* parse.y (parser_tokadd_string, parser_yylex): ignore a backslash

  which prefixes an non-ascii character, which has no escape
  syntax.  [ruby-core:39222] [Ruby 1.9 - Bug #5262]


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@33161 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

3 files changed, 34 insertions, 2 deletions

diff --git a/ChangeLog b/ChangeLog
index 6ee79fd8c0..30b4d166d7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+Fri Sep  2 09:58:08 2011  Nobuyoshi Nakada  <nobu@ruby-lang.org>
+
+	* parse.y (parser_tokadd_string, parser_yylex): ignore a backslash
+	  which prefixes an non-ascii character, which has no escape
+	  syntax.  [ruby-core:39222] [Ruby 1.9 - Bug #5262]
+
 Fri Sep  2 04:05:25 2011  Aaron Patterson <aaron@tenderlovemaking.com>
 
 	* ext/psych/lib/psych/visitors/yaml_tree.rb: emit strings tagged as
diff --git a/parse.y b/parse.y
index eb5f7fa5e6..df8dc75307 100644
--- a/parse.y
+++ b/parse.y
@@ -5411,6 +5411,7 @@ parser_str_new(const char *p, long n, rb_encoding *enc, int func, rb_encoding *e
 }
 
 #define lex_goto_eol(parser) ((parser)->parser_lex_p = (parser)->parser_lex_pend)
+#define lex_eol_p() (lex_p >= lex_pend)
 #define peek(c) peek_n((c), 0)
 #define peek_n(c,n) (lex_p+(n) < lex_pend && (c) == (unsigned char)lex_p[n])
 
@@ -5920,6 +5921,8 @@ parser_tokadd_string(struct parser_params *parser,
 		continue;
 
 	      default:
+		if (c == -1) return -1;
+		if (!ISASCII(c)) goto non_ascii;
 		if (func & STR_FUNC_REGEXP) {
 		    pushback(c);
 		    if ((c = tokadd_escape(&enc)) < 0)
@@ -5945,6 +5948,7 @@ parser_tokadd_string(struct parser_params *parser,
 	    }
 	}
 	else if (!parser_isascii()) {
+	  non_ascii:
 	    has_nonascii = 1;
 	    if (enc != *encp) {
 		mixed_error(enc, *encp);
@@ -7003,6 +7007,10 @@ parser_yylex(struct parser_params *parser)
                     tokadd(c);
                 }
             }
+            else if (!lex_eol_p() && !(c = *lex_p, ISASCII(c))) {
+                nextc();
+                if (tokadd_mbchar(c) == -1) return 0;
+            }
             else {
                 c = read_escape(0, &enc);
                 tokadd(c);
diff --git a/test/ruby/test_literal.rb b/test/ruby/test_literal.rb
index 315fe529ea..23c1c9b800 100644
--- a/test/ruby/test_literal.rb
+++ b/test/ruby/test_literal.rb
@@ -53,15 +53,33 @@ class TestRubyLiteral < Test::Unit::TestCase
     assert_equal "3", "\x33"
     assert_equal "\n", "\n"
     bug2500 = '[ruby-core:27228]'
+    bug5262 = '[ruby-core:39222]'
     %w[c C- M-].each do |pre|
       ["u", %w[u{ }]].each do |open, close|
-        str = "\"\\#{pre}\\#{open}5555#{close}\""
-        assert_raise(SyntaxError, "#{bug2500} eval(#{str})") {eval(str)}
+        ["?", ['"', '"']].each do |qopen, qclose|
+          str = "#{qopen}\\#{pre}\\#{open}5555#{close}#{qclose}"
+          assert_raise(SyntaxError, "#{bug2500} eval(#{str})") {eval(str)}
+
+          str = "#{qopen}\\#{pre}\\#{open}\u201c#{close}#{qclose}"
+          assert_raise(SyntaxError, "#{bug5262} eval(#{str})") {eval(str)}
+
+          str = "#{qopen}\\#{pre}\\#{open}\u201c#{close}#{qclose}".encode("euc-jp")
+          assert_raise(SyntaxError, "#{bug5262} eval(#{str})") {eval(str)}
+
+          str = "#{qopen}\\#{pre}\\#{open}\u201c#{close}#{qclose}".encode("iso-8859-13")
+          assert_raise(SyntaxError, "#{bug5262} eval(#{str})") {eval(str)}
+
+          str = "#{qopen}\\#{pre}\\#{open}\xe2\x7f#{close}#{qclose}".force_encoding("utf-8")
+          assert_raise(SyntaxError, "#{bug5262} eval(#{str})") {eval(str)}
+        end
       end
     end
     assert_equal "\x13", "\c\x33"
     assert_equal "\x13", "\C-\x33"
     assert_equal "\xB3", "\M-\x33"
+    assert_equal "\u201c", eval(%["\\\u{201c}"]), bug5262
+    assert_equal "\u201c".encode("euc-jp"), eval(%["\\\u{201c}"].encode("euc-jp")), bug5262
+    assert_equal "\u201c".encode("iso-8859-13"), eval(%["\\\u{201c}"].encode("iso-8859-13")), bug5262
   end
 
   def test_dstring