* string.c (rb_str_inspect): inspect as a dummy encoding string

  when a UTF-16/32 (not BE/LE) string does not have a BOM.
  Unicode and some RFCs say that a string labeld as UTF-16/32
  doesn't have a BOM, it should be considered big endian.
  But many Windows programs generates little endian UTF-16
  strings without a BOM. So String#inspect treats a string
  labeled UTF-16/32 withaout a BOM as a dummy encoding string.
  patched by Martin Duerst. [ruby-core:33461]

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@29984 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

3 files changed, 38 insertions, 4 deletions

diff --git a/ChangeLog b/ChangeLog
index 2df7ee73a2..440ee3d9e0 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,14 @@
+Wed Dec  1 01:29:15 2010  NARUSE, Yui  <naruse@ruby-lang.org>
+
+	* string.c (rb_str_inspect): inspect as a dummy encoding string
+	  when a UTF-16/32 (not BE/LE) string does not have a BOM.
+	  Unicode and some RFCs say that a string labeld as UTF-16/32
+	  doesn't have a BOM, it should be considered big endian.
+	  But many Windows programs generates little endian UTF-16
+	  strings without a BOM. So String#inspect treats a string
+	  labeled UTF-16/32 withaout a BOM as a dummy encoding string.
+	  patched by Martin Duerst. [ruby-core:33461]
+
 Tue Nov 30 17:04:10 2010  NARUSE, Yui  <naruse@ruby-lang.org>
 
 	* addr2line.c (parse_debug_line_cu): ignore DW_LNE_set_discriminator.
diff --git a/string.c b/string.c
index 385b05f5b1..d55f943a5b 100644
--- a/string.c
+++ b/string.c
@@ -4214,10 +4214,22 @@ rb_str_inspect(VALUE str)
     p = RSTRING_PTR(str); pend = RSTRING_END(str);
     prev = p;
     if (enc == utf16) {
-	enc = *p == (char)0xFF ? rb_enc_find("UTF-16LE") : rb_enc_find("UTF-16BE");
+	const unsigned char *q = (const unsigned char *)p;
+	if (q[0] == 0xFE && q[1] == 0xFF)
+	    enc = rb_enc_find("UTF-16BE");
+	else if (q[0] == 0xFF && q[1] == 0xFD)
+	    enc = rb_enc_find("UTF-16LE");
+	else
+	    unicode_p = 0;
     }
     else if (enc == utf32) {
-	enc = *p == (char)0xFF ? rb_enc_find("UTF-32LE") : rb_enc_find("UTF-32BE");
+	const unsigned char *q = (const unsigned char *)p;
+	if (q[0] == 0 && q[1] == 0 && q[2] == 0xFE && q[3] == 0xFF)
+	    enc = rb_enc_find("UTF-32BE");
+	else if (q[3] == 0 && q[2] == 0 && q[1] == 0xFE && q[0] == 0xFF)
+	    enc = rb_enc_find("UTF-32LE");
+	else
+	    unicode_p = 0;
     }
     while (p < pend) {
 	unsigned int c, cc;
@@ -6004,7 +6016,6 @@ static VALUE
 rb_str_each_codepoint(VALUE str)
 {
     VALUE orig = str;
-    long len;
     int n;
     unsigned int c;
     const char *ptr, *end;
@@ -6014,7 +6025,6 @@ rb_str_each_codepoint(VALUE str)
     RETURN_ENUMERATOR(str, 0, 0);
     str = rb_str_new4(str);
     ptr = RSTRING_PTR(str);
-    len = RSTRING_LEN(str);
     end = RSTRING_END(str);
     enc = STR_ENC_GET(str);
     while (ptr < end) {
diff --git a/test/ruby/test_m17n.rb b/test/ruby/test_m17n.rb
index 3dccc76c91..cfe7ed4d3d 100644
--- a/test/ruby/test_m17n.rb
+++ b/test/ruby/test_m17n.rb
@@ -232,6 +232,19 @@ class TestM17N < Test::Unit::TestCase
     Encoding.default_external = orig_ext
   end
 
+  def test_utf_16_32_inspect
+    str = "\u3042"
+    %w/UTF-16 UTF-32/.each do |enc|
+      %w/BE LE/.each do |endian|
+        s = str.encode(enc + endian)
+        # When a UTF-16/32 string doesn't have a BOM,
+        # inspect as a dummy encoding string.
+        assert_equal(s.dup.force_encoding("ISO-2022-JP").inspect,
+                     s.dup.force_encoding(enc).inspect)
+      end
+    end
+  end
+
   def test_str_dump
     [
       e("\xfe"),