(tibetan-canonicalize-for-unicode-alist)

(tibetan-canonicalize-for-unicode-regexp): New variables.
(tibetan-canonicalize-for-unicode-region): New function.
(tibetan-strict-unicode): New variable.
(tibetan-pre-write-canonicalize-for-unicode): New function.

diff --git a/lisp/language/tibet-util.el b/lisp/language/tibet-util.el
index 5d53989a63f5bc928631decb12933d0552cdf825..14987e226638145404099b88aaa1bc38d2b9d963 100644 (file)
--- a/lisp/language/tibet-util.el
+++ b/lisp/language/tibet-util.el
@@ -367,6 +367,64 @@ See also docstring of the function tibetan-compose-region."
     ;; Should return nil as annotations.
     nil))
 
+\f
+;;;
+;;; Unicode-related definitions.
+;;; 
+
+(defvar tibetan-canonicalize-for-unicode-alist
+  '(("\e$(7"Q\e(B" . "") ;; remove vowel a
+    ("\e$(7"T\e(B" . "\e$(7"R"S\e(B") ;; decompose vowels whose use is ``discouraged'' in Unicode 3.0
+    ("\e$(7"V\e(B" . "\e$(7"R"U\e(B")
+    ("\e$(7"W\e(B" . "\e$(7#C"a\e(B")
+    ("\e$(7"X\e(B" . "\e$(7#C"R"a\e(B")
+    ("\e$(7"Y\e(B" . "\e$(7#D"a\e(B")
+    ("\e$(7"Z\e(B" . "\e$(7#D"R"a\e(B")
+    ("\e$(7"b\e(B" . "\e$(7"R"a\e(B"))
+  "Rules for canonicalizing Tibetan vowels for Unicode.")
+
+(defvar tibetan-canonicalize-for-unicode-regexp
+  "[\e$(7"Q"T"V"W"X"Y"Z"b\e(B]"
+  "Regexp for Tibetan vowels to be canonicalized in Unicode.")
+
+(defun tibetan-canonicalize-for-unicode-region (from to)
+  (save-restriction
+    (narrow-to-region from to)
+    (goto-char from)
+    (while (re-search-forward tibetan-canonicalize-for-unicode-regexp nil t)
+      (let (
+           ;;(from (match-beginning 0))
+           ;;(to (match-end 0))
+           (canonical-form
+            (cdr (assoc (match-string 0)
+                        tibetan-canonicalize-for-unicode-alist))))
+       ;;(goto-char from)
+       ;;(delete-region from to)
+       ;;(insert canonical-form)
+       (replace-match canonical-form)
+       ))))
+
+(defvar tibetan-strict-unicode t
+  "*Flag to control Tibetan canonicalizing for Unicode.
+
+If non-nil, the vowel a is removed and composite vowels are decomposed
+before writing buffer in Unicode.  See also
+`tibetan-canonicalize-for-unicode-regexp' and
+`tibetan-canonicalize-for-unicode-alist'.")
+
+;;;###autoload
+(defun tibetan-pre-write-canonicalize-for-unicode (from to)
+  (let ((old-buf (current-buffer))
+       (strict-unicode tibetan-strict-unicode))
+    (set-buffer (generate-new-buffer " *temp*"))
+    (if (stringp from)
+       (insert from)
+      (insert-buffer-substring old-buf from to))
+    (if strict-unicode
+       (tibetan-canonicalize-for-unicode-region (point-min) (point-max)))
+    ;; Should return nil as annotations.
+    nil))
+
 (provide 'tibet-util)
 
 ;;; tibet-util.el ends here