nobu	2016-12-01 17:26:39 +0900 (Thu, 01 Dec 2016)

  New Revision: 56956

  https://svn.ruby-lang.org/cgi-bin/viewvc.cgi?view=revision&revision=56956

  Log:
    parse.y: reject invalid codepoint
    
    * parse.y (parser_tokadd_codepoint): reject invalid codepoint,
      surrogate blocks and surrogate pair, as well as mruby.

  Modified files:
    trunk/parse.y
    trunk/test/ruby/test_unicode_escape.rb
Index: test/ruby/test_unicode_escape.rb
===================================================================
--- test/ruby/test_unicode_escape.rb	(revision 56955)
+++ test/ruby/test_unicode_escape.rb	(revision 56956)
@@ -264,12 +264,9 @@ EOS https://github.com/ruby/ruby/blob/trunk/test/ruby/test_unicode_escape.rb#L264
      assert_raise(SyntaxError) { eval %q("\u{ 123 456}")}  # extra space
      assert_raise(SyntaxError) { eval %q("\u{123  456}")}  # extra space
 
-# The utf-8 encoding object currently does not object to codepoints
-# in the surrogate blocks, so these do not raise an error.
-#     assert_raise(SyntaxError) { "\uD800" }       # surrogate block
-#     assert_raise(SyntaxError) { "\uDCBA" }       # surrogate block
-#     assert_raise(SyntaxError) { "\uDFFF" }       # surrogate block
-#     assert_raise(SyntaxError) { "\uD847\uDD9A" } # surrogate pair
-
+     assert_raise(SyntaxError) { eval %q("\uD800") }       # surrogate block
+     assert_raise(SyntaxError) { eval %q("\uDCBA") }       # surrogate block
+     assert_raise(SyntaxError) { eval %q("\uDFFF") }       # surrogate block
+     assert_raise(SyntaxError) { eval %q("\uD847\uDD9A") } # surrogate pair
   end
 end
Index: parse.y
===================================================================
--- parse.y	(revision 56955)
+++ parse.y	(revision 56956)
@@ -5757,11 +5757,15 @@ parser_tok_hex(struct parser_params *par https://github.com/ruby/ruby/blob/trunk/parse.y#L5757
 
 #define tokcopy(n) memcpy(tokspace(n), lex_p - (n), (n))
 
-static void
+static int
 parser_tokadd_codepoint(struct parser_params *parser, rb_encoding **encp,
 			int string_literal, int regexp_literal,
 			int codepoint, int numlen)
 {
+    if ((codepoint & 0xfffff800) == 0xd800) {
+	yyerror("invalid Unicode codepoint");
+	return FALSE;
+    }
     lex_p += numlen;
     if (regexp_literal) {
 	tokcopy(numlen);
@@ -5773,6 +5777,7 @@ parser_tokadd_codepoint(struct parser_pa https://github.com/ruby/ruby/blob/trunk/parse.y#L5777
     else if (string_literal) {
 	tokadd(codepoint);
     }
+    return TRUE;
 }
 
 /* return value is for ?\u3042 */
@@ -5806,8 +5811,11 @@ parser_tokadd_utf8(struct parser_params https://github.com/ruby/ruby/blob/trunk/parse.y#L5811
 		yyerror("invalid Unicode codepoint (too large)");
 		return 0;
 	    }
-	    parser_tokadd_codepoint(parser, encp,string_literal, regexp_literal,
-				    codepoint, (int)numlen);
+	    if (!parser_tokadd_codepoint(parser, encp,
+					 string_literal, regexp_literal,
+					 codepoint, (int)numlen)) {
+		return 0;
+	    }
 	    if (ISSPACE(c = nextc())) last = c;
 	} while (string_literal && c != close_brace);
 
@@ -5824,8 +5832,11 @@ parser_tokadd_utf8(struct parser_params https://github.com/ruby/ruby/blob/trunk/parse.y#L5832
 	    yyerror("invalid Unicode escape");
 	    return 0;
 	}
-	parser_tokadd_codepoint(parser, encp, string_literal, regexp_literal,
-				codepoint, 4);
+	if (!parser_tokadd_codepoint(parser, encp,
+				     string_literal, regexp_literal,
+				     codepoint, 4)) {
+	    return 0;
+	}
     }
 
     return codepoint;

--
ML: ruby-changes@q...
Info: http://www.atdot.net/~ko1/quickml/