# HG changeset patch
# User Aidan Kehoe <kehoea@parhasard.net>
# Date 1233408097 0
# Node ID b25f081370e0ea07648c447823f1b97e3d45b33b
# Parent  2669b1b7e33b70dad0f42a94629ab3afeffa9b55# Parent  732e3243f2e4f8be3758ef0820b6a415dfbd1966
Automated merge with file:///Sources/xemacs-21.5-checked-out

diff -r 732e3243f2e4 -r b25f081370e0 src/ChangeLog
--- a/src/ChangeLog	Sat Jan 31 13:21:32 2009 +0000
+++ b/src/ChangeLog	Sat Jan 31 13:21:37 2009 +0000
@@ -1,3 +1,8 @@
+2009-01-31  Aidan Kehoe  <kehoea@parhasard.net>
+
+	* unicode.c (unicode_convert): 
+	Correct little-endian UTF-16 surrogate handling. 
+
 2009-01-16  Aidan Kehoe  <kehoea@parhasard.net>
 
 	* chartab.c (print_table_entry): 
diff -r 732e3243f2e4 -r b25f081370e0 src/unicode.c
--- a/src/unicode.c	Sat Jan 31 13:21:32 2009 +0000
+++ b/src/unicode.c	Sat Jan 31 13:21:37 2009 +0000
@@ -2115,23 +2115,47 @@
 		{
 		  int tempch;
 
-		  if (!valid_utf_16_last_surrogate(ch & 0xFFFF))
-		    {
-                      DECODE_ERROR_OCTET ((ch >> 24) & 0xFF, dst, data,
-                                        ignore_bom);
-                      DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data,
-                                        ignore_bom);
-                      DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data,
-                                        ignore_bom);
-                      DECODE_ERROR_OCTET (ch & 0xFF, dst, data,
-                                        ignore_bom);
-		    }
-                  else 
+                  if (little_endian)
                     {
-                      tempch = utf_16_surrogates_to_code((ch >> 16), 
-                                                         (ch & 0xffff));
-                      decode_unicode_char(tempch, dst, data, ignore_bom);
+                      if (!valid_utf_16_last_surrogate(ch >> 16))
+                        {
+                          DECODE_ERROR_OCTET (ch & 0xFF, dst, data,
+                                              ignore_bom);
+                          DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data,
+                                              ignore_bom);
+                          DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data,
+                                              ignore_bom);
+                          DECODE_ERROR_OCTET ((ch >> 24) & 0xFF, dst, data,
+                                              ignore_bom);
+                        }
+                      else
+                        {
+                          tempch = utf_16_surrogates_to_code((ch & 0xffff),
+                                                             (ch >> 16));
+                          decode_unicode_char(tempch, dst, data, ignore_bom); 
+                        }
                     }
+                  else
+                    {
+                      if (!valid_utf_16_last_surrogate(ch & 0xFFFF))
+                        {
+                          DECODE_ERROR_OCTET ((ch >> 24) & 0xFF, dst, data,
+                                              ignore_bom);
+                          DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data,
+                                              ignore_bom);
+                          DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data,
+                                              ignore_bom);
+                          DECODE_ERROR_OCTET (ch & 0xFF, dst, data,
+                                              ignore_bom);
+                        }
+                      else 
+                        {
+                          tempch = utf_16_surrogates_to_code((ch >> 16), 
+                                                             (ch & 0xffff));
+                          decode_unicode_char(tempch, dst, data, ignore_bom); 
+                        }
+                    }
+
 		  ch = 0;
 		  counter = 0;
                 }
diff -r 732e3243f2e4 -r b25f081370e0 tests/ChangeLog
--- a/tests/ChangeLog	Sat Jan 31 13:21:32 2009 +0000
+++ b/tests/ChangeLog	Sat Jan 31 13:21:37 2009 +0000
@@ -1,3 +1,8 @@
+2009-01-31  Aidan Kehoe  <kehoea@parhasard.net>
+
+	* automated/mule-tests.el: 
+	Test little-endian Unicode surrogates too. 
+
 2009-01-18  Aidan Kehoe  <kehoea@parhasard.net>
 	
 	* automated/lisp-tests.el: (char-table-with-string): 
diff -r 732e3243f2e4 -r b25f081370e0 tests/automated/mule-tests.el
--- a/tests/automated/mule-tests.el	Sat Jan 31 13:21:32 2009 +0000
+++ b/tests/automated/mule-tests.el	Sat Jan 31 13:21:37 2009 +0000
@@ -446,12 +446,17 @@
 		       (encode-coding-string xemacs-character 'ctext))))))
 
   (loop
-    for (code-point encoded) 
-    in '((#x10000 "\xd8\x00\xdc\x00")
-         (#x10FFFD "\xdb\xff\xdf\xfd"))
-    do (Assert (equal (encode-coding-string 
-                       (decode-char 'ucs code-point) 'utf-16)
-                      encoded)))
+    for (code-point utf-16-big-endian utf-16-little-endian) 
+    in '((#x10000 "\xd8\x00\xdc\x00" "\x00\xd8\x00\xdc")
+         (#x10FFFD "\xdb\xff\xdf\xfd" "\xff\xdb\xfd\xdf"))
+    do
+    (Assert (equal (encode-coding-string 
+                    (decode-char 'ucs code-point) 'utf-16)
+                   utf-16-big-endian))
+    (Assert (equal (encode-coding-string 
+                    (decode-char 'ucs code-point) 'utf-16-le)
+                   utf-16-little-endian))
+
          
   ;;---------------------------------------------------------------
   ;; Regression test for a couple of CCL-related bugs.