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 ;;; The first (* 8 395) => 3160 entries in **CHARACTER-DATABASE**
166 ;;; contain entries for the distinct character attributes:
167 ;;; specifically, indexes into the GC kinds, Bidi kinds, CCC kinds,
168 ;;; the decimal digit property, the digit property and the
169 ;;; bidi-mirrored boolean property. (There are two spare bytes for
170 ;;; other information, should that become necessary)
172 ;;; the next (ash #x110000 -8) entries contain single-byte indexes
173 ;;; into a table of 256-element 4-byte-sized entries. These entries
174 ;;; follow directly on, and are of the form
175 ;;; {attribute-index[11b],transformed-code-point[21b]}x256, where the
176 ;;; attribute index is an index into the miscellaneous information
177 ;;; table, and the transformed code point is the code point of the
178 ;;; simple mapping of the character to its lowercase or uppercase
179 ;;; equivalent, as appropriate and if any.
181 ;;; I feel the opacity of the above suggests the need for a diagram:
183 ;;; C _______________________________________
186 ;;; [***************|=============================|--------...]
188 ;;; A \______________________/| B
190 ;;; To look up information about a character, take the high 13 bits of
191 ;;; its code point, and index the character database with that and a
192 ;;; base of 3160 (going past the miscellaneous information[*], so
193 ;;; treating (a) as the start of the array). This, labelled A, gives
194 ;;; us another index into the detailed pages[-], which we can use to
195 ;;; look up the details for the character in question: we add the low
196 ;;; 8 bits of the character, shifted twice (because we have four-byte
197 ;;; table entries) to 1024 times the `page' index, with a base of 6088
198 ;;; to skip over everything else. This gets us to point B. If we're
199 ;;; after a transformed code point (i.e. an upcase or downcase
200 ;;; operation), we can simply read it off now, beginning with an
201 ;;; offset of 11 bits from point B in some endianness; if we're
202 ;;; looking for miscellaneous information, we take the 11-bit value at
203 ;;; B, and index the character database once more to get to the
204 ;;; relevant miscellaneous information.
206 ;;; As an optimization to the common case (pun intended) of looking up
207 ;;; case information for a character, the entries in C above are
208 ;;; sorted such that the characters which are UPPER-CASE-P in CL terms
209 ;;; have index values lower than all others, followed by those which
210 ;;; are LOWER-CASE-P in CL terms; this permits implementation of
211 ;;; character case tests without actually going to the trouble of
212 ;;; looking up the value associated with the index. (Actually, this
213 ;;; isn't just a speed optimization; the information about whether a
214 ;;; character is BOTH-CASE-P is used just in the ordering and not
215 ;;; explicitly recorded in the database).
217 ;;; The moral of all this? Next time, don't just say "FIXME: document
219 (defun ucd-index (char)
220 (let* ((cp (char-code char))
221 (cp-high (ash cp -8))
222 (page (aref **character-database** (+ 3160 cp-high))))
223 (+ 7512 (ash page 10) (ash (ldb (byte 8 0) cp) 2))))
225 (declaim (ftype (sfunction (t) (unsigned-byte 11)) ucd-value-0))
226 (defun ucd-value-0 (char)
227 (let ((index (ucd-index char))
228 (character-database **character-database**))
229 (dpb (aref character-database index)
231 (ldb (byte 3 5) (aref character-database (+ index 1))))))
233 (declaim (ftype (sfunction (t) (unsigned-byte 21)) ucd-value-1))
234 (defun ucd-value-1 (char)
235 (let ((index (ucd-index char))
236 (character-database **character-database**))
237 (dpb (aref character-database (+ index 1))
239 (dpb (aref character-database (+ index 2))
241 (aref character-database (+ index 3))))))
243 (declaim (ftype (sfunction (t) (unsigned-byte 8)) ucd-general-category))
244 (defun ucd-general-category (char)
245 (aref **character-database** (* 8 (ucd-value-0 char))))
247 (defun ucd-decimal-digit (char)
248 (let ((decimal-digit (aref **character-database**
249 (+ 3 (* 8 (ucd-value-0 char))))))
250 (when (< decimal-digit 10)
253 (defun char-code (char)
255 "Return the integer code of CHAR."
258 (defun char-int (char)
260 "Return the integer code of CHAR. (In SBCL this is the same as CHAR-CODE, as
261 there are no character bits or fonts.)"
264 (defun code-char (code)
266 "Return the character with the code CODE."
269 (defun character (object)
271 "Coerce OBJECT into a CHARACTER if possible. Legal inputs are characters,
272 strings and symbols of length 1."
273 (flet ((do-error (control args)
274 (error 'simple-type-error
276 ;;?? how to express "symbol with name of length 1"?
277 :expected-type '(or character (string 1))
278 :format-control control
279 :format-arguments args)))
282 (string (if (= 1 (length (the string object)))
285 "String is not of length one: ~S" (list object))))
286 (symbol (if (= 1 (length (symbol-name object)))
287 (schar (symbol-name object) 0)
289 "Symbol name is not of length one: ~S" (list object))))
290 (t (do-error "~S cannot be coerced to a character." (list object))))))
292 (defun char-name (char)
294 "Return the name (a STRING) for a CHARACTER object."
295 (let ((char-code (char-code char)))
296 (or (second (assoc char-code *base-char-name-alist*))
298 (let ((h-code (cdr (binary-search char-code
299 (car *unicode-character-name-database*)
303 (huffman-decode h-code *unicode-character-name-huffman-tree*))
304 ((< char-code #x10000)
305 (format nil "U~4,'0X" char-code))
307 (format nil "U~8,'0X" char-code)))))))
309 (defun name-char (name)
311 "Given an argument acceptable to STRING, NAME-CHAR returns a character whose
312 name is that string, if one exists. Otherwise, NIL is returned."
313 (or (let ((char-code (car (rassoc-if (lambda (names)
314 (member name names :test #'string-equal))
315 *base-char-name-alist*))))
317 (code-char char-code)))
319 (let ((encoding (huffman-encode (string-upcase name)
320 *unicode-character-name-huffman-tree*)))
323 (car (binary-search encoding
324 (cdr *unicode-character-name-database*)
326 (name-string (string name))
327 (name-length (length name-string)))
330 (code-char char-code))
331 ((and (or (= name-length 9)
333 (char-equal (char name-string 0) #\U)
334 (loop for i from 1 below name-length
335 always (digit-char-p (char name-string i) 16)))
336 (code-char (parse-integer name-string :start 1 :radix 16)))
342 (defun standard-char-p (char)
344 "The argument must be a character object. STANDARD-CHAR-P returns T if the
345 argument is a standard character -- one of the 95 ASCII printing characters or
347 (and (typep char 'base-char)
348 (let ((n (char-code (the base-char char))))
352 (defun %standard-char-p (thing)
354 "Return T if and only if THING is a standard-char. Differs from
355 STANDARD-CHAR-P in that THING doesn't have to be a character."
356 (and (characterp thing) (standard-char-p thing)))
358 (defun graphic-char-p (char)
360 "The argument must be a character object. GRAPHIC-CHAR-P returns T if the
361 argument is a printing character (space through ~ in ASCII), otherwise returns
363 (let ((n (char-code char)))
367 (defun alpha-char-p (char)
369 "The argument must be a character object. ALPHA-CHAR-P returns T if the
370 argument is an alphabetic character, A-Z or a-z; otherwise NIL."
371 (< (ucd-general-category char) 5))
373 (defun upper-case-p (char)
375 "The argument must be a character object; UPPER-CASE-P returns T if the
376 argument is an upper-case character, NIL otherwise."
377 (< (ucd-value-0 char) 4))
379 (defun lower-case-p (char)
381 "The argument must be a character object; LOWER-CASE-P returns T if the
382 argument is a lower-case character, NIL otherwise."
383 (< 3 (ucd-value-0 char) 8))
385 (defun both-case-p (char)
387 "The argument must be a character object. BOTH-CASE-P returns T if the
388 argument is an alphabetic character and if the character exists in both upper
389 and lower case. For ASCII, this is the same as ALPHA-CHAR-P."
390 (< (ucd-value-0 char) 8))
392 (defun digit-char-p (char &optional (radix 10.))
394 "If char is a digit in the specified radix, returns the fixnum for which
395 that digit stands, else returns NIL."
396 (let ((m (- (char-code char) 48)))
398 (cond ((<= radix 10.)
399 ;; Special-case decimal and smaller radices.
400 (if (and (>= m 0) (< m radix)) m nil))
401 ;; Digits 0 - 9 are used as is, since radix is larger.
402 ((and (>= m 0) (< m 10)) m)
403 ;; Check for upper case A - Z.
404 ((and (>= (setq m (- m 7)) 10) (< m radix)) m)
405 ;; Also check lower case a - z.
406 ((and (>= (setq m (- m 32)) 10) (< m radix)) m)
408 (t (let ((number (ucd-decimal-digit char)))
409 (when (and number (< number radix))
412 (defun alphanumericp (char)
414 "Given a character-object argument, ALPHANUMERICP returns T if the argument
415 is either numeric or alphabetic."
416 (let ((gc (ucd-general-category char)))
420 (defun char= (character &rest more-characters)
422 "Return T if all of the arguments are the same character."
423 (declare (truly-dynamic-extent more-characters))
424 (dolist (c more-characters t)
425 (declare (type character c))
426 (unless (eq c character) (return nil))))
428 (defun char/= (character &rest more-characters)
430 "Return T if no two of the arguments are the same character."
431 (declare (truly-dynamic-extent more-characters))
432 (do* ((head character (car list))
433 (list more-characters (cdr list)))
435 (declare (type character head))
437 (declare (type character c))
438 (when (eq head c) (return-from char/= nil)))))
440 (defun char< (character &rest more-characters)
442 "Return T if the arguments are in strictly increasing alphabetic order."
443 (declare (truly-dynamic-extent more-characters))
444 (do* ((c character (car list))
445 (list more-characters (cdr list)))
447 (unless (< (char-int c)
448 (char-int (car list)))
451 (defun char> (character &rest more-characters)
453 "Return T if the arguments are in strictly decreasing alphabetic order."
454 (declare (truly-dynamic-extent more-characters))
455 (do* ((c character (car list))
456 (list more-characters (cdr list)))
458 (unless (> (char-int c)
459 (char-int (car list)))
462 (defun char<= (character &rest more-characters)
464 "Return T if the arguments are in strictly non-decreasing alphabetic order."
465 (declare (truly-dynamic-extent more-characters))
466 (do* ((c character (car list))
467 (list more-characters (cdr list)))
469 (unless (<= (char-int c)
470 (char-int (car list)))
473 (defun char>= (character &rest more-characters)
475 "Return T if the arguments are in strictly non-increasing alphabetic order."
476 (declare (truly-dynamic-extent more-characters))
477 (do* ((c character (car list))
478 (list more-characters (cdr list)))
480 (unless (>= (char-int c)
481 (char-int (car list)))
484 ;;; EQUAL-CHAR-CODE is used by the following functions as a version of CHAR-INT
485 ;;; which loses font, bits, and case info.
487 (defmacro equal-char-code (character)
489 `(let ((,ch ,character))
490 (if (= (ucd-value-0 ,ch) 0)
494 (defun two-arg-char-equal (c1 c2)
495 (= (equal-char-code c1) (equal-char-code c2)))
497 (defun char-equal (character &rest more-characters)
499 "Return T if all of the arguments are the same character.
501 (declare (truly-dynamic-extent more-characters))
502 (do ((clist more-characters (cdr clist)))
504 (unless (two-arg-char-equal (car clist) character)
507 (defun two-arg-char-not-equal (c1 c2)
508 (/= (equal-char-code c1) (equal-char-code c2)))
510 (defun char-not-equal (character &rest more-characters)
512 "Return T if no two of the arguments are the same character.
514 (declare (truly-dynamic-extent more-characters))
515 (do* ((head character (car list))
516 (list more-characters (cdr list)))
518 (unless (do* ((l list (cdr l)))
520 (if (two-arg-char-equal head (car l))
524 (defun two-arg-char-lessp (c1 c2)
525 (< (equal-char-code c1) (equal-char-code c2)))
527 (defun char-lessp (character &rest more-characters)
529 "Return T if the arguments are in strictly increasing alphabetic order.
531 (declare (truly-dynamic-extent more-characters))
532 (do* ((c character (car list))
533 (list more-characters (cdr list)))
535 (unless (two-arg-char-lessp c (car list))
538 (defun two-arg-char-greaterp (c1 c2)
539 (> (equal-char-code c1) (equal-char-code c2)))
541 (defun char-greaterp (character &rest more-characters)
543 "Return T if the arguments are in strictly decreasing alphabetic order.
545 (declare (truly-dynamic-extent more-characters))
546 (do* ((c character (car list))
547 (list more-characters (cdr list)))
549 (unless (two-arg-char-greaterp c (car list))
552 (defun two-arg-char-not-greaterp (c1 c2)
553 (<= (equal-char-code c1) (equal-char-code c2)))
555 (defun char-not-greaterp (character &rest more-characters)
557 "Return T if the arguments are in strictly non-decreasing alphabetic order.
559 (declare (truly-dynamic-extent more-characters))
560 (do* ((c character (car list))
561 (list more-characters (cdr list)))
563 (unless (two-arg-char-not-greaterp c (car list))
566 (defun two-arg-char-not-lessp (c1 c2)
567 (>= (equal-char-code c1) (equal-char-code c2)))
569 (defun char-not-lessp (character &rest more-characters)
571 "Return T if the arguments are in strictly non-increasing alphabetic order.
573 (declare (truly-dynamic-extent more-characters))
574 (do* ((c character (car list))
575 (list more-characters (cdr list)))
577 (unless (two-arg-char-not-lessp c (car list))
580 ;;;; miscellaneous functions
582 (defun char-upcase (char)
584 "Return CHAR converted to upper-case if that is possible. Don't convert
585 lowercase eszet (U+DF)."
586 (if (< 3 (ucd-value-0 char) 8)
587 (code-char (ucd-value-1 char))
590 (defun char-downcase (char)
592 "Return CHAR converted to lower-case if that is possible."
593 (if (< (ucd-value-0 char) 4)
594 (code-char (ucd-value-1 char))
597 (defun digit-char (weight &optional (radix 10))
599 "All arguments must be integers. Returns a character object that represents
600 a digit of the given weight in the specified radix. Returns NIL if no such
602 (and (typep weight 'fixnum)
603 (>= weight 0) (< weight radix) (< weight 36)
604 (code-char (if (< weight 10) (+ 48 weight) (+ 55 weight)))))