handle Hangul syllable decomposition

[sbcl.git] / src / code / target-char.lisp

diff --git a/src/code/target-char.lisp b/src/code/target-char.lisp
index 6d02245..77ec7b2 100644 (file)
--- a/src/code/target-char.lisp
+++ b/src/code/target-char.lisp
@@ -631,16 +631,38 @@ character exists."
                      (dpb v2 (byte 8 8) v3))))
     (if (= length 1)
         (string (code-char entry))
-        (let ((result (make-string length))
-              (e (* 4 entry)))
-          (dotimes (i length result)
-            (let ((code (dpb (aref long-decompositions (+ e 1))
-                             (byte 8 16)
-                             (dpb (aref long-decompositions (+ e 2))
-                                  (byte 8 8)
-                                  (aref long-decompositions (+ e 3))))))
-              (setf (char result i) (code-char code)))
-            (incf e 4))))))
+        (if (<= #xac00 cp #xd7a3)
+            ;; see Unicode 6.2, section 3-12
+            (let* ((sbase #xac00)
+                   (lbase #x1100)
+                   (vbase #x1161)
+                   (tbase #x11a7)
+                   (lcount 19)
+                   (vcount 21)
+                   (tcount 28)
+                   (ncount (* vcount tcount))
+                   (scount (* lcount ncount))
+                   (sindex (- cp sbase))
+                   (lindex (floor sindex ncount))
+                   (vindex (floor (mod sindex ncount) tcount))
+                   (tindex (mod sindex tcount))
+                   (result (make-string length)))
+              (declare (ignore scount))
+              (setf (char result 0) (code-char (+ lbase lindex)))
+              (setf (char result 1) (code-char (+ vbase vindex)))
+              (when (> tindex 0)
+                (setf (char result 2) (code-char (+ tbase tindex))))
+              result)
+            (let ((result (make-string length))
+                  (e (* 4 entry)))
+              (dotimes (i length result)
+                (let ((code (dpb (aref long-decompositions (+ e 1))
+                                 (byte 8 16)
+                                 (dpb (aref long-decompositions (+ e 2))
+                                      (byte 8 8)
+                                      (aref long-decompositions (+ e 3))))))
+                  (setf (char result i) (code-char code)))
+                (incf e 4)))))))
 
 (defun decompose-char (char)
   (if (= (char-decomposition-info char) 0)