(ucs-mule-to-mule-unicode): Define this

translation table name here.
(ucs-mule-cjk-to-unicode): New translation table name.
(font-ccl-encoder-alist): Change the font registry pattern to
"ISO10646.*-*".

diff --git a/lisp/international/fontset.el b/lisp/international/fontset.el
index 7baa89ae66cb1326a8d01419eb215457c70883dd..ab0f32e0690ca041f87d285be172197135dbd677 100644 (file)
--- a/lisp/international/fontset.el
+++ b/lisp/international/fontset.el
@@ -112,35 +112,61 @@
 (set-font-encoding "ISO8859-1" 'ascii 0)
 (set-font-encoding "JISX0201" 'latin-jisx0201 0)
 
+;; Allow display of arbitrary characters with an iso-10646-encoded
+;; (`Unicode') font.
+(define-translation-table 'ucs-mule-to-mule-unicode
+  ucs-mule-to-mule-unicode)
+
 (define-ccl-program ccl-encode-unicode-font
   `(0
-    (if (r0 == ,(charset-id 'ascii))
-       ((r2 = r1)
-        (r1 = 0))
-      (if (r0 == ,(charset-id 'latin-iso8859-1))
-         ((r2 = (r1 + 128))
-          (r1 = 0))
-       (if (r0 == ,(charset-id 'mule-unicode-0100-24ff))
-           ((r1 *= 96)
-            (r1 += r2)
-            (r1 += ,(- #x100 (* 32 96) 32))
-            (r1 >8= 0)
-            (r2 = r7))
-         (if (r0 == ,(charset-id 'mule-unicode-2500-33ff))
-             ((r1 *= 96)
-              (r1 += r2)
-              (r1 += ,(- #x2500 (* 32 96) 32))
-              (r1 >8= 0)
-              (r2 = r7))
-           (if (r0 == ,(charset-id 'mule-unicode-e000-ffff))
-               ((r1 *= 96)
-                (r1 += r2)
-                (r1 += ,(- #xe000 (* 32 96) 32))
-                (r1 >8= 0)
-                (r2 = r7)))))))))
-
+    ;; r0: charset-id
+    ;; r1: 1st position code
+    ;; r2: 2nd position code (if r0 is 2D charset)
+    ((if (r0 == ,(charset-id 'ascii))
+        ((r2 = r1)
+         (r1 = 0))
+       ((if (r2 >= 0)
+           ;; This is a 2D charset.
+           (r1 = ((r1 << 7) | r2)))
+       ;; Look for a translation for non-ASCII chars.
+       (translate-character ucs-mule-to-mule-unicode r0 r1)
+       (if (r0 == ,(charset-id 'latin-iso8859-1))
+           ((r2 = (r1 + 128))
+            (r1 = 0))
+         ((r2 = (r1 & #x7F))
+          (r1 >>= 7)
+          (if (r0 == ,(charset-id 'mule-unicode-0100-24ff))
+              ((r1 *= 96)
+               (r1 += r2)
+               (r1 += ,(- #x100 (* 32 96) 32))
+               (r1 >8= 0)
+               (r2 = r7))
+            (if (r0 == ,(charset-id 'mule-unicode-2500-33ff))
+                ((r1 *= 96)
+                 (r1 += r2)
+                 (r1 += ,(- #x2500 (* 32 96) 32))
+                 (r1 >8= 0)
+                 (r2 = r7))
+              (if (r0 == ,(charset-id 'mule-unicode-e000-ffff))
+                  ((r1 *= 96)
+                   (r1 += r2)
+                   (r1 += ,(- #xe000 (* 32 96) 32))
+                   (r1 >8= 0)
+                   (r2 = r7))
+                ;; No way, use the glyph for U+FFFD.
+                ((r1 = #xFF)
+                 (r2 = #xFD)))))))))))
+  "Encode characters for display with iso10646 font.
+Translate through the translation-table named
+`ucs-mule-to-mule-unicode' initially.")
+
+;; Use the above CCL encoder for Unicode fonts.  Please note that the
+;; regexp is not simply "ISO10646-1" because there exists, for
+;; instance, the following Devanagari Unicode fonts:
+;;     -misc-fixed-medium-r-normal--24-240-72-72-c-120-iso10646.indian-1
+;;     -sibal-devanagari-medium-r-normal--24-240-75-75-P--iso10646-dev
 (setq font-ccl-encoder-alist
-      (cons '("ISO10646-1" . ccl-encode-unicode-font)
+      (cons '("ISO10646.*-*" . ccl-encode-unicode-font)
            font-ccl-encoder-alist))
 
 ;; Setting for suppressing XLoadQueryFont on big fonts.