aboutsummaryrefslogtreecommitdiffstats
path: root/lib/stdlib/src
diff options
context:
space:
mode:
Diffstat (limited to 'lib/stdlib/src')
-rw-r--r--lib/stdlib/src/Makefile1
-rw-r--r--lib/stdlib/src/base64.erl108
-rw-r--r--lib/stdlib/src/c.erl18
-rw-r--r--lib/stdlib/src/ets.erl2
-rw-r--r--lib/stdlib/src/filename.erl17
-rw-r--r--lib/stdlib/src/rand.erl261
-rw-r--r--lib/stdlib/src/stdlib.app.src1
-rw-r--r--lib/stdlib/src/supervisor.erl946
-rw-r--r--lib/stdlib/src/uri_string.erl1842
9 files changed, 2664 insertions, 532 deletions
diff --git a/lib/stdlib/src/Makefile b/lib/stdlib/src/Makefile
index bf836203ec..8b156929d7 100644
--- a/lib/stdlib/src/Makefile
+++ b/lib/stdlib/src/Makefile
@@ -121,6 +121,7 @@ MODULES= \
timer \
unicode \
unicode_util \
+ uri_string \
win32reg \
zip
diff --git a/lib/stdlib/src/base64.erl b/lib/stdlib/src/base64.erl
index 5885745fb1..c8cf6fdffe 100644
--- a/lib/stdlib/src/base64.erl
+++ b/lib/stdlib/src/base64.erl
@@ -113,9 +113,9 @@ encode_binary(Bin) ->
Data :: ascii_binary().
decode(Bin) when is_binary(Bin) ->
- decode_binary(<<>>, Bin);
+ decode_binary(Bin, <<>>);
decode(List) when is_list(List) ->
- list_to_binary(decode_l(List)).
+ decode_list(List, <<>>).
-spec mime_decode(Base64) -> Data when
Base64 :: ascii_string() | ascii_binary(),
@@ -186,31 +186,41 @@ mime_decode_to_string(List) when is_list(List) ->
bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,
bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad}).
-decode_binary(Result0, <<C:8,T0/bits>>) ->
- case element(C, ?DECODE_MAP) of
- bad ->
- erlang:error({badarg,C});
- ws ->
- decode_binary(Result0, T0);
- eq ->
- case strip_ws(T0) of
- <<$=:8,T/binary>> ->
- <<>> = strip_ws(T),
- Split = byte_size(Result0) - 1,
- <<Result:Split/bytes,_:4>> = Result0,
- Result;
- T ->
- <<>> = strip_ws(T),
- Split = byte_size(Result0) - 1,
- <<Result:Split/bytes,_:2>> = Result0,
- Result
- end;
- Bits ->
- decode_binary(<<Result0/bits,Bits:6>>, T0)
+decode_binary(<<C1:8, Cs/bits>>, A) ->
+ case element(C1, ?DECODE_MAP) of
+ ws -> decode_binary(Cs, A);
+ B1 -> decode_binary(Cs, A, B1)
end;
-decode_binary(Result, <<>>) ->
- true = is_binary(Result),
- Result.
+decode_binary(<<>>, A) ->
+ A.
+
+decode_binary(<<C2:8, Cs/bits>>, A, B1) ->
+ case element(C2, ?DECODE_MAP) of
+ ws -> decode_binary(Cs, A, B1);
+ B2 -> decode_binary(Cs, A, B1, B2)
+ end.
+
+decode_binary(<<C3:8, Cs/bits>>, A, B1, B2) ->
+ case element(C3, ?DECODE_MAP) of
+ ws -> decode_binary(Cs, A, B1, B2);
+ B3 -> decode_binary(Cs, A, B1, B2, B3)
+ end.
+
+decode_binary(<<C4:8, Cs/bits>>, A, B1, B2, B3) ->
+ case element(C4, ?DECODE_MAP) of
+ ws -> decode_binary(Cs, A, B1, B2, B3);
+ eq when B3 =:= eq -> only_ws_binary(Cs, <<A/binary,B1:6,(B2 bsr 4):2>>);
+ eq -> only_ws_binary(Cs, <<A/binary,B1:6,B2:6,(B3 bsr 2):4>>);
+ B4 -> decode_binary(Cs, <<A/binary,B1:6,B2:6,B3:6,B4:6>>)
+ end.
+
+only_ws_binary(<<>>, A) ->
+ A;
+only_ws_binary(<<C:8, Cs/bits>>, A) ->
+ case element(C, ?DECODE_MAP) of
+ ws -> only_ws_binary(Cs, A);
+ _ -> erlang:error(function_clause)
+ end.
%% Skipping pad character if not at end of string. Also liberal about
%% excess padding and skipping of other illegal (non-base64 alphabet)
@@ -262,6 +272,42 @@ mime_decode_binary_after_eq(Result0, <<>>, Eq) ->
Result
end.
+decode_list([C1 | Cs], A) ->
+ case element(C1, ?DECODE_MAP) of
+ ws -> decode_list(Cs, A);
+ B1 -> decode_list(Cs, A, B1)
+ end;
+decode_list([], A) ->
+ A.
+
+decode_list([C2 | Cs], A, B1) ->
+ case element(C2, ?DECODE_MAP) of
+ ws -> decode_list(Cs, A, B1);
+ B2 -> decode_list(Cs, A, B1, B2)
+ end.
+
+decode_list([C3 | Cs], A, B1, B2) ->
+ case element(C3, ?DECODE_MAP) of
+ ws -> decode_list(Cs, A, B1, B2);
+ B3 -> decode_list(Cs, A, B1, B2, B3)
+ end.
+
+decode_list([C4 | Cs], A, B1, B2, B3) ->
+ case element(C4, ?DECODE_MAP) of
+ ws -> decode_list(Cs, A, B1, B2, B3);
+ eq when B3 =:= eq -> only_ws(Cs, <<A/binary,B1:6,(B2 bsr 4):2>>);
+ eq -> only_ws(Cs, <<A/binary,B1:6,B2:6,(B3 bsr 2):4>>);
+ B4 -> decode_list(Cs, <<A/binary,B1:6,B2:6,B3:6,B4:6>>)
+ end.
+
+only_ws([], A) ->
+ A;
+only_ws([C | Cs], A) ->
+ case element(C, ?DECODE_MAP) of
+ ws -> only_ws(Cs, A);
+ _ -> erlang:error(function_clause)
+ end.
+
decode([], A) -> A;
decode([$=,$=,C2,C1|Cs], A) ->
Bits2x6 = (b64d(C1) bsl 18) bor (b64d(C2) bsl 12),
@@ -292,16 +338,6 @@ strip_spaces([$\r|Cs], A) -> strip_spaces(Cs, A);
strip_spaces([$\n|Cs], A) -> strip_spaces(Cs, A);
strip_spaces([C|Cs], A) -> strip_spaces(Cs, [C | A]).
-strip_ws(<<$\t,T/binary>>) ->
- strip_ws(T);
-strip_ws(<<$\n,T/binary>>) ->
- strip_ws(T);
-strip_ws(<<$\r,T/binary>>) ->
- strip_ws(T);
-strip_ws(<<$\s,T/binary>>) ->
- strip_ws(T);
-strip_ws(T) -> T.
-
%% Skipping pad character if not at end of string. Also liberal about
%% excess padding and skipping of other illegal (non-base64 alphabet)
%% characters. See section 3.3 of RFC4648
diff --git a/lib/stdlib/src/c.erl b/lib/stdlib/src/c.erl
index c04a201ce1..9a447af5b7 100644
--- a/lib/stdlib/src/c.erl
+++ b/lib/stdlib/src/c.erl
@@ -668,19 +668,23 @@ lm() ->
[l(M) || M <- mm()].
%% erlangrc(Home)
-%% Try to run a ".erlang" file, first in the current directory
-%% else in home directory.
+%% Try to run a ".erlang" file in home directory.
+
+-spec erlangrc() -> {ok, file:filename()} | {error, term()}.
erlangrc() ->
case init:get_argument(home) of
{ok,[[Home]]} ->
erlangrc([Home]);
_ ->
- f_p_e(["."], ".erlang")
+ {error, enoent}
end.
-erlangrc([Home]) ->
- f_p_e([".",Home], ".erlang").
+-spec erlangrc(PathList) -> {ok, file:filename()} | {error, term()}
+ when PathList :: [Dir :: file:name()].
+
+erlangrc([Home|_]=Paths) when is_list(Home) ->
+ f_p_e(Paths, ".erlang").
error(Fmt, Args) ->
error_logger:error_msg(Fmt, Args).
@@ -692,11 +696,11 @@ f_p_e(P, F) ->
{error, E={Line, _Mod, _Term}} ->
error("file:path_eval(~tp,~tp): error on line ~p: ~ts~n",
[P, F, Line, file:format_error(E)]),
- ok;
+ {error, E};
{error, E} ->
error("file:path_eval(~tp,~tp): ~ts~n",
[P, F, file:format_error(E)]),
- ok;
+ {error, E};
Other ->
Other
end.
diff --git a/lib/stdlib/src/ets.erl b/lib/stdlib/src/ets.erl
index 4858c8d13c..b6548626f3 100644
--- a/lib/stdlib/src/ets.erl
+++ b/lib/stdlib/src/ets.erl
@@ -1700,6 +1700,8 @@ choice(Height, Width, P, Mode, Tab, Key, Turn, Opos) ->
io:format("~ts\n", [ErrorString]),
choice(Height, Width, P, Mode, Tab, Key, Turn, Opos)
end;
+ eof ->
+ ok;
_ ->
choice(Height, Width, P, Mode, Tab, Key, Turn, Opos)
end.
diff --git a/lib/stdlib/src/filename.erl b/lib/stdlib/src/filename.erl
index 919f8f20e6..a322bd002d 100644
--- a/lib/stdlib/src/filename.erl
+++ b/lib/stdlib/src/filename.erl
@@ -472,6 +472,10 @@ join(Name1, Name2) when is_atom(Name2) ->
join1([UcLetter, $:|Rest], RelativeName, [], win32)
when is_integer(UcLetter), UcLetter >= $A, UcLetter =< $Z ->
join1(Rest, RelativeName, [$:, UcLetter+$a-$A], win32);
+join1([$\\,$\\|Rest], RelativeName, [], win32) ->
+ join1([$/,$/|Rest], RelativeName, [], win32);
+join1([$/,$/|Rest], RelativeName, [], win32) ->
+ join1(Rest, RelativeName, [$/,$/], win32);
join1([$\\|Rest], RelativeName, Result, win32) ->
join1([$/|Rest], RelativeName, Result, win32);
join1([$/|Rest], RelativeName, [$., $/|Result], OsType) ->
@@ -500,6 +504,10 @@ join1([Atom|Rest], RelativeName, Result, OsType) when is_atom(Atom) ->
join1b(<<UcLetter, $:, Rest/binary>>, RelativeName, [], win32)
when is_integer(UcLetter), UcLetter >= $A, UcLetter =< $Z ->
join1b(Rest, RelativeName, [$:, UcLetter+$a-$A], win32);
+join1b(<<$\\,$\\,Rest/binary>>, RelativeName, [], win32) ->
+ join1b(<<$/,$/,Rest/binary>>, RelativeName, [], win32);
+join1b(<<$/,$/,Rest/binary>>, RelativeName, [], win32) ->
+ join1b(Rest, RelativeName, [$/,$/], win32);
join1b(<<$\\,Rest/binary>>, RelativeName, Result, win32) ->
join1b(<<$/,Rest/binary>>, RelativeName, Result, win32);
join1b(<<$/,Rest/binary>>, RelativeName, [$., $/|Result], OsType) ->
@@ -510,6 +518,8 @@ join1b(<<>>, <<>>, Result, OsType) ->
list_to_binary(maybe_remove_dirsep(Result, OsType));
join1b(<<>>, RelativeName, [$:|Rest], win32) ->
join1b(RelativeName, <<>>, [$:|Rest], win32);
+join1b(<<>>, RelativeName, [$/,$/|Result], win32) ->
+ join1b(RelativeName, <<>>, [$/,$/|Result], win32);
join1b(<<>>, RelativeName, [$/|Result], OsType) ->
join1b(RelativeName, <<>>, [$/|Result], OsType);
join1b(<<>>, RelativeName, [$., $/|Result], OsType) ->
@@ -523,6 +533,8 @@ maybe_remove_dirsep([$/, $:, Letter], win32) ->
[Letter, $:, $/];
maybe_remove_dirsep([$/], _) ->
[$/];
+maybe_remove_dirsep([$/,$/], win32) ->
+ [$/,$/];
maybe_remove_dirsep([$/|Name], _) ->
lists:reverse(Name);
maybe_remove_dirsep(Name, _) ->
@@ -712,6 +724,9 @@ win32_splitb(<<Letter0,$:,Rest/binary>>) when ?IS_DRIVELETTER(Letter0) ->
Letter = fix_driveletter(Letter0),
L = binary:split(Rest,[<<"/">>,<<"\\">>],[global]),
[<<Letter,$:>> | [ X || X <- L, X =/= <<>> ]];
+win32_splitb(<<Slash,Slash,Rest/binary>>) when ((Slash =:= $\\) orelse (Slash =:= $/)) ->
+ L = binary:split(Rest,[<<"/">>,<<"\\">>],[global]),
+ [<<"//">> | [ X || X <- L, X =/= <<>> ]];
win32_splitb(<<Slash,Rest/binary>>) when ((Slash =:= $\\) orelse (Slash =:= $/)) ->
L = binary:split(Rest,[<<"/">>,<<"\\">>],[global]),
[<<$/>> | [ X || X <- L, X =/= <<>> ]];
@@ -723,6 +738,8 @@ win32_splitb(Name) ->
unix_split(Name) ->
split(Name, [], unix).
+win32_split([Slash,Slash|Rest]) when ((Slash =:= $\\) orelse (Slash =:= $/)) ->
+ split(Rest, [[$/,$/]], win32);
win32_split([$\\|Rest]) ->
win32_split([$/|Rest]);
win32_split([X, $\\|Rest]) when is_integer(X) ->
diff --git a/lib/stdlib/src/rand.erl b/lib/stdlib/src/rand.erl
index 7a8a5e6d4a..362e98006e 100644
--- a/lib/stdlib/src/rand.erl
+++ b/lib/stdlib/src/rand.erl
@@ -21,8 +21,8 @@
%% Multiple PRNG module for Erlang/OTP
%% Copyright (c) 2015-2016 Kenji Rikitake
%%
-%% exrop (xoroshiro116+) added and statistical distribution
-%% improvements by the Erlang/OTP team 2017
+%% exrop (xoroshiro116+) added, statistical distribution
+%% improvements and uniform_real added by the Erlang/OTP team 2017
%% =====================================================================
-module(rand).
@@ -30,10 +30,14 @@
-export([seed_s/1, seed_s/2, seed/1, seed/2,
export_seed/0, export_seed_s/1,
uniform/0, uniform/1, uniform_s/1, uniform_s/2,
+ uniform_real/0, uniform_real_s/1,
jump/0, jump/1,
normal/0, normal/2, normal_s/1, normal_s/3
]).
+%% Debug
+-export([make_float/3, float2str/1, bc64/1]).
+
-compile({inline, [exs64_next/1, exsplus_next/1,
exs1024_next/1, exs1024_calc/2,
exrop_next/1, exrop_next_s/2,
@@ -60,6 +64,10 @@
%% N i evaluated 3 times
(?BSL((Bits), (X), (N)) bor ((X) bsr ((Bits)-(N))))).
+-define(
+ BC(V, N),
+ bc((V), ?BIT((N) - 1), N)).
+
%%-define(TWO_POW_MINUS53, (math:pow(2, -53))).
-define(TWO_POW_MINUS53, 1.11022302462515657e-16).
@@ -84,14 +92,21 @@
%% The 'bits' field indicates how many bits the integer
%% returned from 'next' has got, i.e 'next' shall return
%% an random integer in the range 0..(2^Bits - 1).
-%% At least 53 bits is required for the floating point
-%% producing fallbacks. This field is only used when
-%% the 'uniform' or 'uniform_n' fields are not defined.
+%% At least 55 bits is required for the floating point
+%% producing fallbacks, but 56 bits would be more future proof.
%%
%% The fields 'next', 'uniform' and 'uniform_n'
-%% implement the algorithm. If 'uniform' or 'uinform_n'
+%% implement the algorithm. If 'uniform' or 'uniform_n'
%% is not present there is a fallback using 'next' and either
-%% 'bits' or the deprecated 'max'.
+%% 'bits' or the deprecated 'max'. The 'next' function
+%% must generate a word with at least 56 good random bits.
+%%
+%% The 'weak_low_bits' field indicate how many bits are of
+%% lesser quality and they will not be used by the floating point
+%% producing functions, nor by the range producing functions
+%% when more bits are needed, to avoid weak bits in the middle
+%% of the generated bits. The lowest bits from the range
+%% functions still have the generator's quality.
%%
-type alg_handler() ::
#{type := alg(),
@@ -148,11 +163,7 @@
%% For ranges larger than the algorithm bit size
uniform_range(Range, #{next:=Next, bits:=Bits} = Alg, R, V) ->
- WeakLowBits =
- case Alg of
- #{weak_low_bits:=WLB} -> WLB;
- #{} -> 0
- end,
+ WeakLowBits = maps:get(weak_low_bits, Alg, 0),
%% Maybe waste the lowest bit(s) when shifting in new bits
Shift = Bits - WeakLowBits,
ShiftMask = bnot ?MASK(WeakLowBits),
@@ -297,7 +308,7 @@ uniform_s({#{bits:=Bits, next:=Next} = Alg, R0}) ->
{(V bsr (Bits - 53)) * ?TWO_POW_MINUS53, {Alg, R1}};
uniform_s({#{max:=Max, next:=Next} = Alg, R0}) ->
{V, R1} = Next(R0),
- %% Old broken algorithm with non-uniform density
+ %% Old algorithm with non-uniform density
{V / (Max + 1), {Alg, R1}}.
@@ -317,7 +328,7 @@ uniform_s(N, {#{bits:=Bits, next:=Next} = Alg, R0})
?uniform_range(N, Alg, R1, V, MaxMinusN, I);
uniform_s(N, {#{max:=Max, next:=Next} = Alg, R0})
when is_integer(N), 1 =< N ->
- %% Old broken algorithm with skewed probability
+ %% Old algorithm with skewed probability
%% and gap in ranges > Max
{V, R1} = Next(R0),
if
@@ -328,6 +339,189 @@ uniform_s(N, {#{max:=Max, next:=Next} = Alg, R0})
{trunc(F * N) + 1, {Alg, R1}}
end.
+%% uniform_real/0: returns a random float X where 0.0 < X =< 1.0,
+%% updating the state in the process dictionary.
+
+-spec uniform_real() -> X :: float().
+uniform_real() ->
+ {X, Seed} = uniform_real_s(seed_get()),
+ _ = seed_put(Seed),
+ X.
+
+%% uniform_real_s/1: given a state, uniform_s/1
+%% returns a random float X where 0.0 < X =< 1.0,
+%% and a new state.
+%%
+%% This function does not use the same form of uniformity
+%% as the uniform_s/1 function.
+%%
+%% Instead, this function does not generate numbers with equal
+%% distance in the interval, but rather tries to keep all mantissa
+%% bits random also for small numbers, meaning that the distance
+%% between possible numbers decreases when the numbers
+%% approaches 0.0, as does the possibility for a particular
+%% number. Hence uniformity is preserved.
+%%
+%% To generate 56 bits at the time instead of 53 is actually
+%% a speed optimization since the probability to have to
+%% generate a second word decreases by 1/2 for every extra bit.
+%%
+%% This function generates normalized numbers, so the smallest number
+%% that can be generated is 2^-1022 with the distance 2^-1074
+%% to the next to smallest number, compared to 2^-53 for uniform_s/1.
+%%
+%% This concept of uniformity should work better for applications
+%% where you need to calculate 1.0/X or math:log(X) since those
+%% operations benefits from larger precision approaching 0.0,
+%% and that this function does not return 0.0 nor denormalized
+%% numbers very close to 0.0. The log() operation in The Box-Muller
+%% transformation for normal distribution is an example of this.
+%%
+%%-define(TWO_POW_MINUS55, (math:pow(2, -55))).
+%%-define(TWO_POW_MINUS110, (math:pow(2, -110))).
+%%-define(TWO_POW_MINUS55, 2.7755575615628914e-17).
+%%-define(TWO_POW_MINUS110, 7.7037197775489436e-34).
+%%
+-spec uniform_real_s(State :: state()) -> {X :: float(), NewState :: state()}.
+uniform_real_s({#{bits:=Bits, next:=Next} = Alg, R0}) ->
+ %% Generate a 56 bit number without using the weak low bits.
+ %%
+ %% Be sure to use only 53 bits when multiplying with
+ %% math:pow(2.0, -N) to avoid rounding which would make
+ %% "even" floats more probable than "odd".
+ %%
+ {V1, R1} = Next(R0),
+ M1 = V1 bsr (Bits - 56),
+ if
+ ?BIT(55) =< M1 ->
+ %% We have 56 bits - waste 3
+ {(M1 bsr 3) * math:pow(2.0, -53), {Alg, R1}};
+ ?BIT(54) =< M1 ->
+ %% We have 55 bits - waste 2
+ {(M1 bsr 2) * math:pow(2.0, -54), {Alg, R1}};
+ ?BIT(53) =< M1 ->
+ %% We have 54 bits - waste 1
+ {(M1 bsr 1) * math:pow(2.0, -55), {Alg, R1}};
+ ?BIT(52) =< M1 ->
+ %% We have 53 bits - use all
+ {M1 * math:pow(2.0, -56), {Alg, R1}};
+ true ->
+ %% Need more bits
+ {V2, R2} = Next(R1),
+ uniform_real_s(Alg, Next, M1, -56, R2, V2, Bits)
+ end;
+uniform_real_s({#{max:=_, next:=Next} = Alg, R0}) ->
+ %% Generate a 56 bit number.
+ %% Ignore the weak low bits for these old algorithms,
+ %% just produce something reasonable.
+ %%
+ %% Be sure to use only 53 bits when multiplying with
+ %% math:pow(2.0, -N) to avoid rounding which would make
+ %% "even" floats more probable than "odd".
+ %%
+ {V1, R1} = Next(R0),
+ M1 = ?MASK(56, V1),
+ if
+ ?BIT(55) =< M1 ->
+ %% We have 56 bits - waste 3
+ {(M1 bsr 3) * math:pow(2.0, -53), {Alg, R1}};
+ ?BIT(54) =< M1 ->
+ %% We have 55 bits - waste 2
+ {(M1 bsr 2) * math:pow(2.0, -54), {Alg, R1}};
+ ?BIT(53) =< M1 ->
+ %% We have 54 bits - waste 1
+ {(M1 bsr 1) * math:pow(2.0, -55), {Alg, R1}};
+ ?BIT(52) =< M1 ->
+ %% We have 53 bits - use all
+ {M1 * math:pow(2.0, -56), {Alg, R1}};
+ true ->
+ %% Need more bits
+ {V2, R2} = Next(R1),
+ uniform_real_s(Alg, Next, M1, -56, R2, V2, 56)
+ end.
+
+uniform_real_s(Alg, _Next, M0, -1064, R1, V1, Bits) -> % 19*56
+ %% This is a very theoretical bottom case.
+ %% The odds of getting here is about 2^-1008,
+ %% through a white box test case, or thanks to
+ %% a malfunctioning PRNG producing 18 56-bit zeros in a row.
+ %%
+ %% Fill up to 53 bits, we have at most 52
+ B0 = (53 - ?BC(M0, 52)), % Missing bits
+ {(((M0 bsl B0) bor (V1 bsr (Bits - B0))) * math:pow(2.0, -1064 - B0)),
+ {Alg, R1}};
+uniform_real_s(Alg, Next, M0, BitNo, R1, V1, Bits) ->
+ if
+ %% Optimize the most probable.
+ %% Fill up to 53 bits.
+ ?BIT(51) =< M0 ->
+ %% We have 52 bits in M0 - need 1
+ {(((M0 bsl 1) bor (V1 bsr (Bits - 1)))
+ * math:pow(2.0, BitNo - 1)),
+ {Alg, R1}};
+ ?BIT(50) =< M0 ->
+ %% We have 51 bits in M0 - need 2
+ {(((M0 bsl 2) bor (V1 bsr (Bits - 2)))
+ * math:pow(2.0, BitNo - 2)),
+ {Alg, R1}};
+ ?BIT(49) =< M0 ->
+ %% We have 50 bits in M0 - need 3
+ {(((M0 bsl 3) bor (V1 bsr (Bits - 3)))
+ * math:pow(2.0, BitNo - 3)),
+ {Alg, R1}};
+ M0 == 0 ->
+ M1 = V1 bsr (Bits - 56),
+ if
+ ?BIT(55) =< M1 ->
+ %% We have 56 bits - waste 3
+ {(M1 bsr 3) * math:pow(2.0, BitNo - 53), {Alg, R1}};
+ ?BIT(54) =< M1 ->
+ %% We have 55 bits - waste 2
+ {(M1 bsr 2) * math:pow(2.0, BitNo - 54), {Alg, R1}};
+ ?BIT(53) =< M1 ->
+ %% We have 54 bits - waste 1
+ {(M1 bsr 1) * math:pow(2.0, BitNo - 55), {Alg, R1}};
+ ?BIT(52) =< M1 ->
+ %% We have 53 bits - use all
+ {M1 * math:pow(2.0, BitNo - 56), {Alg, R1}};
+ BitNo =:= -1008 ->
+ %% Endgame
+ %% For the last round we can not have 14 zeros or more
+ %% at the top of M1 because then we will underflow,
+ %% so we need at least 43 bits
+ if
+ ?BIT(42) =< M1 ->
+ %% We have 43 bits - get the last bits
+ uniform_real_s(Alg, Next, M1, BitNo - 56, R1);
+ true ->
+ %% Would underflow 2^-1022 - start all over
+ %%
+ %% We could just crash here since the odds for
+ %% the PRNG being broken is much higher than
+ %% for a good PRNG generating this many zeros
+ %% in a row. Maybe we should write an error
+ %% report or call this a system limit...?
+ uniform_real_s({Alg, R1})
+ end;
+ true ->
+ %% Need more bits
+ uniform_real_s(Alg, Next, M1, BitNo - 56, R1)
+ end;
+ true ->
+ %% Fill up to 53 bits
+ B0 = 53 - ?BC(M0, 49), % Number of bits we need to append
+ {(((M0 bsl B0) bor (V1 bsr (Bits - B0)))
+ * math:pow(2.0, BitNo - B0)),
+ {Alg, R1}}
+ end.
+%%
+uniform_real_s(#{bits:=Bits} = Alg, Next, M0, BitNo, R0) ->
+ {V1, R1} = Next(R0),
+ uniform_real_s(Alg, Next, M0, BitNo, R1, V1, Bits);
+uniform_real_s(#{max:=_} = Alg, Next, M0, BitNo, R0) ->
+ {V1, R1} = Next(R0),
+ uniform_real_s(Alg, Next, M0, BitNo, R1, ?MASK(56, V1), 56).
+
%% jump/1: given a state, jump/1
%% returns a new state which is equivalent to that
%% after a large number of call defined for each algorithm.
@@ -1025,3 +1219,42 @@ normal_fi(Indx) ->
1.0214971439701471e-02,8.6165827693987316e-03,7.0508754713732268e-03,
5.5224032992509968e-03,4.0379725933630305e-03,2.6090727461021627e-03,
1.2602859304985975e-03}).
+
+%%%bitcount64(0) -> 0;
+%%%bitcount64(V) -> 1 + bitcount(V, 64).
+%%%
+%%%-define(
+%%% BITCOUNT(V, N),
+%%% bitcount(V, N) ->
+%%% if
+%%% (1 bsl ((N) bsr 1)) =< (V) ->
+%%% ((N) bsr 1) + bitcount((V) bsr ((N) bsr 1), ((N) bsr 1));
+%%% true ->
+%%% bitcount((V), ((N) bsr 1))
+%%% end).
+%%%?BITCOUNT(V, 64);
+%%%?BITCOUNT(V, 32);
+%%%?BITCOUNT(V, 16);
+%%%?BITCOUNT(V, 8);
+%%%?BITCOUNT(V, 4);
+%%%?BITCOUNT(V, 2);
+%%%bitcount(_, 1) -> 0.
+
+bc64(V) -> ?BC(V, 64).
+
+%% Linear from high bit - higher probability first gives faster execution
+bc(V, B, N) when B =< V -> N;
+bc(V, B, N) -> bc(V, B bsr 1, N - 1).
+
+make_float(S, E, M) ->
+ <<F/float>> = <<S:1, E:11, M:52>>,
+ F.
+
+float2str(N) ->
+ <<S:1, E:11, M:52>> = <<(float(N))/float>>,
+ lists:flatten(
+ io_lib:format(
+ "~c~c.~13.16.0bE~b",
+ [case S of 1 -> $-; 0 -> $+ end,
+ case E of 0 -> $0; _ -> $1 end,
+ M, E - 16#3ff])).
diff --git a/lib/stdlib/src/stdlib.app.src b/lib/stdlib/src/stdlib.app.src
index ab0824ca17..5fb48acfab 100644
--- a/lib/stdlib/src/stdlib.app.src
+++ b/lib/stdlib/src/stdlib.app.src
@@ -101,6 +101,7 @@
timer,
unicode,
unicode_util,
+ uri_string,
win32reg,
zip]},
{registered,[timer_server,rsh_starter,take_over_monitor,pool_master,
diff --git a/lib/stdlib/src/supervisor.erl b/lib/stdlib/src/supervisor.erl
index 7920e55930..e56415650f 100644
--- a/lib/stdlib/src/supervisor.erl
+++ b/lib/stdlib/src/supervisor.erl
@@ -31,7 +31,6 @@
%% Internal exports
-export([init/1, handle_call/3, handle_cast/2, handle_info/2,
terminate/2, code_change/3, format_status/2]).
--export([try_again_restart/2]).
%% For release_handler only
-export([get_callback_module/1]).
@@ -79,6 +78,7 @@
| {RestartStrategy :: strategy(),
Intensity :: non_neg_integer(),
Period :: pos_integer()}.
+-type children() :: {Ids :: [child_id()], Db :: #{child_id() => child_rec()}}.
%%--------------------------------------------------------------------------
%% Defaults
@@ -96,7 +96,7 @@
pid = undefined :: child()
| {restarting, pid() | undefined}
| [pid()],
- name :: child_id(),
+ id :: child_id(),
mfargs :: mfargs(),
restart_type :: restart(),
shutdown :: shutdown(),
@@ -104,16 +104,11 @@
modules = [] :: modules()}).
-type child_rec() :: #child{}.
--define(DICTS, dict).
--define(DICT, dict:dict).
--define(SETS, sets).
--define(SET, sets:set).
-
-record(state, {name,
strategy :: strategy() | 'undefined',
- children = [] :: [child_rec()],
- dynamics :: {'dict', ?DICT(pid(), list())}
- | {'set', ?SET(pid())}
+ children = {[],#{}} :: children(), % Ids in start order
+ dynamics :: {'maps', #{pid() => list()}}
+ | {'sets', sets:set(pid())}
| 'undefined',
intensity :: non_neg_integer() | 'undefined',
period :: pos_integer() | 'undefined',
@@ -124,6 +119,9 @@
-type state() :: #state{}.
-define(is_simple(State), State#state.strategy =:= simple_one_for_one).
+-define(is_temporary(_Child_), _Child_#child.restart_type=:=temporary).
+-define(is_transient(_Child_), _Child_#child.restart_type=:=transient).
+-define(is_permanent(_Child_), _Child_#child.restart_type=:=permanent).
-callback init(Args :: term()) ->
{ok, {SupFlags :: sup_flags(), [ChildSpec :: child_spec()]}}
@@ -179,16 +177,16 @@ start_child(Supervisor, ChildSpec) ->
| {'error', Error},
Error :: 'running' | 'restarting' | 'not_found' | 'simple_one_for_one' |
term().
-restart_child(Supervisor, Name) ->
- call(Supervisor, {restart_child, Name}).
+restart_child(Supervisor, Id) ->
+ call(Supervisor, {restart_child, Id}).
-spec delete_child(SupRef, Id) -> Result when
SupRef :: sup_ref(),
Id :: child_id(),
Result :: 'ok' | {'error', Error},
Error :: 'running' | 'restarting' | 'not_found' | 'simple_one_for_one'.
-delete_child(Supervisor, Name) ->
- call(Supervisor, {delete_child, Name}).
+delete_child(Supervisor, Id) ->
+ call(Supervisor, {delete_child, Id}).
%%-----------------------------------------------------------------
%% Func: terminate_child/2
@@ -202,16 +200,16 @@ delete_child(Supervisor, Name) ->
Id :: pid() | child_id(),
Result :: 'ok' | {'error', Error},
Error :: 'not_found' | 'simple_one_for_one'.
-terminate_child(Supervisor, Name) ->
- call(Supervisor, {terminate_child, Name}).
+terminate_child(Supervisor, Id) ->
+ call(Supervisor, {terminate_child, Id}).
-spec get_childspec(SupRef, Id) -> Result when
SupRef :: sup_ref(),
Id :: pid() | child_id(),
Result :: {'ok', child_spec()} | {'error', Error},
Error :: 'not_found'.
-get_childspec(Supervisor, Name) ->
- call(Supervisor, {get_childspec, Name}).
+get_childspec(Supervisor, Id) ->
+ call(Supervisor, {get_childspec, Id}).
-spec which_children(SupRef) -> [{Id,Child,Type,Modules}] when
SupRef :: sup_ref(),
@@ -246,17 +244,6 @@ check_childspecs(ChildSpecs) when is_list(ChildSpecs) ->
check_childspecs(X) -> {error, {badarg, X}}.
%%%-----------------------------------------------------------------
-%%% Called by restart/2
--spec try_again_restart(SupRef, Child) -> ok when
- SupRef :: sup_ref(),
- Child :: child_id() | pid().
-try_again_restart(Supervisor, Child) ->
- cast(Supervisor, {try_again_restart, Child}).
-
-cast(Supervisor, Req) ->
- gen_server:cast(Supervisor, Req).
-
-%%%-----------------------------------------------------------------
%%% Called by release_handler during upgrade
-spec get_callback_module(Pid) -> Module when
Pid :: pid(),
@@ -325,7 +312,7 @@ init_children(State, StartSpec) ->
init_dynamic(State, [StartSpec]) ->
case check_startspec([StartSpec]) of
{ok, Children} ->
- {ok, State#state{children = Children}};
+ {ok, dyn_init(State#state{children = Children})};
Error ->
{stop, {start_spec, Error}}
end;
@@ -334,35 +321,34 @@ init_dynamic(_State, StartSpec) ->
%%-----------------------------------------------------------------
%% Func: start_children/2
-%% Args: Children = [child_rec()] in start order
+%% Args: Children = children() % Ids in start order
%% SupName = {local, atom()} | {global, atom()} | {pid(), Mod}
-%% Purpose: Start all children. The new list contains #child's
+%% Purpose: Start all children. The new map contains #child's
%% with pids.
%% Returns: {ok, NChildren} | {error, NChildren, Reason}
-%% NChildren = [child_rec()] in termination order (reversed
-%% start order)
+%% NChildren = children() % Ids in termination order
+%% (reversed start order)
%%-----------------------------------------------------------------
-start_children(Children, SupName) -> start_children(Children, [], SupName).
-
-start_children([Child|Chs], NChildren, SupName) ->
- case do_start_child(SupName, Child) of
- {ok, undefined} when Child#child.restart_type =:= temporary ->
- start_children(Chs, NChildren, SupName);
- {ok, Pid} ->
- start_children(Chs, [Child#child{pid = Pid}|NChildren], SupName);
- {ok, Pid, _Extra} ->
- start_children(Chs, [Child#child{pid = Pid}|NChildren], SupName);
- {error, Reason} ->
- report_error(start_error, Reason, Child, SupName),
- {error, lists:reverse(Chs) ++ [Child | NChildren],
- {failed_to_start_child,Child#child.name,Reason}}
- end;
-start_children([], NChildren, _SupName) ->
- {ok, NChildren}.
+start_children(Children, SupName) ->
+ Start =
+ fun(Id,Child) ->
+ case do_start_child(SupName, Child) of
+ {ok, undefined} when ?is_temporary(Child) ->
+ remove;
+ {ok, Pid} ->
+ {update,Child#child{pid = Pid}};
+ {ok, Pid, _Extra} ->
+ {update,Child#child{pid = Pid}};
+ {error, Reason} ->
+ report_error(start_error, Reason, Child, SupName),
+ {abort,{failed_to_start_child,Id,Reason}}
+ end
+ end,
+ children_map(Start,Children).
do_start_child(SupName, Child) ->
#child{mfargs = {M, F, Args}} = Child,
- case catch apply(M, F, Args) of
+ case do_start_child_i(M, F, Args) of
{ok, Pid} when is_pid(Pid) ->
NChild = Child#child{pid = Pid},
report_progress(NChild, SupName),
@@ -371,10 +357,8 @@ do_start_child(SupName, Child) ->
NChild = Child#child{pid = Pid},
report_progress(NChild, SupName),
{ok, Pid, Extra};
- ignore ->
- {ok, undefined};
- {error, What} -> {error, What};
- What -> {error, What}
+ Other ->
+ Other
end.
do_start_child_i(M, F, A) ->
@@ -400,17 +384,17 @@ do_start_child_i(M, F, A) ->
-spec handle_call(call(), term(), state()) -> {'reply', term(), state()}.
handle_call({start_child, EArgs}, _From, State) when ?is_simple(State) ->
- Child = hd(State#state.children),
+ Child = get_dynamic_child(State),
#child{mfargs = {M, F, A}} = Child,
Args = A ++ EArgs,
case do_start_child_i(M, F, Args) of
{ok, undefined} ->
{reply, {ok, undefined}, State};
{ok, Pid} ->
- NState = save_dynamic_child(Child#child.restart_type, Pid, Args, State),
+ NState = dyn_store(Pid, Args, State),
{reply, {ok, Pid}, NState};
{ok, Pid, Extra} ->
- NState = save_dynamic_child(Child#child.restart_type, Pid, Args, State),
+ NState = dyn_store(Pid, Args, State),
{reply, {ok, Pid, Extra}, NState};
What ->
{reply, What, State}
@@ -426,121 +410,94 @@ handle_call({start_child, ChildSpec}, _From, State) ->
end;
%% terminate_child for simple_one_for_one can only be done with pid
-handle_call({terminate_child, Name}, _From, State) when not is_pid(Name),
- ?is_simple(State) ->
+handle_call({terminate_child, Id}, _From, State) when not is_pid(Id),
+ ?is_simple(State) ->
{reply, {error, simple_one_for_one}, State};
-handle_call({terminate_child, Name}, _From, State) ->
- case get_child(Name, State, ?is_simple(State)) of
- {value, Child} ->
- case do_terminate(Child, State#state.name) of
- #child{restart_type=RT} when RT=:=temporary; ?is_simple(State) ->
- {reply, ok, state_del_child(Child, State)};
- NChild ->
- {reply, ok, replace_child(NChild, State)}
- end;
- false ->
+handle_call({terminate_child, Id}, _From, State) ->
+ case find_child(Id, State) of
+ {ok, Child} ->
+ do_terminate(Child, State#state.name),
+ {reply, ok, del_child(Child, State)};
+ error ->
{reply, {error, not_found}, State}
end;
%% restart_child request is invalid for simple_one_for_one supervisors
-handle_call({restart_child, _Name}, _From, State) when ?is_simple(State) ->
+handle_call({restart_child, _Id}, _From, State) when ?is_simple(State) ->
{reply, {error, simple_one_for_one}, State};
-handle_call({restart_child, Name}, _From, State) ->
- case get_child(Name, State) of
- {value, Child} when Child#child.pid =:= undefined ->
+handle_call({restart_child, Id}, _From, State) ->
+ case find_child(Id, State) of
+ {ok, Child} when Child#child.pid =:= undefined ->
case do_start_child(State#state.name, Child) of
{ok, Pid} ->
- NState = replace_child(Child#child{pid = Pid}, State),
+ NState = set_pid(Pid, Id, State),
{reply, {ok, Pid}, NState};
{ok, Pid, Extra} ->
- NState = replace_child(Child#child{pid = Pid}, State),
+ NState = set_pid(Pid, Id, State),
{reply, {ok, Pid, Extra}, NState};
Error ->
{reply, Error, State}
end;
- {value, #child{pid=?restarting(_)}} ->
+ {ok, #child{pid=?restarting(_)}} ->
{reply, {error, restarting}, State};
- {value, _} ->
+ {ok, _} ->
{reply, {error, running}, State};
_ ->
{reply, {error, not_found}, State}
end;
%% delete_child request is invalid for simple_one_for_one supervisors
-handle_call({delete_child, _Name}, _From, State) when ?is_simple(State) ->
+handle_call({delete_child, _Id}, _From, State) when ?is_simple(State) ->
{reply, {error, simple_one_for_one}, State};
-handle_call({delete_child, Name}, _From, State) ->
- case get_child(Name, State) of
- {value, Child} when Child#child.pid =:= undefined ->
- NState = remove_child(Child, State),
+handle_call({delete_child, Id}, _From, State) ->
+ case find_child(Id, State) of
+ {ok, Child} when Child#child.pid =:= undefined ->
+ NState = remove_child(Id, State),
{reply, ok, NState};
- {value, #child{pid=?restarting(_)}} ->
+ {ok, #child{pid=?restarting(_)}} ->
{reply, {error, restarting}, State};
- {value, _} ->
+ {ok, _} ->
{reply, {error, running}, State};
_ ->
{reply, {error, not_found}, State}
end;
-handle_call({get_childspec, Name}, _From, State) ->
- case get_child(Name, State, ?is_simple(State)) of
- {value, Child} ->
+handle_call({get_childspec, Id}, _From, State) ->
+ case find_child(Id, State) of
+ {ok, Child} ->
{reply, {ok, child_to_spec(Child)}, State};
- false ->
+ error ->
{reply, {error, not_found}, State}
end;
-handle_call(which_children, _From, #state{children = [#child{restart_type = temporary,
- child_type = CT,
- modules = Mods}]} =
- State) when ?is_simple(State) ->
- Reply = lists:map(fun(Pid) -> {undefined, Pid, CT, Mods} end,
- ?SETS:to_list(dynamics_db(temporary, State#state.dynamics))),
- {reply, Reply, State};
-
-handle_call(which_children, _From, #state{children = [#child{restart_type = RType,
- child_type = CT,
- modules = Mods}]} =
- State) when ?is_simple(State) ->
- Reply = lists:map(fun({?restarting(_),_}) -> {undefined,restarting,CT,Mods};
- ({Pid, _}) -> {undefined, Pid, CT, Mods} end,
- ?DICTS:to_list(dynamics_db(RType, State#state.dynamics))),
+handle_call(which_children, _From, State) when ?is_simple(State) ->
+ #child{child_type = CT,modules = Mods} = get_dynamic_child(State),
+ Reply = dyn_map(fun(?restarting(_)) -> {undefined, restarting, CT, Mods};
+ (Pid) -> {undefined, Pid, CT, Mods}
+ end, State),
{reply, Reply, State};
handle_call(which_children, _From, State) ->
Resp =
- lists:map(fun(#child{pid = ?restarting(_), name = Name,
- child_type = ChildType, modules = Mods}) ->
- {Name, restarting, ChildType, Mods};
- (#child{pid = Pid, name = Name,
- child_type = ChildType, modules = Mods}) ->
- {Name, Pid, ChildType, Mods}
- end,
- State#state.children),
+ children_to_list(
+ fun(Id,#child{pid = ?restarting(_),
+ child_type = ChildType, modules = Mods}) ->
+ {Id, restarting, ChildType, Mods};
+ (Id,#child{pid = Pid,
+ child_type = ChildType, modules = Mods}) ->
+ {Id, Pid, ChildType, Mods}
+ end,
+ State#state.children),
{reply, Resp, State};
-
-handle_call(count_children, _From, #state{children = [#child{restart_type = temporary,
- child_type = CT}]} = State)
- when ?is_simple(State) ->
- Sz = ?SETS:size(dynamics_db(temporary, State#state.dynamics)),
- Reply = case CT of
- supervisor -> [{specs, 1}, {active, Sz},
- {supervisors, Sz}, {workers, 0}];
- worker -> [{specs, 1}, {active, Sz},
- {supervisors, 0}, {workers, Sz}]
- end,
- {reply, Reply, State};
-
-handle_call(count_children, _From, #state{dynamic_restarts = Restarts,
- children = [#child{restart_type = RType,
- child_type = CT}]} = State)
+handle_call(count_children, _From, #state{dynamic_restarts = Restarts} = State)
when ?is_simple(State) ->
- Sz = ?DICTS:size(dynamics_db(RType, State#state.dynamics)),
- Active = Sz - Restarts,
+ #child{child_type = CT} = get_dynamic_child(State),
+ Sz = dyn_size(State),
+ Active = Sz - Restarts, % Restarts is always 0 for temporary children
Reply = case CT of
supervisor -> [{specs, 1}, {active, Active},
{supervisors, Sz}, {workers, 0}];
@@ -552,16 +509,15 @@ handle_call(count_children, _From, #state{dynamic_restarts = Restarts,
handle_call(count_children, _From, State) ->
%% Specs and children are together on the children list...
{Specs, Active, Supers, Workers} =
- lists:foldl(fun(Child, Counts) ->
- count_child(Child, Counts)
- end, {0,0,0,0}, State#state.children),
+ children_fold(fun(_Id, Child, Counts) ->
+ count_child(Child, Counts)
+ end, {0,0,0,0}, State#state.children),
%% Reformat counts to a property list.
Reply = [{specs, Specs}, {active, Active},
{supervisors, Supers}, {workers, Workers}],
{reply, Reply, State}.
-
count_child(#child{pid = Pid, child_type = worker},
{Specs, Active, Supers, Workers}) ->
case is_pid(Pid) andalso is_process_alive(Pid) of
@@ -575,34 +531,15 @@ count_child(#child{pid = Pid, child_type = supervisor},
false -> {Specs+1, Active, Supers+1, Workers}
end.
-
%%% If a restart attempt failed, this message is cast
%%% from restart/2 in order to give gen_server the chance to
%%% check it's inbox before trying again.
--spec handle_cast({try_again_restart, child_id() | pid()}, state()) ->
+-spec handle_cast({try_again_restart, child_id() | {'restarting',pid()}}, state()) ->
{'noreply', state()} | {stop, shutdown, state()}.
-handle_cast({try_again_restart,Pid}, #state{children=[Child]}=State)
- when ?is_simple(State) ->
- RT = Child#child.restart_type,
- RPid = restarting(Pid),
- case dynamic_child_args(RPid, RT, State#state.dynamics) of
- {ok, Args} ->
- {M, F, _} = Child#child.mfargs,
- NChild = Child#child{pid = RPid, mfargs = {M, F, Args}},
- case restart(NChild,State) of
- {ok, State1} ->
- {noreply, State1};
- {shutdown, State1} ->
- {stop, shutdown, State1}
- end;
- error ->
- {noreply, State}
- end;
-
-handle_cast({try_again_restart,Name}, State) ->
- case lists:keyfind(Name,#child.name,State#state.children) of
- Child = #child{pid=?restarting(_)} ->
+handle_cast({try_again_restart,TryAgainId}, State) ->
+ case find_child_and_args(TryAgainId, State) of
+ {ok, Child = #child{pid=?restarting(_)}} ->
case restart(Child,State) of
{ok, State1} ->
{noreply, State1};
@@ -637,10 +574,8 @@ handle_info(Msg, State) ->
%%
-spec terminate(term(), state()) -> 'ok'.
-terminate(_Reason, #state{children=[Child]} = State) when ?is_simple(State) ->
- terminate_dynamic_children(Child, dynamics_db(Child#child.restart_type,
- State#state.dynamics),
- State#state.name);
+terminate(_Reason, State) when ?is_simple(State) ->
+ terminate_dynamic_children(State);
terminate(_Reason, State) ->
terminate_children(State#state.children, State#state.name).
@@ -675,8 +610,8 @@ code_change(_, State, _) ->
update_childspec(State, StartSpec) when ?is_simple(State) ->
case check_startspec(StartSpec) of
- {ok, [Child]} ->
- {ok, State#state{children = [Child]}};
+ {ok, {[_],_}=Children} ->
+ {ok, State#state{children = Children}};
Error ->
{error, Error}
end;
@@ -690,39 +625,36 @@ update_childspec(State, StartSpec) ->
{error, Error}
end.
-update_childspec1([Child|OldC], Children, KeepOld) ->
- case update_chsp(Child, Children) of
- {ok,NewChildren} ->
- update_childspec1(OldC, NewChildren, KeepOld);
+update_childspec1({[Id|OldIds], OldDb}, {Ids,Db}, KeepOld) ->
+ case update_chsp(maps:get(Id,OldDb), Db) of
+ {ok,NewDb} ->
+ update_childspec1({OldIds,OldDb}, {Ids,NewDb}, KeepOld);
false ->
- update_childspec1(OldC, Children, [Child|KeepOld])
+ update_childspec1({OldIds,OldDb}, {Ids,Db}, [Id|KeepOld])
end;
-update_childspec1([], Children, KeepOld) ->
+update_childspec1({[],OldDb}, {Ids,Db}, KeepOld) ->
+ KeepOldDb = maps:with(KeepOld,OldDb),
%% Return them in (kept) reverse start order.
- lists:reverse(Children ++ KeepOld).
-
-update_chsp(OldCh, Children) ->
- case lists:map(fun(Ch) when OldCh#child.name =:= Ch#child.name ->
- Ch#child{pid = OldCh#child.pid};
- (Ch) ->
- Ch
- end,
- Children) of
- Children ->
- false; % OldCh not found in new spec.
- NewC ->
- {ok, NewC}
+ {lists:reverse(Ids ++ KeepOld),maps:merge(KeepOldDb,Db)}.
+
+update_chsp(#child{id=Id}=OldChild, NewDb) ->
+ case maps:find(Id, NewDb) of
+ {ok,Child} ->
+ {ok,NewDb#{Id => Child#child{pid = OldChild#child.pid}}};
+ error -> % Id not found in new spec.
+ false
end.
+
%%% ---------------------------------------------------
%%% Start a new child.
%%% ---------------------------------------------------
handle_start_child(Child, State) ->
- case get_child(Child#child.name, State) of
- false ->
+ case find_child(Child#child.id, State) of
+ error ->
case do_start_child(State#state.name, Child) of
- {ok, undefined} when Child#child.restart_type =:= temporary ->
+ {ok, undefined} when ?is_temporary(Child) ->
{{ok, undefined}, State};
{ok, Pid} ->
{{ok, Pid}, save_child(Child#child{pid = Pid}, State)};
@@ -731,9 +663,9 @@ handle_start_child(Child, State) ->
{error, What} ->
{{error, {What, Child}}, State}
end;
- {value, OldChild} when is_pid(OldChild#child.pid) ->
+ {ok, OldChild} when is_pid(OldChild#child.pid) ->
{{error, {already_started, OldChild#child.pid}}, State};
- {value, _OldChild} ->
+ {ok, _OldChild} ->
{{error, already_present}, State}
end.
@@ -742,63 +674,45 @@ handle_start_child(Child, State) ->
%%% Returns: {ok, state()} | {shutdown, state()}
%%% ---------------------------------------------------
-restart_child(Pid, Reason, #state{children = [Child]} = State) when ?is_simple(State) ->
- RestartType = Child#child.restart_type,
- case dynamic_child_args(Pid, RestartType, State#state.dynamics) of
- {ok, Args} ->
- {M, F, _} = Child#child.mfargs,
- NChild = Child#child{pid = Pid, mfargs = {M, F, Args}},
- do_restart(RestartType, Reason, NChild, State);
- error ->
- {ok, State}
- end;
-
restart_child(Pid, Reason, State) ->
- Children = State#state.children,
- case lists:keyfind(Pid, #child.pid, Children) of
- #child{restart_type = RestartType} = Child ->
- do_restart(RestartType, Reason, Child, State);
- false ->
+ case find_child_and_args(Pid, State) of
+ {ok, Child} ->
+ do_restart(Reason, Child, State);
+ error ->
{ok, State}
end.
-do_restart(permanent, Reason, Child, State) ->
+do_restart(Reason, Child, State) when ?is_permanent(Child) ->
report_error(child_terminated, Reason, Child, State#state.name),
restart(Child, State);
-do_restart(_, normal, Child, State) ->
- NState = state_del_child(Child, State),
+do_restart(normal, Child, State) ->
+ NState = del_child(Child, State),
{ok, NState};
-do_restart(_, shutdown, Child, State) ->
- NState = state_del_child(Child, State),
+do_restart(shutdown, Child, State) ->
+ NState = del_child(Child, State),
{ok, NState};
-do_restart(_, {shutdown, _Term}, Child, State) ->
- NState = state_del_child(Child, State),
+do_restart({shutdown, _Term}, Child, State) ->
+ NState = del_child(Child, State),
{ok, NState};
-do_restart(transient, Reason, Child, State) ->
+do_restart(Reason, Child, State) when ?is_transient(Child) ->
report_error(child_terminated, Reason, Child, State#state.name),
restart(Child, State);
-do_restart(temporary, Reason, Child, State) ->
+do_restart(Reason, Child, State) when ?is_temporary(Child) ->
report_error(child_terminated, Reason, Child, State#state.name),
- NState = state_del_child(Child, State),
+ NState = del_child(Child, State),
{ok, NState}.
restart(Child, State) ->
case add_restart(State) of
{ok, NState} ->
case restart(NState#state.strategy, Child, NState) of
- {try_again,NState2} ->
+ {{try_again, TryAgainId}, NState2} ->
%% Leaving control back to gen_server before
%% trying again. This way other incoming requsts
%% for the supervisor can be handled - e.g. a
%% shutdown request for the supervisor or the
%% child.
- Id = if ?is_simple(State) -> Child#child.pid;
- true -> Child#child.name
- end,
- ok = try_again_restart(self(), Id),
- {ok,NState2};
- {try_again, NState2, #child{name=ChName}} ->
- ok = try_again_restart(self(), ChName),
+ try_again_restart(TryAgainId),
{ok,NState2};
Other ->
Other
@@ -806,124 +720,111 @@ restart(Child, State) ->
{terminate, NState} ->
report_error(shutdown, reached_max_restart_intensity,
Child, State#state.name),
- {shutdown, remove_child(Child, NState)}
+ {shutdown, del_child(Child, NState)}
end.
restart(simple_one_for_one, Child, State0) ->
#child{pid = OldPid, mfargs = {M, F, A}} = Child,
- State = case OldPid of
+ State1 = case OldPid of
?restarting(_) ->
NRes = State0#state.dynamic_restarts - 1,
State0#state{dynamic_restarts = NRes};
_ ->
State0
end,
- Dynamics = ?DICTS:erase(OldPid, dynamics_db(Child#child.restart_type,
- State#state.dynamics)),
+ State2 = dyn_erase(OldPid, State1),
case do_start_child_i(M, F, A) of
{ok, Pid} ->
- DynamicsDb = {dict, ?DICTS:store(Pid, A, Dynamics)},
- NState = State#state{dynamics = DynamicsDb},
+ NState = dyn_store(Pid, A, State2),
{ok, NState};
{ok, Pid, _Extra} ->
- DynamicsDb = {dict, ?DICTS:store(Pid, A, Dynamics)},
- NState = State#state{dynamics = DynamicsDb},
+ NState = dyn_store(Pid, A, State2),
{ok, NState};
{error, Error} ->
- NRestarts = State#state.dynamic_restarts + 1,
- DynamicsDb = {dict, ?DICTS:store(restarting(OldPid), A, Dynamics)},
- NState = State#state{dynamic_restarts = NRestarts,
- dynamics = DynamicsDb},
- report_error(start_error, Error, Child, State#state.name),
- {try_again, NState}
+ ROldPid = restarting(OldPid),
+ NRestarts = State2#state.dynamic_restarts + 1,
+ State3 = State2#state{dynamic_restarts = NRestarts},
+ NState = dyn_store(ROldPid, A, State3),
+ report_error(start_error, Error, Child, NState#state.name),
+ {{try_again, ROldPid}, NState}
end;
-restart(one_for_one, Child, State) ->
+restart(one_for_one, #child{id=Id} = Child, State) ->
OldPid = Child#child.pid,
case do_start_child(State#state.name, Child) of
{ok, Pid} ->
- NState = replace_child(Child#child{pid = Pid}, State),
+ NState = set_pid(Pid, Id, State),
{ok, NState};
{ok, Pid, _Extra} ->
- NState = replace_child(Child#child{pid = Pid}, State),
+ NState = set_pid(Pid, Id, State),
{ok, NState};
{error, Reason} ->
- NState = replace_child(Child#child{pid = restarting(OldPid)}, State),
+ NState = set_pid(restarting(OldPid), Id, State),
report_error(start_error, Reason, Child, State#state.name),
- {try_again, NState}
- end;
-restart(rest_for_one, Child, State) ->
- {ChAfter, ChBefore} = split_child(Child#child.pid, State#state.children),
- ChAfter2 = terminate_children(ChAfter, State#state.name),
- case start_children(ChAfter2, State#state.name) of
- {ok, ChAfter3} ->
- {ok, State#state{children = ChAfter3 ++ ChBefore}};
- {error, ChAfter3, {failed_to_start_child, ChName, _Reason}}
- when ChName =:= Child#child.name ->
- NChild = Child#child{pid=restarting(Child#child.pid)},
- NState = State#state{children = ChAfter3 ++ ChBefore},
- {try_again, replace_child(NChild,NState)};
- {error, ChAfter3, {failed_to_start_child, ChName, _Reason}} ->
- NChild = lists:keyfind(ChName, #child.name, ChAfter3),
- NChild2 = NChild#child{pid=?restarting(undefined)},
- NState = State#state{children = ChAfter3 ++ ChBefore},
- {try_again, replace_child(NChild2,NState), NChild2}
+ {{try_again,Id}, NState}
end;
-restart(one_for_all, Child, State) ->
- Children1 = del_child(Child#child.pid, State#state.children),
- Children2 = terminate_children(Children1, State#state.name),
- case start_children(Children2, State#state.name) of
- {ok, NChs} ->
- {ok, State#state{children = NChs}};
- {error, NChs, {failed_to_start_child, ChName, _Reason}}
- when ChName =:= Child#child.name ->
- NChild = Child#child{pid=restarting(Child#child.pid)},
- NState = State#state{children = NChs},
- {try_again, replace_child(NChild,NState)};
- {error, NChs, {failed_to_start_child, ChName, _Reason}} ->
- NChild = lists:keyfind(ChName, #child.name, NChs),
- NChild2 = NChild#child{pid=?restarting(undefined)},
- NState = State#state{children = NChs},
- {try_again, replace_child(NChild2,NState), NChild2}
+restart(rest_for_one, #child{id=Id} = Child, #state{name=SupName} = State) ->
+ {ChAfter, ChBefore} = split_child(Id, State#state.children),
+ {Return, ChAfter2} = restart_multiple_children(Child, ChAfter, SupName),
+ {Return, State#state{children = append(ChAfter2,ChBefore)}};
+restart(one_for_all, Child, #state{name=SupName} = State) ->
+ Children1 = del_child(Child#child.id, State#state.children),
+ {Return, NChildren} = restart_multiple_children(Child, Children1, SupName),
+ {Return, State#state{children = NChildren}}.
+
+restart_multiple_children(Child, Children, SupName) ->
+ Children1 = terminate_children(Children, SupName),
+ case start_children(Children1, SupName) of
+ {ok, NChildren} ->
+ {ok, NChildren};
+ {error, NChildren, {failed_to_start_child, FailedId, _Reason}} ->
+ NewPid = if FailedId =:= Child#child.id ->
+ restarting(Child#child.pid);
+ true ->
+ ?restarting(undefined)
+ end,
+ {{try_again, FailedId}, set_pid(NewPid,FailedId,NChildren)}
end.
restarting(Pid) when is_pid(Pid) -> ?restarting(Pid);
restarting(RPid) -> RPid.
+-spec try_again_restart(child_id() | {'restarting',pid()}) -> 'ok'.
+try_again_restart(TryAgainId) ->
+ gen_server:cast(self(), {try_again_restart, TryAgainId}).
+
%%-----------------------------------------------------------------
%% Func: terminate_children/2
-%% Args: Children = [child_rec()] in termination order
+%% Args: Children = children() % Ids in termination order
%% SupName = {local, atom()} | {global, atom()} | {pid(),Mod}
-%% Returns: NChildren = [child_rec()] in
-%% startup order (reversed termination order)
+%% Returns: NChildren = children() % Ids in startup order
+%% % (reversed termination order)
%%-----------------------------------------------------------------
terminate_children(Children, SupName) ->
- terminate_children(Children, SupName, []).
-
-%% Temporary children should not be restarted and thus should
-%% be skipped when building the list of terminated children, although
-%% we do want them to be shut down as many functions from this module
-%% use this function to just clear everything.
-terminate_children([Child = #child{restart_type=temporary} | Children], SupName, Res) ->
- _ = do_terminate(Child, SupName),
- terminate_children(Children, SupName, Res);
-terminate_children([Child | Children], SupName, Res) ->
- NChild = do_terminate(Child, SupName),
- terminate_children(Children, SupName, [NChild | Res]);
-terminate_children([], _SupName, Res) ->
- Res.
+ Terminate =
+ fun(_Id,Child) when ?is_temporary(Child) ->
+ %% Temporary children should not be restarted and thus should
+ %% be skipped when building the list of terminated children.
+ do_terminate(Child, SupName),
+ remove;
+ (_Id,Child) ->
+ do_terminate(Child, SupName),
+ {update,Child#child{pid=undefined}}
+ end,
+ {ok,NChildren} = children_map(Terminate, Children),
+ NChildren.
do_terminate(Child, SupName) when is_pid(Child#child.pid) ->
case shutdown(Child#child.pid, Child#child.shutdown) of
ok ->
ok;
- {error, normal} when Child#child.restart_type =/= permanent ->
+ {error, normal} when not (?is_permanent(Child)) ->
ok;
{error, OtherReason} ->
report_error(shutdown_error, OtherReason, Child, SupName)
end,
- Child#child{pid = undefined};
-do_terminate(Child, _SupName) ->
- Child#child{pid = undefined}.
+ ok;
+do_terminate(_Child, _SupName) ->
+ ok.
%%-----------------------------------------------------------------
%% Shutdowns a child. We must check the EXIT value
@@ -996,66 +897,50 @@ monitor_child(Pid) ->
ok
end.
-
%%-----------------------------------------------------------------
-%% Func: terminate_dynamic_children/3
-%% Args: Child = child_rec()
-%% Dynamics = ?DICT() | ?SET()
-%% SupName = {local, atom()} | {global, atom()} | {pid(),Mod}
+%% Func: terminate_dynamic_children/1
+%% Args: State
%% Returns: ok
%%
-%%
%% Shutdown all dynamic children. This happens when the supervisor is
%% stopped. Because the supervisor can have millions of dynamic children, we
-%% can have an significative overhead here.
+%% can have a significative overhead here.
%%-----------------------------------------------------------------
-terminate_dynamic_children(Child, Dynamics, SupName) ->
- {Pids, EStack0} = monitor_dynamic_children(Child, Dynamics),
- Sz = ?SETS:size(Pids),
+terminate_dynamic_children(State) ->
+ Child = get_dynamic_child(State),
+ {Pids, EStack0} = monitor_dynamic_children(Child,State),
+ Sz = sets:size(Pids),
EStack = case Child#child.shutdown of
brutal_kill ->
- ?SETS:fold(fun(P, _) -> exit(P, kill) end, ok, Pids),
+ sets:fold(fun(P, _) -> exit(P, kill) end, ok, Pids),
wait_dynamic_children(Child, Pids, Sz, undefined, EStack0);
infinity ->
- ?SETS:fold(fun(P, _) -> exit(P, shutdown) end, ok, Pids),
+ sets:fold(fun(P, _) -> exit(P, shutdown) end, ok, Pids),
wait_dynamic_children(Child, Pids, Sz, undefined, EStack0);
Time ->
- ?SETS:fold(fun(P, _) -> exit(P, shutdown) end, ok, Pids),
+ sets:fold(fun(P, _) -> exit(P, shutdown) end, ok, Pids),
TRef = erlang:start_timer(Time, self(), kill),
wait_dynamic_children(Child, Pids, Sz, TRef, EStack0)
end,
%% Unroll stacked errors and report them
- ?DICTS:fold(fun(Reason, Ls, _) ->
- report_error(shutdown_error, Reason,
- Child#child{pid=Ls}, SupName)
- end, ok, EStack).
-
-
-monitor_dynamic_children(#child{restart_type=temporary}, Dynamics) ->
- ?SETS:fold(fun(P, {Pids, EStack}) ->
- case monitor_child(P) of
- ok ->
- {?SETS:add_element(P, Pids), EStack};
- {error, normal} ->
- {Pids, EStack};
- {error, Reason} ->
- {Pids, ?DICTS:append(Reason, P, EStack)}
- end
- end, {?SETS:new(), ?DICTS:new()}, Dynamics);
-monitor_dynamic_children(#child{restart_type=RType}, Dynamics) ->
- ?DICTS:fold(fun(P, _, {Pids, EStack}) when is_pid(P) ->
- case monitor_child(P) of
- ok ->
- {?SETS:add_element(P, Pids), EStack};
- {error, normal} when RType =/= permanent ->
- {Pids, EStack};
- {error, Reason} ->
- {Pids, ?DICTS:append(Reason, P, EStack)}
- end;
- (?restarting(_), _, {Pids, EStack}) ->
- {Pids, EStack}
- end, {?SETS:new(), ?DICTS:new()}, Dynamics).
-
+ dict:fold(fun(Reason, Ls, _) ->
+ report_error(shutdown_error, Reason,
+ Child#child{pid=Ls}, State#state.name)
+ end, ok, EStack).
+
+monitor_dynamic_children(Child,State) ->
+ dyn_fold(fun(P,{Pids, EStack}) when is_pid(P) ->
+ case monitor_child(P) of
+ ok ->
+ {sets:add_element(P, Pids), EStack};
+ {error, normal} when not (?is_permanent(Child)) ->
+ {Pids, EStack};
+ {error, Reason} ->
+ {Pids, dict:append(Reason, P, EStack)}
+ end;
+ (?restarting(_), {Pids, EStack}) ->
+ {Pids, EStack}
+ end, {sets:new(), dict:new()}, State).
wait_dynamic_children(_Child, _Pids, 0, undefined, EStack) ->
EStack;
@@ -1073,39 +958,38 @@ wait_dynamic_children(#child{shutdown=brutal_kill} = Child, Pids, Sz,
TRef, EStack) ->
receive
{'DOWN', _MRef, process, Pid, killed} ->
- wait_dynamic_children(Child, ?SETS:del_element(Pid, Pids), Sz-1,
+ wait_dynamic_children(Child, sets:del_element(Pid, Pids), Sz-1,
TRef, EStack);
{'DOWN', _MRef, process, Pid, Reason} ->
- wait_dynamic_children(Child, ?SETS:del_element(Pid, Pids), Sz-1,
- TRef, ?DICTS:append(Reason, Pid, EStack))
+ wait_dynamic_children(Child, sets:del_element(Pid, Pids), Sz-1,
+ TRef, dict:append(Reason, Pid, EStack))
end;
-wait_dynamic_children(#child{restart_type=RType} = Child, Pids, Sz,
- TRef, EStack) ->
+wait_dynamic_children(Child, Pids, Sz, TRef, EStack) ->
receive
{'DOWN', _MRef, process, Pid, shutdown} ->
- wait_dynamic_children(Child, ?SETS:del_element(Pid, Pids), Sz-1,
+ wait_dynamic_children(Child, sets:del_element(Pid, Pids), Sz-1,
TRef, EStack);
{'DOWN', _MRef, process, Pid, {shutdown, _}} ->
- wait_dynamic_children(Child, ?SETS:del_element(Pid, Pids), Sz-1,
+ wait_dynamic_children(Child, sets:del_element(Pid, Pids), Sz-1,
TRef, EStack);
- {'DOWN', _MRef, process, Pid, normal} when RType =/= permanent ->
- wait_dynamic_children(Child, ?SETS:del_element(Pid, Pids), Sz-1,
+ {'DOWN', _MRef, process, Pid, normal} when not (?is_permanent(Child)) ->
+ wait_dynamic_children(Child, sets:del_element(Pid, Pids), Sz-1,
TRef, EStack);
{'DOWN', _MRef, process, Pid, Reason} ->
- wait_dynamic_children(Child, ?SETS:del_element(Pid, Pids), Sz-1,
- TRef, ?DICTS:append(Reason, Pid, EStack));
+ wait_dynamic_children(Child, sets:del_element(Pid, Pids), Sz-1,
+ TRef, dict:append(Reason, Pid, EStack));
{timeout, TRef, kill} ->
- ?SETS:fold(fun(P, _) -> exit(P, kill) end, ok, Pids),
+ sets:fold(fun(P, _) -> exit(P, kill) end, ok, Pids),
wait_dynamic_children(Child, Pids, Sz, undefined, EStack)
end.
%%-----------------------------------------------------------------
-%% Child/State manipulating functions.
+%% Access #state.children
%%-----------------------------------------------------------------
%% Note we do not want to save the parameter list for temporary processes as
@@ -1113,114 +997,184 @@ wait_dynamic_children(#child{restart_type=RType} = Child, Pids, Sz,
%% Especially for dynamic children to simple_one_for_one supervisors
%% it could become very costly as it is not uncommon to spawn
%% very many such processes.
-save_child(#child{restart_type = temporary,
- mfargs = {M, F, _}} = Child, #state{children = Children} = State) ->
- State#state{children = [Child#child{mfargs = {M, F, undefined}} |Children]};
-save_child(Child, #state{children = Children} = State) ->
- State#state{children = [Child |Children]}.
-
-save_dynamic_child(temporary, Pid, _, #state{dynamics = Dynamics} = State) ->
- DynamicsDb = dynamics_db(temporary, Dynamics),
- State#state{dynamics = {set, ?SETS:add_element(Pid, DynamicsDb)}};
-save_dynamic_child(RestartType, Pid, Args, #state{dynamics = Dynamics} = State) ->
- DynamicsDb = dynamics_db(RestartType, Dynamics),
- State#state{dynamics = {dict, ?DICTS:store(Pid, Args, DynamicsDb)}}.
-
-dynamics_db(temporary, undefined) ->
- ?SETS:new();
-dynamics_db(_, undefined) ->
- ?DICTS:new();
-dynamics_db(_, {_Tag, DynamicsDb}) ->
- DynamicsDb.
-
-dynamic_child_args(_Pid, temporary, _DynamicsDb) ->
- {ok, undefined};
-dynamic_child_args(Pid, _RT, {dict, DynamicsDb}) ->
- ?DICTS:find(Pid, DynamicsDb);
-dynamic_child_args(_Pid, _RT, undefined) ->
- error.
-
-state_del_child(#child{pid = Pid, restart_type = temporary}, State) when ?is_simple(State) ->
- NDynamics = ?SETS:del_element(Pid, dynamics_db(temporary, State#state.dynamics)),
- State#state{dynamics = {set, NDynamics}};
-state_del_child(#child{pid = Pid, restart_type = RType}, State) when ?is_simple(State) ->
- NDynamics = ?DICTS:erase(Pid, dynamics_db(RType, State#state.dynamics)),
- State#state{dynamics = {dict, NDynamics}};
-state_del_child(Child, State) ->
- NChildren = del_child(Child#child.name, State#state.children),
- State#state{children = NChildren}.
-
-del_child(Name, [Ch|Chs]) when Ch#child.name =:= Name, Ch#child.restart_type =:= temporary ->
- Chs;
-del_child(Name, [Ch|Chs]) when Ch#child.name =:= Name ->
- [Ch#child{pid = undefined} | Chs];
-del_child(Pid, [Ch|Chs]) when Ch#child.pid =:= Pid, Ch#child.restart_type =:= temporary ->
- Chs;
-del_child(Pid, [Ch|Chs]) when Ch#child.pid =:= Pid ->
- [Ch#child{pid = undefined} | Chs];
-del_child(Name, [Ch|Chs]) ->
- [Ch|del_child(Name, Chs)];
-del_child(_, []) ->
- [].
+-spec save_child(child_rec(), state()) -> state().
+save_child(#child{mfargs = {M, F, _}} = Child, State) when ?is_temporary(Child) ->
+ do_save_child(Child#child{mfargs = {M, F, undefined}}, State);
+save_child(Child, State) ->
+ do_save_child(Child, State).
+
+-spec do_save_child(child_rec(), state()) -> state().
+do_save_child(#child{id = Id} = Child, #state{children = {Ids,Db}} = State) ->
+ State#state{children = {[Id|Ids],Db#{Id => Child}}}.
+
+-spec del_child(child_rec(), state()) -> state();
+ (child_id(), children()) -> children().
+del_child(#child{pid = Pid}, State) when ?is_simple(State) ->
+ dyn_erase(Pid,State);
+del_child(Child, State) when is_record(Child,child), is_record(State,state) ->
+ NChildren = del_child(Child#child.id, State#state.children),
+ State#state{children = NChildren};
+del_child(Id, {Ids,Db}) ->
+ case maps:get(Id, Db) of
+ Child when Child#child.restart_type =:= temporary ->
+ {lists:delete(Id, Ids), maps:remove(Id, Db)};
+ Child ->
+ {Ids, Db#{Id=>Child#child{pid=undefined}}}
+ end.
-%% Chs = [S4, S3, Ch, S1, S0]
-%% Ret: {[S4, S3, Ch], [S1, S0]}
-split_child(Name, Chs) ->
- split_child(Name, Chs, []).
-
-split_child(Name, [Ch|Chs], After) when Ch#child.name =:= Name ->
- {lists:reverse([Ch#child{pid = undefined} | After]), Chs};
-split_child(Pid, [Ch|Chs], After) when Ch#child.pid =:= Pid ->
- {lists:reverse([Ch#child{pid = undefined} | After]), Chs};
-split_child(Name, [Ch|Chs], After) ->
- split_child(Name, Chs, [Ch | After]);
-split_child(_, [], After) ->
- {lists:reverse(After), []}.
-
-get_child(Name, State) ->
- get_child(Name, State, false).
-
-get_child(Pid, State, AllowPid) when AllowPid, is_pid(Pid) ->
- get_dynamic_child(Pid, State);
-get_child(Name, State, _) ->
- lists:keysearch(Name, #child.name, State#state.children).
-
-get_dynamic_child(Pid, #state{children=[Child], dynamics=Dynamics}) ->
- case is_dynamic_pid(Pid, Dynamics) of
- true ->
- {value, Child#child{pid=Pid}};
- false ->
- RPid = restarting(Pid),
- case is_dynamic_pid(RPid, Dynamics) of
- true ->
- {value, Child#child{pid=RPid}};
- false ->
+%% In: {[S4, S3, Ch, S1, S0],Db}
+%% Ret: {{[S4, S3, Ch],Db1}, {[S1, S0],Db2}}
+%% Db1 and Db2 contain the keys in the lists they are associated with.
+-spec split_child(child_id(), children()) -> {children(), children()}.
+split_child(Id, {Ids,Db}) ->
+ {IdsAfter,IdsBefore} = split_ids(Id, Ids, []),
+ DbBefore = maps:with(IdsBefore,Db),
+ #{Id:=Ch} = DbAfter = maps:with(IdsAfter,Db),
+ {{IdsAfter,DbAfter#{Id=>Ch#child{pid=undefined}}},{IdsBefore,DbBefore}}.
+
+split_ids(Id, [Id|Ids], After) ->
+ {lists:reverse([Id|After]), Ids};
+split_ids(Id, [Other|Ids], After) ->
+ split_ids(Id, Ids, [Other | After]).
+
+%% Find the child record for a given Pid (dynamic child) or Id
+%% (non-dynamic child). This is called from the API functions.
+-spec find_child(pid() | child_id(), state()) -> {ok,child_rec()} | error.
+find_child(Pid, State) when is_pid(Pid), ?is_simple(State) ->
+ case find_dynamic_child(Pid, State) of
+ error ->
+ case find_dynamic_child(restarting(Pid), State) of
+ error ->
case erlang:is_process_alive(Pid) of
- true -> false;
- false -> {value, Child}
- end
- end
+ true -> error;
+ false -> {ok, get_dynamic_child(State)}
+ end;
+ Other ->
+ Other
+ end;
+ Other ->
+ Other
+ end;
+find_child(Id, #state{children = {_Ids,Db}}) ->
+ maps:find(Id, Db).
+
+%% Get the child record - either by child id or by pid. If
+%% simple_one_for_one, then insert the pid and args into the returned
+%% child record. This is called when trying to restart the child.
+-spec find_child_and_args(IdOrPid, state()) -> {ok, child_rec()} | error when
+ IdOrPid :: pid() | {restarting,pid()} | child_id().
+find_child_and_args(Pid, State) when ?is_simple(State) ->
+ case find_dynamic_child(Pid, State) of
+ {ok,#child{mfargs={M,F,_}} = Child} ->
+ {ok, Args} = dyn_args(Pid, State),
+ {ok, Child#child{mfargs = {M, F, Args}}};
+ error ->
+ error
+ end;
+find_child_and_args(Pid, State) when is_pid(Pid) ->
+ find_child_by_pid(Pid, State);
+find_child_and_args(Id, #state{children={_Ids,Db}}) ->
+ maps:find(Id, Db).
+
+%% Given the pid, find the child record for a dynamic child, and
+%% include the pid in the returned record.
+-spec find_dynamic_child(IdOrPid, state()) -> {ok, child_rec()} | error when
+ IdOrPid :: pid() | {restarting,pid()} | child_id().
+find_dynamic_child(Pid, State) ->
+ case dyn_exists(Pid, State) of
+ true ->
+ Child = get_dynamic_child(State),
+ {ok, Child#child{pid=Pid}};
+ false ->
+ error
end.
-is_dynamic_pid(Pid, {dict, Dynamics}) ->
- ?DICTS:is_key(Pid, Dynamics);
-is_dynamic_pid(Pid, {set, Dynamics}) ->
- ?SETS:is_element(Pid, Dynamics);
-is_dynamic_pid(_Pid, undefined) ->
- false.
-
-replace_child(Child, State) ->
- Chs = do_replace_child(Child, State#state.children),
- State#state{children = Chs}.
-
-do_replace_child(Child, [Ch|Chs]) when Ch#child.name =:= Child#child.name ->
- [Child | Chs];
-do_replace_child(Child, [Ch|Chs]) ->
- [Ch|do_replace_child(Child, Chs)].
+%% Given the pid, find the child record for a non-dyanamic child.
+-spec find_child_by_pid(IdOrPid, state()) -> {ok,child_rec()} | error when
+ IdOrPid :: pid() | {restarting,pid()}.
+find_child_by_pid(Pid,#state{children={_Ids,Db}}) ->
+ Fun = fun(_Id,#child{pid=P}=Ch,_) when P =:= Pid ->
+ throw(Ch);
+ (_,_,error) ->
+ error
+ end,
+ try maps:fold(Fun,error,Db)
+ catch throw:Child -> {ok,Child}
+ end.
-remove_child(Child, State) ->
- Chs = lists:keydelete(Child#child.name, #child.name, State#state.children),
- State#state{children = Chs}.
+%% Get the child record from a simple_one_for_one supervisor - no pid
+%% It is assumed that the child can always be found
+-spec get_dynamic_child(state()) -> child_rec().
+get_dynamic_child(#state{children={[Id],Db}}) ->
+ #{Id := Child} = Db,
+ Child.
+
+%% Update pid in the given child record and store it in the process state
+-spec set_pid(term(), child_id(), state()) -> state();
+ (term(), child_id(), children()) -> children().
+set_pid(Pid, Id, #state{children=Children} = State) ->
+ State#state{children = set_pid(Pid, Id, Children)};
+set_pid(Pid, Id, {Ids, Db}) ->
+ NewDb = maps:update_with(Id, fun(Child) -> Child#child{pid=Pid} end, Db),
+ {Ids,NewDb}.
+
+%% Remove the Id and the child record from the process state
+-spec remove_child(child_id(), state()) -> state().
+remove_child(Id, #state{children={Ids,Db}} = State) ->
+ NewIds = lists:delete(Id,Ids),
+ NewDb = maps:remove(Id,Db),
+ State#state{children = {NewIds,NewDb}}.
+
+%% In the order of Ids, traverse the children and update each child
+%% according to the return value of the Fun.
+%% On error, abort and return the merge of the old and the updated map.
+%% NOTE: The returned list of Ids is reverted compared to the input.
+-spec children_map(Fun, children()) -> {ok, children()} |
+ {error,children(),Reason} when
+ Fun :: fun((child_id(),child_rec()) -> {update,child_rec()} |
+ remove |
+ {abort, Reason}),
+ Reason :: term().
+children_map(Fun,{Ids,Db}) ->
+ children_map(Fun, Ids, Db, []).
+
+children_map(Fun,[Id|Ids],Db,Acc) ->
+ case Fun(Id,maps:get(Id,Db)) of
+ {update,Child} ->
+ children_map(Fun,Ids,Db#{Id => Child},[Id|Acc]);
+ remove ->
+ children_map(Fun,Ids,maps:remove(Id,Db),Acc);
+ {abort,Reason} ->
+ {error,{lists:reverse(Ids)++[Id|Acc],Db},Reason}
+ end;
+children_map(_Fun,[],Db,Acc) ->
+ {ok,{Acc,Db}}.
+
+%% In the order of Ids, map over all children and return the list
+-spec children_to_list(Fun, children()) -> List when
+ Fun :: fun((child_id(), child_rec()) -> Elem),
+ List :: list(Elem),
+ Elem :: term().
+children_to_list(Fun,{Ids,Db}) ->
+ children_to_list(Fun, Ids, Db, []).
+children_to_list(Fun,[Id|Ids],Db,Acc) ->
+ children_to_list(Fun,Ids,Db,[Fun(Id,maps:get(Id,Db))|Acc]);
+children_to_list(_Fun,[],_Db,Acc) ->
+ lists:reverse(Acc).
+
+%% The order is not important - so ignore Ids
+-spec children_fold(Fun, Acc0, children()) -> Acc1 when
+ Fun :: fun((child_id(), child_rec(), AccIn) -> AccOut),
+ Acc0 :: term(),
+ Acc1 :: term(),
+ AccIn :: term(),
+ AccOut :: term().
+children_fold(Fun,Init,{_Ids,Db}) ->
+ maps:fold(Fun, Init, Db).
+
+-spec append(children(), children()) -> children().
+append({Ids1,Db1},{Ids2,Db2}) ->
+ {Ids1++Ids2,maps:merge(Db1,Db2)}.
%%-----------------------------------------------------------------
%% Func: init_state/4
@@ -1290,27 +1244,27 @@ supname(N, _) -> N.
%%% Returns: {ok, [child_rec()]} | Error
%%% ------------------------------------------------------
-check_startspec(Children) -> check_startspec(Children, []).
+check_startspec(Children) -> check_startspec(Children, [], #{}).
-check_startspec([ChildSpec|T], Res) ->
+check_startspec([ChildSpec|T], Ids, Db) ->
case check_childspec(ChildSpec) of
- {ok, Child} ->
- case lists:keymember(Child#child.name, #child.name, Res) of
+ {ok, #child{id=Id}=Child} ->
+ case maps:is_key(Id, Db) of
%% The error message duplicate_child_name is kept for
%% backwards compatibility, although
%% duplicate_child_id would be more correct.
- true -> {duplicate_child_name, Child#child.name};
- false -> check_startspec(T, [Child | Res])
+ true -> {duplicate_child_name, Id};
+ false -> check_startspec(T, [Id | Ids], Db#{Id=>Child})
end;
Error -> Error
end;
-check_startspec([], Res) ->
- {ok, lists:reverse(Res)}.
+check_startspec([], Ids, Db) ->
+ {ok, {lists:reverse(Ids),Db}}.
check_childspec(ChildSpec) when is_map(ChildSpec) ->
catch do_check_childspec(maps:merge(?default_child_spec,ChildSpec));
-check_childspec({Name, Func, RestartType, Shutdown, ChildType, Mods}) ->
- check_childspec(#{id => Name,
+check_childspec({Id, Func, RestartType, Shutdown, ChildType, Mods}) ->
+ check_childspec(#{id => Id,
start => Func,
restart => RestartType,
shutdown => Shutdown,
@@ -1320,15 +1274,15 @@ check_childspec(X) -> {invalid_child_spec, X}.
do_check_childspec(#{restart := RestartType,
type := ChildType} = ChildSpec)->
- Name = case ChildSpec of
- #{id := N} -> N;
+ Id = case ChildSpec of
+ #{id := I} -> I;
_ -> throw(missing_id)
end,
Func = case ChildSpec of
#{start := F} -> F;
_ -> throw(missing_start)
end,
- validName(Name),
+ validId(Id),
validFunc(Func),
validRestartType(RestartType),
validChildType(ChildType),
@@ -1343,14 +1297,14 @@ do_check_childspec(#{restart := RestartType,
_ -> {M,_,_} = Func, [M]
end,
validMods(Mods),
- {ok, #child{name = Name, mfargs = Func, restart_type = RestartType,
+ {ok, #child{id = Id, mfargs = Func, restart_type = RestartType,
shutdown = Shutdown, child_type = ChildType, modules = Mods}}.
validChildType(supervisor) -> true;
validChildType(worker) -> true;
validChildType(What) -> throw({invalid_child_type, What}).
-validName(_Name) -> true.
+validId(_Id) -> true.
validFunc({M, F, A}) when is_atom(M),
is_atom(F),
@@ -1379,13 +1333,13 @@ validMods(Mods) when is_list(Mods) ->
Mods);
validMods(Mods) -> throw({invalid_modules, Mods}).
-child_to_spec(#child{name = Name,
+child_to_spec(#child{id = Id,
mfargs = Func,
restart_type = RestartType,
shutdown = Shutdown,
child_type = ChildType,
modules = Mods}) ->
- #{id => Name,
+ #{id => Id,
start => Func,
restart => RestartType,
shutdown => Shutdown,
@@ -1439,17 +1393,16 @@ report_error(Error, Reason, Child, SupName) ->
{offender, extract_child(Child)}],
error_logger:error_report(supervisor_report, ErrorMsg).
-
extract_child(Child) when is_list(Child#child.pid) ->
[{nb_children, length(Child#child.pid)},
- {id, Child#child.name},
+ {id, Child#child.id},
{mfargs, Child#child.mfargs},
{restart_type, Child#child.restart_type},
{shutdown, Child#child.shutdown},
{child_type, Child#child.child_type}];
extract_child(Child) ->
[{pid, Child#child.pid},
- {id, Child#child.name},
+ {id, Child#child.id},
{mfargs, Child#child.mfargs},
{restart_type, Child#child.restart_type},
{shutdown, Child#child.shutdown},
@@ -1465,3 +1418,46 @@ format_status(terminate, [_PDict, State]) ->
format_status(_, [_PDict, State]) ->
[{data, [{"State", State}]},
{supervisor, [{"Callback", State#state.module}]}].
+
+%%%-----------------------------------------------------------------
+%%% Dynamics database access
+dyn_size(#state{dynamics = {Mod,Db}}) ->
+ Mod:size(Db).
+
+dyn_erase(Pid,#state{dynamics={sets,Db}}=State) ->
+ State#state{dynamics={sets,sets:del_element(Pid,Db)}};
+dyn_erase(Pid,#state{dynamics={maps,Db}}=State) ->
+ State#state{dynamics={maps,maps:remove(Pid,Db)}}.
+
+dyn_store(Pid,_,#state{dynamics={sets,Db}}=State) ->
+ State#state{dynamics={sets,sets:add_element(Pid,Db)}};
+dyn_store(Pid,Args,#state{dynamics={maps,Db}}=State) ->
+ State#state{dynamics={maps,Db#{Pid => Args}}}.
+
+dyn_fold(Fun,Init,#state{dynamics={sets,Db}}) ->
+ sets:fold(Fun,Init,Db);
+dyn_fold(Fun,Init,#state{dynamics={maps,Db}}) ->
+ maps:fold(fun(Pid,_,Acc) -> Fun(Pid,Acc) end, Init, Db).
+
+dyn_map(Fun, #state{dynamics={sets,Db}}) ->
+ lists:map(Fun, sets:to_list(Db));
+dyn_map(Fun, #state{dynamics={maps,Db}}) ->
+ lists:map(Fun, maps:keys(Db)).
+
+dyn_exists(Pid, #state{dynamics={sets, Db}}) ->
+ sets:is_element(Pid, Db);
+dyn_exists(Pid, #state{dynamics={maps, Db}}) ->
+ maps:is_key(Pid, Db).
+
+dyn_args(_Pid, #state{dynamics={sets, _Db}}) ->
+ {ok,undefined};
+dyn_args(Pid, #state{dynamics={maps, Db}}) ->
+ maps:find(Pid, Db).
+
+dyn_init(State) ->
+ dyn_init(get_dynamic_child(State),State).
+
+dyn_init(Child,State) when ?is_temporary(Child) ->
+ State#state{dynamics={sets,sets:new()}};
+dyn_init(_Child,State) ->
+ State#state{dynamics={maps,maps:new()}}.
diff --git a/lib/stdlib/src/uri_string.erl b/lib/stdlib/src/uri_string.erl
new file mode 100644
index 0000000000..22212da222
--- /dev/null
+++ b/lib/stdlib/src/uri_string.erl
@@ -0,0 +1,1842 @@
+%%
+%% %CopyrightBegin%
+%%
+%% Copyright Ericsson AB 2017. All Rights Reserved.
+%%
+%% Licensed under the Apache License, Version 2.0 (the "License");
+%% you may not use this file except in compliance with the License.
+%% You may obtain a copy of the License at
+%%
+%% http://www.apache.org/licenses/LICENSE-2.0
+%%
+%% Unless required by applicable law or agreed to in writing, software
+%% distributed under the License is distributed on an "AS IS" BASIS,
+%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+%% See the License for the specific language governing permissions and
+%% limitations under the License.
+%%
+%% %CopyrightEnd%
+%%
+%%
+%% [RFC 3986, Chapter 2.2. Reserved Characters]
+%%
+%% reserved = gen-delims / sub-delims
+%%
+%% gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"
+%%
+%% sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
+%% / "*" / "+" / "," / ";" / "="
+%%
+%%
+%% [RFC 3986, Chapter 2.3. Unreserved Characters]
+%%
+%% unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
+%%
+%%
+%% [RFC 3986, Chapter 3. Syntax Components]
+%%
+%% The generic URI syntax consists of a hierarchical sequence of
+%% components referred to as the scheme, authority, path, query, and
+%% fragment.
+%%
+%% URI = scheme ":" hier-part [ "?" query ] [ "#" fragment ]
+%%
+%% hier-part = "//" authority path-abempty
+%% / path-absolute
+%% / path-rootless
+%% / path-empty
+%%
+%% The scheme and path components are required, though the path may be
+%% empty (no characters). When authority is present, the path must
+%% either be empty or begin with a slash ("/") character. When
+%% authority is not present, the path cannot begin with two slash
+%% characters ("//"). These restrictions result in five different ABNF
+%% rules for a path (Section 3.3), only one of which will match any
+%% given URI reference.
+%%
+%% The following are two example URIs and their component parts:
+%%
+%% foo://example.com:8042/over/there?name=ferret#nose
+%% \_/ \______________/\_________/ \_________/ \__/
+%% | | | | |
+%% scheme authority path query fragment
+%% | _____________________|__
+%% / \ / \
+%% urn:example:animal:ferret:nose
+%%
+%%
+%% [RFC 3986, Chapter 3.1. Scheme]
+%%
+%% Each URI begins with a scheme name that refers to a specification for
+%% assigning identifiers within that scheme.
+%%
+%% scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
+%%
+%%
+%% [RFC 3986, Chapter 3.2. Authority]
+%%
+%% Many URI schemes include a hierarchical element for a naming
+%% authority so that governance of the name space defined by the
+%% remainder of the URI is delegated to that authority (which may, in
+%% turn, delegate it further).
+%%
+%% authority = [ userinfo "@" ] host [ ":" port ]
+%%
+%%
+%% [RFC 3986, Chapter 3.2.1. User Information]
+%%
+%% The userinfo subcomponent may consist of a user name and, optionally,
+%% scheme-specific information about how to gain authorization to access
+%% the resource. The user information, if present, is followed by a
+%% commercial at-sign ("@") that delimits it from the host.
+%%
+%% userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
+%%
+%%
+%% [RFC 3986, Chapter 3.2.2. Host]
+%%
+%% The host subcomponent of authority is identified by an IP literal
+%% encapsulated within square brackets, an IPv4 address in dotted-
+%% decimal form, or a registered name.
+%%
+%% host = IP-literal / IPv4address / reg-name
+%%
+%% IP-literal = "[" ( IPv6address / IPvFuture ) "]"
+%%
+%% IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" )
+%%
+%% IPv6address = 6( h16 ":" ) ls32
+%% / "::" 5( h16 ":" ) ls32
+%% / [ h16 ] "::" 4( h16 ":" ) ls32
+%% / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
+%% / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
+%% / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32
+%% / [ *4( h16 ":" ) h16 ] "::" ls32
+%% / [ *5( h16 ":" ) h16 ] "::" h16
+%% / [ *6( h16 ":" ) h16 ] "::"
+%%
+%% ls32 = ( h16 ":" h16 ) / IPv4address
+%% ; least-significant 32 bits of address
+%%
+%% h16 = 1*4HEXDIG
+%% ; 16 bits of address represented in hexadecimal
+%%
+%% IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
+%%
+%% dec-octet = DIGIT ; 0-9
+%% / %x31-39 DIGIT ; 10-99
+%% / "1" 2DIGIT ; 100-199
+%% / "2" %x30-34 DIGIT ; 200-249
+%% / "25" %x30-35 ; 250-255
+%%
+%% reg-name = *( unreserved / pct-encoded / sub-delims )
+%%
+%%
+%% [RFC 3986, Chapter 3.2.2. Port]
+%%
+%% The port subcomponent of authority is designated by an optional port
+%% number in decimal following the host and delimited from it by a
+%% single colon (":") character.
+%%
+%% port = *DIGIT
+%%
+%%
+%% [RFC 3986, Chapter 3.3. Path]
+%%
+%% The path component contains data, usually organized in hierarchical
+%% form, that, along with data in the non-hierarchical query component
+%% (Section 3.4), serves to identify a resource within the scope of the
+%% URI's scheme and naming authority (if any). The path is terminated
+%% by the first question mark ("?") or number sign ("#") character, or
+%% by the end of the URI.
+%%
+%% path = path-abempty ; begins with "/" or is empty
+%% / path-absolute ; begins with "/" but not "//"
+%% / path-noscheme ; begins with a non-colon segment
+%% / path-rootless ; begins with a segment
+%% / path-empty ; zero characters
+%%
+%% path-abempty = *( "/" segment )
+%% path-absolute = "/" [ segment-nz *( "/" segment ) ]
+%% path-noscheme = segment-nz-nc *( "/" segment )
+%% path-rootless = segment-nz *( "/" segment )
+%% path-empty = 0<pchar>
+%% segment = *pchar
+%% segment-nz = 1*pchar
+%% segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
+%% ; non-zero-length segment without any colon ":"
+%%
+%% pchar = unreserved / pct-encoded / sub-delims / ":" / "@"
+%%
+%%
+%% [RFC 3986, Chapter 3.4. Query]
+%%
+%% The query component contains non-hierarchical data that, along with
+%% data in the path component (Section 3.3), serves to identify a
+%% resource within the scope of the URI's scheme and naming authority
+%% (if any). The query component is indicated by the first question
+%% mark ("?") character and terminated by a number sign ("#") character
+%% or by the end of the URI.
+%%
+%% query = *( pchar / "/" / "?" )
+%%
+%%
+%% [RFC 3986, Chapter 3.5. Fragment]
+%%
+%% The fragment identifier component of a URI allows indirect
+%% identification of a secondary resource by reference to a primary
+%% resource and additional identifying information.
+%%
+%% fragment = *( pchar / "/" / "?" )
+%%
+%%
+%% [RFC 3986, Chapter 4.1. URI Reference]
+%%
+%% URI-reference is used to denote the most common usage of a resource
+%% identifier.
+%%
+%% URI-reference = URI / relative-ref
+%%
+%%
+%% [RFC 3986, Chapter 4.2. Relative Reference]
+%%
+%% A relative reference takes advantage of the hierarchical syntax
+%% (Section 1.2.3) to express a URI reference relative to the name space
+%% of another hierarchical URI.
+%%
+%% relative-ref = relative-part [ "?" query ] [ "#" fragment ]
+%%
+%% relative-part = "//" authority path-abempty
+%% / path-absolute
+%% / path-noscheme
+%% / path-empty
+%%
+%%
+%% [RFC 3986, Chapter 4.3. Absolute URI]
+%%
+%% Some protocol elements allow only the absolute form of a URI without
+%% a fragment identifier. For example, defining a base URI for later
+%% use by relative references calls for an absolute-URI syntax rule that
+%% does not allow a fragment.
+%%
+%% absolute-URI = scheme ":" hier-part [ "?" query ]
+%%
+-module(uri_string).
+
+%%-------------------------------------------------------------------------
+%% External API
+%%-------------------------------------------------------------------------
+-export([normalize/1, parse/1,
+ recompose/1, transcode/2]).
+-export_type([error/0, uri_map/0, uri_string/0]).
+
+
+%%-------------------------------------------------------------------------
+%% Internal API
+%%-------------------------------------------------------------------------
+-export([is_host/1, is_path/1]). % suppress warnings
+
+
+%%-------------------------------------------------------------------------
+%% Macros
+%%-------------------------------------------------------------------------
+-define(CHAR(Char), <<Char/utf8>>).
+-define(STRING_EMPTY, <<>>).
+-define(STRING(MatchStr), <<MatchStr/binary>>).
+-define(STRING_REST(MatchStr, Rest), <<MatchStr/utf8, Rest/binary>>).
+
+-define(DEC2HEX(X),
+ if ((X) >= 0) andalso ((X) =< 9) -> (X) + $0;
+ ((X) >= 10) andalso ((X) =< 15) -> (X) + $A - 10
+ end).
+
+-define(HEX2DEC(X),
+ if ((X) >= $0) andalso ((X) =< $9) -> (X) - $0;
+ ((X) >= $A) andalso ((X) =< $F) -> (X) - $A + 10;
+ ((X) >= $a) andalso ((X) =< $f) -> (X) - $a + 10
+ end).
+
+
+%%%=========================================================================
+%%% API
+%%%=========================================================================
+
+%%-------------------------------------------------------------------------
+%% URI compliant with RFC 3986
+%% ASCII %x21 - %x7A ("!" - "z") except
+%% %x34 " double quote
+%% %x60 < less than
+%% %x62 > greater than
+%% %x92 \ backslash
+%% %x94 ^ caret / circumflex
+%% %x96 ` grave / accent
+%%-------------------------------------------------------------------------
+-type uri_string() :: iodata().
+-type error() :: {error, atom(), term()}.
+
+
+%%-------------------------------------------------------------------------
+%% RFC 3986, Chapter 3. Syntax Components
+%%-------------------------------------------------------------------------
+-type uri_map() ::
+ #{fragment => unicode:chardata(),
+ host => unicode:chardata(),
+ path => unicode:chardata(),
+ port => non_neg_integer() | undefined,
+ query => unicode:chardata(),
+ scheme => unicode:chardata(),
+ userinfo => unicode:chardata()} | #{}.
+
+
+%%-------------------------------------------------------------------------
+%% Normalize URIs
+%%-------------------------------------------------------------------------
+-spec normalize(URIString) -> NormalizedURI when
+ URIString :: uri_string(),
+ NormalizedURI :: uri_string().
+normalize(URIString) ->
+ %% Percent-encoding normalization and case normalization for
+ %% percent-encoded triplets are achieved by running parse and
+ %% recompose on the input URI string.
+ recompose(
+ normalize_path_segment(
+ normalize_scheme_based(
+ normalize_case(
+ parse(URIString))))).
+
+
+%%-------------------------------------------------------------------------
+%% Parse URIs
+%%-------------------------------------------------------------------------
+-spec parse(URIString) -> URIMap when
+ URIString :: uri_string(),
+ URIMap :: uri_map()
+ | error().
+parse(URIString) when is_binary(URIString) ->
+ try parse_uri_reference(URIString, #{})
+ catch
+ throw:{error, Atom, RestData} -> {error, Atom, RestData}
+ end;
+parse(URIString) when is_list(URIString) ->
+ try
+ Binary = unicode:characters_to_binary(URIString),
+ Map = parse_uri_reference(Binary, #{}),
+ convert_mapfields_to_list(Map)
+ catch
+ throw:{error, Atom, RestData} -> {error, Atom, RestData}
+ end.
+
+
+%%-------------------------------------------------------------------------
+%% Recompose URIs
+%%-------------------------------------------------------------------------
+-spec recompose(URIMap) -> URIString when
+ URIMap :: uri_map(),
+ URIString :: uri_string()
+ | error().
+recompose(Map) ->
+ case is_valid_map(Map) of
+ false ->
+ {error, invalid_map, Map};
+ true ->
+ try
+ T0 = update_scheme(Map, empty),
+ T1 = update_userinfo(Map, T0),
+ T2 = update_host(Map, T1),
+ T3 = update_port(Map, T2),
+ T4 = update_path(Map, T3),
+ T5 = update_query(Map, T4),
+ update_fragment(Map, T5)
+ catch
+ throw:{error, Atom, RestData} -> {error, Atom, RestData}
+ end
+ end.
+
+
+%%-------------------------------------------------------------------------
+%% Transcode URIs
+%%-------------------------------------------------------------------------
+-spec transcode(URIString, Options) -> Result when
+ URIString :: uri_string(),
+ Options :: [{in_encoding, unicode:encoding()}|{out_encoding, unicode:encoding()}],
+ Result :: uri_string()
+ | error().
+transcode(URIString, Options) when is_binary(URIString) ->
+ try
+ InEnc = proplists:get_value(in_encoding, Options, utf8),
+ OutEnc = proplists:get_value(out_encoding, Options, utf8),
+ List = convert_to_list(URIString, InEnc),
+ Output = transcode(List, [], InEnc, OutEnc),
+ convert_to_binary(Output, utf8, OutEnc)
+ catch
+ throw:{error, Atom, RestData} -> {error, Atom, RestData}
+ end;
+transcode(URIString, Options) when is_list(URIString) ->
+ InEnc = proplists:get_value(in_encoding, Options, utf8),
+ OutEnc = proplists:get_value(out_encoding, Options, utf8),
+ Flattened = flatten_list(URIString, InEnc),
+ try transcode(Flattened, [], InEnc, OutEnc)
+ catch
+ throw:{error, Atom, RestData} -> {error, Atom, RestData}
+ end.
+
+
+%%%========================================================================
+%%% Internal functions
+%%%========================================================================
+
+%%-------------------------------------------------------------------------
+%% Converts Map fields to lists
+%%-------------------------------------------------------------------------
+convert_mapfields_to_list(Map) ->
+ Fun = fun (_, V) when is_binary(V) -> unicode:characters_to_list(V);
+ (_, V) -> V end,
+ maps:map(Fun, Map).
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 4.1. URI Reference]
+%%
+%% URI-reference is used to denote the most common usage of a resource
+%% identifier.
+%%
+%% URI-reference = URI / relative-ref
+%%-------------------------------------------------------------------------
+-spec parse_uri_reference(binary(), uri_map()) -> uri_map().
+parse_uri_reference(<<>>, _) -> #{path => <<>>};
+parse_uri_reference(URIString, URI) ->
+ try parse_scheme_start(URIString, URI)
+ catch
+ throw:{_,_,_} ->
+ parse_relative_part(URIString, URI)
+ end.
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 4.2. Relative Reference]
+%%
+%% A relative reference takes advantage of the hierarchical syntax
+%% (Section 1.2.3) to express a URI reference relative to the name space
+%% of another hierarchical URI.
+%%
+%% relative-ref = relative-part [ "?" query ] [ "#" fragment ]
+%%
+%% relative-part = "//" authority path-abempty
+%% / path-absolute
+%% / path-noscheme
+%% / path-empty
+%%-------------------------------------------------------------------------
+-spec parse_relative_part(binary(), uri_map()) -> uri_map().
+parse_relative_part(?STRING_REST("//", Rest), URI) ->
+ %% Parse userinfo - "//" is NOT part of authority
+ try parse_userinfo(Rest, URI) of
+ {T, URI1} ->
+ Userinfo = calculate_parsed_userinfo(Rest, T),
+ URI2 = maybe_add_path(URI1),
+ URI2#{userinfo => decode_userinfo(Userinfo)}
+ catch
+ throw:{_,_,_} ->
+ {T, URI1} = parse_host(Rest, URI),
+ Host = calculate_parsed_host_port(Rest, T),
+ URI2 = maybe_add_path(URI1),
+ URI2#{host => decode_host(remove_brackets(Host))}
+ end;
+parse_relative_part(?STRING_REST($/, Rest), URI) ->
+ {T, URI1} = parse_segment(Rest, URI), % path-absolute
+ Path = calculate_parsed_part(Rest, T),
+ URI1#{path => decode_path(?STRING_REST($/, Path))};
+parse_relative_part(?STRING_REST($?, Rest), URI) ->
+ {T, URI1} = parse_query(Rest, URI), % path-empty ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ URI2 = maybe_add_path(URI1),
+ URI2#{query => decode_query(Query)};
+parse_relative_part(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI), % path-empty
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ URI2 = maybe_add_path(URI1),
+ URI2#{fragment => decode_fragment(Fragment)};
+parse_relative_part(?STRING_REST(Char, Rest), URI) ->
+ case is_segment_nz_nc(Char) of
+ true ->
+ {T, URI1} = parse_segment_nz_nc(Rest, URI), % path-noscheme
+ Path = calculate_parsed_part(Rest, T),
+ URI1#{path => decode_path(?STRING_REST(Char, Path))};
+ false -> throw({error,invalid_uri,[Char]})
+ end.
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 3.3. Path]
+%%
+%% The path component contains data, usually organized in hierarchical
+%% form, that, along with data in the non-hierarchical query component
+%% (Section 3.4), serves to identify a resource within the scope of the
+%% URI's scheme and naming authority (if any). The path is terminated
+%% by the first question mark ("?") or number sign ("#") character, or
+%% by the end of the URI.
+%%
+%% path = path-abempty ; begins with "/" or is empty
+%% / path-absolute ; begins with "/" but not "//"
+%% / path-noscheme ; begins with a non-colon segment
+%% / path-rootless ; begins with a segment
+%% / path-empty ; zero characters
+%%
+%% path-abempty = *( "/" segment )
+%% path-absolute = "/" [ segment-nz *( "/" segment ) ]
+%% path-noscheme = segment-nz-nc *( "/" segment )
+%% path-rootless = segment-nz *( "/" segment )
+%% path-empty = 0<pchar>
+%% segment = *pchar
+%% segment-nz = 1*pchar
+%% segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
+%% ; non-zero-length segment without any colon ":"
+%%
+%% pchar = unreserved / pct-encoded / sub-delims / ":" / "@"
+%%-------------------------------------------------------------------------
+
+%%-------------------------------------------------------------------------
+%% path-abempty
+%%-------------------------------------------------------------------------
+-spec parse_segment(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_segment(?STRING_REST($/, Rest), URI) ->
+ parse_segment(Rest, URI); % segment
+parse_segment(?STRING_REST($?, Rest), URI) ->
+ {T, URI1} = parse_query(Rest, URI), % ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{query => decode_query(Query)}};
+parse_segment(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI),
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_segment(?STRING_REST(Char, Rest), URI) ->
+ case is_pchar(Char) of
+ true -> parse_segment(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_segment(?STRING_EMPTY, URI) ->
+ {?STRING_EMPTY, URI}.
+
+
+%%-------------------------------------------------------------------------
+%% path-noscheme
+%%-------------------------------------------------------------------------
+-spec parse_segment_nz_nc(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_segment_nz_nc(?STRING_REST($/, Rest), URI) ->
+ parse_segment(Rest, URI); % segment
+parse_segment_nz_nc(?STRING_REST($?, Rest), URI) ->
+ {T, URI1} = parse_query(Rest, URI), % ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{query => decode_query(Query)}};
+parse_segment_nz_nc(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI),
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_segment_nz_nc(?STRING_REST(Char, Rest), URI) ->
+ case is_segment_nz_nc(Char) of
+ true -> parse_segment_nz_nc(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_segment_nz_nc(?STRING_EMPTY, URI) ->
+ {?STRING_EMPTY, URI}.
+
+
+%% Check if char is pchar.
+-spec is_pchar(char()) -> boolean().
+is_pchar($%) -> true; % pct-encoded
+is_pchar($:) -> true;
+is_pchar($@) -> true;
+is_pchar(Char) -> is_unreserved(Char) orelse is_sub_delim(Char).
+
+%% Check if char is segment_nz_nc.
+-spec is_segment_nz_nc(char()) -> boolean().
+is_segment_nz_nc($%) -> true; % pct-encoded
+is_segment_nz_nc($@) -> true;
+is_segment_nz_nc(Char) -> is_unreserved(Char) orelse is_sub_delim(Char).
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 3.1. Scheme]
+%%
+%% Each URI begins with a scheme name that refers to a specification for
+%% assigning identifiers within that scheme.
+%%
+%% scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
+%%-------------------------------------------------------------------------
+-spec parse_scheme_start(binary(), uri_map()) -> uri_map().
+parse_scheme_start(?STRING_REST(Char, Rest), URI) ->
+ case is_alpha(Char) of
+ true -> {T, URI1} = parse_scheme(Rest, URI),
+ Scheme = calculate_parsed_scheme(Rest, T),
+ URI2 = maybe_add_path(URI1),
+ URI2#{scheme => ?STRING_REST(Char, Scheme)};
+ false -> throw({error,invalid_uri,[Char]})
+ end.
+
+%% Add path component if it missing after parsing the URI.
+%% According to the URI specification there is always a
+%% path component in every URI-reference and it can be
+%% empty.
+maybe_add_path(Map) ->
+ case maps:is_key(path, Map) of
+ false ->
+ Map#{path => <<>>};
+ _Else ->
+ Map
+ end.
+
+
+-spec parse_scheme(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_scheme(?STRING_REST($:, Rest), URI) ->
+ {_, URI1} = parse_hier(Rest, URI),
+ {Rest, URI1};
+parse_scheme(?STRING_REST(Char, Rest), URI) ->
+ case is_scheme(Char) of
+ true -> parse_scheme(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_scheme(?STRING_EMPTY, _URI) ->
+ throw({error,invalid_uri,<<>>}).
+
+
+%% Check if char is allowed in scheme
+-spec is_scheme(char()) -> boolean().
+is_scheme($+) -> true;
+is_scheme($-) -> true;
+is_scheme($.) -> true;
+is_scheme(Char) -> is_alpha(Char) orelse is_digit(Char).
+
+
+%%-------------------------------------------------------------------------
+%% hier-part = "//" authority path-abempty
+%% / path-absolute
+%% / path-rootless
+%% / path-empty
+%%-------------------------------------------------------------------------
+-spec parse_hier(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_hier(?STRING_REST("//", Rest), URI) ->
+ % Parse userinfo - "//" is NOT part of authority
+ try parse_userinfo(Rest, URI) of
+ {T, URI1} ->
+ Userinfo = calculate_parsed_userinfo(Rest, T),
+ {Rest, URI1#{userinfo => decode_userinfo(Userinfo)}}
+ catch
+ throw:{_,_,_} ->
+ {T, URI1} = parse_host(Rest, URI),
+ Host = calculate_parsed_host_port(Rest, T),
+ {Rest, URI1#{host => decode_host(remove_brackets(Host))}}
+ end;
+parse_hier(?STRING_REST($/, Rest), URI) ->
+ {T, URI1} = parse_segment(Rest, URI), % path-absolute
+ Path = calculate_parsed_part(Rest, T),
+ {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}};
+parse_hier(?STRING_REST($?, Rest), URI) ->
+ {T, URI1} = parse_query(Rest, URI), % path-empty ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{query => decode_query(Query)}};
+parse_hier(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI), % path-empty
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_hier(?STRING_REST(Char, Rest), URI) -> % path-rootless
+ case is_pchar(Char) of
+ true -> % segment_nz
+ {T, URI1} = parse_segment(Rest, URI),
+ Path = calculate_parsed_part(Rest, T),
+ {Rest, URI1#{path => decode_path(?STRING_REST(Char, Path))}};
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_hier(?STRING_EMPTY, URI) ->
+ {<<>>, URI}.
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 3.2. Authority]
+%%
+%% Many URI schemes include a hierarchical element for a naming
+%% authority so that governance of the name space defined by the
+%% remainder of the URI is delegated to that authority (which may, in
+%% turn, delegate it further).
+%%
+%% The authority component is preceded by a double slash ("//") and is
+%% terminated by the next slash ("/"), question mark ("?"), or number
+%% sign ("#") character, or by the end of the URI.
+%%
+%% authority = [ userinfo "@" ] host [ ":" port ]
+%%
+%%
+%% [RFC 3986, Chapter 3.2.1. User Information]
+%%
+%% The userinfo subcomponent may consist of a user name and, optionally,
+%% scheme-specific information about how to gain authorization to access
+%% the resource. The user information, if present, is followed by a
+%% commercial at-sign ("@") that delimits it from the host.
+%%
+%% userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
+%%-------------------------------------------------------------------------
+-spec parse_userinfo(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_userinfo(?CHAR($@), URI) ->
+ {?STRING_EMPTY, URI#{host => <<>>}};
+parse_userinfo(?STRING_REST($@, Rest), URI) ->
+ {T, URI1} = parse_host(Rest, URI),
+ Host = calculate_parsed_host_port(Rest, T),
+ {Rest, URI1#{host => decode_host(remove_brackets(Host))}};
+parse_userinfo(?STRING_REST(Char, Rest), URI) ->
+ case is_userinfo(Char) of
+ true -> parse_userinfo(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_userinfo(?STRING_EMPTY, _URI) ->
+ %% URI cannot end in userinfo state
+ throw({error,invalid_uri,<<>>}).
+
+
+%% Check if char is allowed in userinfo
+-spec is_userinfo(char()) -> boolean().
+is_userinfo($%) -> true; % pct-encoded
+is_userinfo($:) -> true;
+is_userinfo(Char) -> is_unreserved(Char) orelse is_sub_delim(Char).
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 3.2.2. Host]
+%%
+%% The host subcomponent of authority is identified by an IP literal
+%% encapsulated within square brackets, an IPv4 address in dotted-
+%% decimal form, or a registered name.
+%%
+%% host = IP-literal / IPv4address / reg-name
+%%
+%% IP-literal = "[" ( IPv6address / IPvFuture ) "]"
+%%
+%% IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" )
+%%
+%% IPv6address = 6( h16 ":" ) ls32
+%% / "::" 5( h16 ":" ) ls32
+%% / [ h16 ] "::" 4( h16 ":" ) ls32
+%% / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
+%% / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
+%% / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32
+%% / [ *4( h16 ":" ) h16 ] "::" ls32
+%% / [ *5( h16 ":" ) h16 ] "::" h16
+%% / [ *6( h16 ":" ) h16 ] "::"
+%%
+%% ls32 = ( h16 ":" h16 ) / IPv4address
+%% ; least-significant 32 bits of address
+%%
+%% h16 = 1*4HEXDIG
+%% ; 16 bits of address represented in hexadecimal
+%%
+%% IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
+%%
+%% dec-octet = DIGIT ; 0-9
+%% / %x31-39 DIGIT ; 10-99
+%% / "1" 2DIGIT ; 100-199
+%% / "2" %x30-34 DIGIT ; 200-249
+%% / "25" %x30-35 ; 250-255
+%%
+%% reg-name = *( unreserved / pct-encoded / sub-delims )
+%%-------------------------------------------------------------------------
+-spec parse_host(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_host(?STRING_REST($:, Rest), URI) ->
+ {T, URI1} = parse_port(Rest, URI),
+ H = calculate_parsed_host_port(Rest, T),
+ Port = get_port(H),
+ {Rest, URI1#{port => Port}};
+parse_host(?STRING_REST($/, Rest), URI) ->
+ {T, URI1} = parse_segment(Rest, URI), % path-abempty
+ Path = calculate_parsed_part(Rest, T),
+ {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}};
+parse_host(?STRING_REST($?, Rest), URI) ->
+ {T, URI1} = parse_query(Rest, URI), % path-empty ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{query => decode_query(Query)}};
+parse_host(?STRING_REST($[, Rest), URI) ->
+ parse_ipv6_bin(Rest, [], URI);
+parse_host(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI), % path-empty
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_host(?STRING_REST(Char, Rest), URI) ->
+ case is_digit(Char) of
+ true -> parse_ipv4_bin(Rest, [Char], URI);
+ false -> parse_reg_name(?STRING_REST(Char, Rest), URI)
+ end;
+parse_host(?STRING_EMPTY, URI) ->
+ {?STRING_EMPTY, URI}.
+
+
+-spec parse_reg_name(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_reg_name(?STRING_REST($:, Rest), URI) ->
+ {T, URI1} = parse_port(Rest, URI),
+ H = calculate_parsed_host_port(Rest, T),
+ Port = get_port(H),
+ {Rest, URI1#{port => Port}};
+parse_reg_name(?STRING_REST($/, Rest), URI) ->
+ {T, URI1} = parse_segment(Rest, URI), % path-abempty
+ Path = calculate_parsed_part(Rest, T),
+ {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}};
+parse_reg_name(?STRING_REST($?, Rest), URI) ->
+ {T, URI1} = parse_query(Rest, URI), % path-empty ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{query => decode_query(Query)}};
+parse_reg_name(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI), % path-empty
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_reg_name(?STRING_REST(Char, Rest), URI) ->
+ case is_reg_name(Char) of
+ true -> parse_reg_name(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_reg_name(?STRING_EMPTY, URI) ->
+ {?STRING_EMPTY, URI}.
+
+%% Check if char is allowed in reg-name
+-spec is_reg_name(char()) -> boolean().
+is_reg_name($%) -> true;
+is_reg_name(Char) -> is_unreserved(Char) orelse is_sub_delim(Char).
+
+
+-spec parse_ipv4_bin(binary(), list(), uri_map()) -> {binary(), uri_map()}.
+parse_ipv4_bin(?STRING_REST($:, Rest), Acc, URI) ->
+ _ = validate_ipv4_address(lists:reverse(Acc)),
+ {T, URI1} = parse_port(Rest, URI),
+ H = calculate_parsed_host_port(Rest, T),
+ Port = get_port(H),
+ {Rest, URI1#{port => Port}};
+parse_ipv4_bin(?STRING_REST($/, Rest), Acc, URI) ->
+ _ = validate_ipv4_address(lists:reverse(Acc)),
+ {T, URI1} = parse_segment(Rest, URI), % path-abempty
+ Path = calculate_parsed_part(Rest, T),
+ {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}};
+parse_ipv4_bin(?STRING_REST($?, Rest), Acc, URI) ->
+ _ = validate_ipv4_address(lists:reverse(Acc)),
+ {T, URI1} = parse_query(Rest, URI), % path-empty ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{query => decode_query(Query)}};
+parse_ipv4_bin(?STRING_REST($#, Rest), Acc, URI) ->
+ _ = validate_ipv4_address(lists:reverse(Acc)),
+ {T, URI1} = parse_fragment(Rest, URI), % path-empty
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_ipv4_bin(?STRING_REST(Char, Rest), Acc, URI) ->
+ case is_ipv4(Char) of
+ true -> parse_ipv4_bin(Rest, [Char|Acc], URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_ipv4_bin(?STRING_EMPTY, Acc, URI) ->
+ _ = validate_ipv4_address(lists:reverse(Acc)),
+ {?STRING_EMPTY, URI}.
+
+
+%% Check if char is allowed in IPv4 addresses
+-spec is_ipv4(char()) -> boolean().
+is_ipv4($.) -> true;
+is_ipv4(Char) -> is_digit(Char).
+
+-spec validate_ipv4_address(list()) -> list().
+validate_ipv4_address(Addr) ->
+ case inet:parse_ipv4strict_address(Addr) of
+ {ok, _} -> Addr;
+ {error, _} -> throw({error,invalid_uri,Addr})
+ end.
+
+
+-spec parse_ipv6_bin(binary(), list(), uri_map()) -> {binary(), uri_map()}.
+parse_ipv6_bin(?STRING_REST($], Rest), Acc, URI) ->
+ _ = validate_ipv6_address(lists:reverse(Acc)),
+ parse_ipv6_bin_end(Rest, URI);
+parse_ipv6_bin(?STRING_REST(Char, Rest), Acc, URI) ->
+ case is_ipv6(Char) of
+ true -> parse_ipv6_bin(Rest, [Char|Acc], URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_ipv6_bin(?STRING_EMPTY, _Acc, _URI) ->
+ throw({error,invalid_uri,<<>>}).
+
+%% Check if char is allowed in IPv6 addresses
+-spec is_ipv6(char()) -> boolean().
+is_ipv6($:) -> true;
+is_ipv6($.) -> true;
+is_ipv6(Char) -> is_hex_digit(Char).
+
+
+-spec parse_ipv6_bin_end(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_ipv6_bin_end(?STRING_REST($:, Rest), URI) ->
+ {T, URI1} = parse_port(Rest, URI),
+ H = calculate_parsed_host_port(Rest, T),
+ Port = get_port(H),
+ {Rest, URI1#{port => Port}};
+parse_ipv6_bin_end(?STRING_REST($/, Rest), URI) ->
+ {T, URI1} = parse_segment(Rest, URI), % path-abempty
+ Path = calculate_parsed_part(Rest, T),
+ {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}};
+parse_ipv6_bin_end(?STRING_REST($?, Rest), URI) ->
+ {T, URI1} = parse_query(Rest, URI), % path-empty ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{query => decode_query(Query)}};
+parse_ipv6_bin_end(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI), % path-empty
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_ipv6_bin_end(?STRING_REST(Char, Rest), URI) ->
+ case is_ipv6(Char) of
+ true -> parse_ipv6_bin_end(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_ipv6_bin_end(?STRING_EMPTY, URI) ->
+ {?STRING_EMPTY, URI}.
+
+-spec validate_ipv6_address(list()) -> list().
+validate_ipv6_address(Addr) ->
+ case inet:parse_ipv6strict_address(Addr) of
+ {ok, _} -> Addr;
+ {error, _} -> throw({error,invalid_uri,Addr})
+ end.
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 3.2.2. Port]
+%%
+%% The port subcomponent of authority is designated by an optional port
+%% number in decimal following the host and delimited from it by a
+%% single colon (":") character.
+%%
+%% port = *DIGIT
+%%-------------------------------------------------------------------------
+-spec parse_port(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_port(?STRING_REST($/, Rest), URI) ->
+ {T, URI1} = parse_segment(Rest, URI), % path-abempty
+ Path = calculate_parsed_part(Rest, T),
+ {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}};
+parse_port(?STRING_REST($?, Rest), URI) ->
+ {T, URI1} = parse_query(Rest, URI), % path-empty ?query
+ Query = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{query => decode_query(Query)}};
+parse_port(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI), % path-empty
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_port(?STRING_REST(Char, Rest), URI) ->
+ case is_digit(Char) of
+ true -> parse_port(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_port(?STRING_EMPTY, URI) ->
+ {?STRING_EMPTY, URI}.
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 3.4. Query]
+%%
+%% The query component contains non-hierarchical data that, along with
+%% data in the path component (Section 3.3), serves to identify a
+%% resource within the scope of the URI's scheme and naming authority
+%% (if any). The query component is indicated by the first question
+%% mark ("?") character and terminated by a number sign ("#") character
+%% or by the end of the URI.
+%%
+%% query = *( pchar / "/" / "?" )
+%%-------------------------------------------------------------------------
+-spec parse_query(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_query(?STRING_REST($#, Rest), URI) ->
+ {T, URI1} = parse_fragment(Rest, URI),
+ Fragment = calculate_parsed_query_fragment(Rest, T),
+ {Rest, URI1#{fragment => decode_fragment(Fragment)}};
+parse_query(?STRING_REST(Char, Rest), URI) ->
+ case is_query(Char) of
+ true -> parse_query(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_query(?STRING_EMPTY, URI) ->
+ {?STRING_EMPTY, URI}.
+
+
+%% Check if char is allowed in query
+-spec is_query(char()) -> boolean().
+is_query($/) -> true;
+is_query($?) -> true;
+is_query(Char) -> is_pchar(Char).
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 3.5. Fragment]
+%%
+%% The fragment identifier component of a URI allows indirect
+%% identification of a secondary resource by reference to a primary
+%% resource and additional identifying information.
+%%
+%% fragment = *( pchar / "/" / "?" )
+%%-------------------------------------------------------------------------
+-spec parse_fragment(binary(), uri_map()) -> {binary(), uri_map()}.
+parse_fragment(?STRING_REST(Char, Rest), URI) ->
+ case is_fragment(Char) of
+ true -> parse_fragment(Rest, URI);
+ false -> throw({error,invalid_uri,[Char]})
+ end;
+parse_fragment(?STRING_EMPTY, URI) ->
+ {?STRING_EMPTY, URI}.
+
+
+%% Check if char is allowed in fragment
+-spec is_fragment(char()) -> boolean().
+is_fragment($/) -> true;
+is_fragment($?) -> true;
+is_fragment(Char) -> is_pchar(Char).
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 2.2. Reserved Characters]
+%%
+%% reserved = gen-delims / sub-delims
+%%
+%% gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"
+%%
+%% sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
+%% / "*" / "+" / "," / ";" / "="
+%%
+%%-------------------------------------------------------------------------
+
+%% Check if char is sub-delim.
+-spec is_sub_delim(char()) -> boolean().
+is_sub_delim($!) -> true;
+is_sub_delim($$) -> true;
+is_sub_delim($&) -> true;
+is_sub_delim($') -> true;
+is_sub_delim($() -> true;
+is_sub_delim($)) -> true;
+
+is_sub_delim($*) -> true;
+is_sub_delim($+) -> true;
+is_sub_delim($,) -> true;
+is_sub_delim($;) -> true;
+is_sub_delim($=) -> true;
+is_sub_delim(_) -> false.
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 2.3. Unreserved Characters]
+%%
+%% unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
+%%
+%%-------------------------------------------------------------------------
+-spec is_unreserved(char()) -> boolean().
+is_unreserved($-) -> true;
+is_unreserved($.) -> true;
+is_unreserved($_) -> true;
+is_unreserved($~) -> true;
+is_unreserved(Char) -> is_alpha(Char) orelse is_digit(Char).
+
+-spec is_alpha(char()) -> boolean().
+is_alpha(C)
+ when $A =< C, C =< $Z;
+ $a =< C, C =< $z -> true;
+is_alpha(_) -> false.
+
+-spec is_digit(char()) -> boolean().
+is_digit(C)
+ when $0 =< C, C =< $9 -> true;
+is_digit(_) -> false.
+
+-spec is_hex_digit(char()) -> boolean().
+is_hex_digit(C)
+ when $0 =< C, C =< $9;$a =< C, C =< $f;$A =< C, C =< $F -> true;
+is_hex_digit(_) -> false.
+
+
+%% Remove enclosing brackets from binary
+-spec remove_brackets(binary()) -> binary().
+remove_brackets(<<$[/utf8, Rest/binary>>) ->
+ {H,T} = split_binary(Rest, byte_size(Rest) - 1),
+ case T =:= <<$]/utf8>> of
+ true -> H;
+ false -> Rest
+ end;
+remove_brackets(Addr) -> Addr.
+
+
+%%-------------------------------------------------------------------------
+%% Helper functions for calculating the parsed binary.
+%%-------------------------------------------------------------------------
+-spec calculate_parsed_scheme(binary(), binary()) -> binary().
+calculate_parsed_scheme(Input, <<>>) ->
+ strip_last_char(Input, [$:]);
+calculate_parsed_scheme(Input, Unparsed) ->
+ get_parsed_binary(Input, Unparsed).
+
+
+-spec calculate_parsed_part(binary(), binary()) -> binary().
+calculate_parsed_part(Input, <<>>) ->
+ strip_last_char(Input, [$?,$#]);
+calculate_parsed_part(Input, Unparsed) ->
+ get_parsed_binary(Input, Unparsed).
+
+
+-spec calculate_parsed_userinfo(binary(), binary()) -> binary().
+calculate_parsed_userinfo(Input, <<>>) ->
+ strip_last_char(Input, [$?,$#,$@]);
+calculate_parsed_userinfo(Input, Unparsed) ->
+ get_parsed_binary(Input, Unparsed).
+
+
+-spec calculate_parsed_host_port(binary(), binary()) -> binary().
+calculate_parsed_host_port(Input, <<>>) ->
+ strip_last_char(Input, [$:,$?,$#,$/]);
+calculate_parsed_host_port(Input, Unparsed) ->
+ get_parsed_binary(Input, Unparsed).
+
+
+calculate_parsed_query_fragment(Input, <<>>) ->
+ strip_last_char(Input, [$#]);
+calculate_parsed_query_fragment(Input, Unparsed) ->
+ get_parsed_binary(Input, Unparsed).
+
+
+get_port(<<>>) ->
+ undefined;
+get_port(B) ->
+ try binary_to_integer(B)
+ catch
+ error:badarg ->
+ throw({error, invalid_uri, B})
+ end.
+
+
+%% Strip last char if it is in list
+%%
+%% This function is optimized for speed: parse/1 is about 10% faster than
+%% with an alternative implementation based on lists and sets.
+strip_last_char(<<>>, _) -> <<>>;
+strip_last_char(Input, [C0]) ->
+ case binary:last(Input) of
+ C0 ->
+ init_binary(Input);
+ _Else ->
+ Input
+ end;
+strip_last_char(Input, [C0,C1]) ->
+ case binary:last(Input) of
+ C0 ->
+ init_binary(Input);
+ C1 ->
+ init_binary(Input);
+ _Else ->
+ Input
+ end;
+strip_last_char(Input, [C0,C1,C2]) ->
+ case binary:last(Input) of
+ C0 ->
+ init_binary(Input);
+ C1 ->
+ init_binary(Input);
+ C2 ->
+ init_binary(Input);
+ _Else ->
+ Input
+ end;
+strip_last_char(Input, [C0,C1,C2,C3]) ->
+ case binary:last(Input) of
+ C0 ->
+ init_binary(Input);
+ C1 ->
+ init_binary(Input);
+ C2 ->
+ init_binary(Input);
+ C3 ->
+ init_binary(Input);
+ _Else ->
+ Input
+ end.
+
+
+%% Get parsed binary
+get_parsed_binary(Input, Unparsed) ->
+ {First, _} = split_binary(Input, byte_size(Input) - byte_size_exl_head(Unparsed)),
+ First.
+
+
+%% Return all bytes of the binary except the last one. The binary must be non-empty.
+init_binary(B) ->
+ {Init, _} =
+ split_binary(B, byte_size(B) - 1),
+ Init.
+
+
+%% Returns the size of a binary exluding the first element.
+%% Used in calls to split_binary().
+-spec byte_size_exl_head(binary()) -> number().
+byte_size_exl_head(<<>>) -> 0;
+byte_size_exl_head(Binary) -> byte_size(Binary) + 1.
+
+
+%%-------------------------------------------------------------------------
+%% [RFC 3986, Chapter 2.1. Percent-Encoding]
+%%
+%% A percent-encoding mechanism is used to represent a data octet in a
+%% component when that octet's corresponding character is outside the
+%% allowed set or is being used as a delimiter of, or within, the
+%% component. A percent-encoded octet is encoded as a character
+%% triplet, consisting of the percent character "%" followed by the two
+%% hexadecimal digits representing that octet's numeric value. For
+%% example, "%20" is the percent-encoding for the binary octet
+%% "00100000" (ABNF: %x20), which in US-ASCII corresponds to the space
+%% character (SP). Section 2.4 describes when percent-encoding and
+%% decoding is applied.
+%%
+%% pct-encoded = "%" HEXDIG HEXDIG
+%%-------------------------------------------------------------------------
+-spec decode_userinfo(binary()) -> binary().
+decode_userinfo(Cs) ->
+ check_utf8(decode(Cs, fun is_userinfo/1, <<>>)).
+
+-spec decode_host(binary()) -> binary().
+decode_host(Cs) ->
+ check_utf8(decode(Cs, fun is_host/1, <<>>)).
+
+-spec decode_path(binary()) -> binary().
+decode_path(Cs) ->
+ check_utf8(decode(Cs, fun is_path/1, <<>>)).
+
+-spec decode_query(binary()) -> binary().
+decode_query(Cs) ->
+ check_utf8(decode(Cs, fun is_query/1, <<>>)).
+
+-spec decode_fragment(binary()) -> binary().
+decode_fragment(Cs) ->
+ check_utf8(decode(Cs, fun is_fragment/1, <<>>)).
+
+
+%% Returns Cs if it is utf8 encoded.
+check_utf8(Cs) ->
+ case unicode:characters_to_list(Cs) of
+ {incomplete,_,_} ->
+ throw({error,invalid_utf8,Cs});
+ {error,_,_} ->
+ throw({error,invalid_utf8,Cs});
+ _ -> Cs
+ end.
+
+%%-------------------------------------------------------------------------
+%% Percent-encode
+%%-------------------------------------------------------------------------
+
+%% Only validates as scheme cannot have percent-encoded characters
+-spec encode_scheme(list()|binary()) -> list() | binary().
+encode_scheme([]) ->
+ throw({error,invalid_scheme,""});
+encode_scheme(<<>>) ->
+ throw({error,invalid_scheme,<<>>});
+encode_scheme(Scheme) ->
+ case validate_scheme(Scheme) of
+ true -> Scheme;
+ false -> throw({error,invalid_scheme,Scheme})
+ end.
+
+-spec encode_userinfo(list()|binary()) -> list() | binary().
+encode_userinfo(Cs) ->
+ encode(Cs, fun is_userinfo/1).
+
+-spec encode_host(list()|binary()) -> list() | binary().
+encode_host(Cs) ->
+ case classify_host(Cs) of
+ regname -> Cs;
+ ipv4 -> Cs;
+ ipv6 -> bracket_ipv6(Cs);
+ other -> encode(Cs, fun is_reg_name/1)
+ end.
+
+-spec encode_path(list()|binary()) -> list() | binary().
+encode_path(Cs) ->
+ encode(Cs, fun is_path/1).
+
+-spec encode_query(list()|binary()) -> list() | binary().
+encode_query(Cs) ->
+ encode(Cs, fun is_query/1).
+
+-spec encode_fragment(list()|binary()) -> list() | binary().
+encode_fragment(Cs) ->
+ encode(Cs, fun is_fragment/1).
+
+%%-------------------------------------------------------------------------
+%% Helper funtions for percent-decode
+%%-------------------------------------------------------------------------
+decode(<<$%,C0,C1,Cs/binary>>, Fun, Acc) ->
+ case is_hex_digit(C0) andalso is_hex_digit(C1) of
+ true ->
+ B = ?HEX2DEC(C0)*16+?HEX2DEC(C1),
+ decode(Cs, Fun, <<Acc/binary, B>>);
+ false -> throw({error,invalid_percent_encoding,<<$%,C0,C1>>})
+ end;
+decode(<<C,Cs/binary>>, Fun, Acc) ->
+ case Fun(C) of
+ true -> decode(Cs, Fun, <<Acc/binary, C>>);
+ false -> throw({error,invalid_percent_encoding,<<C,Cs/binary>>})
+ end;
+decode(<<>>, _Fun, Acc) ->
+ Acc.
+
+%% Check if char is allowed in host
+-spec is_host(char()) -> boolean().
+is_host($:) -> true;
+is_host(Char) -> is_unreserved(Char) orelse is_sub_delim(Char).
+
+%% Check if char is allowed in path
+-spec is_path(char()) -> boolean().
+is_path($/) -> true;
+is_path(Char) -> is_pchar(Char).
+
+
+%%-------------------------------------------------------------------------
+%% Helper functions for percent-encode
+%%-------------------------------------------------------------------------
+-spec encode(list()|binary(), fun()) -> list() | binary().
+encode(Component, Fun) when is_list(Component) ->
+ B = unicode:characters_to_binary(Component),
+ unicode:characters_to_list(encode(B, Fun, <<>>));
+encode(Component, Fun) when is_binary(Component) ->
+ encode(Component, Fun, <<>>).
+%%
+encode(<<Char/utf8, Rest/binary>>, Fun, Acc) ->
+ C = encode_codepoint_binary(Char, Fun),
+ encode(Rest, Fun, <<Acc/binary,C/binary>>);
+encode(<<Char, Rest/binary>>, _Fun, _Acc) ->
+ throw({error,invalid_input,<<Char,Rest/binary>>});
+encode(<<>>, _Fun, Acc) ->
+ Acc.
+
+
+-spec encode_codepoint_binary(integer(), fun()) -> binary().
+encode_codepoint_binary(C, Fun) ->
+ case Fun(C) of
+ false -> percent_encode_binary(C);
+ true -> <<C>>
+ end.
+
+
+-spec percent_encode_binary(integer()) -> binary().
+percent_encode_binary(Code) ->
+ percent_encode_binary(<<Code/utf8>>, <<>>).
+
+
+percent_encode_binary(<<A:4,B:4,Rest/binary>>, Acc) ->
+ percent_encode_binary(Rest, <<Acc/binary,$%,(?DEC2HEX(A)),(?DEC2HEX(B))>>);
+percent_encode_binary(<<>>, Acc) ->
+ Acc.
+
+
+%%-------------------------------------------------------------------------
+%%-------------------------------------------------------------------------
+validate_scheme([]) -> true;
+validate_scheme([H|T]) ->
+ case is_scheme(H) of
+ true -> validate_scheme(T);
+ false -> false
+ end;
+validate_scheme(<<>>) -> true;
+validate_scheme(<<H, Rest/binary>>) ->
+ case is_scheme(H) of
+ true -> validate_scheme(Rest);
+ false -> false
+ end.
+
+
+%%-------------------------------------------------------------------------
+%% Classifies hostname into the following categories:
+%% regname, ipv4 - address does not contain reserved characters to be
+%% percent-encoded
+%% ipv6 - address does not contain reserved characters but it shall be
+%% encolsed in brackets
+%% other - address shall be percent-encoded
+%%-------------------------------------------------------------------------
+classify_host([]) -> other;
+classify_host(Addr) when is_binary(Addr) ->
+ A = unicode:characters_to_list(Addr),
+ classify_host_ipv6(A);
+classify_host(Addr) ->
+ classify_host_ipv6(Addr).
+
+classify_host_ipv6(Addr) ->
+ case is_ipv6_address(Addr) of
+ true -> ipv6;
+ false -> classify_host_ipv4(Addr)
+ end.
+
+classify_host_ipv4(Addr) ->
+ case is_ipv4_address(Addr) of
+ true -> ipv4;
+ false -> classify_host_regname(Addr)
+ end.
+
+classify_host_regname([]) -> regname;
+classify_host_regname([H|T]) ->
+ case is_reg_name(H) of
+ true -> classify_host_regname(T);
+ false -> other
+ end.
+
+is_ipv4_address(Addr) ->
+ case inet:parse_ipv4strict_address(Addr) of
+ {ok, _} -> true;
+ {error, _} -> false
+ end.
+
+is_ipv6_address(Addr) ->
+ case inet:parse_ipv6strict_address(Addr) of
+ {ok, _} -> true;
+ {error, _} -> false
+ end.
+
+bracket_ipv6(Addr) when is_binary(Addr) ->
+ concat(<<$[,Addr/binary>>,<<$]>>);
+bracket_ipv6(Addr) when is_list(Addr) ->
+ [$[|Addr] ++ "]".
+
+
+%%-------------------------------------------------------------------------
+%% Helper funtions for recompose
+%%-------------------------------------------------------------------------
+
+%%-------------------------------------------------------------------------
+%% Checks if input Map has valid combination of fields that can be
+%% recomposed into a URI.
+%%
+%% The implementation is based on a decision tree that fulfills the
+%% following rules:
+%% - 'path' shall always be present in the input map
+%% URI = scheme ":" hier-part [ "?" query ] [ "#" fragment ]
+%% hier-part = "//" authority path-abempty
+%% / path-absolute
+%% / path-rootless
+%% / path-empty
+%% - 'host' shall be present in the input map when 'path' starts with
+%% two slashes ("//")
+%% path = path-abempty ; begins with "/" or is empty
+%% / path-absolute ; begins with "/" but not "//"
+%% / path-noscheme ; begins with a non-colon segment
+%% / path-rootless ; begins with a segment
+%% / path-empty ; zero characters
+%% path-abempty = *( "/" segment )
+%% segment = *pchar
+%% - 'host' shall be present if userinfo or port is present in input map
+%% authority = [ userinfo "@" ] host [ ":" port ]
+%% - All fields shall be valid (scheme, userinfo, host, port, path, query
+%% or fragment).
+%%-------------------------------------------------------------------------
+is_valid_map(#{path := Path} = Map) ->
+ ((starts_with_two_slash(Path) andalso is_valid_map_host(Map))
+ orelse
+ (maps:is_key(userinfo, Map) andalso is_valid_map_host(Map))
+ orelse
+ (maps:is_key(port, Map) andalso is_valid_map_host(Map))
+ orelse
+ all_fields_valid(Map));
+is_valid_map(#{}) ->
+ false.
+
+
+is_valid_map_host(Map) ->
+ maps:is_key(host, Map) andalso all_fields_valid(Map).
+
+
+all_fields_valid(Map) ->
+ Fun = fun(scheme, _, Acc) -> Acc;
+ (userinfo, _, Acc) -> Acc;
+ (host, _, Acc) -> Acc;
+ (port, _, Acc) -> Acc;
+ (path, _, Acc) -> Acc;
+ (query, _, Acc) -> Acc;
+ (fragment, _, Acc) -> Acc;
+ (_, _, _) -> false
+ end,
+ maps:fold(Fun, true, Map).
+
+
+starts_with_two_slash([$/,$/|_]) ->
+ true;
+starts_with_two_slash(?STRING_REST("//", _)) ->
+ true;
+starts_with_two_slash(_) -> false.
+
+
+update_scheme(#{scheme := Scheme}, _) ->
+ add_colon_postfix(encode_scheme(Scheme));
+update_scheme(#{}, _) ->
+ empty.
+
+
+update_userinfo(#{userinfo := Userinfo}, empty) ->
+ add_auth_prefix(encode_userinfo(Userinfo));
+update_userinfo(#{userinfo := Userinfo}, URI) ->
+ concat(URI,add_auth_prefix(encode_userinfo(Userinfo)));
+update_userinfo(#{}, empty) ->
+ empty;
+update_userinfo(#{}, URI) ->
+ URI.
+
+
+update_host(#{host := Host}, empty) ->
+ add_auth_prefix(encode_host(Host));
+update_host(#{host := Host} = Map, URI) ->
+ concat(URI,add_host_prefix(Map, encode_host(Host)));
+update_host(#{}, empty) ->
+ empty;
+update_host(#{}, URI) ->
+ URI.
+
+
+%% URI cannot be empty for ports. E.g. ":8080" is not a valid URI
+update_port(#{port := undefined}, URI) ->
+ concat(URI, <<":">>);
+update_port(#{port := Port}, URI) ->
+ concat(URI,add_colon(encode_port(Port)));
+update_port(#{}, URI) ->
+ URI.
+
+
+update_path(#{path := Path}, empty) ->
+ encode_path(Path);
+update_path(#{path := Path}, URI) ->
+ concat(URI,encode_path(Path));
+update_path(#{}, empty) ->
+ empty;
+update_path(#{}, URI) ->
+ URI.
+
+
+update_query(#{query := Query}, empty) ->
+ encode_query(Query);
+update_query(#{query := Query}, URI) ->
+ concat(URI,add_question_mark(encode_query(Query)));
+update_query(#{}, empty) ->
+ empty;
+update_query(#{}, URI) ->
+ URI.
+
+
+update_fragment(#{fragment := Fragment}, empty) ->
+ add_hashmark(encode_fragment(Fragment));
+update_fragment(#{fragment := Fragment}, URI) ->
+ concat(URI,add_hashmark(encode_fragment(Fragment)));
+update_fragment(#{}, empty) ->
+ "";
+update_fragment(#{}, URI) ->
+ URI.
+
+%%-------------------------------------------------------------------------
+%% Concatenates its arguments that can be lists and binaries.
+%% The result is a list if at least one of its argument is a list and
+%% binary otherwise.
+%%-------------------------------------------------------------------------
+concat(A, B) when is_binary(A), is_binary(B) ->
+ <<A/binary, B/binary>>;
+concat(A, B) when is_binary(A), is_list(B) ->
+ unicode:characters_to_list(A) ++ B;
+concat(A, B) when is_list(A) ->
+ A ++ maybe_to_list(B).
+
+add_hashmark(Comp) when is_binary(Comp) ->
+ <<$#, Comp/binary>>;
+add_hashmark(Comp) when is_list(Comp) ->
+ [$#|Comp].
+
+add_question_mark(Comp) when is_binary(Comp) ->
+ <<$?, Comp/binary>>;
+add_question_mark(Comp) when is_list(Comp) ->
+ [$?|Comp].
+
+add_colon(Comp) when is_binary(Comp) ->
+ <<$:, Comp/binary>>.
+
+add_colon_postfix(Comp) when is_binary(Comp) ->
+ <<Comp/binary,$:>>;
+add_colon_postfix(Comp) when is_list(Comp) ->
+ Comp ++ ":".
+
+add_auth_prefix(Comp) when is_binary(Comp) ->
+ <<"//", Comp/binary>>;
+add_auth_prefix(Comp) when is_list(Comp) ->
+ [$/,$/|Comp].
+
+add_host_prefix(#{userinfo := _}, Host) when is_binary(Host) ->
+ <<$@,Host/binary>>;
+add_host_prefix(#{}, Host) when is_binary(Host) ->
+ <<"//",Host/binary>>;
+add_host_prefix(#{userinfo := _}, Host) when is_list(Host) ->
+ [$@|Host];
+add_host_prefix(#{}, Host) when is_list(Host) ->
+ [$/,$/|Host].
+
+maybe_to_list(Comp) when is_binary(Comp) -> unicode:characters_to_list(Comp);
+maybe_to_list(Comp) -> Comp.
+
+encode_port(Port) ->
+ integer_to_binary(Port).
+
+%%-------------------------------------------------------------------------
+%% Helper functions for transcode
+%%-------------------------------------------------------------------------
+
+%%-------------------------------------------------------------------------
+%% uri_string:transcode(<<"x%00%00%00%F6"/utf32>>).
+%% 1. Convert (transcode/2) input to list form (list of unicode codepoints)
+%% "x%00%00%00%F6"
+%% 2. Accumulate characters until percent-encoded segment (transcode/4).
+%% Acc = "x"
+%% 3. Convert percent-encoded triplets to binary form (transcode_pct/4)
+%% <<0,0,0,246>>
+%% 4. Transcode in-encoded binary to out-encoding (utf32 -> utf8):
+%% <<195,182>>
+%% 5. Percent-encode out-encoded binary:
+%% <<"%C3%B6"/utf8>> = <<37,67,51,37,66,54>>
+%% 6. Convert binary to list form, reverse it and append the accumulator
+%% "6B%3C%" + "x"
+%% 7. Reverse Acc and return it
+%%-------------------------------------------------------------------------
+transcode([$%,_C0,_C1|_Rest] = L, Acc, InEnc, OutEnc) ->
+ transcode_pct(L, Acc, <<>>, InEnc, OutEnc);
+transcode([_C|_Rest] = L, Acc, InEnc, OutEnc) ->
+ transcode(L, Acc, [], InEnc, OutEnc).
+%%
+transcode([$%,_C0,_C1|_Rest] = L, Acc, List, InEncoding, OutEncoding) ->
+ transcode_pct(L, List ++ Acc, <<>>, InEncoding, OutEncoding);
+transcode([C|Rest], Acc, List, InEncoding, OutEncoding) ->
+ transcode(Rest, Acc, [C|List], InEncoding, OutEncoding);
+transcode([], Acc, List, _InEncoding, _OutEncoding) ->
+ lists:reverse(List ++ Acc).
+
+
+%% Transcode percent-encoded segment
+transcode_pct([$%,C0,C1|Rest] = L, Acc, B, InEncoding, OutEncoding) ->
+ case is_hex_digit(C0) andalso is_hex_digit(C1) of
+ true ->
+ Int = ?HEX2DEC(C0)*16+?HEX2DEC(C1),
+ transcode_pct(Rest, Acc, <<B/binary, Int>>, InEncoding, OutEncoding);
+ false -> throw({error, invalid_percent_encoding,L})
+ end;
+transcode_pct([_C|_Rest] = L, Acc, B, InEncoding, OutEncoding) ->
+ OutBinary = convert_to_binary(B, InEncoding, OutEncoding),
+ PctEncUtf8 = percent_encode_segment(OutBinary),
+ Out = lists:reverse(convert_to_list(PctEncUtf8, utf8)),
+ transcode(L, Out ++ Acc, [], InEncoding, OutEncoding);
+transcode_pct([], Acc, B, InEncoding, OutEncoding) ->
+ OutBinary = convert_to_binary(B, InEncoding, OutEncoding),
+ PctEncUtf8 = percent_encode_segment(OutBinary),
+ Out = convert_to_list(PctEncUtf8, utf8),
+ lists:reverse(Acc) ++ Out.
+
+
+%% Convert to binary
+convert_to_binary(Binary, InEncoding, OutEncoding) ->
+ case unicode:characters_to_binary(Binary, InEncoding, OutEncoding) of
+ {error, _List, RestData} ->
+ throw({error, invalid_input, RestData});
+ {incomplete, _List, RestData} ->
+ throw({error, invalid_input, RestData});
+ Result ->
+ Result
+ end.
+
+
+%% Convert to list
+convert_to_list(Binary, InEncoding) ->
+ case unicode:characters_to_list(Binary, InEncoding) of
+ {error, _List, RestData} ->
+ throw({error, invalid_input, RestData});
+ {incomplete, _List, RestData} ->
+ throw({error, invalid_input, RestData});
+ Result ->
+ Result
+ end.
+
+
+%% Flatten input list
+flatten_list([], _) ->
+ [];
+flatten_list(L, InEnc) ->
+ flatten_list(L, InEnc, []).
+%%
+flatten_list([H|T], InEnc, Acc) when is_binary(H) ->
+ L = convert_to_list(H, InEnc),
+ flatten_list(T, InEnc, lists:reverse(L) ++ Acc);
+flatten_list([H|T], InEnc, Acc) when is_list(H) ->
+ flatten_list(H ++ T, InEnc, Acc);
+flatten_list([H|T], InEnc, Acc) ->
+ flatten_list(T, InEnc, [H|Acc]);
+flatten_list([], _InEnc, Acc) ->
+ lists:reverse(Acc);
+flatten_list(Arg, _, _) ->
+ throw({error, invalid_input, Arg}).
+
+
+percent_encode_segment(Segment) ->
+ percent_encode_binary(Segment, <<>>).
+
+
+%%-------------------------------------------------------------------------
+%% Helper functions for normalize
+%%-------------------------------------------------------------------------
+
+%% 6.2.2.1. Case Normalization
+normalize_case(#{scheme := Scheme, host := Host} = Map) ->
+ Map#{scheme => to_lower(Scheme),
+ host => to_lower(Host)};
+normalize_case(#{host := Host} = Map) ->
+ Map#{host => to_lower(Host)};
+normalize_case(#{scheme := Scheme} = Map) ->
+ Map#{scheme => to_lower(Scheme)};
+normalize_case(#{} = Map) ->
+ Map.
+
+
+to_lower(Cs) when is_list(Cs) ->
+ B = convert_to_binary(Cs, utf8, utf8),
+ convert_to_list(to_lower(B), utf8);
+to_lower(Cs) when is_binary(Cs) ->
+ to_lower(Cs, <<>>).
+%%
+to_lower(<<C,Cs/binary>>, Acc) when $A =< C, C =< $Z ->
+ to_lower(Cs, <<Acc/binary,(C + 32)>>);
+to_lower(<<C,Cs/binary>>, Acc) ->
+ to_lower(Cs, <<Acc/binary,C>>);
+to_lower(<<>>, Acc) ->
+ Acc.
+
+
+%% 6.2.2.3. Path Segment Normalization
+%% 5.2.4. Remove Dot Segments
+normalize_path_segment(Map) ->
+ Path = maps:get(path, Map, undefined),
+ Map#{path => remove_dot_segments(Path)}.
+
+
+remove_dot_segments(Path) when is_binary(Path) ->
+ remove_dot_segments(Path, <<>>);
+remove_dot_segments(Path) when is_list(Path) ->
+ B = convert_to_binary(Path, utf8, utf8),
+ B1 = remove_dot_segments(B, <<>>),
+ convert_to_list(B1, utf8).
+%%
+remove_dot_segments(<<>>, Output) ->
+ Output;
+remove_dot_segments(<<"../",T/binary>>, Output) ->
+ remove_dot_segments(T, Output);
+remove_dot_segments(<<"./",T/binary>>, Output) ->
+ remove_dot_segments(T, Output);
+remove_dot_segments(<<"/./",T/binary>>, Output) ->
+ remove_dot_segments(<<$/,T/binary>>, Output);
+remove_dot_segments(<<"/.">>, Output) ->
+ remove_dot_segments(<<$/>>, Output);
+remove_dot_segments(<<"/../",T/binary>>, Output) ->
+ Out1 = remove_last_segment(Output),
+ remove_dot_segments(<<$/,T/binary>>, Out1);
+remove_dot_segments(<<"/..">>, Output) ->
+ Out1 = remove_last_segment(Output),
+ remove_dot_segments(<<$/>>, Out1);
+remove_dot_segments(<<$.>>, Output) ->
+ remove_dot_segments(<<>>, Output);
+remove_dot_segments(<<"..">>, Output) ->
+ remove_dot_segments(<<>>, Output);
+remove_dot_segments(Input, Output) ->
+ {First, Rest} = first_path_segment(Input),
+ remove_dot_segments(Rest, <<Output/binary,First/binary>>).
+
+
+first_path_segment(Input) ->
+ F = first_path_segment(Input, <<>>),
+ split_binary(Input, byte_size(F)).
+%%
+first_path_segment(<<$/,T/binary>>, Acc) ->
+ first_path_segment_end(<<T/binary>>, <<Acc/binary,$/>>);
+first_path_segment(<<C,T/binary>>, Acc) ->
+ first_path_segment_end(<<T/binary>>, <<Acc/binary,C>>).
+
+
+first_path_segment_end(<<>>, Acc) ->
+ Acc;
+first_path_segment_end(<<$/,_/binary>>, Acc) ->
+ Acc;
+first_path_segment_end(<<C,T/binary>>, Acc) ->
+ first_path_segment_end(<<T/binary>>, <<Acc/binary,C>>).
+
+
+remove_last_segment(<<>>) ->
+ <<>>;
+remove_last_segment(B) ->
+ {Init, Last} = split_binary(B, byte_size(B) - 1),
+ case Last of
+ <<$/>> ->
+ Init;
+ _Char ->
+ remove_last_segment(Init)
+ end.
+
+
+%% RFC 3986, 6.2.3. Scheme-Based Normalization
+normalize_scheme_based(Map) ->
+ Scheme = maps:get(scheme, Map, undefined),
+ Port = maps:get(port, Map, undefined),
+ Path= maps:get(path, Map, undefined),
+ normalize_scheme_based(Map, Scheme, Port, Path).
+%%
+normalize_scheme_based(Map, Scheme, Port, Path)
+ when Scheme =:= "http"; Scheme =:= <<"http">> ->
+ normalize_http(Map, Port, Path);
+normalize_scheme_based(Map, Scheme, Port, Path)
+ when Scheme =:= "https"; Scheme =:= <<"https">> ->
+ normalize_https(Map, Port, Path);
+normalize_scheme_based(Map, Scheme, Port, _Path)
+ when Scheme =:= "ftp"; Scheme =:= <<"ftp">> ->
+ normalize_ftp(Map, Port);
+normalize_scheme_based(Map, Scheme, Port, _Path)
+ when Scheme =:= "ssh"; Scheme =:= <<"ssh">> ->
+ normalize_ssh_sftp(Map, Port);
+normalize_scheme_based(Map, Scheme, Port, _Path)
+ when Scheme =:= "sftp"; Scheme =:= <<"sftp">> ->
+ normalize_ssh_sftp(Map, Port);
+normalize_scheme_based(Map, Scheme, Port, _Path)
+ when Scheme =:= "tftp"; Scheme =:= <<"tftp">> ->
+ normalize_tftp(Map, Port);
+normalize_scheme_based(Map, _, _, _) ->
+ Map.
+
+
+normalize_http(Map, Port, Path) ->
+ M1 = normalize_port(Map, Port, 80),
+ normalize_http_path(M1, Path).
+
+
+normalize_https(Map, Port, Path) ->
+ M1 = normalize_port(Map, Port, 443),
+ normalize_http_path(M1, Path).
+
+
+normalize_ftp(Map, Port) ->
+ normalize_port(Map, Port, 21).
+
+
+normalize_ssh_sftp(Map, Port) ->
+ normalize_port(Map, Port, 22).
+
+
+normalize_tftp(Map, Port) ->
+ normalize_port(Map, Port, 69).
+
+
+normalize_port(Map, Port, Default) ->
+ case Port of
+ Default ->
+ maps:remove(port, Map);
+ _Else ->
+ Map
+ end.
+
+
+normalize_http_path(Map, Path) ->
+ case Path of
+ "" ->
+ Map#{path => "/"};
+ <<>> ->
+ Map#{path => <<"/">>};
+ _Else ->
+ Map
+ end.
generated by cgit v1.2.3 (git 2.25.1) at 2024-10-15 21:03:49 +0000