1%% @author Bob Ippolito <bob@mochimedia.com>
2%% @copyright 2007 Mochi Media, Inc.
3
4%% @doc Utilities for parsing and quoting.
5
6-module(rabbit_http_util).
7-author('bob@mochimedia.com').
8-export([join/2, quote_plus/1, urlencode/1, parse_qs/1, unquote/1]).
9-export([path_split/1]).
10-export([urlsplit/1, urlsplit_path/1, urlunsplit/1, urlunsplit_path/1]).
11-export([parse_header/1]).
12-export([shell_quote/1, cmd/1, cmd_string/1, cmd_port/2, cmd_status/1, cmd_status/2]).
13-export([record_to_proplist/2, record_to_proplist/3]).
14-export([safe_relative_path/1, partition/2]).
15-export([parse_qvalues/1, pick_accepted_encodings/3]).
16-export([make_io/1]).
17
18-define(PERCENT, 37).  % $\%
19-define(FULLSTOP, 46). % $\.
20-define(IS_HEX(C), ((C >= $0 andalso C =< $9) orelse
21                    (C >= $a andalso C =< $f) orelse
22                    (C >= $A andalso C =< $F))).
23-define(QS_SAFE(C), ((C >= $a andalso C =< $z) orelse
24                     (C >= $A andalso C =< $Z) orelse
25                     (C >= $0 andalso C =< $9) orelse
26                     (C =:= ?FULLSTOP orelse C =:= $- orelse C =:= $~ orelse
27                      C =:= $_))).
28
29hexdigit(C) when C < 10 -> $0 + C;
30hexdigit(C) when C < 16 -> $A + (C - 10).
31
32unhexdigit(C) when C >= $0, C =< $9 -> C - $0;
33unhexdigit(C) when C >= $a, C =< $f -> C - $a + 10;
34unhexdigit(C) when C >= $A, C =< $F -> C - $A + 10.
35
36%% @spec partition(String, Sep) -> {String, [], []} | {Prefix, Sep, Postfix}
37%% @doc Inspired by Python 2.5's str.partition:
38%%      partition("foo/bar", "/") = {"foo", "/", "bar"},
39%%      partition("foo", "/") = {"foo", "", ""}.
40partition(String, Sep) ->
41    case partition(String, Sep, []) of
42        undefined ->
43            {String, "", ""};
44        Result ->
45            Result
46    end.
47
48partition("", _Sep, _Acc) ->
49    undefined;
50partition(S, Sep, Acc) ->
51    case partition2(S, Sep) of
52        undefined ->
53            [C | Rest] = S,
54            partition(Rest, Sep, [C | Acc]);
55        Rest ->
56            {lists:reverse(Acc), Sep, Rest}
57    end.
58
59partition2(Rest, "") ->
60    Rest;
61partition2([C | R1], [C | R2]) ->
62    partition2(R1, R2);
63partition2(_S, _Sep) ->
64    undefined.
65
66
67
68%% @spec safe_relative_path(string()) -> string() | undefined
69%% @doc Return the reduced version of a relative path or undefined if it
70%%      is not safe. safe relative paths can be joined with an absolute path
71%%      and will result in a subdirectory of the absolute path. Safe paths
72%%      never contain a backslash character.
73safe_relative_path("/" ++ _) ->
74    undefined;
75safe_relative_path(P) ->
76    case string:chr(P, $\\) of
77        0 ->
78           safe_relative_path(P, []);
79        _ ->
80           undefined
81    end.
82
83safe_relative_path("", Acc) ->
84    case Acc of
85        [] ->
86            "";
87        _ ->
88            string:join(lists:reverse(Acc), "/")
89    end;
90safe_relative_path(P, Acc) ->
91    case partition(P, "/") of
92        {"", "/", _} ->
93            %% /foo or foo//bar
94            undefined;
95        {"..", _, _} when Acc =:= [] ->
96            undefined;
97        {"..", _, Rest} ->
98            safe_relative_path(Rest, tl(Acc));
99        {Part, "/", ""} ->
100            safe_relative_path("", ["", Part | Acc]);
101        {Part, _, Rest} ->
102            safe_relative_path(Rest, [Part | Acc])
103    end.
104
105%% @spec shell_quote(string()) -> string()
106%% @doc Quote a string according to UNIX shell quoting rules, returns a string
107%%      surrounded by double quotes.
108shell_quote(L) ->
109    shell_quote(L, [$\"]).
110
111%% @spec cmd_port([string()], Options) -> port()
112%% @doc open_port({spawn, mochiweb_util:cmd_string(Argv)}, Options).
113cmd_port(Argv, Options) ->
114    open_port({spawn, cmd_string(Argv)}, Options).
115
116%% @spec cmd([string()]) -> string()
117%% @doc os:cmd(cmd_string(Argv)).
118cmd(Argv) ->
119    os:cmd(cmd_string(Argv)).
120
121%% @spec cmd_string([string()]) -> string()
122%% @doc Create a shell quoted command string from a list of arguments.
123cmd_string(Argv) ->
124    string:join([shell_quote(X) || X <- Argv], " ").
125
126%% @spec cmd_status([string()]) -> {ExitStatus::integer(), Stdout::binary()}
127%% @doc Accumulate the output and exit status from the given application,
128%%      will be spawned with cmd_port/2.
129cmd_status(Argv) ->
130    cmd_status(Argv, []).
131
132%% @spec cmd_status([string()], [atom()]) -> {ExitStatus::integer(), Stdout::binary()}
133%% @doc Accumulate the output and exit status from the given application,
134%%      will be spawned with cmd_port/2.
135cmd_status(Argv, Options) ->
136    Port = cmd_port(Argv, [exit_status, stderr_to_stdout,
137                           use_stdio, binary | Options]),
138    try cmd_loop(Port, [])
139    after catch port_close(Port)
140    end.
141
142%% @spec cmd_loop(port(), list()) -> {ExitStatus::integer(), Stdout::binary()}
143%% @doc Accumulate the output and exit status from a port.
144cmd_loop(Port, Acc) ->
145    receive
146        {Port, {exit_status, Status}} ->
147            {Status, iolist_to_binary(lists:reverse(Acc))};
148        {Port, {data, Data}} ->
149            cmd_loop(Port, [Data | Acc])
150    end.
151
152%% @spec join([iolist()], iolist()) -> iolist()
153%% @doc Join a list of strings or binaries together with the given separator
154%%      string or char or binary. The output is flattened, but may be an
155%%      iolist() instead of a string() if any of the inputs are binary().
156join([], _Separator) ->
157    [];
158join([S], _Separator) ->
159    lists:flatten(S);
160join(Strings, Separator) ->
161    lists:flatten(revjoin(lists:reverse(Strings), Separator, [])).
162
163revjoin([], _Separator, Acc) ->
164    Acc;
165revjoin([S | Rest], Separator, []) ->
166    revjoin(Rest, Separator, [S]);
167revjoin([S | Rest], Separator, Acc) ->
168    revjoin(Rest, Separator, [S, Separator | Acc]).
169
170%% @spec quote_plus(atom() | integer() | float() | string() | binary()) -> string()
171%% @doc URL safe encoding of the given term.
172quote_plus(Atom) when is_atom(Atom) ->
173    quote_plus(atom_to_list(Atom));
174quote_plus(Int) when is_integer(Int) ->
175    quote_plus(integer_to_list(Int));
176quote_plus(Binary) when is_binary(Binary) ->
177    quote_plus(binary_to_list(Binary));
178quote_plus(Float) when is_float(Float) ->
179    quote_plus(rabbit_numerical:digits(Float));
180quote_plus(String) ->
181    quote_plus(String, []).
182
183quote_plus([], Acc) ->
184    lists:reverse(Acc);
185quote_plus([C | Rest], Acc) when ?QS_SAFE(C) ->
186    quote_plus(Rest, [C | Acc]);
187quote_plus([$\s | Rest], Acc) ->
188    quote_plus(Rest, [$+ | Acc]);
189quote_plus([C | Rest], Acc) ->
190    <<Hi:4, Lo:4>> = <<C>>,
191    quote_plus(Rest, [hexdigit(Lo), hexdigit(Hi), ?PERCENT | Acc]).
192
193%% @spec urlencode([{Key, Value}]) -> string()
194%% @doc URL encode the property list.
195urlencode(Props) ->
196    Pairs = lists:foldr(
197              fun ({K, V}, Acc) ->
198                      [quote_plus(K) ++ "=" ++ quote_plus(V) | Acc]
199              end, [], Props),
200    string:join(Pairs, "&").
201
202%% @spec parse_qs(string() | binary()) -> [{Key, Value}]
203%% @doc Parse a query string or application/x-www-form-urlencoded.
204parse_qs(Binary) when is_binary(Binary) ->
205    parse_qs(binary_to_list(Binary));
206parse_qs(String) ->
207    parse_qs(String, []).
208
209parse_qs([], Acc) ->
210    lists:reverse(Acc);
211parse_qs(String, Acc) ->
212    {Key, Rest} = parse_qs_key(String),
213    {Value, Rest1} = parse_qs_value(Rest),
214    parse_qs(Rest1, [{Key, Value} | Acc]).
215
216parse_qs_key(String) ->
217    parse_qs_key(String, []).
218
219parse_qs_key([], Acc) ->
220    {qs_revdecode(Acc), ""};
221parse_qs_key([$= | Rest], Acc) ->
222    {qs_revdecode(Acc), Rest};
223parse_qs_key(Rest=[$; | _], Acc) ->
224    {qs_revdecode(Acc), Rest};
225parse_qs_key(Rest=[$& | _], Acc) ->
226    {qs_revdecode(Acc), Rest};
227parse_qs_key([C | Rest], Acc) ->
228    parse_qs_key(Rest, [C | Acc]).
229
230parse_qs_value(String) ->
231    parse_qs_value(String, []).
232
233parse_qs_value([], Acc) ->
234    {qs_revdecode(Acc), ""};
235parse_qs_value([$; | Rest], Acc) ->
236    {qs_revdecode(Acc), Rest};
237parse_qs_value([$& | Rest], Acc) ->
238    {qs_revdecode(Acc), Rest};
239parse_qs_value([C | Rest], Acc) ->
240    parse_qs_value(Rest, [C | Acc]).
241
242%% @spec unquote(string() | binary()) -> string()
243%% @doc Unquote a URL encoded string.
244unquote(Binary) when is_binary(Binary) ->
245    unquote(binary_to_list(Binary));
246unquote(String) ->
247    qs_revdecode(lists:reverse(String)).
248
249qs_revdecode(S) ->
250    qs_revdecode(S, []).
251
252qs_revdecode([], Acc) ->
253    Acc;
254qs_revdecode([$+ | Rest], Acc) ->
255    qs_revdecode(Rest, [$\s | Acc]);
256qs_revdecode([Lo, Hi, ?PERCENT | Rest], Acc) when ?IS_HEX(Lo), ?IS_HEX(Hi) ->
257    qs_revdecode(Rest, [(unhexdigit(Lo) bor (unhexdigit(Hi) bsl 4)) | Acc]);
258qs_revdecode([C | Rest], Acc) ->
259    qs_revdecode(Rest, [C | Acc]).
260
261%% @spec urlsplit(Url) -> {Scheme, Netloc, Path, Query, Fragment}
262%% @doc Return a 5-tuple, does not expand % escapes. Only supports HTTP style
263%%      URLs.
264urlsplit(Url) ->
265    {Scheme, Url1} = urlsplit_scheme(Url),
266    {Netloc, Url2} = urlsplit_netloc(Url1),
267    {Path, Query, Fragment} = urlsplit_path(Url2),
268    {Scheme, Netloc, Path, Query, Fragment}.
269
270urlsplit_scheme(Url) ->
271    case urlsplit_scheme(Url, []) of
272        no_scheme ->
273            {"", Url};
274        Res ->
275            Res
276    end.
277
278urlsplit_scheme([C | Rest], Acc) when ((C >= $a andalso C =< $z) orelse
279                                       (C >= $A andalso C =< $Z) orelse
280                                       (C >= $0 andalso C =< $9) orelse
281                                       C =:= $+ orelse C =:= $- orelse
282                                       C =:= $.) ->
283    urlsplit_scheme(Rest, [C | Acc]);
284urlsplit_scheme([$: | Rest], Acc=[_ | _]) ->
285    {string:to_lower(lists:reverse(Acc)), Rest};
286urlsplit_scheme(_Rest, _Acc) ->
287    no_scheme.
288
289urlsplit_netloc("//" ++ Rest) ->
290    urlsplit_netloc(Rest, []);
291urlsplit_netloc(Path) ->
292    {"", Path}.
293
294urlsplit_netloc("", Acc) ->
295    {lists:reverse(Acc), ""};
296urlsplit_netloc(Rest=[C | _], Acc) when C =:= $/; C =:= $?; C =:= $# ->
297    {lists:reverse(Acc), Rest};
298urlsplit_netloc([C | Rest], Acc) ->
299    urlsplit_netloc(Rest, [C | Acc]).
300
301
302%% @spec path_split(string()) -> {Part, Rest}
303%% @doc Split a path starting from the left, as in URL traversal.
304%%      path_split("foo/bar") = {"foo", "bar"},
305%%      path_split("/foo/bar") = {"", "foo/bar"}.
306path_split(S) ->
307    path_split(S, []).
308
309path_split("", Acc) ->
310    {lists:reverse(Acc), ""};
311path_split("/" ++ Rest, Acc) ->
312    {lists:reverse(Acc), Rest};
313path_split([C | Rest], Acc) ->
314    path_split(Rest, [C | Acc]).
315
316
317%% @spec urlunsplit({Scheme, Netloc, Path, Query, Fragment}) -> string()
318%% @doc Assemble a URL from the 5-tuple. Path must be absolute.
319urlunsplit({Scheme, Netloc, Path, Query, Fragment}) ->
320    lists:flatten([case Scheme of "" -> "";  _ -> [Scheme, "://"] end,
321                   Netloc,
322                   urlunsplit_path({Path, Query, Fragment})]).
323
324%% @spec urlunsplit_path({Path, Query, Fragment}) -> string()
325%% @doc Assemble a URL path from the 3-tuple.
326urlunsplit_path({Path, Query, Fragment}) ->
327    lists:flatten([Path,
328                   case Query of "" -> ""; _ -> [$? | Query] end,
329                   case Fragment of "" -> ""; _ -> [$# | Fragment] end]).
330
331%% @spec urlsplit_path(Url) -> {Path, Query, Fragment}
332%% @doc Return a 3-tuple, does not expand % escapes. Only supports HTTP style
333%%      paths.
334urlsplit_path(Path) ->
335    urlsplit_path(Path, []).
336
337urlsplit_path("", Acc) ->
338    {lists:reverse(Acc), "", ""};
339urlsplit_path("?" ++ Rest, Acc) ->
340    {Query, Fragment} = urlsplit_query(Rest),
341    {lists:reverse(Acc), Query, Fragment};
342urlsplit_path("#" ++ Rest, Acc) ->
343    {lists:reverse(Acc), "", Rest};
344urlsplit_path([C | Rest], Acc) ->
345    urlsplit_path(Rest, [C | Acc]).
346
347urlsplit_query(Query) ->
348    urlsplit_query(Query, []).
349
350urlsplit_query("", Acc) ->
351    {lists:reverse(Acc), ""};
352urlsplit_query("#" ++ Rest, Acc) ->
353    {lists:reverse(Acc), Rest};
354urlsplit_query([C | Rest], Acc) ->
355    urlsplit_query(Rest, [C | Acc]).
356
357%% @spec parse_header(string()) -> {Type, [{K, V}]}
358%% @doc  Parse a Content-Type like header, return the main Content-Type
359%%       and a property list of options.
360parse_header(String) ->
361    %% TODO: This is exactly as broken as Python's cgi module.
362    %%       Should parse properly like mochiweb_cookies.
363    [Type | Parts] = [string:strip(S) || S <- string:tokens(String, ";")],
364    F = fun (S, Acc) ->
365                case lists:splitwith(fun (C) -> C =/= $= end, S) of
366                    {"", _} ->
367                        %% Skip anything with no name
368                        Acc;
369                    {_, ""} ->
370                        %% Skip anything with no value
371                        Acc;
372                    {Name, [$\= | Value]} ->
373                        [{string:to_lower(string:strip(Name)),
374                          unquote_header(string:strip(Value))} | Acc]
375                end
376        end,
377    {string:to_lower(Type),
378     lists:foldr(F, [], Parts)}.
379
380unquote_header("\"" ++ Rest) ->
381    unquote_header(Rest, []);
382unquote_header(S) ->
383    S.
384
385unquote_header("", Acc) ->
386    lists:reverse(Acc);
387unquote_header("\"", Acc) ->
388    lists:reverse(Acc);
389unquote_header([$\\, C | Rest], Acc) ->
390    unquote_header(Rest, [C | Acc]);
391unquote_header([C | Rest], Acc) ->
392    unquote_header(Rest, [C | Acc]).
393
394%% @spec record_to_proplist(Record, Fields) -> proplist()
395%% @doc calls record_to_proplist/3 with a default TypeKey of '__record'
396record_to_proplist(Record, Fields) ->
397    record_to_proplist(Record, Fields, '__record').
398
399%% @spec record_to_proplist(Record, Fields, TypeKey) -> proplist()
400%% @doc Return a proplist of the given Record with each field in the
401%%      Fields list set as a key with the corresponding value in the Record.
402%%      TypeKey is the key that is used to store the record type
403%%      Fields should be obtained by calling record_info(fields, record_type)
404%%      where record_type is the record type of Record
405record_to_proplist(Record, Fields, TypeKey)
406  when tuple_size(Record) - 1 =:= length(Fields) ->
407    lists:zip([TypeKey | Fields], tuple_to_list(Record)).
408
409
410shell_quote([], Acc) ->
411    lists:reverse([$\" | Acc]);
412shell_quote([C | Rest], Acc) when C =:= $\" orelse C =:= $\` orelse
413                                  C =:= $\\ orelse C =:= $\$ ->
414    shell_quote(Rest, [C, $\\ | Acc]);
415shell_quote([C | Rest], Acc) ->
416    shell_quote(Rest, [C | Acc]).
417
418%% @spec parse_qvalues(string()) -> [qvalue()] | invalid_qvalue_string
419%% @type qvalue() = {media_type() | encoding() , float()}.
420%% @type media_type() = string().
421%% @type encoding() = string().
422%%
423%% @doc Parses a list (given as a string) of elements with Q values associated
424%%      to them. Elements are separated by commas and each element is separated
425%%      from its Q value by a semicolon. Q values are optional but when missing
426%%      the value of an element is considered as 1.0. A Q value is always in the
427%%      range [0.0, 1.0]. A Q value list is used for example as the value of the
428%%      HTTP "Accept" and "Accept-Encoding" headers.
429%%
430%%      Q values are described in section 2.9 of the RFC 2616 (HTTP 1.1).
431%%
432%%      Example:
433%%
434%%      parse_qvalues("gzip; q=0.5, deflate, identity;q=0.0") ->
435%%          [{"gzip", 0.5}, {"deflate", 1.0}, {"identity", 0.0}]
436%%
437parse_qvalues(QValuesStr) ->
438    try
439        lists:map(
440            fun(Pair) ->
441                [Type | Params] = string:tokens(Pair, ";"),
442                NormParams = normalize_media_params(Params),
443                {Q, NonQParams} = extract_q(NormParams),
444                {string:join([string:strip(Type) | NonQParams], ";"), Q}
445            end,
446            string:tokens(string:to_lower(QValuesStr), ",")
447        )
448    catch
449        _Type:_Error ->
450            invalid_qvalue_string
451    end.
452
453normalize_media_params(Params) ->
454    {ok, Re} = re:compile("\\s"),
455    normalize_media_params(Re, Params, []).
456
457normalize_media_params(_Re, [], Acc) ->
458    lists:reverse(Acc);
459normalize_media_params(Re, [Param | Rest], Acc) ->
460    NormParam = re:replace(Param, Re, "", [global, {return, list}]),
461    normalize_media_params(Re, Rest, [NormParam | Acc]).
462
463extract_q(NormParams) ->
464    {ok, KVRe} = re:compile("^([^=]+)=([^=]+)$"),
465    {ok, QRe} = re:compile("^((?:0|1)(?:\\.\\d{1,3})?)$"),
466    extract_q(KVRe, QRe, NormParams, []).
467
468extract_q(_KVRe, _QRe, [], Acc) ->
469    {1.0, lists:reverse(Acc)};
470extract_q(KVRe, QRe, [Param | Rest], Acc) ->
471    case re:run(Param, KVRe, [{capture, [1, 2], list}]) of
472        {match, [Name, Value]} ->
473            case Name of
474            "q" ->
475                {match, [Q]} = re:run(Value, QRe, [{capture, [1], list}]),
476                QVal = case Q of
477                    "0" ->
478                        0.0;
479                    "1" ->
480                        1.0;
481                    Else ->
482                        list_to_float(Else)
483                end,
484                case QVal < 0.0 orelse QVal > 1.0 of
485                false ->
486                    {QVal, lists:reverse(Acc) ++ Rest}
487                end;
488            _ ->
489                extract_q(KVRe, QRe, Rest, [Param | Acc])
490            end
491    end.
492
493%% @spec pick_accepted_encodings([qvalue()], [encoding()], encoding()) ->
494%%    [encoding()]
495%%
496%% @doc Determines which encodings specified in the given Q values list are
497%%      valid according to a list of supported encodings and a default encoding.
498%%
499%%      The returned list of encodings is sorted, descendingly, according to the
500%%      Q values of the given list. The last element of this list is the given
501%%      default encoding unless this encoding is explicitly or implicitly
502%%      marked with a Q value of 0.0 in the given Q values list.
503%%      Note: encodings with the same Q value are kept in the same order as
504%%            found in the input Q values list.
505%%
506%%      This encoding picking process is described in section 14.3 of the
507%%      RFC 2616 (HTTP 1.1).
508%%
509%%      Example:
510%%
511%%      pick_accepted_encodings(
512%%          [{"gzip", 0.5}, {"deflate", 1.0}],
513%%          ["gzip", "identity"],
514%%          "identity"
515%%      ) ->
516%%          ["gzip", "identity"]
517%%
518pick_accepted_encodings(AcceptedEncs, SupportedEncs, DefaultEnc) ->
519    SortedQList = lists:reverse(
520        lists:sort(fun({_, Q1}, {_, Q2}) -> Q1 < Q2 end, AcceptedEncs)
521    ),
522    {Accepted, Refused} = lists:foldr(
523        fun({E, Q}, {A, R}) ->
524            case Q > 0.0 of
525                true ->
526                    {[E | A], R};
527                false ->
528                    {A, [E | R]}
529            end
530        end,
531        {[], []},
532        SortedQList
533    ),
534    Refused1 = lists:foldr(
535        fun(Enc, Acc) ->
536            case Enc of
537                "*" ->
538                    lists:subtract(SupportedEncs, Accepted) ++ Acc;
539                _ ->
540                    [Enc | Acc]
541            end
542        end,
543        [],
544        Refused
545    ),
546    Accepted1 = lists:foldr(
547        fun(Enc, Acc) ->
548            case Enc of
549                "*" ->
550                    lists:subtract(SupportedEncs, Accepted ++ Refused1) ++ Acc;
551                _ ->
552                    [Enc | Acc]
553            end
554        end,
555        [],
556        Accepted
557    ),
558    Accepted2 = case lists:member(DefaultEnc, Accepted1) of
559        true ->
560            Accepted1;
561        false ->
562            Accepted1 ++ [DefaultEnc]
563    end,
564    [E || E <- Accepted2, lists:member(E, SupportedEncs),
565        not lists:member(E, Refused1)].
566
567make_io(Atom) when is_atom(Atom) ->
568    atom_to_list(Atom);
569make_io(Integer) when is_integer(Integer) ->
570    integer_to_list(Integer);
571make_io(Io) when is_list(Io); is_binary(Io) ->
572    Io.
573
574%%
575%% Tests
576%%
577-ifdef(TEST).
578-include_lib("eunit/include/eunit.hrl").
579
580make_io_test() ->
581    ?assertEqual(
582       <<"atom">>,
583       iolist_to_binary(make_io(atom))),
584    ?assertEqual(
585       <<"20">>,
586       iolist_to_binary(make_io(20))),
587    ?assertEqual(
588       <<"list">>,
589       iolist_to_binary(make_io("list"))),
590    ?assertEqual(
591       <<"binary">>,
592       iolist_to_binary(make_io(<<"binary">>))),
593    ok.
594
595-record(test_record, {field1=f1, field2=f2}).
596record_to_proplist_test() ->
597    ?assertEqual(
598       [{'__record', test_record},
599        {field1, f1},
600        {field2, f2}],
601       record_to_proplist(#test_record{}, record_info(fields, test_record))),
602    ?assertEqual(
603       [{'typekey', test_record},
604        {field1, f1},
605        {field2, f2}],
606       record_to_proplist(#test_record{},
607                          record_info(fields, test_record),
608                          typekey)),
609    ok.
610
611shell_quote_test() ->
612    ?assertEqual(
613       "\"foo \\$bar\\\"\\`' baz\"",
614       shell_quote("foo $bar\"`' baz")),
615    ok.
616
617cmd_port_test_spool(Port, Acc) ->
618    receive
619        {Port, eof} ->
620            Acc;
621        {Port, {data, {eol, Data}}} ->
622            cmd_port_test_spool(Port, ["\n", Data | Acc]);
623        {Port, Unknown} ->
624            throw({unknown, Unknown})
625    after 1000 ->
626            throw(timeout)
627    end.
628
629cmd_port_test() ->
630    Port = cmd_port(["echo", "$bling$ `word`!"],
631                    [eof, stream, {line, 4096}]),
632    Res = try lists:append(lists:reverse(cmd_port_test_spool(Port, [])))
633          after catch port_close(Port)
634          end,
635    self() ! {Port, wtf},
636    try cmd_port_test_spool(Port, [])
637    catch throw:{unknown, wtf} -> ok
638    end,
639    try cmd_port_test_spool(Port, [])
640    catch throw:timeout -> ok
641    end,
642    ?assertEqual(
643       "$bling$ `word`!\n",
644       Res).
645
646cmd_test() ->
647    ?assertEqual(
648       "$bling$ `word`!\n",
649       cmd(["echo", "$bling$ `word`!"])),
650    ok.
651
652cmd_string_test() ->
653    ?assertEqual(
654       "\"echo\" \"\\$bling\\$ \\`word\\`!\"",
655       cmd_string(["echo", "$bling$ `word`!"])),
656    ok.
657
658cmd_status_test() ->
659    ?assertEqual(
660       {0, <<"$bling$ `word`!\n">>},
661       cmd_status(["echo", "$bling$ `word`!"])),
662    ok.
663
664
665parse_header_test() ->
666    ?assertEqual(
667       {"multipart/form-data", [{"boundary", "AaB03x"}]},
668       parse_header("multipart/form-data; boundary=AaB03x")),
669    %% This tests (currently) intentionally broken behavior
670    ?assertEqual(
671       {"multipart/form-data",
672        [{"b", ""},
673         {"cgi", "is"},
674         {"broken", "true\"e"}]},
675       parse_header("multipart/form-data;b=;cgi=\"i\\s;broken=true\"e;=z;z")),
676    ok.
677
678path_split_test() ->
679    {"", "foo/bar"} = path_split("/foo/bar"),
680    {"foo", "bar"} = path_split("foo/bar"),
681    {"bar", ""} = path_split("bar"),
682    ok.
683
684urlsplit_test() ->
685    {"", "", "/foo", "", "bar?baz"} = urlsplit("/foo#bar?baz"),
686    {"https", "host:port", "/foo", "", "bar?baz"} =
687        urlsplit("https://host:port/foo#bar?baz"),
688    {"https", "host", "", "", ""} = urlsplit("https://host"),
689    {"", "", "/wiki/Category:Fruit", "", ""} =
690        urlsplit("/wiki/Category:Fruit"),
691    ok.
692
693urlsplit_path_test() ->
694    {"/foo/bar", "", ""} = urlsplit_path("/foo/bar"),
695    {"/foo", "baz", ""} = urlsplit_path("/foo?baz"),
696    {"/foo", "", "bar?baz"} = urlsplit_path("/foo#bar?baz"),
697    {"/foo", "", "bar?baz#wibble"} = urlsplit_path("/foo#bar?baz#wibble"),
698    {"/foo", "bar", "baz"} = urlsplit_path("/foo?bar#baz"),
699    {"/foo", "bar?baz", "baz"} = urlsplit_path("/foo?bar?baz#baz"),
700    ok.
701
702urlunsplit_test() ->
703    "/foo#bar?baz" = urlunsplit({"", "", "/foo", "", "bar?baz"}),
704    "https://host:port/foo#bar?baz" =
705        urlunsplit({"https", "host:port", "/foo", "", "bar?baz"}),
706    ok.
707
708urlunsplit_path_test() ->
709    "/foo/bar" = urlunsplit_path({"/foo/bar", "", ""}),
710    "/foo?baz" = urlunsplit_path({"/foo", "baz", ""}),
711    "/foo#bar?baz" = urlunsplit_path({"/foo", "", "bar?baz"}),
712    "/foo#bar?baz#wibble" = urlunsplit_path({"/foo", "", "bar?baz#wibble"}),
713    "/foo?bar#baz" = urlunsplit_path({"/foo", "bar", "baz"}),
714    "/foo?bar?baz#baz" = urlunsplit_path({"/foo", "bar?baz", "baz"}),
715    ok.
716
717join_test() ->
718    ?assertEqual("foo,bar,baz",
719                  join(["foo", "bar", "baz"], $,)),
720    ?assertEqual("foo,bar,baz",
721                  join(["foo", "bar", "baz"], ",")),
722    ?assertEqual("foo bar",
723                  join([["foo", " bar"]], ",")),
724    ?assertEqual("foo bar,baz",
725                  join([["foo", " bar"], "baz"], ",")),
726    ?assertEqual("foo",
727                  join(["foo"], ",")),
728    ?assertEqual("foobarbaz",
729                  join(["foo", "bar", "baz"], "")),
730    ?assertEqual("foo" ++ [<<>>] ++ "bar" ++ [<<>>] ++ "baz",
731                 join(["foo", "bar", "baz"], <<>>)),
732    ?assertEqual("foobar" ++ [<<"baz">>],
733                 join(["foo", "bar", <<"baz">>], "")),
734    ?assertEqual("",
735                 join([], "any")),
736    ok.
737
738quote_plus_test() ->
739    "foo" = quote_plus(foo),
740    "1" = quote_plus(1),
741    "1.1" = quote_plus(1.1),
742    "foo" = quote_plus("foo"),
743    "foo+bar" = quote_plus("foo bar"),
744    "foo%0A" = quote_plus("foo\n"),
745    "foo%0A" = quote_plus("foo\n"),
746    "foo%3B%26%3D" = quote_plus("foo;&="),
747    "foo%3B%26%3D" = quote_plus(<<"foo;&=">>),
748    ok.
749
750unquote_test() ->
751    ?assertEqual("foo bar",
752                 unquote("foo+bar")),
753    ?assertEqual("foo bar",
754                 unquote("foo%20bar")),
755    ?assertEqual("foo\r\n",
756                 unquote("foo%0D%0A")),
757    ?assertEqual("foo\r\n",
758                 unquote(<<"foo%0D%0A">>)),
759    ok.
760
761urlencode_test() ->
762    "foo=bar&baz=wibble+%0D%0A&z=1" = urlencode([{foo, "bar"},
763                                                 {"baz", "wibble \r\n"},
764                                                 {z, 1}]),
765    ok.
766
767parse_qs_test() ->
768    ?assertEqual(
769       [{"foo", "bar"}, {"baz", "wibble \r\n"}, {"z", "1"}],
770       parse_qs("foo=bar&baz=wibble+%0D%0a&z=1")),
771    ?assertEqual(
772       [{"", "bar"}, {"baz", "wibble \r\n"}, {"z", ""}],
773       parse_qs("=bar&baz=wibble+%0D%0a&z=")),
774    ?assertEqual(
775       [{"foo", "bar"}, {"baz", "wibble \r\n"}, {"z", "1"}],
776       parse_qs(<<"foo=bar&baz=wibble+%0D%0a&z=1">>)),
777    ?assertEqual(
778       [],
779       parse_qs("")),
780    ?assertEqual(
781       [{"foo", ""}, {"bar", ""}, {"baz", ""}],
782       parse_qs("foo;bar&baz")),
783    ok.
784
785partition_test() ->
786    {"foo", "", ""} = partition("foo", "/"),
787    {"foo", "/", "bar"} = partition("foo/bar", "/"),
788    {"foo", "/", ""} = partition("foo/", "/"),
789    {"", "/", "bar"} = partition("/bar", "/"),
790    {"f", "oo/ba", "r"} = partition("foo/bar", "oo/ba"),
791    ok.
792
793safe_relative_path_test() ->
794    "foo" = safe_relative_path("foo"),
795    "foo/" = safe_relative_path("foo/"),
796    "foo" = safe_relative_path("foo/bar/.."),
797    "bar" = safe_relative_path("foo/../bar"),
798    "bar/" = safe_relative_path("foo/../bar/"),
799    "" = safe_relative_path("foo/.."),
800    "" = safe_relative_path("foo/../"),
801    undefined = safe_relative_path("/foo"),
802    undefined = safe_relative_path("../foo"),
803    undefined = safe_relative_path("foo/../.."),
804    undefined = safe_relative_path("foo//"),
805    undefined = safe_relative_path("foo\\bar"),
806    ok.
807
808parse_qvalues_test() ->
809    [] = parse_qvalues(""),
810    [{"identity", 0.0}] = parse_qvalues("identity;q=0"),
811    [{"identity", 0.0}] = parse_qvalues("identity ;q=0"),
812    [{"identity", 0.0}] = parse_qvalues(" identity; q =0 "),
813    [{"identity", 0.0}] = parse_qvalues("identity ; q = 0"),
814    [{"identity", 0.0}] = parse_qvalues("identity ; q= 0.0"),
815    [{"gzip", 1.0}, {"deflate", 1.0}, {"identity", 0.0}] = parse_qvalues(
816        "gzip,deflate,identity;q=0.0"
817    ),
818    [{"deflate", 1.0}, {"gzip", 1.0}, {"identity", 0.0}] = parse_qvalues(
819        "deflate,gzip,identity;q=0.0"
820    ),
821    [{"gzip", 1.0}, {"deflate", 1.0}, {"gzip", 1.0}, {"identity", 0.0}] =
822        parse_qvalues("gzip,deflate,gzip,identity;q=0"),
823    [{"gzip", 1.0}, {"deflate", 1.0}, {"identity", 0.0}] = parse_qvalues(
824        "gzip, deflate , identity; q=0.0"
825    ),
826    [{"gzip", 1.0}, {"deflate", 1.0}, {"identity", 0.0}] = parse_qvalues(
827        "gzip; q=1, deflate;q=1.0, identity;q=0.0"
828    ),
829    [{"gzip", 0.5}, {"deflate", 1.0}, {"identity", 0.0}] = parse_qvalues(
830        "gzip; q=0.5, deflate;q=1.0, identity;q=0"
831    ),
832    [{"gzip", 0.5}, {"deflate", 1.0}, {"identity", 0.0}] = parse_qvalues(
833        "gzip; q=0.5, deflate , identity;q=0.0"
834    ),
835    [{"gzip", 0.5}, {"deflate", 0.8}, {"identity", 0.0}] = parse_qvalues(
836        "gzip; q=0.5, deflate;q=0.8, identity;q=0.0"
837    ),
838    [{"gzip", 0.5}, {"deflate", 1.0}, {"identity", 1.0}] = parse_qvalues(
839        "gzip; q=0.5,deflate,identity"
840    ),
841    [{"gzip", 0.5}, {"deflate", 1.0}, {"identity", 1.0}, {"identity", 1.0}] =
842        parse_qvalues("gzip; q=0.5,deflate,identity, identity "),
843    [{"text/html;level=1", 1.0}, {"text/plain", 0.5}] =
844        parse_qvalues("text/html;level=1, text/plain;q=0.5"),
845    [{"text/html;level=1", 0.3}, {"text/plain", 1.0}] =
846        parse_qvalues("text/html;level=1;q=0.3, text/plain"),
847    [{"text/html;level=1", 0.3}, {"text/plain", 1.0}] =
848        parse_qvalues("text/html; level = 1; q = 0.3, text/plain"),
849    [{"text/html;level=1", 0.3}, {"text/plain", 1.0}] =
850        parse_qvalues("text/html;q=0.3;level=1, text/plain"),
851    invalid_qvalue_string = parse_qvalues("gzip; q=1.1, deflate"),
852    invalid_qvalue_string = parse_qvalues("gzip; q=0.5, deflate;q=2"),
853    invalid_qvalue_string = parse_qvalues("gzip, deflate;q=AB"),
854    invalid_qvalue_string = parse_qvalues("gzip; q=2.1, deflate"),
855    invalid_qvalue_string = parse_qvalues("gzip; q=0.1234, deflate"),
856    invalid_qvalue_string = parse_qvalues("text/html;level=1;q=0.3, text/html;level"),
857    ok.
858
859pick_accepted_encodings_test() ->
860    ["identity"] = pick_accepted_encodings(
861        [],
862        ["gzip", "identity"],
863        "identity"
864    ),
865    ["gzip", "identity"] = pick_accepted_encodings(
866        [{"gzip", 1.0}],
867        ["gzip", "identity"],
868        "identity"
869    ),
870    ["identity"] = pick_accepted_encodings(
871        [{"gzip", 0.0}],
872        ["gzip", "identity"],
873        "identity"
874    ),
875    ["gzip", "identity"] = pick_accepted_encodings(
876        [{"gzip", 1.0}, {"deflate", 1.0}],
877        ["gzip", "identity"],
878        "identity"
879    ),
880    ["gzip", "identity"] = pick_accepted_encodings(
881        [{"gzip", 0.5}, {"deflate", 1.0}],
882        ["gzip", "identity"],
883        "identity"
884    ),
885    ["identity"] = pick_accepted_encodings(
886        [{"gzip", 0.0}, {"deflate", 0.0}],
887        ["gzip", "identity"],
888        "identity"
889    ),
890    ["gzip"] = pick_accepted_encodings(
891        [{"gzip", 1.0}, {"deflate", 1.0}, {"identity", 0.0}],
892        ["gzip", "identity"],
893        "identity"
894    ),
895    ["gzip", "deflate", "identity"] = pick_accepted_encodings(
896        [{"gzip", 1.0}, {"deflate", 1.0}],
897        ["gzip", "deflate", "identity"],
898        "identity"
899    ),
900    ["gzip", "deflate"] = pick_accepted_encodings(
901        [{"gzip", 1.0}, {"deflate", 1.0}, {"identity", 0.0}],
902        ["gzip", "deflate", "identity"],
903        "identity"
904    ),
905    ["deflate", "gzip", "identity"] = pick_accepted_encodings(
906        [{"gzip", 0.2}, {"deflate", 1.0}],
907        ["gzip", "deflate", "identity"],
908        "identity"
909    ),
910    ["deflate", "deflate", "gzip", "identity"] = pick_accepted_encodings(
911        [{"gzip", 0.2}, {"deflate", 1.0}, {"deflate", 1.0}],
912        ["gzip", "deflate", "identity"],
913        "identity"
914    ),
915    ["deflate", "gzip", "gzip", "identity"] = pick_accepted_encodings(
916        [{"gzip", 0.2}, {"deflate", 1.0}, {"gzip", 1.0}],
917        ["gzip", "deflate", "identity"],
918        "identity"
919    ),
920    ["gzip", "deflate", "gzip", "identity"] = pick_accepted_encodings(
921        [{"gzip", 0.2}, {"deflate", 0.9}, {"gzip", 1.0}],
922        ["gzip", "deflate", "identity"],
923        "identity"
924    ),
925    [] = pick_accepted_encodings(
926        [{"*", 0.0}],
927        ["gzip", "deflate", "identity"],
928        "identity"
929    ),
930    ["gzip", "deflate", "identity"] = pick_accepted_encodings(
931        [{"*", 1.0}],
932        ["gzip", "deflate", "identity"],
933        "identity"
934    ),
935    ["gzip", "deflate", "identity"] = pick_accepted_encodings(
936        [{"*", 0.6}],
937        ["gzip", "deflate", "identity"],
938        "identity"
939    ),
940    ["gzip"] = pick_accepted_encodings(
941        [{"gzip", 1.0}, {"*", 0.0}],
942        ["gzip", "deflate", "identity"],
943        "identity"
944    ),
945    ["gzip", "deflate"] = pick_accepted_encodings(
946        [{"gzip", 1.0}, {"deflate", 0.6}, {"*", 0.0}],
947        ["gzip", "deflate", "identity"],
948        "identity"
949    ),
950    ["deflate", "gzip"] = pick_accepted_encodings(
951        [{"gzip", 0.5}, {"deflate", 1.0}, {"*", 0.0}],
952        ["gzip", "deflate", "identity"],
953        "identity"
954    ),
955    ["gzip", "identity"] = pick_accepted_encodings(
956        [{"deflate", 0.0}, {"*", 1.0}],
957        ["gzip", "deflate", "identity"],
958        "identity"
959    ),
960    ["gzip", "identity"] = pick_accepted_encodings(
961        [{"*", 1.0}, {"deflate", 0.0}],
962        ["gzip", "deflate", "identity"],
963        "identity"
964    ),
965    ok.
966
967-endif.
968