fix CL case conversions of characters involving iota subscript

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

diff --git a/src/code/target-char.lisp b/src/code/target-char.lisp
index fdd7b9c..55b4bcc 100644 (file)
--- a/src/code/target-char.lisp
+++ b/src/code/target-char.lisp
@@ -62,11 +62,13 @@
                                      (dpb (aref info (+ (* 4 j) 2))
                                           (byte 8 8)
                                           (aref info (+ (* 4 j) 3)))))))
-                    (dotimes (i (/ (length info) 12) table)
+                    #!+sb-unicode
+                    (dotimes (i (/ (length info) 12))
                       (setf (gethash (dpb (code (* 3 i)) (byte 21 21)
                                           (code (1+ (* 3 i))))
                                      table)
-                            (code-char (code (+ (* 3 i) 2))))))))
+                            (code-char (code (+ (* 3 i) 2)))))
+                    table)))
               (defun !character-database-cold-init ()
                 (setf **character-database** ,character-database))
               ,(with-open-file (stream (file "ucd-names" "lisp-expr")
@@ -398,20 +400,20 @@ argument is an alphabetic character, A-Z or a-z; otherwise NIL."
   #!+sb-doc
   "The argument must be a character object; UPPER-CASE-P returns T if the
 argument is an upper-case character, NIL otherwise."
-  (< (ucd-value-0 char) 4))
+  (< (ucd-value-0 char) 5))
 
 (defun lower-case-p (char)
   #!+sb-doc
   "The argument must be a character object; LOWER-CASE-P returns T if the
 argument is a lower-case character, NIL otherwise."
-  (< 3 (ucd-value-0 char) 8))
+  (< 4 (ucd-value-0 char) 9))
 
 (defun both-case-p (char)
   #!+sb-doc
   "The argument must be a character object. BOTH-CASE-P returns T if the
 argument is an alphabetic character and if the character exists in both upper
 and lower case. For ASCII, this is the same as ALPHA-CHAR-P."
-  (< (ucd-value-0 char) 8))
+  (< (ucd-value-0 char) 9))
 
 (defun digit-char-p (char &optional (radix 10.))
   #!+sb-doc
@@ -511,7 +513,7 @@ is either numeric or alphabetic."
 (defmacro equal-char-code (character)
   (let ((ch (gensym)))
     `(let ((,ch ,character))
-      (if (= (ucd-value-0 ,ch) 0)
+      (if (< (ucd-value-0 ,ch) 5)
           (ucd-value-1 ,ch)
           (char-code ,ch)))))
 
@@ -607,14 +609,14 @@ Case is ignored."
   #!+sb-doc
   "Return CHAR converted to upper-case if that is possible. Don't convert
 lowercase eszet (U+DF)."
-  (if (< 3 (ucd-value-0 char) 8)
+  (if (< 4 (ucd-value-0 char) 9)
       (code-char (ucd-value-1 char))
       char))
 
 (defun char-downcase (char)
   #!+sb-doc
   "Return CHAR converted to lower-case if that is possible."
-  (if (< (ucd-value-0 char) 4)
+  (if (< (ucd-value-0 char) 5)
       (code-char (ucd-value-1 char))
       char))
 
@@ -836,10 +838,17 @@ character exists."
 
 (defun normalize-string (string &optional (form :nfd))
   (declare (type (member :nfd :nfkd :nfc :nfkc) form))
+  #!-sb-unicode
+  (etypecase string
+    ((array nil (*)) string)
+    (string
+     (ecase form
+       ((:nfc :nfkc) string)
+       ((:nfd :nfkd) (error "Cannot normalize to ~A form in #-SB-UNICODE builds" form)))))
+  #!+sb-unicode
   (etypecase string
-    #!+sb-unicode
     (base-string string)
-    ((or (array character (*)) #!-sb-unicode base-string)
+    ((array character (*))
      (ecase form
        ((:nfc)
         (canonically-compose (sort-combiners (decompose-string string))))