1 ;;;; character functions
3 ;;;; This software is part of the SBCL system. See the README file for
6 ;;;; This software is derived from the CMU CL system, which was
7 ;;;; written at Carnegie Mellon University and released into the
8 ;;;; public domain. The software is in the public domain and is
9 ;;;; provided with absolutely no warranty. See the COPYING and CREDITS
10 ;;;; files for more information.
12 (in-package "SB!IMPL")
14 ;;; We compile some trivial character operations via inline expansion.
16 (declaim (inline standard-char-p graphic-char-p alpha-char-p
17 upper-case-p lower-case-p both-case-p alphanumericp
19 (declaim (maybe-inline digit-char-p digit-weight))
22 `(integer 0 (,char-code-limit)))
26 (defvar *unicode-character-name-database*)
27 (defvar *unicode-character-name-huffman-tree*))
31 (flet ((file (name type)
32 (merge-pathnames (make-pathname
34 '(:relative :up :up "output")
35 :name name :type type)
36 sb!xc:*compile-file-truename*)))
37 (let ((character-database
38 (with-open-file (stream (file "ucd" "dat")
40 :element-type '(unsigned-byte 8))
41 (let* ((length (file-length stream))
43 length :element-type '(unsigned-byte 8))))
44 (read-sequence array stream)
47 (declaim (type (simple-array (unsigned-byte 8) (*)) **character-database**))
48 (defglobal **character-database** ,character-database)
49 (defun !character-database-cold-init ()
50 (setf **character-database** ,character-database))
51 ,(with-open-file (stream (file "ucd-names" "lisp-expr")
53 :element-type 'character)
54 (let ((names (make-hash-table)))
57 for code-point = (read stream nil nil)
58 for char-name = (string-upcase (read stream nil nil))
60 do (setf (gethash code-point names) char-name))
65 (maphash (lambda (code name)
66 (declare (ignore code))
71 (make-array (hash-table-count names)
74 (maphash (lambda (code name)
76 (cons code (huffman-encode name tree))
80 (sort (copy-seq code->name) #'< :key #'cdr))
82 (sort (copy-seq name->code) #'< :key #'car))
84 `(defun !character-name-database-cold-init ()
86 (setq *unicode-character-name-database*
87 (cons ',code->name ',name->code)
88 *unicode-character-name-huffman-tree* ',tree))))))))))
90 #+sb-xc-host (!character-name-database-cold-init)
92 (defparameter *base-char-name-alist*
93 ;; Note: The *** markers here indicate character names which are
94 ;; required by the ANSI specification of #'CHAR-NAME. For the others,
95 ;; we prefer the ASCII standard name.
96 '((#x00 "Nul" "Null" "^@")
103 (#x07 "Bel" "Bell" "^g")
104 (#x08 "Backspace" "^h" "Bs") ; *** See Note above
105 (#x09 "Tab" "^i" "Ht") ; *** See Note above
106 (#x0A "Newline" "Linefeed" "^j" "Lf" "Nl") ; *** See Note above
108 (#x0C "Page" "^l" "Form" "Formfeed" "Ff" "Np") ; *** See Note above
109 (#x0D "Return" "^m" "Cr") ; *** See Note above
123 (#x1B "Esc" "Escape" "^[" "Altmode" "Alt")
128 (#x20 "Space" "Sp") ; *** See Note above
129 (#x7f "Rubout" "Delete" "Del")
132 (#x82 "Break-Permitted")
133 (#x83 "No-Break-Permitted")
136 (#x86 "Start-Selected-Area")
137 (#x87 "End-Selected-Area")
138 (#x88 "Character-Tabulation-Set")
139 (#x89 "Character-Tabulation-With-Justification")
140 (#x8A "Line-Tabulation-Set")
141 (#x8B "Partial-Line-Forward")
142 (#x8C "Partial-Line-Backward")
143 (#x8D "Reverse-Linefeed")
144 (#x8E "Single-Shift-Two")
145 (#x8F "Single-Shift-Three")
146 (#x90 "Device-Control-String")
147 (#x91 "Private-Use-One")
148 (#x92 "Private-Use-Two")
149 (#x93 "Set-Transmit-State")
150 (#x94 "Cancel-Character")
151 (#x95 "Message-Waiting")
152 (#x96 "Start-Guarded-Area")
153 (#x97 "End-Guarded-Area")
154 (#x98 "Start-String")
156 (#x9A "Single-Character-Introducer")
157 (#x9B "Control-Sequence-Introducer")
158 (#x9C "String-Terminator")
159 (#x9D "Operating-System-Command")
160 (#x9E "Privacy-Message")
161 (#x9F "Application-Program-Command"))) ; *** See Note above
163 ;;;; UCD accessor functions
165 ;;;; FIXME: Document the format of the character database.
168 ;; (+ 1488 (ash #x110000 -8)) => 5840
169 (defun ucd-index (char)
170 (let* ((cp (char-code char))
171 (cp-high (ash cp -8))
172 (page (aref **character-database** (+ 1488 cp-high))))
173 (+ 5840 (ash page 10) (ash (ldb (byte 8 0) cp) 2))))
175 (declaim (ftype (sfunction (t) (unsigned-byte 8)) ucd-value-0))
176 (defun ucd-value-0 (char)
177 (aref **character-database** (ucd-index char)))
179 (declaim (ftype (sfunction (t) (unsigned-byte 24)) ucd-value-1))
180 (defun ucd-value-1 (char)
181 (let ((index (ucd-index char))
182 (character-database **character-database**))
183 (dpb (aref character-database (+ index 3))
185 (dpb (aref character-database (+ index 2))
187 (aref character-database (1+ index))))))
189 (declaim (ftype (sfunction (t) (unsigned-byte 8)) ucd-general-category))
190 (defun ucd-general-category (char)
191 (aref **character-database** (* 8 (ucd-value-0 char))))
193 (defun ucd-decimal-digit (char)
194 (let ((decimal-digit (aref **character-database**
195 (+ 3 (* 8 (ucd-value-0 char))))))
196 (when (< decimal-digit 10)
199 (defun char-code (char)
201 "Return the integer code of CHAR."
204 (defun char-int (char)
206 "Return the integer code of CHAR. (In SBCL this is the same as CHAR-CODE, as
207 there are no character bits or fonts.)"
210 (defun code-char (code)
212 "Return the character with the code CODE."
215 (defun character (object)
217 "Coerce OBJECT into a CHARACTER if possible. Legal inputs are characters,
218 strings and symbols of length 1."
219 (flet ((do-error (control args)
220 (error 'simple-type-error
222 ;;?? how to express "symbol with name of length 1"?
223 :expected-type '(or character (string 1))
224 :format-control control
225 :format-arguments args)))
228 (string (if (= 1 (length (the string object)))
231 "String is not of length one: ~S" (list object))))
232 (symbol (if (= 1 (length (symbol-name object)))
233 (schar (symbol-name object) 0)
235 "Symbol name is not of length one: ~S" (list object))))
236 (t (do-error "~S cannot be coerced to a character." (list object))))))
238 (defun char-name (char)
240 "Return the name (a STRING) for a CHARACTER object."
241 (let ((char-code (char-code char)))
242 (or (second (assoc char-code *base-char-name-alist*))
244 (let ((h-code (cdr (binary-search char-code
245 (car *unicode-character-name-database*)
249 (huffman-decode h-code *unicode-character-name-huffman-tree*))
250 ((< char-code #x10000)
251 (format nil "U~4,'0X" char-code))
253 (format nil "U~8,'0X" char-code)))))))
255 (defun name-char (name)
257 "Given an argument acceptable to STRING, NAME-CHAR returns a character whose
258 name is that string, if one exists. Otherwise, NIL is returned."
259 (or (let ((char-code (car (rassoc-if (lambda (names)
260 (member name names :test #'string-equal))
261 *base-char-name-alist*))))
263 (code-char char-code)))
265 (let ((encoding (huffman-encode (string-upcase name)
266 *unicode-character-name-huffman-tree*)))
269 (car (binary-search encoding
270 (cdr *unicode-character-name-database*)
272 (name-string (string name))
273 (name-length (length name-string)))
276 (code-char char-code))
277 ((and (or (= name-length 9)
279 (char-equal (char name-string 0) #\U)
280 (loop for i from 1 below name-length
281 always (digit-char-p (char name-string i) 16)))
282 (code-char (parse-integer name-string :start 1 :radix 16)))
288 (defun standard-char-p (char)
290 "The argument must be a character object. STANDARD-CHAR-P returns T if the
291 argument is a standard character -- one of the 95 ASCII printing characters or
293 (and (typep char 'base-char)
294 (let ((n (char-code (the base-char char))))
298 (defun %standard-char-p (thing)
300 "Return T if and only if THING is a standard-char. Differs from
301 STANDARD-CHAR-P in that THING doesn't have to be a character."
302 (and (characterp thing) (standard-char-p thing)))
304 (defun graphic-char-p (char)
306 "The argument must be a character object. GRAPHIC-CHAR-P returns T if the
307 argument is a printing character (space through ~ in ASCII), otherwise returns
309 (let ((n (char-code char)))
313 (defun alpha-char-p (char)
315 "The argument must be a character object. ALPHA-CHAR-P returns T if the
316 argument is an alphabetic character, A-Z or a-z; otherwise NIL."
317 (< (ucd-general-category char) 5))
319 (defun upper-case-p (char)
321 "The argument must be a character object; UPPER-CASE-P returns T if the
322 argument is an upper-case character, NIL otherwise."
323 (= (ucd-value-0 char) 0))
325 (defun lower-case-p (char)
327 "The argument must be a character object; LOWER-CASE-P returns T if the
328 argument is a lower-case character, NIL otherwise."
329 (= (ucd-value-0 char) 1))
331 (defun both-case-p (char)
333 "The argument must be a character object. BOTH-CASE-P returns T if the
334 argument is an alphabetic character and if the character exists in both upper
335 and lower case. For ASCII, this is the same as ALPHA-CHAR-P."
336 (< (ucd-value-0 char) 2))
338 (defun digit-char-p (char &optional (radix 10.))
340 "If char is a digit in the specified radix, returns the fixnum for which
341 that digit stands, else returns NIL."
342 (let ((m (- (char-code char) 48)))
344 (cond ((<= radix 10.)
345 ;; Special-case decimal and smaller radices.
346 (if (and (>= m 0) (< m radix)) m nil))
347 ;; Digits 0 - 9 are used as is, since radix is larger.
348 ((and (>= m 0) (< m 10)) m)
349 ;; Check for upper case A - Z.
350 ((and (>= (setq m (- m 7)) 10) (< m radix)) m)
351 ;; Also check lower case a - z.
352 ((and (>= (setq m (- m 32)) 10) (< m radix)) m)
354 (t (let ((number (ucd-decimal-digit char)))
355 (when (and number (< number radix))
358 (defun alphanumericp (char)
360 "Given a character-object argument, ALPHANUMERICP returns T if the argument
361 is either numeric or alphabetic."
362 (let ((gc (ucd-general-category char)))
366 (defun char= (character &rest more-characters)
368 "Return T if all of the arguments are the same character."
369 (declare (truly-dynamic-extent more-characters))
370 (dolist (c more-characters t)
371 (declare (type character c))
372 (unless (eq c character) (return nil))))
374 (defun char/= (character &rest more-characters)
376 "Return T if no two of the arguments are the same character."
377 (declare (truly-dynamic-extent more-characters))
378 (do* ((head character (car list))
379 (list more-characters (cdr list)))
381 (declare (type character head))
383 (declare (type character c))
384 (when (eq head c) (return-from char/= nil)))))
386 (defun char< (character &rest more-characters)
388 "Return T if the arguments are in strictly increasing alphabetic order."
389 (declare (truly-dynamic-extent more-characters))
390 (do* ((c character (car list))
391 (list more-characters (cdr list)))
393 (unless (< (char-int c)
394 (char-int (car list)))
397 (defun char> (character &rest more-characters)
399 "Return T if the arguments are in strictly decreasing alphabetic order."
400 (declare (truly-dynamic-extent more-characters))
401 (do* ((c character (car list))
402 (list more-characters (cdr list)))
404 (unless (> (char-int c)
405 (char-int (car list)))
408 (defun char<= (character &rest more-characters)
410 "Return T if the arguments are in strictly non-decreasing alphabetic order."
411 (declare (truly-dynamic-extent more-characters))
412 (do* ((c character (car list))
413 (list more-characters (cdr list)))
415 (unless (<= (char-int c)
416 (char-int (car list)))
419 (defun char>= (character &rest more-characters)
421 "Return T if the arguments are in strictly non-increasing alphabetic order."
422 (declare (truly-dynamic-extent more-characters))
423 (do* ((c character (car list))
424 (list more-characters (cdr list)))
426 (unless (>= (char-int c)
427 (char-int (car list)))
430 ;;; EQUAL-CHAR-CODE is used by the following functions as a version of CHAR-INT
431 ;;; which loses font, bits, and case info.
433 (defmacro equal-char-code (character)
435 `(let ((,ch ,character))
436 (if (= (ucd-value-0 ,ch) 0)
440 (defun two-arg-char-equal (c1 c2)
441 (= (equal-char-code c1) (equal-char-code c2)))
443 (defun char-equal (character &rest more-characters)
445 "Return T if all of the arguments are the same character.
447 (declare (truly-dynamic-extent more-characters))
448 (do ((clist more-characters (cdr clist)))
450 (unless (two-arg-char-equal (car clist) character)
453 (defun two-arg-char-not-equal (c1 c2)
454 (/= (equal-char-code c1) (equal-char-code c2)))
456 (defun char-not-equal (character &rest more-characters)
458 "Return T if no two of the arguments are the same character.
460 (declare (truly-dynamic-extent more-characters))
461 (do* ((head character (car list))
462 (list more-characters (cdr list)))
464 (unless (do* ((l list (cdr l)))
466 (if (two-arg-char-equal head (car l))
470 (defun two-arg-char-lessp (c1 c2)
471 (< (equal-char-code c1) (equal-char-code c2)))
473 (defun char-lessp (character &rest more-characters)
475 "Return T if the arguments are in strictly increasing alphabetic order.
477 (declare (truly-dynamic-extent more-characters))
478 (do* ((c character (car list))
479 (list more-characters (cdr list)))
481 (unless (two-arg-char-lessp c (car list))
484 (defun two-arg-char-greaterp (c1 c2)
485 (> (equal-char-code c1) (equal-char-code c2)))
487 (defun char-greaterp (character &rest more-characters)
489 "Return T if the arguments are in strictly decreasing alphabetic order.
491 (declare (truly-dynamic-extent more-characters))
492 (do* ((c character (car list))
493 (list more-characters (cdr list)))
495 (unless (two-arg-char-greaterp c (car list))
498 (defun two-arg-char-not-greaterp (c1 c2)
499 (<= (equal-char-code c1) (equal-char-code c2)))
501 (defun char-not-greaterp (character &rest more-characters)
503 "Return T if the arguments are in strictly non-decreasing alphabetic order.
505 (declare (truly-dynamic-extent more-characters))
506 (do* ((c character (car list))
507 (list more-characters (cdr list)))
509 (unless (two-arg-char-not-greaterp c (car list))
512 (defun two-arg-char-not-lessp (c1 c2)
513 (>= (equal-char-code c1) (equal-char-code c2)))
515 (defun char-not-lessp (character &rest more-characters)
517 "Return T if the arguments are in strictly non-increasing alphabetic order.
519 (declare (truly-dynamic-extent more-characters))
520 (do* ((c character (car list))
521 (list more-characters (cdr list)))
523 (unless (two-arg-char-not-lessp c (car list))
526 ;;;; miscellaneous functions
528 (defun char-upcase (char)
530 "Return CHAR converted to upper-case if that is possible. Don't convert
531 lowercase eszet (U+DF)."
532 (if (= (ucd-value-0 char) 1)
533 (code-char (ucd-value-1 char))
536 (defun char-downcase (char)
538 "Return CHAR converted to lower-case if that is possible."
539 (if (= (ucd-value-0 char) 0)
540 (code-char (ucd-value-1 char))
543 (defun digit-char (weight &optional (radix 10))
545 "All arguments must be integers. Returns a character object that represents
546 a digit of the given weight in the specified radix. Returns NIL if no such
548 (and (typep weight 'fixnum)
549 (>= weight 0) (< weight radix) (< weight 36)
550 (code-char (if (< weight 10) (+ 48 weight) (+ 55 weight)))))