From b5e6b46e90101dbd092469e183070635634d9394 Mon Sep 17 00:00:00 2001
From: nobu <nobu@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>
Date: Mon, 17 Aug 2009 05:35:10 +0000
Subject: * parse.y (lex_get_str, lex_io_gets, rb_parser_compile_string):  
 must be ascii compatible.

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@24569 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
---
 ChangeLog              |  5 +++++
 parse.y                | 51 +++++++++++++++++++++++++++++++++++---------------
 test/ruby/test_eval.rb |  7 +++++++
 3 files changed, 48 insertions(+), 15 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 30322c1890..5dc0e0bb7d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,8 @@
+Mon Aug 17 14:35:03 2009  Nobuyoshi Nakada  <nobu@ruby-lang.org>
+
+	* parse.y (lex_get_str, lex_io_gets, rb_parser_compile_string):
+	  must be ascii compatible.
+
 Mon Aug 17 10:37:41 2009  NARUSE, Yui  <naruse@ruby-lang.org>
 
 	* regparse.c (add_code_range_to_buf0): added with checkdup argument.
diff --git a/parse.y b/parse.y
index 41706eaf34..6084bcaa84 100644
--- a/parse.y
+++ b/parse.y
@@ -5031,10 +5031,21 @@ yycompile(struct parser_params *parser, const char *f, int line)
 }
 #endif /* !RIPPER */
 
+static rb_encoding *
+must_be_ascii_compatible(VALUE s)
+{
+    rb_encoding *enc = rb_enc_get(s);
+    if (!rb_enc_asciicompat(enc)) {
+	rb_raise(rb_eArgError, "invalid source encoding");
+    }
+    return enc;
+}
+
 static VALUE
 lex_get_str(struct parser_params *parser, VALUE s)
 {
     char *beg, *end, *pend;
+    rb_encoding *enc = must_be_ascii_compatible(s);
 
     beg = RSTRING_PTR(s);
     if (lex_gets_ptr) {
@@ -5047,18 +5058,20 @@ lex_get_str(struct parser_params *parser, VALUE s)
 	if (*end++ == '\n') break;
     }
     lex_gets_ptr = end - RSTRING_PTR(s);
-    return rb_enc_str_new(beg, end - beg, rb_enc_get(s));
+    return rb_enc_str_new(beg, end - beg, enc);
 }
 
 static VALUE
 lex_getline(struct parser_params *parser)
 {
     VALUE line = (*parser->parser_lex_gets)(parser, parser->parser_lex_input);
+    if (NIL_P(line)) return line;
+    must_be_ascii_compatible(line);
 #ifndef RIPPER
-    if (ruby_debug_lines && !NIL_P(line)) {
+    if (ruby_debug_lines) {
 	rb_ary_push(ruby_debug_lines, line);
     }
-    if (ruby_coverage && !NIL_P(line)) {
+    if (ruby_coverage) {
 	rb_ary_push(ruby_coverage, Qnil);
     }
 #endif
@@ -5068,16 +5081,8 @@ lex_getline(struct parser_params *parser)
 static const rb_data_type_t parser_data_type;
 
 #ifndef RIPPER
-NODE*
-rb_compile_string(const char *f, VALUE s, int line)
-{
-    VALUE volatile vparser = rb_parser_new();
-
-    return rb_parser_compile_string(vparser, f, s, line);
-}
-
-NODE*
-rb_parser_compile_string(volatile VALUE vparser, const char *f, VALUE s, int line)
+static NODE*
+parser_compile_string(volatile VALUE vparser, const char *f, VALUE s, int line)
 {
     struct parser_params *parser;
     NODE *node;
@@ -5096,16 +5101,32 @@ rb_parser_compile_string(volatile VALUE vparser, const char *f, VALUE s, int lin
     return node;
 }
 
+NODE*
+rb_compile_string(const char *f, VALUE s, int line)
+{
+    must_be_ascii_compatible(s);
+    return parser_compile_string(rb_parser_new(), f, s, line);
+}
+
+NODE*
+rb_parser_compile_string(volatile VALUE vparser, const char *f, VALUE s, int line)
+{
+    must_be_ascii_compatible(s);
+    return parser_compile_string(vparser, f, s, line);
+}
+
 NODE*
 rb_compile_cstr(const char *f, const char *s, int len, int line)
 {
-    return rb_compile_string(f, rb_str_new(s, len), line);
+    VALUE str = rb_str_new(s, len);
+    return parser_compile_string(rb_parser_new(), f, str, line);
 }
 
 NODE*
 rb_parser_compile_cstr(volatile VALUE vparser, const char *f, const char *s, int len, int line)
 {
-    return rb_parser_compile_string(vparser, f, rb_str_new(s, len), line);
+    VALUE str = rb_str_new(s, len);
+    return parser_compile_string(vparser, f, str, line);
 }
 
 static VALUE
diff --git a/test/ruby/test_eval.rb b/test/ruby/test_eval.rb
index 6dd40bed84..a6900e075e 100644
--- a/test/ruby/test_eval.rb
+++ b/test/ruby/test_eval.rb
@@ -408,4 +408,11 @@ class TestEval < Test::Unit::TestCase
       assert_equal("0", f.read.chomp)
     end
   end
+
+  def test_eval_ascii_incompatible
+    assert_raise(ArgumentError) {eval("__ENCODING__".encode("utf-16be"))}
+    assert_raise(ArgumentError) {eval("__ENCODING__".encode("utf-16le"))}
+    assert_raise(ArgumentError) {eval("__ENCODING__".encode("utf-32be"))}
+    assert_raise(ArgumentError) {eval("__ENCODING__".encode("utf-32le"))}
+  end
 end
-- 
cgit v1.2.3