(detect_coding_utf_16): Fix previous change.

diff --git a/src/ChangeLog b/src/ChangeLog
index bc4ce2aa0afa67af88cd226119c9d82d5ccb6fd9..6a6e90aafb947fce327c2d3a0ec2c92ad848420c 100644 (file)
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,8 @@
+2009-06-16  Kenichi Handa  <handa@m17n.org>
+
+       * coding.c (detect_coding_utf_16): Fix the logic of rejecting
+       UTF-16 by checking the dispersion of Eth and Oth bytes.
+
 2009-06-15  Andreas Schwab  <schwab@linux-m68k.org>
 
        * coding.c (detect_coding_utf_16): Fix typo counting odd bytes.

diff --git a/src/coding.c b/src/coding.c
index 8d90297db0f26a667e42161aca7672b53a32a24b..cde56d9af424302a82f6bd4fbcd315bbecf0cec1 100644 (file)
--- a/src/coding.c
+++ b/src/coding.c
@@ -1665,10 +1665,11 @@ detect_coding_utf_16 (coding, detect_info)
       e[c1] = 1;
       o[c2] = 1;
 
-      detect_info->rejected
-       |= (CATEGORY_MASK_UTF_16_BE | CATEGORY_MASK_UTF_16_LE);
+      detect_info->rejected |= (CATEGORY_MASK_UTF_16_AUTO
+                               |CATEGORY_MASK_UTF_16_BE
+                               | CATEGORY_MASK_UTF_16_LE);
 
-      while (1)
+      while (detect_info->rejected != CATEGORY_MASK_UTF_16)
        {
          TWO_MORE_BYTES (c1, c2);
          if (c2 < 0)
@@ -1677,18 +1678,17 @@ detect_coding_utf_16 (coding, detect_info)
            {
              e[c1] = 1;
              e_num++;
-             if (e_num >= 128 && o_num >= 128)
-               break;
+             if (e_num >= 128)
+               detect_info->rejected |= CATEGORY_MASK_UTF_16_BE_NOSIG;
            }
          if (! o[c2])
            {
              o[c2] = 1;
              o_num++;
-             if (e_num >= 128 && o_num >= 128)
-               break;
+             if (o_num >= 128)
+               detect_info->rejected |= CATEGORY_MASK_UTF_16_LE_NOSIG;
            }
        }
-      detect_info->rejected |= CATEGORY_MASK_UTF_16;
       return 0;
     }