolve_uri(Scheme, Host, Port, Path, Query, URI, #{}). +%% +resolve_uri(Scheme, Host, Port, Path, Query, URI, Map0) -> + case maps:is_key(scheme, URI) of + true -> + Port = get_port(URI), + maybe_add_query( + Map0#{scheme => maps:get(scheme, URI), + host => maps:get(host, URI), + port => Port, + path => maps:get(path, URI)}, + URI); + false -> + Map = Map0#{scheme => Scheme}, + resolve_authority(Host, Port, Path, Query, URI, Map) + end. + + +get_port(URI) -> + Scheme = maps:get(scheme, URI), + case maps:get(port, URI, undefined) of + undefined -> + get_default_port(Scheme); + Port -> + Port + end. + + +get_default_port("http") -> + 80; +get_default_port("https") -> + 443. + + +resolve_authority(Host, Port, Path, Query, RelURI, Map) -> + case maps:is_key(host, RelURI) of + true -> + Port = get_port(RelURI), + maybe_add_query( + Map#{host => maps:get(host, RelURI), + port => Port, + path => maps:get(path, RelURI)}, + RelURI); + false -> + Map1 = Map#{host => Host, + port => Port}, + resolve_path(Path, Query, RelURI, Map1) + end. + + +maybe_add_query(Map, RelURI) -> + case maps:is_key(query, RelURI) of + true -> + Map#{query => maps:get(query, RelURI)}; + false -> + Map + end. + + +resolve_path(Path, Query, RelURI, Map) -> + case maps:is_key(path, RelURI) of + true -> + Path1 = calculate_path(Path, maps:get(path, RelURI)), + maybe_add_query( + Map#{path => Path1}, + RelURI); + false -> + Map1 = Map#{path => Path}, + resolve_query(Query, RelURI, Map1) + end. + + +calculate_path(BaseP, RelP) -> + case starts_with_slash(RelP) of + true -> + RelP; + false -> + merge_paths(BaseP, RelP) + end. + + +starts_with_slash([$/|_]) -> + true; +starts_with_slash(<<$/,_/binary>>) -> + true; +starts_with_slash(_) -> + false. + + +%% RFC3986 - 5.2.3. Merge Paths +merge_paths("", RelP) -> + [$/|RelP]; +merge_paths(BaseP, RelP) when is_list(BaseP) -> + do_merge_paths(lists:reverse(BaseP), RelP); +merge_paths(BaseP, RelP) when is_binary(BaseP) -> + B = binary_to_list(BaseP), + R = binary_to_list(RelP), + Res = merge_paths(B, R), + list_to_binary(Res). + + +do_merge_paths([$/|_] = L, RelP) -> + lists:reverse(L) ++ RelP; +do_merge_paths([_|T], RelP) -> + do_merge_paths(T, RelP). + + +resolve_query(Query, RelURI, Map) -> + case maps:is_key(query, RelURI) of + true -> + Map#{query => maps:get(query, RelURI)}; + false -> + Map#{query => Query} end. -maybe_to_list(Port) when is_integer(Port) -> - integer_to_list(Port); -maybe_to_list(Port) when is_list(Port) -> + +maybe_to_integer(Port) when is_list(Port) -> + {Port1, _} = string:to_integer(Port), + Port1; +maybe_to_integer(Port) when is_integer(Port) -> Port. -%%% Guessing that we received a relative URI, fix it to become an absoluteURI -fix_relative_uri(Request, RedirUrl) -> - {Server, Port0} = Request#request.address, - Port = maybe_to_list(Port0), - Path = Request#request.path, - atom_to_list(Request#request.scheme) ++ "://" ++ Server ++ ":" ++ Port - ++ Path ++ RedirUrl. - + error(#request{id = Id}, Reason) -> {Id, {error, Reason}}. @@ -478,18 +593,3 @@ format_response({StatusLine, Headers, Body}) -> {Body, <<>>} end, {{StatusLine, http_response:header_list(Headers), NewBody}, Data}. - -%%-------------------------------------------------------------------------- -%% These functions is just simple wrappers to parse specifically HTTP URIs -%%-------------------------------------------------------------------------- - -scheme_defaults() -> - [{http, 80}, {https, 443}]. - -uri_parse(URI, Opts) -> - http_uri:parse(URI, [{scheme_defaults, scheme_defaults()} | Opts]). - - -%%-------------------------------------------------------------------------- - - diff --git a/lib/inets/src/http_lib/http_util.erl b/lib/inets/src/http_lib/http_util.erl index 487d04f7aa..5577b00cc8 100644 --- a/lib/inets/src/http_lib/http_util.erl +++ b/lib/inets/src/http_lib/http_util.erl @@ -27,7 +27,8 @@ convert_month/1, is_hostname/1, timestamp/0, timeout/2, - html_encode/1 + html_encode/1, + maybe_add_brackets/2 ]). @@ -194,6 +195,24 @@ html_encode(Chars) -> lists:append([char_to_html_entity(Char, Reserved) || Char <- Chars]). +maybe_add_brackets(Addr, false) -> + Addr; +maybe_add_brackets(Addr, true) when is_list(Addr) -> + case is_ipv6_address(Addr) of + true -> + [$[|Addr] ++ "]"; + false -> + Addr + end; +maybe_add_brackets(Addr, true) when is_binary(Addr) -> + case is_ipv6_address(Addr) of + true -> + <<$[,Addr/binary,$]>>; + false -> + Addr + end. + + %%%======================================================================== %%% Internal functions %%%======================================================================== @@ -205,3 +224,14 @@ char_to_html_entity(Char, Reserved) -> false -> [Char] end. + +is_ipv6_address(Addr) when is_binary(Addr) -> + B = binary_to_list(Addr), + is_ipv6_address(B); +is_ipv6_address(Addr) when is_list(Addr) -> + case inet:parse_ipv6strict_address(Addr) of + {ok, _ } -> + true; + {error, _} -> + false + end. diff --git a/lib/inets/test/httpc_SUITE.erl b/lib/inets/test/httpc_SUITE.erl index cc166d522e..143fa14ee0 100644 --- a/lib/inets/test/httpc_SUITE.erl +++ b/lib/inets/test/httpc_SUITE.erl @@ -128,6 +128,7 @@ only_simulated() -> redirect_found, redirect_see_other, redirect_temporary_redirect, + redirect_relative_uri, port_in_host_header, redirect_port_in_host_header, relaxed, @@ -577,7 +578,26 @@ redirect_temporary_redirect(Config) when is_list(Config) -> {ok, {{_,200,_}, [_ | _], [_|_]}} = httpc:request(post, {URL307, [],"text/plain", "foobar"}, [], []). +%%------------------------------------------------------------------------- +redirect_relative_uri() -> + [{doc, "The server SHOULD generate a Location header field in the response " + "containing a preferred URI reference for the new permanent URI. The user " + "agent MAY use the Location field value for automatic redirection. The server's " + "response payload usually contains a short hypertext note with a " + "hyperlink to the new URI(s)."}]. +redirect_relative_uri(Config) when is_list(Config) -> + + URL301 = url(group_name(Config), "/301_rel_uri.html", Config), + + {ok, {{_,200,_}, [_ | _], [_|_]}} + = httpc:request(get, {URL301, []}, [], []), + {ok, {{_,200,_}, [_ | _], []}} + = httpc:request(head, {URL301, []}, [], []), + + {ok, {{_,200,_}, [_ | _], [_|_]}} + = httpc:request(post, {URL301, [],"text/plain", "foobar"}, + [], []). %%------------------------------------------------------------------------- redirect_loop() -> [{"doc, Test redirect loop detection"}]. @@ -1778,6 +1798,23 @@ handle_uri(_,"/301.html",Port,_,Socket,_) -> "Content-Length:" ++ integer_to_list(length(Body)) ++ "\r\n\r\n" ++ Body; + +handle_uri("HEAD","/301_rel_uri.html",_,_,_,_) -> + NewUri = "/dummy.html", + "HTTP/1.1 301 Moved Permanently\r\n" ++ + "Location:" ++ NewUri ++ "\r\n" ++ + "Content-Length:0\r\n\r\n"; + +handle_uri(_,"/301_rel_uri.html",_,_,_,_) -> + NewUri = "/dummy.html", + Body = "New place", + "HTTP/1.1 301 Moved Permanently\r\n" ++ + "Location:" ++ NewUri ++ "\r\n" ++ + "Content-Length:" ++ integer_to_list(length(Body)) + ++ "\r\n\r\n" ++ Body; + + handle_uri("HEAD","/302.html",Port,_,Socket,_) -> NewUri = url_start(Socket) ++ integer_to_list(Port) ++ "/dummy.html", -- cgit v1.2.3