* parse.y (rb_check_id_cstr): new function to check if ID is

registered with NUL-terminated C string.
* sprintf.c (rb_str_format): avoid inadvertent symbol creation.

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@35300 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

ChangeLog

@@ -1,3 +1,10 @@
+Wed Apr 11 22:31:19 2012  Nobuyoshi Nakada  <nobu@ruby-lang.org>
+
+	* parse.y (rb_check_id_cstr): new function to check if ID is
+	  registered with NUL-terminated C string.
+
+	* sprintf.c (rb_str_format): avoid inadvertent symbol creation.
+
 Wed Apr 11 20:28:36 2012  Hiroshi Shirosaki  <h.shirosaki@gmail.com>
 
 	* io.c (rb_io_eof): use eof() instead of io_fillbuf(). It's because

README.EXT

@@ -452,12 +452,14 @@ argument by using
 
   rb_to_id(VALUE symbol)
   rb_check_id(volatile VALUE *name)
+  rb_check_id_cstr(const char *name, long len, rb_encoding *enc)
 
 These functions try to convert the argument to a String if it was not
-a Symbol nor a String.  The latter function stores the converted
+a Symbol nor a String.  The second function stores the converted
 result into *name, and returns 0 if the string is not a known symbol.
 After this function returned a non-zero value, *name is always a
 Symbol or a String, otherwise it is a String if the result is 0.
+The third function takes NUL-terminated C string, not Ruby VALUE.
 
 You can convert C ID to Ruby Symbol by using
 

README.EXT.ja

@@ -500,12 +500,14 @@ IDとは変数名，メソッド名を表す整数です．Rubyの中では
 
   rb_to_id(VALUE symbol)
   rb_check_id(volatile VALUE *name)
+  rb_check_id_cstr(const char *name, long len, rb_encoding *enc)
 
 もし引数がシンボルでも文字列でもなければ、to_strメソッドで文
-字列に変換しようとします．後者の関数はその変換結果を*nameに保
+字列に変換しようとします．第二の関数はその変換結果を*nameに保
 存し,その名前が既知のシンボルでない場合は0を返します．この関
 数が0以外を返した場合は*nameは常にシンボルか文字列であり、0を
-返した場合は常に文字列です．
+返した場合は常に文字列です．第三の関数はRubyの文字列ではなく
+NUL終端されたCの文字列を使います．
 
 2.2.3 CからRubyのメソッドを呼び出す
 

include/ruby/encoding.h

@@ -220,6 +220,7 @@ char *rb_enc_path_last_separator(const char *,const char *,rb_encoding*);
 char *rb_enc_path_end(const char *,const char *,rb_encoding*);
 const char *ruby_enc_find_basename(const char *name, long *baselen, long *alllen, rb_encoding *enc);
 const char *ruby_enc_find_extname(const char *name, long *len, rb_encoding *enc);
+ID rb_check_id_cstr(const char *ptr, long len, rb_encoding *enc);
 
 RUBY_EXTERN VALUE rb_cEncoding;
 #define ENC_DUMMY_FLAG (1<<24)

parse.y

@@ -10336,6 +10336,34 @@ rb_check_id(volatile VALUE *namep)
     return (ID)0;
 }
 
+ID
+rb_check_id_cstr(const char *ptr, long len, rb_encoding *enc)
+{
+    st_data_t id;
+    struct RString fake_str;
+    const VALUE name = (VALUE)&fake_str;
+    fake_str.basic.flags = T_STRING|RSTRING_NOEMBED;
+    fake_str.basic.klass = rb_cString;
+    fake_str.as.heap.len = len;
+    fake_str.as.heap.ptr = (char *)ptr;
+    fake_str.as.heap.aux.capa = len;
+    rb_enc_associate(name, enc);
+
+    sym_check_asciionly(name);
+
+    if (st_lookup(global_symbols.sym_id, (st_data_t)name, &id))
+	return (ID)id;
+
+    if (rb_is_attrset_name(name)) {
+	fake_str.as.heap.len = len - 1;
+	if (st_lookup(global_symbols.sym_id, (st_data_t)name, &id)) {
+	    return rb_id_attrset((ID)id);
+	}
+    }
+
+    return (ID)0;
+}
+
 int
 rb_is_const_name(VALUE name)
 {

sprintf.c

@@ -588,8 +588,11 @@ rb_str_format(int argc, const VALUE *argv, VALUE fmt)
 		    rb_enc_raise(enc, rb_eArgError, "named%.*s after <%s>",
 				 len, start, rb_id2name(id));
 		}
-		id = rb_intern3(start + 1, len - 2 /* without parenthesis */, enc);
-		nextvalue = GETNAMEARG(ID2SYM(id), start, len, enc);
+		nextvalue = GETNAMEARG((id = rb_check_id_cstr(start + 1,
+							      len - 2 /* without parenthesis */,
+							      enc),
+					ID2SYM(id)),
+				       start, len, enc);
 		if (nextvalue == Qundef) {
 		    rb_enc_raise(enc, rb_eKeyError, "key%.*s not found", len, start);
 		}

test/ruby/test_sprintf.rb

@@ -331,6 +331,8 @@ class TestSprintf < Test::Unit::TestCase
     assert_equal("named<key2> after unnumbered(2)", e.message)
     e = assert_raise(ArgumentError) {sprintf("%<key><key2>s", :key => "value")}
     assert_equal("named<key2> after <key>", e.message)
+    e = assert_raise(KeyError) {sprintf("%<key>s", {})}
+    assert_equal("key<key> not found", e.message)
   end
 
   def test_named_untyped_enc
@@ -349,6 +351,9 @@ class TestSprintf < Test::Unit::TestCase
       e = assert_raise(ArgumentError) {sprintf("%<#{k}><key>s", k.to_sym => "value")}
       assert_equal(enc, e.message.encoding)
       assert_equal("named<key> after <#{k}>", e.message)
+      e = assert_raise(KeyError) {sprintf("%<#{k}>s", {})}
+      assert_equal(enc, e.message.encoding)
+      assert_equal("key<#{k}> not found", e.message)
     end
   end
 
@@ -361,6 +366,8 @@ class TestSprintf < Test::Unit::TestCase
     e = assert_raise(ArgumentError) {sprintf("%<key>{key2}", :key => "value")}
     assert_equal("named{key2} after <key>", e.message)
     assert_equal("value{key2}", sprintf("%{key}{key2}", :key => "value"))
+    e = assert_raise(KeyError) {sprintf("%{key}", {})}
+    assert_equal("key{key} not found", e.message)
   end
 
   def test_named_typed_enc
@@ -379,6 +386,9 @@ class TestSprintf < Test::Unit::TestCase
       e = assert_raise(ArgumentError) {sprintf("%<#{k}>{key}s", k.to_sym => "value")}
       assert_equal(enc, e.message.encoding)
       assert_equal("named{key} after <#{k}>", e.message)
+      e = assert_raise(KeyError) {sprintf("%{#{k}}", {})}
+      assert_equal(enc, e.message.encoding)
+      assert_equal("key{#{k}} not found", e.message)
     end
   end
 end