diff options
Diffstat (limited to 'lib/stdlib')
| -rw-r--r-- | lib/stdlib/doc/src/Makefile | 3 | ||||
| -rw-r--r-- | lib/stdlib/doc/src/maps.xml | 94 | ||||
| -rw-r--r-- | lib/stdlib/doc/src/uri_string.xml | 98 | ||||
| -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 | ||||
| -rw-r--r-- | lib/stdlib/test/ets_SUITE.erl | 2 | ||||
| -rw-r--r-- | lib/stdlib/test/io_SUITE.erl | 4 | ||||
| -rw-r--r-- | lib/stdlib/test/maps_SUITE.erl | 42 | ||||
| -rw-r--r-- | lib/stdlib/test/uri_string_SUITE.erl | 86 |
11 files changed, 659 insertions, 45 deletions
diff --git a/lib/stdlib/doc/src/Makefile b/lib/stdlib/doc/src/Makefile index aeed79408b..af27efa6c1 100644 --- a/lib/stdlib/doc/src/Makefile +++ b/lib/stdlib/doc/src/Makefile @@ -105,7 +105,8 @@ XML_REF3_FILES = \ XML_REF6_FILES = stdlib_app.xml XML_PART_FILES = part.xml -XML_CHAPTER_FILES = io_protocol.xml unicode_usage.xml notes.xml notes_history.xml assert_hrl.xml +XML_CHAPTER_FILES = introduction.xml io_protocol.xml unicode_usage.xml \ + notes.xml assert_hrl.xml BOOK_FILES = book.xml diff --git a/lib/stdlib/doc/src/maps.xml b/lib/stdlib/doc/src/maps.xml index 8c7270816b..987d92989d 100644 --- a/lib/stdlib/doc/src/maps.xml +++ b/lib/stdlib/doc/src/maps.xml @@ -33,16 +33,31 @@ <p>This module contains functions for maps processing.</p> </description> + <datatypes> + <datatype> + <name name="iterator"/> + <desc> + <p>An iterator representing the key value associations in a map.</p> + <p>Created using <seealso marker="#iterator-1"><c>maps:iterator/1</c></seealso>.</p> + <p>Consumed by <seealso marker="#next-1"><c>maps:next/1</c></seealso>, + <seealso marker="#filter-2"><c>maps:filter/2</c></seealso>, + <seealso marker="#fold-3"><c>maps:fold/3</c></seealso> and + <seealso marker="#map-2"><c>maps:map/2</c></seealso>.</p> + </desc> + </datatype> + </datatypes> + <funcs> <func> <name name="filter" arity="2"/> <fsummary>Select pairs that satisfy a predicate.</fsummary> <desc> - <p>Returns a map <c><anno>Map2</anno></c> for which predicate - <c><anno>Pred</anno></c> holds true in <c><anno>Map1</anno></c>.</p> + <p>Returns a map <c><anno>Map</anno></c> for which predicate + <c><anno>Pred</anno></c> holds true in <c><anno>MapOrIter</anno></c>.</p> <p>The call fails with a <c>{badmap,Map}</c> exception if - <c><anno>Map1</anno></c> is not a map, or with <c>badarg</c> if - <c><anno>Pred</anno></c> is not a function of arity 2.</p> + <c><anno>MapOrIter</anno></c> is not a map or valid iterator, + or with <c>badarg</c> if <c><anno>Pred</anno></c> is not a + function of arity 2.</p> <p><em>Example:</em></p> <code type="none"> > M = #{a => 2, b => 3, c=> 4, "a" => 1, "b" => 2, "c" => 4}, @@ -76,12 +91,16 @@ <fsummary></fsummary> <desc> <p>Calls <c>F(K, V, AccIn)</c> for every <c><anno>K</anno></c> to value - <c><anno>V</anno></c> association in <c><anno>Map</anno></c> in + <c><anno>V</anno></c> association in <c><anno>MapOrIter</anno></c> in any order. Function <c>fun F/3</c> must return a new accumulator, which is passed to the next successive call. This function returns the final value of the accumulator. The initial accumulator value <c><anno>Init</anno></c> is returned if the map is empty.</p> + <p>The call fails with a <c>{badmap,Map}</c> exception if + <c><anno>MapOrIter</anno></c> is not a map or valid iterator, + or with <c>badarg</c> if <c><anno>Fun</anno></c> is not a + function of arity 3.</p> <p><em>Example:</em></p> <code type="none"> > Fun = fun(K,V,AccIn) when is_list(K) -> AccIn + V end, @@ -169,6 +188,32 @@ false</code> </func> <func> + <name name="iterator" arity="1"/> + <fsummary>Create a map iterator.</fsummary> + <desc> + <p>Returns a map iterator <c><anno>Iterator</anno></c> that can + be used by <seealso marker="#next-1"><c>maps:next/1</c></seealso> + to traverse the key-value associations in a map. When iterating + over a map, the memory usage is guaranteed to be bounded no matter + the size of the map.</p> + <p>The call fails with a <c>{badmap,Map}</c> exception if + <c><anno>Map</anno></c> is not a map.</p> + <p><em>Example:</em></p> + <code type="none"> +> M = #{ a => 1, b => 2 }. +#{a => 1,b => 2} +> I = maps:iterator(M). +[{a,1},{b,2}] +> {K1, V1, I2} = maps:next(I). +{a,1,[{b,2}]} +> {K2, V2, I3} = maps:next(I2). +{b,2,[]} +> maps:next(I3). +none</code> + </desc> + </func> + + <func> <name name="keys" arity="1"/> <fsummary></fsummary> <desc> @@ -188,12 +233,16 @@ false</code> <name name="map" arity="2"/> <fsummary></fsummary> <desc> - <p>Produces a new map <c><anno>Map2</anno></c> by calling function + <p>Produces a new map <c><anno>Map</anno></c> by calling function <c>fun F(K, V1)</c> for every <c><anno>K</anno></c> to value - <c><anno>V1</anno></c> association in <c><anno>Map1</anno></c> in + <c><anno>V1</anno></c> association in <c><anno>MapOrIter</anno></c> in any order. Function <c>fun F/2</c> must return value <c><anno>V2</anno></c> to be associated with key <c><anno>K</anno></c> - for the new map <c><anno>Map2</anno></c>.</p> + for the new map <c><anno>Map</anno></c>.</p> + <p>The call fails with a <c>{badmap,Map}</c> exception if + <c><anno>MapOrIter</anno></c> is not a map or valid iterator, + or with <c>badarg</c> if <c><anno>Fun</anno></c> is not a + function of arity 2.</p> <p><em>Example:</em></p> <code type="none"> > Fun = fun(K,V1) when is_list(K) -> V1*2 end, @@ -234,6 +283,35 @@ false</code> </func> <func> + <name name="next" arity="1"/> + <fsummary>Get the next key and value from an iterator.</fsummary> + <desc> + <p>Returns the next key-value association in + <c><anno>Iterator</anno></c> and a new iterator for the + remaining associations in the iterator. + </p> + <p> + If there are no more associations in the iterator, + <c>none</c> is returned. + </p> + <p><em>Example:</em></p> + <code type="none"> +> Map = #{a => 1, b => 2, c => 3}. +#{a => 1,b => 2,c => 3} +> Iter = maps:iterator(Map). +[{a,1},{b,2},{c,3}] +> {_, _, Iter1} = maps:next(Iter). +{a,1,[{b,2},{c,3}]} +> {_, _, Iter2} = maps:next(Iter1). +{b,2,[{c,3}]} +> {_, _, Iter3} = maps:next(Iter2). +{c,3,[]} +> maps:next(Iter3). +none</code> + </desc> + </func> + + <func> <name name="put" arity="3"/> <fsummary></fsummary> <desc> diff --git a/lib/stdlib/doc/src/uri_string.xml b/lib/stdlib/doc/src/uri_string.xml index 9ace2b0a05..21f470e763 100644 --- a/lib/stdlib/doc/src/uri_string.xml +++ b/lib/stdlib/doc/src/uri_string.xml @@ -31,7 +31,8 @@ <modulesummary>URI processing functions.</modulesummary> <description> <p>This module contains functions for parsing and handling URIs - (<url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url>). + (<url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url>) and + form-urlencoded query strings (<url href="https://www.w3.org/TR/html5/forms.html">HTML5</url>). </p> <p>A URI is an identifier consisting of a sequence of characters matching the syntax rule named <em>URI</em> in <url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url>. @@ -71,6 +72,13 @@ <item>Transforming URIs into a normalized form<br></br> <seealso marker="#normalize/1"><c>normalize/1</c></seealso> </item> + <item>Composing form-urlencoded query strings from a list of key-value pairs<br></br> + <seealso marker="#compose_query/1"><c>compose_query/1</c></seealso><br></br> + <seealso marker="#compose_query/2"><c>compose_query/2</c></seealso> + </item> + <item>Dissecting form-urlencoded query strings into a list of key-value pairs<br></br> + <seealso marker="#dissect_query/1"><c>dissect_query/1</c></seealso> + </item> </list> <p>There are four different encodings present during the handling of URIs:</p> <list type="bulleted"> @@ -102,12 +110,15 @@ <desc> <p>Error tuple indicating the type of error. Possible values of the second component:</p> <list type="bulleted"> + <item><c>invalid_character</c></item> + <item><c>invalid_encoding</c></item> <item><c>invalid_input</c></item> <item><c>invalid_map</c></item> <item><c>invalid_percent_encoding</c></item> <item><c>invalid_scheme</c></item> <item><c>invalid_uri</c></item> <item><c>invalid_utf8</c></item> + <item><c>missing_value</c></item> </list> <p>The third component is a term providing additional information about the cause of the error.</p> @@ -134,6 +145,91 @@ <funcs> <func> + <name name="compose_query" arity="1"/> + <fsummary>Compose urlencoded query string.</fsummary> + <desc> + <p>Composes a form-urlencoded <c><anno>QueryString</anno></c> based on a + <c><anno>QueryList</anno></c>, a list of non-percent-encoded key-value pairs. + Form-urlencoding is defined in section + 4.10.22.6 of the <url href="https://www.w3.org/TR/html5/forms.html">HTML5</url> + specification. + </p> + <p>See also the opposite operation <seealso marker="#dissect_query/1"> + <c>dissect_query/1</c></seealso>. + </p> + <p><em>Example:</em></p> + <pre> +1> <input>uri_string:compose_query([{"foo bar","1"},{"city","örebro"}]).</input> +<![CDATA["foo+bar=1&city=%C3%B6rebro"]]> +2> <![CDATA[uri_string:compose_query([{<<"foo bar">>,<<"1">>}, +2> {<<"city">>,<<"örebro"/utf8>>}]).]]> +<![CDATA[<<"foo+bar=1&city=%C3%B6rebro">>]]> + </pre> + </desc> + </func> + + <func> + <name name="compose_query" arity="2"/> + <fsummary>Compose urlencoded query string.</fsummary> + <desc> + <p>Same as <c>compose_query/1</c> but with an additional + <c><anno>Options</anno></c> parameter, that controls the encoding ("charset") + used by the encoding algorithm. There are two supported encodings: <c>utf8</c> + (or <c>unicode</c>) and <c>latin1</c>. + </p> + <p>Each character in the entry's name and value that cannot be expressed using + the selected character encoding, is replaced by a string consisting of a U+0026 + AMPERSAND character (<![CDATA[&]]>), 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. + </p> + <p>Bytes that are out of the range 0x2A, 0x2D, 0x2E, 0x30 to 0x39, 0x41 to 0x5A, 0x5F, + 0x61 to 0x7A, are percent-encoded (U+0025 PERCENT SIGN character (%) followed by + uppercase ASCII hex digits representing the hexadecimal value of the byte). + </p> + <p>See also the opposite operation <seealso marker="#dissect_query/1"> + <c>dissect_query/1</c></seealso>. + </p> + <p><em>Example:</em></p> + <pre> +1> <input>uri_string:compose_query([{"foo bar","1"},{"city","örebro"}],</input> +1> [{encoding, latin1}]). +<![CDATA["foo+bar=1&city=%F6rebro" +2> uri_string:compose_query([{<<"foo bar">>,<<"1">>}, +2> {<<"city">>,<<"東京"/utf8>>}], [{encoding, latin1}]).]]> +<![CDATA[<<"foo+bar=1&city=%26%2326481%3B%26%2320140%3B">>]]> + </pre> + </desc> + </func> + + <func> + <name name="dissect_query" arity="1"/> + <fsummary>Dissect query string.</fsummary> + <desc> + <p>Dissects an urlencoded <c><anno>QueryString</anno></c> and returns a + <c><anno>QueryList</anno></c>, a list of non-percent-encoded key-value pairs. + Form-urlencoding is defined in section + 4.10.22.6 of the <url href="https://www.w3.org/TR/html5/forms.html">HTML5</url> + specification. + </p> + <p>It is not as strict for its input as the decoding algorithm defined by + <url href="https://www.w3.org/TR/html5/forms.html">HTML5</url> + and accepts all unicode characters.</p> + <p>See also the opposite operation <seealso marker="#compose_query/1"> + <c>compose_query/1</c></seealso>. + </p> + <p><em>Example:</em></p> + <pre> +1> <input><![CDATA[uri_string:dissect_query("foo+bar=1&city=%C3%B6rebro").]]></input> +[{"foo bar","1"},{"city","örebro"}] +2> <![CDATA[uri_string:dissect_query(<<"foo+bar=1&city=%26%2326481%3B%26%2320140%3B">>).]]> +<![CDATA[[{<<"foo bar">>,<<"1">>}, + {<<"city">>,<<230,157,177,228,186,172>>}] ]]> + </pre> + </desc> + </func> + + <func> <name name="normalize" arity="1"/> <fsummary>Syntax-based normalization.</fsummary> <desc> 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 %%------------------------------------------------------------------------- diff --git a/lib/stdlib/test/ets_SUITE.erl b/lib/stdlib/test/ets_SUITE.erl index f329a07b7a..07c8b60cbd 100644 --- a/lib/stdlib/test/ets_SUITE.erl +++ b/lib/stdlib/test/ets_SUITE.erl @@ -6369,7 +6369,7 @@ very_big_num(0, Result) -> Result. make_port() -> - open_port({spawn, "efile"}, [eof]). + hd(erlang:ports()). make_pid() -> spawn_link(fun sleeper/0). diff --git a/lib/stdlib/test/io_SUITE.erl b/lib/stdlib/test/io_SUITE.erl index e2c73371cd..13929bdbb6 100644 --- a/lib/stdlib/test/io_SUITE.erl +++ b/lib/stdlib/test/io_SUITE.erl @@ -2138,8 +2138,8 @@ otp_14175(_Config) -> "#{}" = p(#{}, 1), "#{...}" = p(#{a => 1}, 1), "#{#{} => a}" = p(#{#{} => a}, 2), - "#{a => 1,...}" = p(#{a => 1, b => 2}, 2), - "#{a => 1,b => 2}" = p(#{a => 1, b => 2}, -1), + mt("#{a => 1,...}", p(#{a => 1, b => 2}, 2)), + mt("#{a => 1,b => 2}", p(#{a => 1, b => 2}, -1)), M = #{kaaaaaaaaaaaaaaaaaaa => v1,kbbbbbbbbbbbbbbbbbbb => v2, kccccccccccccccccccc => v3,kddddddddddddddddddd => v4, diff --git a/lib/stdlib/test/maps_SUITE.erl b/lib/stdlib/test/maps_SUITE.erl index 42e669a799..a75751b31d 100644 --- a/lib/stdlib/test/maps_SUITE.erl +++ b/lib/stdlib/test/maps_SUITE.erl @@ -30,6 +30,7 @@ -export([t_update_with_3/1, t_update_with_4/1, t_get_3/1, t_filter_2/1, t_fold_3/1,t_map_2/1,t_size_1/1, + t_iterator_1/1, t_with_2/1,t_without_2/1]). %%-define(badmap(V,F,Args), {'EXIT', {{badmap,V}, [{maps,F,Args,_}|_]}}). @@ -47,6 +48,7 @@ all() -> [t_update_with_3,t_update_with_4, t_get_3,t_filter_2, t_fold_3,t_map_2,t_size_1, + t_iterator_1, t_with_2,t_without_2]. t_update_with_3(Config) when is_list(Config) -> @@ -127,6 +129,8 @@ t_filter_2(Config) when is_list(Config) -> Pred2 = fun(K,V) -> is_list(K) andalso (V rem 2) =:= 0 end, #{a := 2,c := 4} = maps:filter(Pred1,M), #{"b" := 2,"c" := 4} = maps:filter(Pred2,M), + #{a := 2,c := 4} = maps:filter(Pred1,maps:iterator(M)), + #{"b" := 2,"c" := 4} = maps:filter(Pred2,maps:iterator(M)), %% error case ?badmap(a,filter,[_,a]) = (catch maps:filter(fun(_,_) -> ok end,id(a))), ?badarg(filter,[<<>>,#{}]) = (catch maps:filter(id(<<>>),#{})), @@ -139,6 +143,8 @@ t_fold_3(Config) when is_list(Config) -> Tot0 = lists:sum(Vs), Tot1 = maps:fold(fun({k,_},V,A) -> A + V end, 0, M0), true = Tot0 =:= Tot1, + Tot2 = maps:fold(fun({k,_},V,A) -> A + V end, 0, maps:iterator(M0)), + true = Tot0 =:= Tot2, %% error case ?badmap(a,fold,[_,0,a]) = (catch maps:fold(fun(_,_,_) -> ok end,0,id(a))), @@ -151,12 +157,48 @@ t_map_2(Config) when is_list(Config) -> #{ {k,1} := 1, {k,200} := 200} = M0, M1 = maps:map(fun({k,_},V) -> V + 42 end, M0), #{ {k,1} := 43, {k,200} := 242} = M1, + M2 = maps:map(fun({k,_},V) -> V + 42 end, maps:iterator(M0)), + #{ {k,1} := 43, {k,200} := 242} = M2, %% error case ?badmap(a,map,[_,a]) = (catch maps:map(fun(_,_) -> ok end, id(a))), ?badarg(map,[<<>>,#{}]) = (catch maps:map(id(<<>>),#{})), ok. +t_iterator_1(Config) when is_list(Config) -> + + %% Small map test + M0 = #{ a => 1, b => 2 }, + I0 = maps:iterator(M0), + {K1,V1,I1} = maps:next(I0), + {K2,V2,I2} = maps:next(I1), + none = maps:next(I2), + + KVList = lists:sort([{K1,V1},{K2,V2}]), + KVList = lists:sort(maps:to_list(M0)), + + %% Large map test + + Vs2 = lists:seq(1,200), + M2 = maps:from_list([{{k,I},I}||I<-Vs2]), + KVList2 = lists:sort(iter_kv(maps:iterator(M2))), + KVList2 = lists:sort(maps:to_list(M2)), + + %% Larger map test + + Vs3 = lists:seq(1,10000), + M3 = maps:from_list([{{k,I},I}||I<-Vs3]), + KVList3 = lists:sort(iter_kv(maps:iterator(M3))), + KVList3 = lists:sort(maps:to_list(M3)), + ok. + +iter_kv(I) -> + case maps:next(I) of + none -> + []; + {K,V,NI} -> + [{K,V} | iter_kv(NI)] + end. t_size_1(Config) when is_list(Config) -> 0 = maps:size(#{}), diff --git a/lib/stdlib/test/uri_string_SUITE.erl b/lib/stdlib/test/uri_string_SUITE.erl index c625da56c6..fef356355c 100644 --- a/lib/stdlib/test/uri_string_SUITE.erl +++ b/lib/stdlib/test/uri_string_SUITE.erl @@ -38,7 +38,10 @@ recompose_query/1, recompose_parse_query/1, recompose_path/1, recompose_parse_path/1, recompose_autogen/1, parse_recompose_autogen/1, - transcode_basic/1, transcode_options/1, transcode_mixed/1, transcode_negative/1 + transcode_basic/1, transcode_options/1, transcode_mixed/1, transcode_negative/1, + compose_query/1, compose_query_latin1/1, compose_query_negative/1, + dissect_query/1, dissect_query_negative/1, + interop_query_latin1/1, interop_query_utf8/1 ]). @@ -107,7 +110,14 @@ all() -> transcode_basic, transcode_options, transcode_mixed, - transcode_negative + transcode_negative, + compose_query, + compose_query_latin1, + compose_query_negative, + dissect_query, + dissect_query_negative, + interop_query_latin1, + interop_query_utf8 ]. groups() -> @@ -823,6 +833,65 @@ transcode_negative(_Config) -> {error,invalid_input,<<"ö">>} = uri_string:transcode("foo%F6bar", [{in_encoding, utf8},{out_encoding, utf8}]). +compose_query(_Config) -> + [] = uri_string:compose_query([]), + "foo=1&bar=2" = uri_string:compose_query([{<<"foo">>,"1"}, {"bar", "2"}]), + "foo=1&b%C3%A4r=2" = uri_string:compose_query([{"foo","1"}, {"bär", "2"}],[{encoding,utf8}]), + "foo=1&b%C3%A4r=2" = uri_string:compose_query([{"foo","1"}, {"bär", "2"}],[{encoding,unicode}]), + "foo=1&b%E4r=2" = uri_string:compose_query([{"foo","1"}, {"bär", "2"}],[{encoding,latin1}]), + "foo+bar=1&%E5%90%88=2" = uri_string:compose_query([{"foo bar","1"}, {"合", "2"}]), + "foo+bar=1&%26%2321512%3B=2" = + uri_string:compose_query([{"foo bar","1"}, {"合", "2"}],[{encoding,latin1}]), + "foo+bar=1&%C3%B6=2" = uri_string:compose_query([{<<"foo bar">>,<<"1">>}, {"ö", <<"2">>}]), + <<"foo+bar=1&%C3%B6=2">> = + uri_string:compose_query([{<<"foo bar">>,<<"1">>}, {<<"ö"/utf8>>, <<"2">>}]). + +compose_query_latin1(_Config) -> + Q = uri_string:compose_query([{"合foö bar","1"}, {"合", "合"}],[{encoding,latin1}]), + Q1 = uri_string:transcode(Q, [{in_encoding, latin1}]), + [{"合foö bar","1"}, {"合", "合"}] = uri_string:dissect_query(Q1), + Q2 = uri_string:compose_query([{<<"合foö bar"/utf8>>,<<"1">>}, {<<"合"/utf8>>, <<"合"/utf8>>}], + [{encoding,latin1}]), + Q3 = uri_string:transcode(Q2, [{in_encoding, latin1}]), + [{<<"合foö bar"/utf8>>,<<"1">>}, {<<"合"/utf8>>, <<"合"/utf8>>}] = + uri_string:dissect_query(Q3). + +compose_query_negative(_Config) -> + {error,invalid_input,4} = uri_string:compose_query([{"",4}]), + {error,invalid_input,5} = uri_string:compose_query([{5,""}]), + {error,invalid_encoding,utf16} = + uri_string:compose_query([{"foo bar","1"}, {<<"ö">>, "2"}],[{encoding,utf16}]). + +dissect_query(_Config) -> + [] = uri_string:dissect_query(""), + [{"foo","1"}, {"amp;bar", "2"}] = uri_string:dissect_query("foo=1&bar=2"), + [{"foo","1"}, {"bar", "2"}] = uri_string:dissect_query("foo=1&bar=2"), + [{"foo","1;bar=2"}] = uri_string:dissect_query("foo=1;bar=2"), + [{"foo","1"}, {"bar", "222"}] = uri_string:dissect_query([<<"foo=1&bar=2">>,"22"]), + [{"foo","ö"}, {"bar", "2"}] = uri_string:dissect_query("foo=%C3%B6&bar=2"), + [{<<"foo">>,<<"ö"/utf8>>}, {<<"bar">>, <<"2">>}] = + uri_string:dissect_query(<<"foo=%C3%B6&bar=2">>), + [{"foo bar","1"},{"ö","2"}] = + uri_string:dissect_query([<<"foo+bar=1&">>,<<"%C3%B6=2">>]), + [{"foo bar","1"},{[21512],"2"}] = + uri_string:dissect_query("foo+bar=1&%26%2321512%3B=2"), + [{<<"foo bar">>,<<"1">>},{<<"合"/utf8>>,<<"2">>}] = + uri_string:dissect_query(<<"foo+bar=1&%26%2321512%3B=2">>), + [{"föo bar","1"},{"ö","2"}] = + uri_string:dissect_query("föo+bar=1&%C3%B6=2"), + [{<<"föo bar"/utf8>>,<<"1">>},{<<"ö"/utf8>>,<<"2">>}] = + uri_string:dissect_query(<<"föo+bar=1&%C3%B6=2"/utf8>>). + +dissect_query_negative(_Config) -> + {error,missing_value,"&"} = + uri_string:dissect_query("foo1&bar=2"), + {error,invalid_percent_encoding,"%XX%B6"} = uri_string:dissect_query("foo=%XX%B6&bar=2"), + {error,invalid_input,[153]} = + uri_string:dissect_query("foo=%99%B6&bar=2"), + {error,invalid_character,"ö"} = uri_string:dissect_query(<<"föo+bar=1&%C3%B6=2">>), + {error,invalid_input,<<"ö">>} = + uri_string:dissect_query([<<"foo+bar=1&">>,<<"%C3%B6=2ö">>]). + normalize(_Config) -> "/a/g" = uri_string:normalize("/a/b/c/./../../g"), <<"mid/6">> = uri_string:normalize(<<"mid/content=5/../6">>), @@ -842,3 +911,16 @@ normalize(_Config) -> uri_string:normalize(<<"sftp://localhost:22">>), <<"tftp://localhost">> = uri_string:normalize(<<"tftp://localhost:69">>). + +interop_query_utf8(_Config) -> + Q = uri_string:compose_query([{"foo bar","1"}, {"合", "2"}]), + Uri = uri_string:recompose(#{path => "/", query => Q}), + #{query := Q1} = uri_string:parse(Uri), + [{"foo bar","1"}, {"合", "2"}] = uri_string:dissect_query(Q1). + +interop_query_latin1(_Config) -> + Q = uri_string:compose_query([{"foo bar","1"}, {"合", "2"}], [{encoding,latin1}]), + Uri = uri_string:recompose(#{path => "/", query => Q}), + Uri1 = uri_string:transcode(Uri, [{in_encoding, latin1}]), + #{query := Q1} = uri_string:parse(Uri1), + [{"foo bar","1"}, {"合", "2"}] = uri_string:dissect_query(Q1). |
