stdlib: Minor unicode_util opts

Exit early for Latin-1

3 files changed, 33 insertions, 16 deletions

diff --git a/lib/stdlib/test/string_SUITE.erl b/lib/stdlib/test/string_SUITE.erl
index 90f980c0e5..3e50a0e3c1 100644
--- a/lib/stdlib/test/string_SUITE.erl
+++ b/lib/stdlib/test/string_SUITE.erl
@@ -92,7 +92,7 @@ end_per_testcase(_Case, _Config) ->
     ok.
 
 debug() ->
-    Config = [{data_dir, ?MODULE_STRING++"_data"}],
+    Config = [{data_dir, "./" ++ ?MODULE_STRING++"_data"}],
     [io:format("~p:~p~n",[Test,?MODULE:Test(Config)]) ||
         {_,Tests} <- groups(), Test <- Tests].
 
@@ -717,19 +717,19 @@ meas(Config) ->
         {_,{_,Scale}} when Scale > 1 ->
             {skip,{will_not_run_in_debug,Scale}};
         _ -> % No scaling
-            DataDir = proplists:get_value(data_dir, Config),
-            TestDir = filename:dirname(string:trim(DataDir, trailing, "/")),
-            do_measure(TestDir)
+            DataDir0 = proplists:get_value(data_dir, Config),
+            DataDir = filename:join(lists:droplast(filename:split(DataDir0))),
+            do_measure(DataDir)
     end.
 
-do_measure(TestDir) ->
-    File =  filename:join(TestDir, ?MODULE_STRING ++ ".erl"),
+do_measure(DataDir) ->
+    File =  filename:join([DataDir,"unicode_util_SUITE_data","NormalizationTest.txt"]),
     io:format("File ~s ",[File]),
     {ok, Bin} = file:read_file(File),
     io:format("~p~n",[byte_size(Bin)]),
     Do = fun(Name, Func, Mode) ->
                  {N, Mean, Stddev, _} = time_func(Func, Mode, Bin),
-                 io:format("~10w ~6w ~6.2fms ±~4.2fms #~.2w gc included~n",
+                 io:format("~10w ~6w ~6.2fms ±~5.2fms #~.2w gc included~n",
                            [Name, Mode, Mean/1000, Stddev/1000, N])
          end,
     io:format("----------------------~n"),
@@ -947,7 +947,7 @@ time_func(Fun, Mode, Bin) ->
                      end),
     receive {Pid,Msg} -> Msg end.
 
-time_func(N,Sum,SumSq, Fun, Str, _) when N < 50 ->
+time_func(N,Sum,SumSq, Fun, Str, _) when N < 20 ->
     {Time, Res} = timer:tc(fun() -> Fun(Str) end),
     time_func(N+1,Sum+Time,SumSq+Time*Time, Fun, Str, Res);
 time_func(N,Sum,SumSq, _, _, Res) ->
diff --git a/lib/stdlib/test/unicode_util_SUITE.erl b/lib/stdlib/test/unicode_util_SUITE.erl
index 03c24c7027..f23e6b8b9a 100644
--- a/lib/stdlib/test/unicode_util_SUITE.erl
+++ b/lib/stdlib/test/unicode_util_SUITE.erl
@@ -312,7 +312,7 @@ get(_) ->
 count(Config) ->
     ct:timetrap({minutes,5}),
     case ct:get_timetrap_info() of
-        {_,{_,Scale}} ->
+        {_,{_,Scale}} when Scale > 1 ->
             {skip,{measurments_skipped_debug,Scale}};
         _ -> % No scaling
             do_measure(Config)
diff --git a/lib/stdlib/uc_spec/gen_unicode_mod.escript b/lib/stdlib/uc_spec/gen_unicode_mod.escript
index fefd7d3b70..73c351e1af 100755
--- a/lib/stdlib/uc_spec/gen_unicode_mod.escript
+++ b/lib/stdlib/uc_spec/gen_unicode_mod.escript
@@ -170,7 +170,7 @@ gen_header(Fd) ->
     io:put_chars(Fd, "-export([spec_version/0, lookup/1, get_case/1]).\n"),
     io:put_chars(Fd, "-inline([class/1]).\n"),
     io:put_chars(Fd, "-compile(nowarn_unused_vars).\n"),
-    io:put_chars(Fd, "-dialyzer({no_improper_lists, [cp/1, gc_prepend/2, gc_e_cont/2]}).\n"),
+    io:put_chars(Fd, "-dialyzer({no_improper_lists, [cp/1, gc/1, gc_prepend/2, gc_e_cont/2]}).\n"),
     io:put_chars(Fd, "-type gc() :: char()|[char()].\n\n\n"),
     ok.
 
@@ -240,7 +240,7 @@ gen_norm(Fd) ->
                  "-spec nfd(unicode:chardata()) -> maybe_improper_list(gc(),unicode:chardata()) | {error, unicode:chardata()}.\n"
                  "nfd(Str0) ->\n"
                  "    case gc(Str0) of\n"
-                 "        [GC|R] when GC < 127 -> [GC|R];\n"
+                 "        [GC|R] when GC < 128 -> [GC|R];\n"
                  "        [GC|Str] -> [decompose(GC)|Str];\n"
                  "        [] -> [];\n"
                  "        {error,_}=Error -> Error\n    end.\n\n"
@@ -250,7 +250,7 @@ gen_norm(Fd) ->
                  "-spec nfkd(unicode:chardata()) -> maybe_improper_list(gc(),unicode:chardata()) | {error, unicode:chardata()}.\n"
                  "nfkd(Str0) ->\n"
                  "    case gc(Str0) of\n"
-                 "        [GC|R] when GC < 127 -> [GC|R];\n"
+                 "        [GC|R] when GC < 128 -> [GC|R];\n"
                  "        [GC|Str] -> [decompose_compat(GC)|Str];\n"
                  "        [] -> [];\n"
                  "        {error,_}=Error -> Error\n    end.\n\n"
@@ -260,7 +260,7 @@ gen_norm(Fd) ->
                  "-spec nfc(unicode:chardata()) -> maybe_improper_list(gc(),unicode:chardata()) | {error, unicode:chardata()}.\n"
                  "nfc(Str0) ->\n"
                  "    case gc(Str0) of\n"
-                 "        [GC|R] when GC < 255 -> [GC|R];\n"
+                 "        [GC|R] when GC < 256 -> [GC|R];\n"
                  "        [GC|Str] -> [compose(decompose(GC))|Str];\n"
                  "        [] -> [];\n"
                  "        {error,_}=Error -> Error\n    end.\n\n"
@@ -270,7 +270,7 @@ gen_norm(Fd) ->
                  "-spec nfkc(unicode:chardata()) -> maybe_improper_list(gc(),unicode:chardata()) | {error, unicode:chardata()}.\n"
                  "nfkc(Str0) ->\n"
                  "    case gc(Str0) of\n"
-                 "        [GC|R] when GC < 127 -> [GC|R];\n"
+                 "        [GC|R] when GC < 128 -> [GC|R];\n"
                  "        [GC|Str] -> [compose_compat_0(decompose_compat(GC))|Str];\n"
                  "        [] -> [];\n"
                  "        {error,_}=Error -> Error\n    end.\n\n"
@@ -476,13 +476,30 @@ gen_gc(Fd, GBP) ->
                  "-spec gc(String::unicode:chardata()) ->"
                  " maybe_improper_list() | {error, unicode:chardata()}.\n"),
     io:put_chars(Fd,
+                 "gc([CP1, CP2|_]=T)\n"
+                 "  when CP1 < 256, CP2 < 256, CP1 =/= $\r -> %% Ascii Fast path\n"
+                 "       T;\n"
+                 "gc(<<CP1/utf8, Rest/binary>>) ->\n"
+                 "    if CP1 < 256, CP1 =/= $\r ->\n"
+                 "           case Rest of\n"
+                 "               <<CP2/utf8, _/binary>> when CP2 < 256 -> %% Ascii Fast path\n"
+                 "                   [CP1|Rest];\n"
+                 "               _ -> gc_1([CP1|Rest])\n"
+                 "           end;\n"
+                 "      true -> gc_1([CP1|Rest])\n"
+                 "    end;\n"
                  "gc(Str) ->\n"
                  "    gc_1(cp(Str)).\n\n"
                  "gc_1([$\\r|R0] = R) ->\n"
                  "    case cp(R0) of % Don't break CRLF\n"
                  "        [$\\n|R1] -> [[$\\r,$\\n]|R1];\n"
                  "        _ -> R\n"
-                 "    end;\n"),
+                 "    end;\n"
+                 %% "gc_1([CP1, CP2|_]=T) when CP1 < 256, CP2 < 256 ->\n"
+                 %% "    T;  %% Fast path\n"
+                 %% "gc_1([CP1|<<CP2/utf8, _/binary>>]=T) when CP1 < 256, CP2 < 256 ->\n"
+                 %% "    T;  %% Fast path\n"
+                ),
 
     io:put_chars(Fd, "%% Handle control\n"),
     GenControl = fun(Range) -> io:format(Fd, "gc_1~s R0;\n", [gen_clause(Range)]) end,
@@ -490,7 +507,7 @@ gen_gc(Fd, GBP) ->
     [R1,R2,R3|Crs] = CRs0,
     [GenControl(CP) || CP <- merge_ranges([R1,R2,R3], split), CP =/= {$\r, undefined}],
     %%GenControl(R1),GenControl(R2),GenControl(R3),
-    io:format(Fd, "gc_1([CP|R]) when CP < 255 -> gc_extend(R,CP);\n", []),
+    io:format(Fd, "gc_1([CP|R]) when CP < 256 -> gc_extend(R,CP);\n", []),
     [GenControl(CP) || CP <- Crs],
     %% One clause per CP
     %% CRs0 = merge_ranges(maps:get(cr, GBP) ++ maps:get(lf, GBP) ++ maps:get(control, GBP)),