diff options
Diffstat (limited to 'lib/stdlib/src')
-rw-r--r-- | lib/stdlib/src/io_lib.erl | 13 | ||||
-rw-r--r-- | lib/stdlib/src/io_lib_pretty.erl | 12 | ||||
-rw-r--r-- | lib/stdlib/src/maps.erl | 121 | ||||
-rw-r--r-- | lib/stdlib/src/uri_string.erl | 229 |
4 files changed, 345 insertions, 30 deletions
diff --git a/lib/stdlib/src/io_lib.erl b/lib/stdlib/src/io_lib.erl index 9d447418f8..3c8430b820 100644 --- a/lib/stdlib/src/io_lib.erl +++ b/lib/stdlib/src/io_lib.erl @@ -963,7 +963,18 @@ limit_tail(Other, D) -> %% maps:from_list() creates a map with the same internal ordering of %% the selected associations as in Map. limit_map(Map, D) -> - maps:from_list(erts_internal:maps_to_list(Map, D)). + limit_map(maps:iterator(Map), D, []). + +limit_map(_I, 0, Acc) -> + maps:from_list(Acc); +limit_map(I, D, Acc) -> + case maps:next(I) of + {K, V, NextI} -> + limit_map(NextI, D-1, [{K,V} | Acc]); + none -> + maps:from_list(Acc) + end. + %% maps:from_list(limit_map_body(erts_internal:maps_to_list(Map, D), D)). %% limit_map_body(_, 0) -> [{'...', '...'}]; diff --git a/lib/stdlib/src/io_lib_pretty.erl b/lib/stdlib/src/io_lib_pretty.erl index 505613b80e..89e1931d2d 100644 --- a/lib/stdlib/src/io_lib_pretty.erl +++ b/lib/stdlib/src/io_lib_pretty.erl @@ -478,9 +478,19 @@ print_length(Term, _D, _RF, _Enc, _Str) -> print_length_map(_Map, 1, _RF, _Enc, _Str) -> {"#{...}", 6}; print_length_map(Map, D, RF, Enc, Str) when is_map(Map) -> - Pairs = print_length_map_pairs(erts_internal:maps_to_list(Map, D), D, RF, Enc, Str), + Pairs = print_length_map_pairs(limit_map(maps:iterator(Map), D, []), D, RF, Enc, Str), {{map, Pairs}, list_length(Pairs, 3)}. +limit_map(_I, 0, Acc) -> + Acc; +limit_map(I, D, Acc) -> + case maps:next(I) of + {K, V, NextI} -> + limit_map(NextI, D-1, [{K,V} | Acc]); + none -> + Acc + end. + print_length_map_pairs([], _D, _RF, _Enc, _Str) -> []; print_length_map_pairs(_Pairs, 1, _RF, _Enc, _Str) -> diff --git a/lib/stdlib/src/maps.erl b/lib/stdlib/src/maps.erl index 5dafdb282a..a13f340709 100644 --- a/lib/stdlib/src/maps.erl +++ b/lib/stdlib/src/maps.erl @@ -23,7 +23,8 @@ -export([get/3, filter/2,fold/3, map/2, size/1, update_with/3, update_with/4, - without/2, with/2]). + without/2, with/2, + iterator/1, next/1]). %% BIFs -export([get/2, find/2, from_list/1, @@ -31,6 +32,15 @@ new/0, put/3, remove/2, take/2, to_list/1, update/3, values/1]). +-opaque iterator() :: {term(), term(), iterator()} + | none | nonempty_improper_list(integer(),map()). + +-export_type([iterator/0]). + +-dialyzer({no_improper_lists, iterator/1}). + +-define(IS_ITERATOR(I), is_tuple(I) andalso tuple_size(I) == 3; I == none; is_integer(hd(I)) andalso is_map(tl(I))). + %% Shadowed by erl_bif_types: maps:get/2 -spec get(Key,Map) -> Value when Key :: term(), @@ -39,7 +49,6 @@ get(_,_) -> erlang:nif_error(undef). - -spec find(Key,Map) -> {ok, Value} | error when Key :: term(), Map :: map(), @@ -114,14 +123,20 @@ remove(_,_) -> erlang:nif_error(undef). take(_,_) -> erlang:nif_error(undef). -%% Shadowed by erl_bif_types: maps:to_list/1 -spec to_list(Map) -> [{Key,Value}] when Map :: map(), Key :: term(), Value :: term(). -to_list(_) -> erlang:nif_error(undef). +to_list(Map) when is_map(Map) -> + to_list_internal(erts_internal:map_next(0, Map, [])); +to_list(Map) -> + erlang:error({badmap,Map},[Map]). +to_list_internal([Iter, Map | Acc]) when is_integer(Iter) -> + to_list_internal(erts_internal:map_next(Iter, Map, Acc)); +to_list_internal(Acc) -> + Acc. %% Shadowed by erl_bif_types: maps:update/3 -spec update(Key,Value,Map1) -> Map2 when @@ -192,47 +207,80 @@ get(Key,Map,Default) -> erlang:error({badmap,Map},[Key,Map,Default]). --spec filter(Pred,Map1) -> Map2 when +-spec filter(Pred,MapOrIter) -> Map when Pred :: fun((Key, Value) -> boolean()), Key :: term(), Value :: term(), - Map1 :: map(), - Map2 :: map(). + MapOrIter :: map() | iterator(), + Map :: map(). filter(Pred,Map) when is_function(Pred,2), is_map(Map) -> - maps:from_list([{K,V}||{K,V}<-maps:to_list(Map),Pred(K,V)]); + maps:from_list(filter_1(Pred, iterator(Map))); +filter(Pred,Iterator) when is_function(Pred,2), ?IS_ITERATOR(Iterator) -> + maps:from_list(filter_1(Pred, Iterator)); filter(Pred,Map) -> erlang:error(error_type(Map),[Pred,Map]). - --spec fold(Fun,Init,Map) -> Acc when +filter_1(Pred, Iter) -> + case next(Iter) of + {K, V, NextIter} -> + case Pred(K,V) of + true -> + [{K,V} | filter_1(Pred, NextIter)]; + false -> + filter_1(Pred, NextIter) + end; + none -> + [] + end. + +-spec fold(Fun,Init,MapOrIter) -> Acc when Fun :: fun((K, V, AccIn) -> AccOut), Init :: term(), Acc :: term(), AccIn :: term(), AccOut :: term(), - Map :: map(), + MapOrIter :: map() | iterator(), K :: term(), V :: term(). fold(Fun,Init,Map) when is_function(Fun,3), is_map(Map) -> - lists:foldl(fun({K,V},A) -> Fun(K,V,A) end,Init,maps:to_list(Map)); + fold_1(Fun,Init,iterator(Map)); +fold(Fun,Init,Iterator) when is_function(Fun,3), ?IS_ITERATOR(Iterator) -> + fold_1(Fun,Init,Iterator); fold(Fun,Init,Map) -> erlang:error(error_type(Map),[Fun,Init,Map]). --spec map(Fun,Map1) -> Map2 when +fold_1(Fun, Acc, Iter) -> + case next(Iter) of + {K, V, NextIter} -> + fold_1(Fun, Fun(K,V,Acc), NextIter); + none -> + Acc + end. + +-spec map(Fun,MapOrIter) -> Map when Fun :: fun((K, V1) -> V2), - Map1 :: map(), - Map2 :: map(), + MapOrIter :: map() | iterator(), + Map :: map(), K :: term(), V1 :: term(), V2 :: term(). map(Fun,Map) when is_function(Fun, 2), is_map(Map) -> - maps:from_list([{K,Fun(K,V)}||{K,V}<-maps:to_list(Map)]); + maps:from_list(map_1(Fun, iterator(Map))); +map(Fun,Iterator) when is_function(Fun, 2), ?IS_ITERATOR(Iterator) -> + maps:from_list(map_1(Fun, Iterator)); map(Fun,Map) -> erlang:error(error_type(Map),[Fun,Map]). +map_1(Fun, Iter) -> + case next(Iter) of + {K, V, NextIter} -> + [{K, Fun(K, V)} | map_1(Fun, NextIter)]; + none -> + [] + end. -spec size(Map) -> non_neg_integer() when Map :: map(). @@ -242,6 +290,26 @@ size(Map) when is_map(Map) -> size(Val) -> erlang:error({badmap,Val},[Val]). +-spec iterator(Map) -> Iterator when + Map :: map(), + Iterator :: iterator(). + +iterator(M) when is_map(M) -> [0 | M]; +iterator(M) -> erlang:error({badmap, M}, [M]). + +-spec next(Iterator) -> {Key, Value, NextIterator} | 'none' when + Iterator :: iterator(), + Key :: term(), + Value :: term(), + NextIterator :: iterator(). +next({K, V, I}) -> + {K, V, I}; +next([Path | Map]) when is_integer(Path), is_map(Map) -> + erts_internal:map_next(Path, Map, iterator); +next(none) -> + none; +next(Iter) -> + erlang:error(badarg, [Iter]). -spec without(Ks,Map1) -> Map2 when Ks :: [K], @@ -250,11 +318,10 @@ size(Val) -> K :: term(). without(Ks,M) when is_list(Ks), is_map(M) -> - lists:foldl(fun(K, M1) -> ?MODULE:remove(K, M1) end, M, Ks); + lists:foldl(fun(K, M1) -> maps:remove(K, M1) end, M, Ks); without(Ks,M) -> erlang:error(error_type(M),[Ks,M]). - -spec with(Ks, Map1) -> Map2 when Ks :: [K], Map1 :: map(), @@ -263,17 +330,17 @@ without(Ks,M) -> with(Ks,Map1) when is_list(Ks), is_map(Map1) -> Fun = fun(K, List) -> - case ?MODULE:find(K, Map1) of - {ok, V} -> - [{K, V} | List]; - error -> - List - end - end, - ?MODULE:from_list(lists:foldl(Fun, [], Ks)); + case maps:find(K, Map1) of + {ok, V} -> + [{K, V} | List]; + error -> + List + end + end, + maps:from_list(lists:foldl(Fun, [], Ks)); with(Ks,M) -> erlang:error(error_type(M),[Ks,M]). -error_type(M) when is_map(M) -> badarg; +error_type(M) when is_map(M); ?IS_ITERATOR(M) -> badarg; error_type(V) -> {badmap, V}. diff --git a/lib/stdlib/src/uri_string.erl b/lib/stdlib/src/uri_string.erl index 22212da222..a84679c595 100644 --- a/lib/stdlib/src/uri_string.erl +++ b/lib/stdlib/src/uri_string.erl @@ -226,7 +226,8 @@ %%------------------------------------------------------------------------- %% External API %%------------------------------------------------------------------------- --export([normalize/1, parse/1, +-export([compose_query/1, compose_query/2, + dissect_query/1, normalize/1, parse/1, recompose/1, transcode/2]). -export_type([error/0, uri_map/0, uri_string/0]). @@ -381,6 +382,76 @@ transcode(URIString, Options) when is_list(URIString) -> end. +%%------------------------------------------------------------------------- +%% Functions for working with the query part of a URI as a list +%% of key/value pairs. +%% HTML5 - 4.10.22.6 URL-encoded form data +%%------------------------------------------------------------------------- + +%%------------------------------------------------------------------------- +%% Compose urlencoded query string from a list of unescaped key/value pairs. +%% (application/x-www-form-urlencoded encoding algorithm) +%%------------------------------------------------------------------------- +-spec compose_query(QueryList) -> QueryString when + QueryList :: [{uri_string(), uri_string()}], + QueryString :: uri_string() + | error(). +compose_query(List) -> + compose_query(List, [{encoding, utf8}]). + + +-spec compose_query(QueryList, Options) -> QueryString when + QueryList :: [{uri_string(), uri_string()}], + Options :: [{encoding, atom()}], + QueryString :: uri_string() + | error(). +compose_query([],_Options) -> + []; +compose_query(List, Options) -> + try compose_query(List, Options, false, <<>>) + catch + throw:{error, Atom, RestData} -> {error, Atom, RestData} + end. +%% +compose_query([{Key,Value}|Rest], Options, IsList, Acc) -> + Separator = get_separator(Rest), + K = form_urlencode(Key, Options), + V = form_urlencode(Value, Options), + IsListNew = IsList orelse is_list(Key) orelse is_list(Value), + compose_query(Rest, Options, IsListNew, <<Acc/binary,K/binary,"=",V/binary,Separator/binary>>); +compose_query([], _Options, IsList, Acc) -> + case IsList of + true -> convert_to_list(Acc, utf8); + false -> Acc + end. + + +%%------------------------------------------------------------------------- +%% Dissect a query string into a list of unescaped key/value pairs. +%% (application/x-www-form-urlencoded decoding algorithm) +%%------------------------------------------------------------------------- +-spec dissect_query(QueryString) -> QueryList when + QueryString :: uri_string(), + QueryList :: [{uri_string(), uri_string()}] + | error(). +dissect_query(<<>>) -> + []; +dissect_query([]) -> + []; +dissect_query(QueryString) when is_list(QueryString) -> + try + B = convert_to_binary(QueryString, utf8, utf8), + dissect_query_key(B, true, [], <<>>, <<>>) + catch + throw:{error, Atom, RestData} -> {error, Atom, RestData} + end; +dissect_query(QueryString) -> + try dissect_query_key(QueryString, false, [], <<>>, <<>>) + catch + throw:{error, Atom, RestData} -> {error, Atom, RestData} + end. + + %%%======================================================================== %%% Internal functions %%%======================================================================== @@ -585,6 +656,7 @@ maybe_add_path(Map) -> end. + -spec parse_scheme(binary(), uri_map()) -> {binary(), uri_map()}. parse_scheme(?STRING_REST($:, Rest), URI) -> {_, URI1} = parse_hier(Rest, URI), @@ -1673,6 +1745,161 @@ percent_encode_segment(Segment) -> %%------------------------------------------------------------------------- +%% Helper functions for compose_query +%%------------------------------------------------------------------------- + +%% Returns separator to be used between key-value pairs +get_separator(L) when length(L) =:= 0 -> + <<>>; +get_separator(_L) -> + <<"&">>. + + +%% HTML5 - 4.10.22.6 URL-encoded form data - encoding +form_urlencode(Cs, [{encoding, latin1}]) when is_list(Cs) -> + B = convert_to_binary(Cs, utf8, utf8), + html5_byte_encode(base10_encode(B)); +form_urlencode(Cs, [{encoding, latin1}]) when is_binary(Cs) -> + html5_byte_encode(base10_encode(Cs)); +form_urlencode(Cs, [{encoding, Encoding}]) + when is_list(Cs), Encoding =:= utf8; Encoding =:= unicode -> + B = convert_to_binary(Cs, utf8, Encoding), + html5_byte_encode(B); +form_urlencode(Cs, [{encoding, Encoding}]) + when is_binary(Cs), Encoding =:= utf8; Encoding =:= unicode -> + html5_byte_encode(Cs); +form_urlencode(Cs, [{encoding, Encoding}]) when is_list(Cs); is_binary(Cs) -> + throw({error,invalid_encoding, Encoding}); +form_urlencode(Cs, _) -> + throw({error,invalid_input, Cs}). + + +%% For each character in the entry's name and value that cannot be expressed using +%% the selected character encoding, replace the character by a string consisting of +%% a U+0026 AMPERSAND character (&), a "#" (U+0023) character, one or more ASCII +%% digits representing the Unicode code point of the character in base ten, and +%% finally a ";" (U+003B) character. +base10_encode(Cs) -> + base10_encode(Cs, <<>>). +%% +base10_encode(<<>>, Acc) -> + Acc; +base10_encode(<<H/utf8,T/binary>>, Acc) when H > 255 -> + Base10 = convert_to_binary(integer_to_list(H,10), utf8, utf8), + base10_encode(T, <<Acc/binary,"&#",Base10/binary,$;>>); +base10_encode(<<H/utf8,T/binary>>, Acc) -> + base10_encode(T, <<Acc/binary,H>>). + + +html5_byte_encode(B) -> + html5_byte_encode(B, <<>>). +%% +html5_byte_encode(<<>>, Acc) -> + Acc; +html5_byte_encode(<<$ ,T/binary>>, Acc) -> + html5_byte_encode(T, <<Acc/binary,$+>>); +html5_byte_encode(<<H,T/binary>>, Acc) -> + case is_url_char(H) of + true -> + html5_byte_encode(T, <<Acc/binary,H>>); + false -> + <<A:4,B:4>> = <<H>>, + html5_byte_encode(T, <<Acc/binary,$%,(?DEC2HEX(A)),(?DEC2HEX(B))>>) + end; +html5_byte_encode(H, _Acc) -> + throw({error,invalid_input, H}). + + +%% Return true if input char can appear in form-urlencoded string +%% Allowed chararacters: +%% 0x2A, 0x2D, 0x2E, 0x30 to 0x39, 0x41 to 0x5A, +%% 0x5F, 0x61 to 0x7A +is_url_char(C) + when C =:= 16#2A; C =:= 16#2D; + C =:= 16#2E; C =:= 16#5F; + 16#30 =< C, C =< 16#39; + 16#41 =< C, C =< 16#5A; + 16#61 =< C, C =< 16#7A -> true; +is_url_char(_) -> false. + + +%%------------------------------------------------------------------------- +%% Helper functions for dissect_query +%%------------------------------------------------------------------------- +dissect_query_key(<<$=,T/binary>>, IsList, Acc, Key, Value) -> + dissect_query_value(T, IsList, Acc, Key, Value); +dissect_query_key(<<"&#",T/binary>>, IsList, Acc, Key, Value) -> + dissect_query_key(T, IsList, Acc, <<Key/binary,"&#">>, Value); +dissect_query_key(<<$&,_T/binary>>, _IsList, _Acc, _Key, _Value) -> + throw({error, missing_value, "&"}); +dissect_query_key(<<H,T/binary>>, IsList, Acc, Key, Value) -> + dissect_query_key(T, IsList, Acc, <<Key/binary,H>>, Value); +dissect_query_key(B, _, _, _, _) -> + throw({error, missing_value, B}). + + +dissect_query_value(<<$&,T/binary>>, IsList, Acc, Key, Value) -> + K = form_urldecode(IsList, Key), + V = form_urldecode(IsList, Value), + dissect_query_key(T, IsList, [{K,V}|Acc], <<>>, <<>>); +dissect_query_value(<<H,T/binary>>, IsList, Acc, Key, Value) -> + dissect_query_value(T, IsList, Acc, Key, <<Value/binary,H>>); +dissect_query_value(<<>>, IsList, Acc, Key, Value) -> + K = form_urldecode(IsList, Key), + V = form_urldecode(IsList, Value), + lists:reverse([{K,V}|Acc]). + + +%% Form-urldecode input based on RFC 1866 [8.2.1] +form_urldecode(true, B) -> + Result = base10_decode(form_urldecode(B, <<>>)), + convert_to_list(Result, utf8); +form_urldecode(false, B) -> + base10_decode(form_urldecode(B, <<>>)); +form_urldecode(<<>>, Acc) -> + Acc; +form_urldecode(<<$+,T/binary>>, Acc) -> + form_urldecode(T, <<Acc/binary,$ >>); +form_urldecode(<<$%,C0,C1,T/binary>>, Acc) -> + case is_hex_digit(C0) andalso is_hex_digit(C1) of + true -> + V = ?HEX2DEC(C0)*16+?HEX2DEC(C1), + form_urldecode(T, <<Acc/binary, V>>); + false -> + L = convert_to_list(<<$%,C0,C1,T/binary>>, utf8), + throw({error, invalid_percent_encoding, L}) + end; +form_urldecode(<<H/utf8,T/binary>>, Acc) -> + form_urldecode(T, <<Acc/binary,H/utf8>>); +form_urldecode(<<H,_/binary>>, _Acc) -> + throw({error, invalid_character, [H]}). + +base10_decode(Cs) -> + base10_decode(Cs, <<>>). +% +base10_decode(<<>>, Acc) -> + Acc; +base10_decode(<<"&#",T/binary>>, Acc) -> + base10_decode_unicode(T, Acc); +base10_decode(<<H/utf8,T/binary>>, Acc) -> + base10_decode(T,<<Acc/binary,H/utf8>>); +base10_decode(<<H,_/binary>>, _) -> + throw({error, invalid_input, [H]}). + + +base10_decode_unicode(B, Acc) -> + base10_decode_unicode(B, 0, Acc). +%% +base10_decode_unicode(<<H/utf8,T/binary>>, Codepoint, Acc) when $0 =< H, H =< $9 -> + Res = Codepoint * 10 + (H - $0), + base10_decode_unicode(T, Res, Acc); +base10_decode_unicode(<<$;,T/binary>>, Codepoint, Acc) -> + base10_decode(T, <<Acc/binary,Codepoint/utf8>>); +base10_decode_unicode(<<H,_/binary>>, _, _) -> + throw({error, invalid_input, [H]}). + + +%%------------------------------------------------------------------------- %% Helper functions for normalize %%------------------------------------------------------------------------- |