(cp-make-coding-system): Use

utf-translation-table-for-decode, not
utf-8-translation-table-for-decode.
(mule-diag): Require.

diff --git a/lisp/international/code-pages.el b/lisp/international/code-pages.el
index b681d4d7d6632938b58f409466c34f50c28ada43..f7609659e438d588277ed484f97df01e878b2395 100644 (file)
--- a/lisp/international/code-pages.el
+++ b/lisp/international/code-pages.el
@@ -54,6 +54,8 @@
 
 ;;; Code:
 
+(require 'mule-diag)                   ; for non-iso-charset-alist
+
 (defun cp-make-translation-table (v)
   "Return a translation table made from 128-long vector V.
 V comprises characters encodable by mule-utf-8."
@@ -174,7 +176,7 @@ corresponding args of `make-coding-system'.  If MNEMONIC isn't given,
               (translate-character ,decoder r0 r1)
               ;; Allow fragmentation on decoding -- relevant for
               ;; Cyrillic, Greek and, possibly Arabic and Hebrew.
-              (translate-character utf-8-translation-table-for-decode r0 r1)
+              (translate-character utf-translation-table-for-decode r0 r1)
               (write-multibyte-character r0 r1)
               (repeat))))))
         (ccl-encoder