diff options
30 files changed, 183 insertions, 50 deletions
diff --git a/bootstrap/lib/compiler/ebin/beam_disasm.beam b/bootstrap/lib/compiler/ebin/beam_disasm.beam Binary files differindex 3fe71cfe1b..eb2a13d620 100644 --- a/bootstrap/lib/compiler/ebin/beam_disasm.beam +++ b/bootstrap/lib/compiler/ebin/beam_disasm.beam diff --git a/bootstrap/lib/compiler/ebin/beam_ssa_pre_codegen.beam b/bootstrap/lib/compiler/ebin/beam_ssa_pre_codegen.beam Binary files differindex 329f31f546..f7c9c34c9a 100644 --- a/bootstrap/lib/compiler/ebin/beam_ssa_pre_codegen.beam +++ b/bootstrap/lib/compiler/ebin/beam_ssa_pre_codegen.beam diff --git a/bootstrap/lib/compiler/ebin/beam_validator.beam b/bootstrap/lib/compiler/ebin/beam_validator.beam Binary files differindex 07daf4ed81..79d64b800b 100644 --- a/bootstrap/lib/compiler/ebin/beam_validator.beam +++ b/bootstrap/lib/compiler/ebin/beam_validator.beam diff --git a/bootstrap/lib/kernel/ebin/disk_log_1.beam b/bootstrap/lib/kernel/ebin/disk_log_1.beam Binary files differindex 36beba1f57..17f65d6ceb 100644 --- a/bootstrap/lib/kernel/ebin/disk_log_1.beam +++ b/bootstrap/lib/kernel/ebin/disk_log_1.beam diff --git a/bootstrap/lib/kernel/ebin/erl_distribution.beam b/bootstrap/lib/kernel/ebin/erl_distribution.beam Binary files differindex 6a096155d6..4490f9c81f 100644 --- a/bootstrap/lib/kernel/ebin/erl_distribution.beam +++ b/bootstrap/lib/kernel/ebin/erl_distribution.beam diff --git a/bootstrap/lib/kernel/ebin/file.beam b/bootstrap/lib/kernel/ebin/file.beam Binary files differindex 9c2759adc9..a7a82a1f12 100644 --- a/bootstrap/lib/kernel/ebin/file.beam +++ b/bootstrap/lib/kernel/ebin/file.beam diff --git a/bootstrap/lib/kernel/ebin/user_drv.beam b/bootstrap/lib/kernel/ebin/user_drv.beam Binary files differindex ce4d4328de..1b59423b71 100644 --- a/bootstrap/lib/kernel/ebin/user_drv.beam +++ b/bootstrap/lib/kernel/ebin/user_drv.beam diff --git a/bootstrap/lib/stdlib/ebin/beam_lib.beam b/bootstrap/lib/stdlib/ebin/beam_lib.beam Binary files differindex 44d783a6d0..a96ccb6f8c 100644 --- a/bootstrap/lib/stdlib/ebin/beam_lib.beam +++ b/bootstrap/lib/stdlib/ebin/beam_lib.beam diff --git a/bootstrap/lib/stdlib/ebin/dets_utils.beam b/bootstrap/lib/stdlib/ebin/dets_utils.beam Binary files differindex 25bb217db1..e38eb384a1 100644 --- a/bootstrap/lib/stdlib/ebin/dets_utils.beam +++ b/bootstrap/lib/stdlib/ebin/dets_utils.beam diff --git a/bootstrap/lib/stdlib/ebin/dets_v9.beam b/bootstrap/lib/stdlib/ebin/dets_v9.beam Binary files differindex 1532c51f8e..046de3bd5a 100644 --- a/bootstrap/lib/stdlib/ebin/dets_v9.beam +++ b/bootstrap/lib/stdlib/ebin/dets_v9.beam diff --git a/bootstrap/lib/stdlib/ebin/epp.beam b/bootstrap/lib/stdlib/ebin/epp.beam Binary files differindex a507106263..e3d8c21edf 100644 --- a/bootstrap/lib/stdlib/ebin/epp.beam +++ b/bootstrap/lib/stdlib/ebin/epp.beam diff --git a/bootstrap/lib/stdlib/ebin/erl_pp.beam b/bootstrap/lib/stdlib/ebin/erl_pp.beam Binary files differindex be2fe2c4c0..1965b9e656 100644 --- a/bootstrap/lib/stdlib/ebin/erl_pp.beam +++ b/bootstrap/lib/stdlib/ebin/erl_pp.beam diff --git a/bootstrap/lib/stdlib/ebin/erl_tar.beam b/bootstrap/lib/stdlib/ebin/erl_tar.beam Binary files differindex 526d151a12..3c5004871d 100644 --- a/bootstrap/lib/stdlib/ebin/erl_tar.beam +++ b/bootstrap/lib/stdlib/ebin/erl_tar.beam diff --git a/bootstrap/lib/stdlib/ebin/filename.beam b/bootstrap/lib/stdlib/ebin/filename.beam Binary files differindex 2b76d5bce8..d353a11f3e 100644 --- a/bootstrap/lib/stdlib/ebin/filename.beam +++ b/bootstrap/lib/stdlib/ebin/filename.beam diff --git a/bootstrap/lib/stdlib/ebin/re.beam b/bootstrap/lib/stdlib/ebin/re.beam Binary files differindex d6920d54ce..5662594d4c 100644 --- a/bootstrap/lib/stdlib/ebin/re.beam +++ b/bootstrap/lib/stdlib/ebin/re.beam diff --git a/bootstrap/lib/stdlib/ebin/string.beam b/bootstrap/lib/stdlib/ebin/string.beam Binary files differindex 437f63357c..e6d68ea84e 100644 --- a/bootstrap/lib/stdlib/ebin/string.beam +++ b/bootstrap/lib/stdlib/ebin/string.beam diff --git a/bootstrap/lib/stdlib/ebin/unicode_util.beam b/bootstrap/lib/stdlib/ebin/unicode_util.beam Binary files differindex c9db0483ca..c0523f2e52 100644 --- a/bootstrap/lib/stdlib/ebin/unicode_util.beam +++ b/bootstrap/lib/stdlib/ebin/unicode_util.beam diff --git a/bootstrap/lib/stdlib/ebin/uri_string.beam b/bootstrap/lib/stdlib/ebin/uri_string.beam Binary files differindex 001118f0b9..b57c84a302 100644 --- a/bootstrap/lib/stdlib/ebin/uri_string.beam +++ b/bootstrap/lib/stdlib/ebin/uri_string.beam diff --git a/erts/doc/src/erl.xml b/erts/doc/src/erl.xml index 471d7caa5a..ed1b0880b4 100644 --- a/erts/doc/src/erl.xml +++ b/erts/doc/src/erl.xml @@ -460,7 +460,10 @@ <tag><c><![CDATA[-remsh Node]]></c></tag> <item> <p>Starts Erlang with a remote shell connected to - <c><![CDATA[Node]]></c>.</p> + <c><![CDATA[Node]]></c>. Requires either <c><![CDATA[-name]]></c> + or <c><![CDATA[-sname]]></c> to be given. If <c><![CDATA[Node]]></c> + does not contain a hostname, one is automatically taken from + <c><![CDATA[-name]]></c> or <c><![CDATA[-sname]]></c></p> </item> <tag><c><![CDATA[-rsh Program]]></c></tag> <item> diff --git a/erts/emulator/beam/beam_emu.c b/erts/emulator/beam/beam_emu.c index 8e93e53003..bae64afb97 100644 --- a/erts/emulator/beam/beam_emu.c +++ b/erts/emulator/beam/beam_emu.c @@ -3272,7 +3272,7 @@ erts_current_reductions(Process *c_p, Process *p) } else { reds_left = c_p->fcalls; } - return REDS_IN(c_p) - reds_left; + return REDS_IN(c_p) - reds_left - erts_proc_sched_data(p)->virtual_reds; } int diff --git a/erts/emulator/beam/erl_bif_info.c b/erts/emulator/beam/erl_bif_info.c index 0339589b79..2704b99aa4 100644 --- a/erts/emulator/beam/erl_bif_info.c +++ b/erts/emulator/beam/erl_bif_info.c @@ -768,7 +768,7 @@ static ErtsProcessInfoArgs pi_args[] = { {am_memory, 0, ERTS_PI_FLAG_NEED_MSGQ_LEN|ERTS_PI_FLAG_FORCE_SIG_SEND, ERTS_PROC_LOCK_MAIN}, {am_garbage_collection, 3+2 + 3+2 + 3+2 + 3+2 + 3+2 + ERTS_MAX_HEAP_SIZE_MAP_SZ, 0, ERTS_PROC_LOCK_MAIN}, {am_group_leader, 0, 0, ERTS_PROC_LOCK_MAIN}, - {am_reductions, 0, 0, ERTS_PROC_LOCK_MAIN}, + {am_reductions, 0, ERTS_PI_FLAG_FORCE_SIG_SEND, ERTS_PROC_LOCK_MAIN}, {am_priority, 0, 0, 0}, {am_trace, 0, 0, ERTS_PROC_LOCK_MAIN}, {am_binary, 0, ERTS_PI_FLAG_FORCE_SIG_SEND, ERTS_PROC_LOCK_MAIN}, diff --git a/lib/erl_interface/include/ei.h b/lib/erl_interface/include/ei.h index ed0420300d..b138118f04 100644 --- a/lib/erl_interface/include/ei.h +++ b/lib/erl_interface/include/ei.h @@ -213,12 +213,12 @@ extern volatile int __erl_errno; * library and when using the library we set a value that we use */ -#define EI_MAXHOSTNAMELEN 64 -#define EI_MAXALIVELEN 63 #define EI_MAX_COOKIE_SIZE 512 #define MAXATOMLEN (255 + 1) #define MAXATOMLEN_UTF8 (255*4 + 1) -#define MAXNODELEN EI_MAXALIVELEN+1+EI_MAXHOSTNAMELEN +#define EI_MAXHOSTNAMELEN (MAXATOMLEN - 2) +#define EI_MAXALIVELEN (MAXATOMLEN - 2) +#define MAXNODELEN MAXATOMLEN typedef enum { ERLANG_ASCII = 1, diff --git a/lib/erl_interface/src/connect/ei_connect.c b/lib/erl_interface/src/connect/ei_connect.c index 0cbad235cc..1b1479d2e9 100644 --- a/lib/erl_interface/src/connect/ei_connect.c +++ b/lib/erl_interface/src/connect/ei_connect.c @@ -791,14 +791,17 @@ int ei_connect_init_ussi(ei_cnode* ec, const char* this_node_name, if (strcmp(hp->h_name, "localhost") == 0) { /* We use a short node name */ if ((ct = strchr(thishostname, '.')) != NULL) *ct = '\0'; - sprintf(thisnodename, "%s@%s", this_node_name, thishostname); } else { /* We use a short node name */ if ((ct = strchr(hp->h_name, '.')) != NULL) *ct = '\0'; strcpy(thishostname, hp->h_name); - sprintf(thisnodename, "%s@%s", this_node_name, hp->h_name); } } + if (strlen(this_node_name) + 1 + strlen(thishostname) > MAXNODELEN) { + EI_TRACE_ERR0("ei_connect_init_ussi","this node name is too long"); + return ERL_ERROR; + } + sprintf(thisnodename, "%s@%s", this_node_name, thishostname); res = ei_connect_xinit_ussi(ec, thishostname, thisalivename, thisnodename, (struct in_addr *)*hp->h_addr_list, cookie, creation, cbs, cbs_sz, setup_context); @@ -889,6 +892,11 @@ int ei_connect_tmo(ei_cnode* ec, char *nodename, unsigned ms) int ei_h_errno; #endif /* !win32 */ int res; + + if (strlen(nodename) > MAXNODELEN) { + EI_TRACE_ERR0("ei_connect","Too long nodename"); + return ERL_ERROR; + } /* extract the host and alive parts from nodename */ if (!(hostname = strchr(nodename,'@'))) { diff --git a/lib/kernel/src/user_drv.erl b/lib/kernel/src/user_drv.erl index 08286dd476..69ff8e7971 100644 --- a/lib/kernel/src/user_drv.erl +++ b/lib/kernel/src/user_drv.erl @@ -120,7 +120,7 @@ server1(Iport, Oport, Shell) -> {Curr,Shell1} = case init:get_argument(remsh) of {ok,[[Node]]} -> - ANode = list_to_atom(Node), + ANode = list_to_atom(append_hostname(Node)), RShell = {ANode,shell,start,[]}, RGr = group:start(self(), RShell, rem_sh_opts(ANode)), {RGr,RShell}; @@ -139,6 +139,12 @@ server1(Iport, Oport, Shell) -> %% Enter the server loop. server_loop(Iport, Oport, Curr, User, Gr, {false, queue:new()}). +append_hostname(Node) -> + case string:find(Node, "@") of + nomatch -> Node ++ string:find(atom_to_list(node()), "@"); + _ -> Node + end. + rem_sh_opts(Node) -> [{expand_fun,fun(B)-> rpc:call(Node,edlin_expand,expand,[B]) end}]. diff --git a/lib/ssl/src/tls_connection.erl b/lib/ssl/src/tls_connection.erl index a05858221a..872a557e67 100644 --- a/lib/ssl/src/tls_connection.erl +++ b/lib/ssl/src/tls_connection.erl @@ -171,19 +171,21 @@ next_record(#state{protocol_buffers = connection_states = ConnectionStates, ssl_options = #ssl_options{padding_check = Check}} = State) -> next_record(State, CipherTexts, ConnectionStates, Check); -next_record(#state{protocol_buffers = #protocol_buffers{tls_cipher_texts = []}, - protocol_specific = #{active_n_toggle := true, active_n := N} = ProtocolSpec, +next_record(#state{user_data_buffer = {_,0,_}, + protocol_buffers = #protocol_buffers{tls_cipher_texts = []}, + protocol_specific = #{active_n_toggle := true, + active_n := N} = ProtocolSpec, static_env = #static_env{socket = Socket, close_tag = CloseTag, transport_cb = Transport} - } = State) -> + } = State) -> case tls_socket:setopts(Transport, Socket, [{active, N}]) of ok -> - {no_record, State#state{protocol_specific = ProtocolSpec#{active_n_toggle => false}}}; + {no_record, State#state{protocol_specific = ProtocolSpec#{active_n_toggle => false}}}; _ -> - self() ! {CloseTag, Socket}, - {no_record, State} - end; + self() ! {CloseTag, Socket}, + {no_record, State} + end; next_record(State) -> {no_record, State}. diff --git a/lib/stdlib/src/string.erl b/lib/stdlib/src/string.erl index 1f8bdc5432..a418754caf 100644 --- a/lib/stdlib/src/string.erl +++ b/lib/stdlib/src/string.erl @@ -128,7 +128,8 @@ length(CD) -> to_graphemes(CD0) -> case unicode_util:gc(CD0) of [GC|CD] -> [GC|to_graphemes(CD)]; - [] -> [] + [] -> []; + {error, Err} -> error({badarg, Err}) end. %% Compare two strings return boolean, assumes that the input are @@ -332,7 +333,10 @@ uppercase(<<CP1/utf8, Rest/binary>>=Orig) -> catch unchanged -> Orig end; uppercase(<<>>) -> - <<>>. + <<>>; +uppercase(Bin) -> + error({badarg, Bin}). + %% Lowercase all chars in Str -spec lowercase(String::unicode:chardata()) -> unicode:chardata(). @@ -346,7 +350,10 @@ lowercase(<<CP1/utf8, Rest/binary>>=Orig) -> catch unchanged -> Orig end; lowercase(<<>>) -> - <<>>. + <<>>; +lowercase(Bin) -> + error({badarg, Bin}). + %% Make a titlecase of the first char in Str -spec titlecase(String::unicode:chardata()) -> unicode:chardata(). @@ -375,7 +382,9 @@ casefold(<<CP1/utf8, Rest/binary>>=Orig) -> catch unchanged -> Orig end; casefold(<<>>) -> - <<>>. + <<>>; +casefold(Bin) -> + error({badarg, Bin}). -spec to_integer(String) -> {Int, Rest} | {'error', Reason} when String :: unicode:chardata(), @@ -544,7 +553,8 @@ length_1([CP1|[CP2|_]=Cont], N) when ?ASCII_LIST(CP1,CP2) -> length_1(Str, N) -> case unicode_util:gc(Str) of [] -> N; - [_|Rest] -> length_1(Rest, N+1) + [_|Rest] -> length_1(Rest, N+1); + {error, Err} -> error({badarg, Err}) end. length_b(<<CP2/utf8, Rest/binary>>, CP1, N) @@ -554,7 +564,8 @@ length_b(Bin0, CP1, N) -> [_|Bin1] = unicode_util:gc([CP1|Bin0]), case unicode_util:cp(Bin1) of [] -> N+1; - [CP3|Bin] -> length_b(Bin, CP3, N+1) + [CP3|Bin] -> length_b(Bin, CP3, N+1); + {error, Err} -> error({badarg, Err}) end. equal_1([A|AR], [B|BR]) when is_integer(A), is_integer(B) -> @@ -599,7 +610,8 @@ reverse_1([CP1|[CP2|_]=Cont], Acc) when ?ASCII_LIST(CP1,CP2) -> reverse_1(CD, Acc) -> case unicode_util:gc(CD) of [GC|Rest] -> reverse_1(Rest, [GC|Acc]); - [] -> Acc + [] -> Acc; + {error, Err} -> error({badarg, Err}) end. reverse_b(<<CP2/utf8, Rest/binary>>, CP1, Acc) @@ -609,7 +621,8 @@ reverse_b(Bin0, CP1, Acc) -> [GC|Bin1] = unicode_util:gc([CP1|Bin0]), case unicode_util:cp(Bin1) of [] -> [GC|Acc]; - [CP3|Bin] -> reverse_b(Bin, CP3, [GC|Acc]) + [CP3|Bin] -> reverse_b(Bin, CP3, [GC|Acc]); + {error, Err} -> error({badarg, Err}) end. slice_l0(<<CP1/utf8, Bin/binary>>, N) when N > 0 -> @@ -622,7 +635,8 @@ slice_l([CP1|[CP2|_]=Cont], N) when ?ASCII_LIST(CP1,CP2),N > 0 -> slice_l(CD, N) when N > 0 -> case unicode_util:gc(CD) of [_|Cont] -> slice_l(Cont, N-1); - [] -> [] + [] -> []; + {error, Err} -> error({badarg, Err}) end; slice_l(Cont, 0) -> Cont. @@ -634,7 +648,8 @@ slice_lb(Bin, CP1, N) -> if N > 1 -> case unicode_util:cp(Rest) of [CP2|Cont] -> slice_lb(Cont, CP2, N-1); - [] -> <<>> + [] -> <<>>; + {error, Err} -> error({badarg, Err}) end; N =:= 1 -> Rest @@ -647,7 +662,10 @@ slice_trail(Orig, N) when is_binary(Orig) -> Sz = byte_size(Orig) - Length, <<Keep:Sz/binary, _/binary>> = Orig, Keep; - _ -> <<>> + <<_, _/binary>> when N > 0 -> + error({badarg, Orig}); + _ -> + <<>> end; slice_trail(CD, N) when is_list(CD) -> slice_list(CD, N). @@ -657,7 +675,8 @@ slice_list([CP1|[CP2|_]=Cont], N) when ?ASCII_LIST(CP1,CP2),N > 0 -> slice_list(CD, N) when N > 0 -> case unicode_util:gc(CD) of [GC|Cont] -> append(GC, slice_list(Cont, N-1)); - [] -> [] + [] -> []; + {error, Err} -> error({badarg, Err}) end; slice_list(_, 0) -> []. @@ -668,7 +687,8 @@ slice_bin(CD, CP1, N) when N > 0 -> [_|Bin] = unicode_util:gc([CP1|CD]), case unicode_util:cp(Bin) of [CP2|Cont] -> slice_bin(Cont, CP2, N-1); - [] -> 0 + [] -> 0; + {error, Err} -> error({badarg, Err}) end; slice_bin(CD, CP1, 0) -> byte_size(CD)+byte_size(<<CP1/utf8>>). @@ -703,14 +723,18 @@ uppercase_bin(CP1, Bin, Changed) -> [] when Changed -> [CP1]; [] -> - throw(unchanged) + throw(unchanged); + {error, Err} -> + error({badarg, Err}) end; [Char|CPs] -> case unicode_util:cp(CPs) of [Next|Rest] -> [Char|uppercase_bin(Next, Rest, true)]; [] -> - [Char] + [Char]; + {error, Err} -> + error({badarg, Err}) end end. @@ -744,14 +768,18 @@ lowercase_bin(CP1, Bin, Changed) -> [] when Changed -> [CP1]; [] -> - throw(unchanged) + throw(unchanged); + {error, Err} -> + error({badarg, Err}) end; [Char|CPs] -> case unicode_util:cp(CPs) of [Next|Rest] -> [Char|lowercase_bin(Next, Rest, true)]; [] -> - [Char] + [Char]; + {error, Err} -> + error({badarg, Err}) end end. @@ -785,14 +813,18 @@ casefold_bin(CP1, Bin, Changed) -> [] when Changed -> [CP1]; [] -> - throw(unchanged) + throw(unchanged); + {error, Err} -> + error({badarg, Err}) end; [Char|CPs] -> case unicode_util:cp(CPs) of [Next|Rest] -> [Char|casefold_bin(Next, Rest, true)]; [] -> - [Char] + [Char]; + {error, Err} -> + error({badarg, Err}) end end. @@ -1634,7 +1666,9 @@ bin_search_inv_1(<<CP1/utf8, BinRest/binary>>=Bin0, Cont, Sep) -> bin_search_inv_1(<<>>, Cont, _Sep) -> {nomatch, Cont}; bin_search_inv_1([], Cont, _Sep) -> - {nomatch, Cont}. + {nomatch, Cont}; +bin_search_inv_1(Bin, _, _) -> + error({badarg, Bin}). bin_search_inv_n(<<CP1/utf8, BinRest/binary>>=Bin0, Cont, Seps) -> @@ -1666,7 +1700,9 @@ bin_search_inv_n(<<CP1/utf8, BinRest/binary>>=Bin0, Cont, Seps) -> bin_search_inv_n(<<>>, Cont, _Sep) -> {nomatch, Cont}; bin_search_inv_n([], Cont, _Sep) -> - {nomatch, Cont}. + {nomatch, Cont}; +bin_search_inv_n(Bin, _, _) -> + error({badarg, Bin}). bin_search_str(Bin0, Start, [], SearchCPs) -> Compiled = binary:compile_pattern(unicode:characters_to_binary(SearchCPs)), diff --git a/lib/stdlib/test/string_SUITE.erl b/lib/stdlib/test/string_SUITE.erl index 248912c3f2..c9aadd7f10 100644 --- a/lib/stdlib/test/string_SUITE.erl +++ b/lib/stdlib/test/string_SUITE.erl @@ -103,6 +103,15 @@ debug() -> test(?LINE,?FUNCTION_NAME,B,C,D, false), test(?LINE,?FUNCTION_NAME,hd(C),[B|tl(C)],D, false)). +-define(TRY(Exp), + fun() -> + try Exp + catch _E:Reason:_ST -> + %% io:format("~p:~w: ~p: ~.0p ~p~n", + %% [?FUNCTION_NAME, ?LINE,_E,Reason, hd(_ST)]), + {'EXIT', Reason} + end + end()). is_empty(_) -> ?TEST("", [], true), @@ -126,6 +135,10 @@ length(_) -> ?TEST(["abc"|<<"abc">>], [], 6), ?TEST(["abc",["def"]], [], 6), ?TEST([<<97/utf8, 778/utf8, 98/utf8>>, [776,111,776]], [], 3), %% åäö in nfd + + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:length(InvalidUTF8)), + {'EXIT', {badarg, _}} = ?TRY(string:length(<<$a, InvalidUTF8/binary, $z>>)), ok. equal(_) -> @@ -226,6 +239,8 @@ to_graphemes(_) -> true = erlang:length(GCs) =:= erlang:length(string:to_graphemes(NFD)), true = erlang:length(GCs) =:= erlang:length(string:to_graphemes(unicode:characters_to_nfc_list(String))), + + {'EXIT', {badarg, _}} = ?TRY(string:to_graphemes(<<$a,192,192,$z>>)), ok. reverse(_) -> @@ -238,6 +253,11 @@ reverse(_) -> ?TEST(Str2, [], lists:reverse(Str2)), ?TEST(Str3, [], lists:reverse(Str3)), true = string:reverse(Str3) =:= lists:reverse(string:to_graphemes(Str3)), + + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:reverse(InvalidUTF8)), + {'EXIT', {badarg, _}} = ?TRY(string:reverse(<<$a, InvalidUTF8/binary, $z>>)), + ok. slice(_) -> @@ -258,6 +278,14 @@ slice(_) -> ?TEST([<<"aå"/utf8>>,"äöbcd"], [3,3], "öbc"), ?TEST([<<"aåä"/utf8>>,"öbcd"], [3,10], "öbcd"), + InvalidUTF8 = <<192,192>>, + [$b, $c|InvalidUTF8] = string:slice(["abc", InvalidUTF8], 1), + InvalidUTF8 = string:slice(["abc", InvalidUTF8], 3), + {'EXIT', {badarg, _}} = ?TRY(string:slice(["abc", InvalidUTF8], 1, 5)), + BadUtf8 = <<$a, InvalidUTF8/binary, "teststring">>, + {'EXIT', {badarg, _}} = ?TRY(string:slice(BadUtf8, 2)), + {'EXIT', {badarg, _}} = ?TRY(string:slice(BadUtf8, 1, 5)), + {'EXIT', {badarg, _}} = ?TRY(string:slice(BadUtf8, 0, 5)), ok. pad(_) -> @@ -270,6 +298,10 @@ pad(_) -> ?TEST(Str, [10, trailing, $.], "Hallå....."), ?TEST(Str++["f"], [10, trailing, $.], "Hallåf...."), ?TEST(Str++[" flåwer"], [10, trailing, $.], "Hallå flåwer"), + + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:pad(InvalidUTF8, 10, both, $.)), + {'EXIT', {badarg, _}} = ?TRY(string:pad(<<$a, InvalidUTF8/binary, $z>>, 10, both, $.)), ok. trim(_) -> @@ -300,6 +332,11 @@ trim(_) -> ?TEST([[<<"!v">>|<<204,128,$v,204,129>>]],[trailing, [[$v,769]]], [$!,$v,768]), ?TEST([[[<<"v">>|<<204,129,118,204,128,118>>],769,118,769]], [trailing, [[118,769]]], [$v,769,$v,768]), ?TEST([<<"vv">>|<<204,128,118,204,128>>], [trailing, [[118,768]]], "v"), + + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:trim(InvalidUTF8, both, "az")), + %% Not checked (using binary search) + %% {'EXIT', {badarg, _}} = ?TRY(string:trim(<<$a, $b, InvalidUTF8/binary, $z>>, both, "az")), ok. chomp(_) -> @@ -400,6 +437,13 @@ take(_) -> ?TEST([<<"e">>,778,"åäöe", <<778/utf8>>, $e, 779], [[[$e,778]], true, trailing], {[$e,778]++"åäöe"++[778], [$e,779]}), + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:take(InvalidUTF8, [$.], false, leading)), + %% Not checked (using binary search) + %% {'EXIT', {badarg, _}} = ?TRY(string:take(InvalidUTF8, [$.], true, leading)), + %% {'EXIT', {badarg, _}} = ?TRY(string:take(InvalidUTF8, [$.], false, trailing)), + {'EXIT', {badarg, _}} = ?TRY(string:take(InvalidUTF8, [$.], true, trailing)), + ok. @@ -416,6 +460,11 @@ uppercase(_) -> ?TEST("ljLJ", [], "LJLJ"), ?TEST("LJlj", [], "LJLJ"), ?TEST("ß sharp s", [], "SS SHARP S"), + + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:uppercase(InvalidUTF8)), + {'EXIT', {badarg, _}} = ?TRY(string:uppercase(<<$a, InvalidUTF8/binary, $z>>)), + ok. lowercase(_) -> @@ -429,6 +478,10 @@ lowercase(_) -> ?TEST(["Mic",<<"HAŁ"/utf8>>], [], "michał"), ?TEST("ß SHARP S", [], "ß sharp s"), ?TEST("İ I WITH DOT ABOVE", [], "i̇ i with dot above"), + + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:lowercase(InvalidUTF8)), + {'EXIT', {badarg, _}} = ?TRY(string:lowercase(<<$a, InvalidUTF8/binary, $z>>)), ok. titlecase(_) -> @@ -442,6 +495,10 @@ titlecase(_) -> ?TEST("ljLJ", [], "LjLJ"), ?TEST("LJlj", [], "Ljlj"), ?TEST("ß sharp s", [], "Ss sharp s"), + + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:titlecase(InvalidUTF8)), + <<$A, _/binary>> = ?TRY(string:titlecase(<<$a, InvalidUTF8/binary, $z>>)), ok. casefold(_) -> @@ -456,6 +513,10 @@ casefold(_) -> ?TEST("ß SHARP S", [], "ss sharp s"), ?TEST("ẞ SHARP S", [], "ss sharp s"), ?TEST("İ I WITH DOT ABOVE", [], "i̇ i with dot above"), + + InvalidUTF8 = <<192,192>>, + {'EXIT', {badarg, _}} = ?TRY(string:casefold(InvalidUTF8)), + {'EXIT', {badarg, _}} = ?TRY(string:casefold(<<$a, InvalidUTF8/binary, $z>>)), ok. @@ -740,7 +801,7 @@ meas(Config) -> _ -> % No scaling, run at most 1.5 min Tester = spawn(Exec), receive {test_done, Tester} -> ok - after 90000 -> + after 118000 -> io:format("Timelimit reached stopping~n",[]), exit(Tester, die) end, diff --git a/lib/stdlib/test/unicode_util_SUITE.erl b/lib/stdlib/test/unicode_util_SUITE.erl index 044b4e5834..6f55f204f4 100644 --- a/lib/stdlib/test/unicode_util_SUITE.erl +++ b/lib/stdlib/test/unicode_util_SUITE.erl @@ -428,7 +428,15 @@ mode(deep_l, Bin) -> [unicode:characters_to_list(Bin)]. fetch(Str, F) -> case F(Str) of [] -> []; - [CP|R] -> [CP|fetch(R,F)] + [CP|R] -> + %% If input is a binary R should be binary + if is_binary(Str) == false -> ok; + is_binary(R); R =:= [] -> ok; + true -> + io:format("Char: ~tc Tail:~tP~n", [CP,R,10]), + exit({bug, F}) + end, + [CP|fetch(R,F)] end. %% *Test.txt file helpers diff --git a/lib/stdlib/uc_spec/gen_unicode_mod.escript b/lib/stdlib/uc_spec/gen_unicode_mod.escript index 8636c69a0d..de67b18afc 100644 --- a/lib/stdlib/uc_spec/gen_unicode_mod.escript +++ b/lib/stdlib/uc_spec/gen_unicode_mod.escript @@ -202,7 +202,8 @@ gen_static(Fd) -> io:put_chars(Fd, " {Upper,_} -> [Upper|Str];\n"), io:put_chars(Fd, " {Upper,_,_,_} -> [Upper|Str]\n"), io:put_chars(Fd, " end;\n"), - io:put_chars(Fd, " [] -> []\n"), + io:put_chars(Fd, " [] -> [];\n"), + io:put_chars(Fd, " {error,Err} -> error({badarg, Err})\n"), io:put_chars(Fd, " end.\n\n"), io:put_chars(Fd, "-spec lowercase(unicode:chardata()) -> " "maybe_improper_list(gc(),unicode:chardata()).\n"), @@ -213,7 +214,8 @@ gen_static(Fd) -> io:put_chars(Fd, " {_,Lower} -> [Lower|Str];\n"), io:put_chars(Fd, " {_,Lower,_,_} -> [Lower|Str]\n"), io:put_chars(Fd, " end;\n"), - io:put_chars(Fd, " [] -> []\n"), + io:put_chars(Fd, " [] -> [];\n"), + io:put_chars(Fd, " {error,Err} -> error({badarg, Err})\n"), io:put_chars(Fd, " end.\n\n"), io:put_chars(Fd, "-spec titlecase(unicode:chardata()) -> " "maybe_improper_list(gc(),unicode:chardata()).\n"), @@ -224,7 +226,8 @@ gen_static(Fd) -> io:put_chars(Fd, " {_,_,Title,_} -> [Title|Str];\n"), io:put_chars(Fd, " {Upper,_} -> [Upper|Str]\n"), io:put_chars(Fd, " end;\n"), - io:put_chars(Fd, " [] -> []\n"), + io:put_chars(Fd, " [] -> [];\n"), + io:put_chars(Fd, " {error,Err} -> error({badarg, Err})\n"), io:put_chars(Fd, " end.\n\n"), io:put_chars(Fd, "-spec casefold(unicode:chardata()) -> " "maybe_improper_list(gc(),unicode:chardata()).\n"), @@ -235,7 +238,8 @@ gen_static(Fd) -> io:put_chars(Fd, " {_,_,_,Fold} -> [Fold|Str];\n"), io:put_chars(Fd, " {_,Lower} -> [Lower|Str]\n"), io:put_chars(Fd, " end;\n"), - io:put_chars(Fd, " [] -> []\n"), + io:put_chars(Fd, " [] -> [];\n"), + io:put_chars(Fd, " {error,Err} -> error({badarg, Err})\n"), io:put_chars(Fd, " end.\n\n"), ok. @@ -619,7 +623,7 @@ gen_gc(Fd, GBP) -> GenHangulT = fun(Range) -> io:format(Fd, "gc_1~s gc_h_T(R1,[CP]);\n", [gen_clause(Range)]) end, [GenHangulT(CP) || CP <- merge_ranges(maps:get(t,GBP))], io:put_chars(Fd, "%% Handle Hangul LV and LVT special, since they are large\n"), - io:put_chars(Fd, "gc_1([CP|_]=R0) when 44000 < CP, CP < 56000 -> gc_h_lv_lvt(R0, []);\n"), + io:put_chars(Fd, "gc_1([CP|_]=R0) when 44000 < CP, CP < 56000 -> gc_h_lv_lvt(R0, R0, []);\n"), io:put_chars(Fd, "\n%% Handle Regional\n"), GenRegional = fun(Range) -> io:format(Fd, "gc_1~s gc_regional(R1,CP);\n", [gen_clause(Range)]) end, @@ -748,7 +752,7 @@ gen_gc(Fd, GBP) -> GenHangulL_2 = fun(Range) -> io:format(Fd, "~8c~s gc_h_V(R1,[CP|Acc]);\n", [$\s,gen_case_clause(Range)]) end, [GenHangulL_2(CP) || CP <- merge_ranges(maps:get(v,GBP))], - io:put_chars(Fd, " R1 -> gc_h_lv_lvt(R1, Acc)\n end.\n\n"), + io:put_chars(Fd, " R1 -> gc_h_lv_lvt(R1, R0, Acc)\n end.\n\n"), io:put_chars(Fd, "%% Handle Hangul V\n"), io:put_chars(Fd, "gc_h_V(R0, Acc) ->\n case cp(R0) of\n"), @@ -783,10 +787,10 @@ gen_gc(Fd, GBP) -> GenHangulLVT = fun(Range) -> io:format(Fd, "gc_h_lv_lvt~s gc_h_T(R1,[CP|Acc]);\n", [gen_clause2(Range)]) end, [GenHangulLVT(CP) || CP <- merge_ranges(maps:get(lvt,GBP))], - io:put_chars(Fd, "gc_h_lv_lvt([CP|R], []) -> gc_extend(cp(R), R, CP);\n"), %% From gc_1/1 + io:put_chars(Fd, "gc_h_lv_lvt([CP|R1], _, []) -> gc_extend(cp(R1), R1, CP);\n"), %% From gc_1/1 io:put_chars(Fd, "%% Also handles error tuples\n"), - io:put_chars(Fd, "gc_h_lv_lvt(R, [CP]) -> gc_extend(R, R, CP);\n"), - io:put_chars(Fd, "gc_h_lv_lvt(R, Acc) -> gc_extend2(R, R, Acc).\n\n"), + io:put_chars(Fd, "gc_h_lv_lvt(R1, R0, [CP]) -> gc_extend(R1, R0, CP);\n"), + io:put_chars(Fd, "gc_h_lv_lvt(R1, R0, Acc) -> gc_extend2(R1, R0, Acc).\n\n"), ok. gen_compose_pairs(Fd, ExclData, Data) -> @@ -887,9 +891,9 @@ gen_clause({R0, R1}) -> io_lib:format("([CP|R1]=R0) when ~w =< CP, CP =< ~w ->", [R0,R1]). gen_clause2({R0, undefined}) -> - io_lib:format("([~w=CP|R1], Acc) ->", [R0]); + io_lib:format("([~w=CP|R1], R0, Acc) ->", [R0]); gen_clause2({R0, R1}) -> - io_lib:format("([CP|R1], Acc) when ~w =< CP, CP =< ~w ->", [R0,R1]). + io_lib:format("([CP|R1], R0, Acc) when ~w =< CP, CP =< ~w ->", [R0,R1]). gen_case_clause({R0, undefined}) -> io_lib:format("[~w=CP|R1] ->", [R0]); diff --git a/lib/tools/doc/src/make.xml b/lib/tools/doc/src/make.xml index af2404707f..322d77323f 100644 --- a/lib/tools/doc/src/make.xml +++ b/lib/tools/doc/src/make.xml @@ -145,5 +145,10 @@ Modules. {'*',[debug_info]}. </code> <p></p> </section> + + <section> + <title>See Also</title> + <p><seealso marker="compiler:compile"><c>compile(3)</c></seealso></p> + </section> </erlref> |