From cabcd537c30e2c0a4496ee9291cfd4e7713645c1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bj=C3=B6rn=20Gustavsson?= <bjorn@erlang.org>
Date: Mon, 16 Feb 2015 13:24:39 +0100
Subject: Modernize and strengthen the test case for string:tokens/2

---
 lib/stdlib/test/string_SUITE.erl | 40 +++++++++++++++++++++++++++++-----------
 1 file changed, 29 insertions(+), 11 deletions(-)

(limited to 'lib')

diff --git a/lib/stdlib/test/string_SUITE.erl b/lib/stdlib/test/string_SUITE.erl
index fccd1bef95..a55c710d50 100644
--- a/lib/stdlib/test/string_SUITE.erl
+++ b/lib/stdlib/test/string_SUITE.erl
@@ -217,21 +217,39 @@ substr(Config) when is_list(Config) ->
     ?line {'EXIT',_} = (catch string:substr("1234", "1")),
     ok.
 
-tokens(suite) ->
-    [];
-tokens(doc) ->
-    [];
 tokens(Config) when is_list(Config) ->
-    ?line [] = string:tokens("",""),
-    ?line [] = string:tokens("abc","abc"),
-    ?line ["abc"] = string:tokens("abc", ""),
-    ?line ["1","2 34","4","5"] = string:tokens("1,2 34,4;5", ";,"),
-    %% invalid arg type
-    ?line {'EXIT',_} = (catch string:tokens('x,y', ",")), 
+    [] = string:tokens("",""),
+    [] = string:tokens("abc","abc"),
+    ["abc"] = string:tokens("abc", ""),
+    ["1","2 34","45","5","6","7"] = do_tokens("1,2 34,45;5,;6;,7", ";,"),
+
     %% invalid arg type
-    ?line {'EXIT',_} = (catch string:tokens("x,y", ',')), 
+    {'EXIT',_} = (catch string:tokens('x,y', ",")),
+    {'EXIT',_} = (catch string:tokens("x,y", ',')),
     ok.
 
+do_tokens(S0, Sep0) ->
+    [H|T] = Sep0,
+    S = [replace_sep(C, T, H) || C <- S0],
+    Sep = [H],
+    io:format("~p ~p\n", [S0,Sep0]),
+    io:format("~p ~p\n", [S,Sep]),
+
+    Res = string:tokens(S0, Sep0),
+    Res = string:tokens(Sep0++S0, Sep0),
+    Res = string:tokens(S0++Sep0, Sep0),
+
+    Res = string:tokens(S, Sep),
+    Res = string:tokens(Sep++S, Sep),
+    Res = string:tokens(S++Sep, Sep),
+
+    Res.
+
+replace_sep(C, Seps, New) ->
+    case lists:member(C, Seps) of
+	true -> New;
+	false -> C
+    end.
 
 chars(suite) ->
     [];
-- 
cgit v1.2.3


From 53288b441ec721ce3bbdcc4ad65b75e11acc5e1b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bj=C3=B6rn=20Gustavsson?= <bjorn@erlang.org>
Date: Mon, 16 Feb 2015 13:53:25 +0100
Subject: Optimize string:tokens/2

We can save some time by reversing the original string
before starting the tokenization.

When there is only one separator, we can save even more time
by treating that case specially so that we don't have to call
lists:member/2 for each character.
---
 lib/stdlib/src/string.erl | 46 +++++++++++++++++++++++++++++++++++-----------
 1 file changed, 35 insertions(+), 11 deletions(-)

(limited to 'lib')

diff --git a/lib/stdlib/src/string.erl b/lib/stdlib/src/string.erl
index f9b083a56d..f6903d1c3d 100644
--- a/lib/stdlib/src/string.erl
+++ b/lib/stdlib/src/string.erl
@@ -221,23 +221,47 @@ substr2([_|String], S) -> substr2(String, S-1).
       Tokens :: [Token :: nonempty_string()].
 
 tokens(S, Seps) ->
-    tokens1(S, Seps, []).
+    case Seps of
+	[] ->
+	    case S of
+		[] -> [];
+		[_|_] -> [S]
+	    end;
+	[C] ->
+	    tokens_single_1(reverse(S), C, []);
+	[_|_] ->
+	    tokens_multiple_1(reverse(S), Seps, [])
+    end.
 
-tokens1([C|S], Seps, Toks) ->
+tokens_single_1([Sep|S], Sep, Toks) ->
+    tokens_single_1(S, Sep, Toks);
+tokens_single_1([C|S], Sep, Toks) ->
+    tokens_single_2(S, Sep, Toks, [C]);
+tokens_single_1([], _, Toks) ->
+    Toks.
+
+tokens_single_2([Sep|S], Sep, Toks, Tok) ->
+    tokens_single_1(S, Sep, [Tok|Toks]);
+tokens_single_2([C|S], Sep, Toks, Tok) ->
+    tokens_single_2(S, Sep, Toks, [C|Tok]);
+tokens_single_2([], _Sep, Toks, Tok) ->
+    [Tok|Toks].
+
+tokens_multiple_1([C|S], Seps, Toks) ->
     case member(C, Seps) of
-	true -> tokens1(S, Seps, Toks);
-	false -> tokens2(S, Seps, Toks, [C])
+	true -> tokens_multiple_1(S, Seps, Toks);
+	false -> tokens_multiple_2(S, Seps, Toks, [C])
     end;
-tokens1([], _Seps, Toks) ->
-    reverse(Toks).
+tokens_multiple_1([], _Seps, Toks) ->
+    Toks.
 
-tokens2([C|S], Seps, Toks, Cs) ->
+tokens_multiple_2([C|S], Seps, Toks, Tok) ->
     case member(C, Seps) of
-	true -> tokens1(S, Seps, [reverse(Cs)|Toks]);
-	false -> tokens2(S, Seps, Toks, [C|Cs])
+	true -> tokens_multiple_1(S, Seps, [Tok|Toks]);
+	false -> tokens_multiple_2(S, Seps, Toks, [C|Tok])
     end;
-tokens2([], _Seps, Toks, Cs) ->
-    reverse([reverse(Cs)|Toks]).
+tokens_multiple_2([], _Seps, Toks, Tok) ->
+    [Tok|Toks].
 
 -spec chars(Character, Number) -> String when
       Character :: char(),
-- 
cgit v1.2.3