marshal.c: register symbol strings first

* marshal.c (r_symreal): register symbol names as strings first so
  that r_symlink always returns valid names.
  [ruby-core:68587] [Bug #10991]
* marshal.c (r_ivar, r_object0): now need to intern symbol names.
* marshal.c (r_object0): compare with symbol names.

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50057 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

3 files changed, 50 insertions, 12 deletions

diff --git a/ChangeLog b/ChangeLog
index acb63351df..465ce345f1 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,23 @@
+Mon Mar 23 02:03:28 2015  Nobuyoshi Nakada  <nobu@ruby-lang.org>
+
+	* marshal.c (r_symreal): register symbol names as strings first so
+	  that r_symlink always returns valid names.
+	  [ruby-core:68587] [Bug #10991]
+
+	* marshal.c (r_ivar, r_object0): now need to intern symbol names.
+
+	* marshal.c (r_object0): compare with symbol names.
+
+Mon Mar 23 01:44:35 2015  Nobuyoshi Nakada  <nobu@ruby-lang.org>
+
+	* marshal.c (r_symreal): register symbol names as strings first so
+	  that r_symlink always returns valid names.
+	  [ruby-core:68587] [Bug #10991]
+
+	* marshal.c (r_ivar, r_object0): now need to intern symbol names.
+
+	* marshal.c (r_object0): compare with symbol names.
+
 Sun Mar 22 22:07:40 2015  Kouhei Sutou  <kou@cozmixng.org>
 
 	* doc/etc.rd.ja: Fix wrong coding for Emacs.
diff --git a/marshal.c b/marshal.c
index 931c349089..1fae54530a 100644
--- a/marshal.c
+++ b/marshal.c
@@ -1295,11 +1295,18 @@ r_bytes0(long len, struct load_arg *arg)
 static int
 sym2encidx(VALUE sym, VALUE val)
 {
-    if (sym == ID2SYM(rb_id_encoding())) {
+    static const char name_encoding[8] = "encoding";
+    const char *p;
+    long l;
+    if (rb_enc_get_index(sym) != ENCINDEX_US_ASCII) return -1;
+    RSTRING_GETMEM(sym, p, l);
+    if (l <= 0) return -1;
+    if (l == sizeof(name_encoding) &&
+	memcmp(p, name_encoding, sizeof(name_encoding)) == 0) {
 	int idx = rb_enc_find_index(StringValueCStr(val));
 	return idx;
     }
-    else if (sym == ID2SYM(rb_intern("E"))) {
+    else if (l == 1 && *p == 'E') {
 	if (val == Qfalse) return rb_usascii_encindex();
 	else if (val == Qtrue) return rb_utf8_encindex();
 	/* bogus ignore */
@@ -1327,7 +1334,8 @@ r_symreal(struct load_arg *arg, int ivar)
     int idx = -1;
     st_index_t n = arg->symbols->num_entries;
 
-    st_insert(arg->symbols, (st_data_t)n, (st_data_t)0);
+    if (rb_enc_str_asciionly_p(s)) rb_enc_associate_index(s, ENCINDEX_US_ASCII);
+    st_insert(arg->symbols, (st_data_t)n, (st_data_t)s);
     if (ivar) {
 	long num = r_long(arg);
 	while (num-- > 0) {
@@ -1336,10 +1344,8 @@ r_symreal(struct load_arg *arg, int ivar)
 	}
     }
     if (idx > 0) rb_enc_associate_index(s, idx);
-    sym = rb_str_intern(s);
-    st_insert(arg->symbols, (st_data_t)n, (st_data_t)sym);
 
-    return sym;
+    return s;
 }
 
 static VALUE
@@ -1367,7 +1373,7 @@ r_symbol(struct load_arg *arg)
 static VALUE
 r_unique(struct load_arg *arg)
 {
-    return rb_sym2str(r_symbol(arg));
+    return r_symbol(arg);
 }
 
 static VALUE
@@ -1465,7 +1471,7 @@ r_ivar(VALUE obj, int *has_encoding, struct load_arg *arg)
 		if (has_encoding) *has_encoding = TRUE;
 	    }
 	    else {
-		rb_ivar_set(obj, SYM2ID(sym), val);
+		rb_ivar_set(obj, rb_intern_str(sym), val);
 	    }
 	} while (--len > 0);
     }
@@ -1790,13 +1796,13 @@ r_object0(struct load_arg *arg, int *ivp, VALUE extmod)
 	    v = r_entry0(v, idx, arg);
 	    values = rb_ary_new2(len);
 	    for (i=0; i<len; i++) {
+		VALUE n = rb_sym2str(RARRAY_AREF(mem, i));
 		slot = r_symbol(arg);
 
-		if (RARRAY_AREF(mem, i) != slot) {
+		if (!rb_str_equal(n, slot)) {
 		    rb_raise(rb_eTypeError, "struct %"PRIsVALUE" not compatible (:%"PRIsVALUE" for :%"PRIsVALUE")",
 			     rb_class_name(klass),
-			     rb_sym2str(slot),
-			     rb_sym2str(RARRAY_AREF(mem, i)));
+			     slot, n);
 		}
                 rb_ary_push(values, r_object(arg));
 		arg->readable -= 2;
@@ -1934,11 +1940,12 @@ r_object0(struct load_arg *arg, int *ivp, VALUE extmod)
 	else {
 	    v = r_symreal(arg, 0);
 	}
+	v = rb_str_intern(v);
 	v = r_leave(v, arg);
 	break;
 
       case TYPE_SYMLINK:
-	v = r_symlink(arg);
+	v = rb_str_intern(r_symlink(arg));
 	break;
 
       default:
diff --git a/test/ruby/test_marshal.rb b/test/ruby/test_marshal.rb
index 9b82d52233..3d5d6c9b1f 100644
--- a/test/ruby/test_marshal.rb
+++ b/test/ruby/test_marshal.rb
@@ -252,6 +252,17 @@ class TestMarshal < Test::Unit::TestCase
     assert_include(Marshal.dump([:a, :a]), ';')
   end
 
+  def test_symlink_in_ivar
+    bug10991 = '[ruby-core:68587] [Bug #10991]'
+    sym = Marshal.load("\x04\x08" +
+                       "I" ":\x0bKernel" +
+                       ("\x06" +
+                        ("I" ":\x07@a" +
+                         ("\x06" ":\x07@b" "e;\x0""o:\x0bObject""\x0")) +
+                        "0"))
+    assert_equal(:Kernel, sym, bug10991)
+  end
+
   ClassUTF8 = eval("class R\u{e9}sum\u{e9}; self; end")
 
   iso_8859_1 = Encoding::ISO_8859_1