X-Git-Url: http://repo.macrolet.net/gitweb/?a=blobdiff_plain;f=src%2Fcode%2Ftarget-char.lisp;h=6d0224582628a98198c4e99509268501daf43b4d;hb=5b6e02e435453eddace1a36d30aaf04d6ebd2f1d;hp=d6a4ec6800988d2a90d31119840fb5d00a92ea5b;hpb=30c596bd5ca6305812598f42ae408b60a4c5f5c5;p=sbcl.git diff --git a/src/code/target-char.lisp b/src/code/target-char.lisp index d6a4ec6..6d02245 100644 --- a/src/code/target-char.lisp +++ b/src/code/target-char.lisp @@ -33,9 +33,9 @@ :directory '(:relative :up :up "output") :name name :type type) - sb!xc:*compile-file-truename*))) - (let ((character-database - (with-open-file (stream (file "ucd" "dat") + sb!xc:*compile-file-truename*)) + (read-ub8-vector (pathname) + (with-open-file (stream pathname :direction :input :element-type '(unsigned-byte 8)) (let* ((length (file-length stream)) @@ -43,9 +43,14 @@ length :element-type '(unsigned-byte 8)))) (read-sequence array stream) array)))) + (let ((character-database (read-ub8-vector (file "ucd" "dat"))) + (decompositions (read-ub8-vector (file "decomp" "dat"))) + (long-decompositions (read-ub8-vector (file "ldecomp" "dat")))) `(progn - (declaim (type (simple-array (unsigned-byte 8) (*)) **character-database**)) + (declaim (type (simple-array (unsigned-byte 8) (*)) **character-database** **character-decompositions** **character-long-decompositions**)) (defglobal **character-database** ,character-database) + (defglobal **character-decompositions** ,decompositions) + (defglobal **character-long-decompositions** ,long-decompositions) (defun !character-database-cold-init () (setf **character-database** ,character-database)) ,(with-open-file (stream (file "ucd-names" "lisp-expr") @@ -161,30 +166,84 @@ (#x9F "Application-Program-Command"))) ; *** See Note above ;;;; UCD accessor functions -;;;; -;;;; FIXME: Document the format of the character database. -;; (* 8 186) => 1488 -;; (+ 1488 (ash #x110000 -8)) => 5840 +;;; The first (* 8 395) => 3160 entries in **CHARACTER-DATABASE** +;;; contain entries for the distinct character attributes: +;;; specifically, indexes into the GC kinds, Bidi kinds, CCC kinds, +;;; the decimal digit property, the digit property and the +;;; bidi-mirrored boolean property. (There are two spare bytes for +;;; other information, should that become necessary) +;;; +;;; the next (ash #x110000 -8) entries contain single-byte indexes +;;; into a table of 256-element 4-byte-sized entries. These entries +;;; follow directly on, and are of the form +;;; {attribute-index[11b],transformed-code-point[21b]}x256, where the +;;; attribute index is an index into the miscellaneous information +;;; table, and the transformed code point is the code point of the +;;; simple mapping of the character to its lowercase or uppercase +;;; equivalent, as appropriate and if any. +;;; +;;; I feel the opacity of the above suggests the need for a diagram: +;;; +;;; C _______________________________________ +;;; / \ +;;; L \ +;;; [***************|=============================|--------...] +;;; (a) \ _ +;;; A \______________________/| B +;;; +;;; To look up information about a character, take the high 13 bits of +;;; its code point, and index the character database with that and a +;;; base of 3160 (going past the miscellaneous information[*], so +;;; treating (a) as the start of the array). This, labelled A, gives +;;; us another index into the detailed pages[-], which we can use to +;;; look up the details for the character in question: we add the low +;;; 8 bits of the character, shifted twice (because we have four-byte +;;; table entries) to 1024 times the `page' index, with a base of 6088 +;;; to skip over everything else. This gets us to point B. If we're +;;; after a transformed code point (i.e. an upcase or downcase +;;; operation), we can simply read it off now, beginning with an +;;; offset of 11 bits from point B in some endianness; if we're +;;; looking for miscellaneous information, we take the 11-bit value at +;;; B, and index the character database once more to get to the +;;; relevant miscellaneous information. +;;; +;;; As an optimization to the common case (pun intended) of looking up +;;; case information for a character, the entries in C above are +;;; sorted such that the characters which are UPPER-CASE-P in CL terms +;;; have index values lower than all others, followed by those which +;;; are LOWER-CASE-P in CL terms; this permits implementation of +;;; character case tests without actually going to the trouble of +;;; looking up the value associated with the index. (Actually, this +;;; isn't just a speed optimization; the information about whether a +;;; character is BOTH-CASE-P is used just in the ordering and not +;;; explicitly recorded in the database). +;;; +;;; The moral of all this? Next time, don't just say "FIXME: document +;;; this" (defun ucd-index (char) (let* ((cp (char-code char)) (cp-high (ash cp -8)) - (page (aref **character-database** (+ 1488 cp-high)))) - (+ 5840 (ash page 10) (ash (ldb (byte 8 0) cp) 2)))) + (page (aref **character-database** (+ 3160 cp-high)))) + (+ 7512 (ash page 10) (ash (ldb (byte 8 0) cp) 2)))) -(declaim (ftype (sfunction (t) (unsigned-byte 8)) ucd-value-0)) +(declaim (ftype (sfunction (t) (unsigned-byte 11)) ucd-value-0)) (defun ucd-value-0 (char) - (aref **character-database** (ucd-index char))) + (let ((index (ucd-index char)) + (character-database **character-database**)) + (dpb (aref character-database index) + (byte 8 3) + (ldb (byte 3 5) (aref character-database (+ index 1)))))) -(declaim (ftype (sfunction (t) (unsigned-byte 24)) ucd-value-1)) +(declaim (ftype (sfunction (t) (unsigned-byte 21)) ucd-value-1)) (defun ucd-value-1 (char) (let ((index (ucd-index char)) (character-database **character-database**)) - (dpb (aref character-database (+ index 3)) - (byte 8 16) + (dpb (aref character-database (+ index 1)) + (byte 5 16) (dpb (aref character-database (+ index 2)) (byte 8 8) - (aref character-database (1+ index)))))) + (aref character-database (+ index 3)))))) (declaim (ftype (sfunction (t) (unsigned-byte 8)) ucd-general-category)) (defun ucd-general-category (char) @@ -195,6 +254,9 @@ (+ 3 (* 8 (ucd-value-0 char)))))) (when (< decimal-digit 10) decimal-digit))) +(declaim (ftype (sfunction (t) (unsigned-byte 8)) ucd-ccc)) +(defun ucd-ccc (char) + (aref **character-database** (+ 2 (* 8 (ucd-value-0 char))))) (defun char-code (char) #!+sb-doc @@ -320,20 +382,20 @@ 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) 0)) + (< (ucd-value-0 char) 4)) (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." - (= (ucd-value-0 char) 1)) + (< 3 (ucd-value-0 char) 8)) (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) 2)) + (< (ucd-value-0 char) 8)) (defun digit-char-p (char &optional (radix 10.)) #!+sb-doc @@ -529,14 +591,14 @@ Case is ignored." #!+sb-doc "Return CHAR converted to upper-case if that is possible. Don't convert lowercase eszet (U+DF)." - (if (= (ucd-value-0 char) 1) + (if (< 3 (ucd-value-0 char) 8) (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) 0) + (if (< (ucd-value-0 char) 4) (code-char (ucd-value-1 char)) char)) @@ -548,3 +610,164 @@ character exists." (and (typep weight 'fixnum) (>= weight 0) (< weight radix) (< weight 36) (code-char (if (< weight 10) (+ 48 weight) (+ 55 weight))))) + +(defun char-decomposition-info (char) + (aref **character-database** (+ 6 (* 8 (ucd-value-0 char))))) + +(defun char-decomposition (char) + (let* ((cp (char-code char)) + (cp-high (ash cp -8)) + (decompositions **character-decompositions**) + (long-decompositions **character-long-decompositions**) + (index (+ #x1100 + (ash (aref decompositions cp-high) 10) + (ash (ldb (byte 8 0) cp) 2))) + (v0 (aref decompositions index)) + (v1 (aref decompositions (+ index 1))) + (v2 (aref decompositions (+ index 2))) + (v3 (aref decompositions (+ index 3))) + (length (dpb v0 (byte 8 3) (ldb (byte 3 5) v1))) + (entry (dpb (ldb (byte 5 0) v1) (byte 5 16) + (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)))))) + +(defun decompose-char (char) + (if (= (char-decomposition-info char) 0) + (string char) + (char-decomposition char))) + +(defun decompose-string (string &optional (kind :canonical)) + (declare (type (member :canonical :compatibility) kind)) + (flet ((canonical (char) + (= 1 (char-decomposition-info char))) + (compat (char) + (/= 0 (char-decomposition-info char)))) + (let (result + (fun (ecase kind + (:canonical #'canonical) + (:compatibility #'compat)))) + (do* ((start 0 (1+ end)) + (end (position-if fun string :start start) + (position-if fun string :start start))) + ((null end) (push (subseq string start end) result)) + (unless (= start end) + (push (subseq string start end) result)) + (push (decompose-char (char string end)) result)) + (apply 'concatenate 'string (nreverse result))))) + +(defun sort-combiners (string) + (let (result (start 0) first-cc first-non-cc) + (tagbody + again + (setf first-cc (position 0 string :key #'ucd-ccc :test #'/= :start start)) + (when first-cc + (setf first-non-cc (position 0 string :key #'ucd-ccc :test #'= :start first-cc))) + (push (subseq string start first-cc) result) + (when first-cc + (push (stable-sort (subseq string first-cc first-non-cc) #'< :key #'ucd-ccc) result)) + (when first-non-cc + (setf start first-non-cc first-cc nil first-non-cc nil) + (go again))) + (apply 'concatenate 'string (nreverse result)))) + +#+nil +(defun primary-composition (char1 char2) + (when (and (char= char1 #\e) + (char= char2 #\combining_acute_accent)) + #\latin_small_letter_e_with_acute)) + +;;; generic sequences. *sigh*. +(defun lref (lstring index) + (dolist (l lstring) + (when (and (<= (first l) index) + (< index (second l))) + (return (aref (third l) (- index (first l))))))) +(defun (setf lref) (newchar lstring index) + (dolist (l lstring) + (when (and (<= (first l) index) + (< index (second l))) + (return (setf (aref (third l) (- index (first l))) newchar))))) +(defun llength (lstring) + (second (first (last lstring)))) +(defun lstring (lstring) + (let ((result (make-string (llength lstring)))) + (dolist (l lstring result) + (replace result (third l) :start1 (first l) :end1 (second l))))) +(defun ldelete (lstring index) + (do* ((ls lstring (cdr ls)) + (l (car ls) (car ls)) + so-fars) + ((and (<= (first l) index) + (< index (second l))) + (append + (nreverse so-fars) + (cond + ((= (first l) index) + (list (list (first l) (1- (second l)) (subseq (third l) 1)))) + ((= index (1- (second l))) + (list (list (first l) (1- (second l)) (subseq (third l) 0 (1- (length (third l))))))) + (t + (list + (list (first l) index + (subseq (third l) 0 (- index (first l)))) + (list index (1- (second l)) + (subseq (third l) (1+ (- index (first l)))))))) + (mapcar (lambda (x) (list (1- (first x)) (1- (second x)) (third x))) + (cdr ls)))) + (push l so-fars))) + +(defun canonically-compose (string) + (labels () + (let* ((result (list (list 0 (length string) string))) + (previous-starter-index (position 0 string :key #'ucd-ccc)) + (i (1+ previous-starter-index))) + (when (= i (length string)) + (return-from canonically-compose string)) + (tagbody + again + (when (and (> (- i previous-starter-index) 2) + (= (ucd-ccc (lref result i)) (ucd-ccc (lref result (1- i))))) + (when (= (ucd-ccc (lref result i)) 0) + (setf previous-starter-index i)) + (incf i) + (go next)) + + (let ((comp (primary-composition (lref result previous-starter-index) + (lref result i)))) + (cond + (comp + (setf (lref result previous-starter-index) comp) + (setf result (ldelete result i))) + (t + (when (= (ucd-ccc (lref result i)) 0) + (setf previous-starter-index i)) + (incf i)))) + next + (unless (= i (llength result)) + (go again))) + (if (= i (length string)) + string + (lstring result))))) + +(defun normalize-string (string &optional (form :nfd)) + (declare (type (member :nfd :nfkd :nfc :nfkc) form)) + (etypecase string + (simple-base-string string) + ((simple-array character (*)) + (ecase form + ((:nfd) + (sort-combiners (decompose-string string))) + ((:nfkd) + (sort-combiners (decompose-string string :compatibility))))) + ((simple-array nil (*)) string)))