Mercurial > hg > xemacs-beta

comparison lisp/mule/mule-category.el @ 428:3ecd8885ac67 r21-2-22

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Import from CVS: tag r21-2-22
author cvs
date Mon, 13 Aug 2007 11:28:15 +0200
parents
children 9d177e8d4150
comparison
equal deleted inserted replaced
427:0a0253eac470 428:3ecd8885ac67
1 ;;; mule-category.el --- category functions for XEmacs/Mule.
2
3 ;; Copyright (C) 1992,93,94,95 Free Software Foundation, Inc.
4 ;; Copyright (C) 1995, 1997, 1999 Electrotechnical Laboratory, JAPAN.
5 ;; Licensed to the Free Software Foundation.
6 ;; Copyright (C) 1995 Amdahl Corporation.
7 ;; Copyright (C) 1995 Sun Microsystems.
8
9 ;; This file is part of XEmacs.
10
11 ;; XEmacs is free software; you can redistribute it and/or modify it
12 ;; under the terms of the GNU General Public License as published by
13 ;; the Free Software Foundation; either version 2, or (at your option)
14 ;; any later version.
15
16 ;; XEmacs is distributed in the hope that it will be useful, but
17 ;; WITHOUT ANY WARRANTY; without even the implied warranty of
18 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19 ;; General Public License for more details.
20
21 ;; You should have received a copy of the GNU General Public License
22 ;; along with XEmacs; see the file COPYING. If not, write to the
23 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
24 ;; Boston, MA 02111-1307, USA.
25
26 ;;; Commentary:
27
28 ;; Functions for working with category tables, which are a particular
29 ;; type of char table. Some function names / arguments should be
30 ;; parallel with syntax tables.
31
32 ;; Written by Ben Wing <ben@xemacs.org>. The initialization code
33 ;; at the end of this file comes from Mule.
34 ;; Some bugfixes by Jareth Hein <jhod@po.iijnet.or.jp>
35
36 ;;; Code:
37
38 (defvar defined-category-hashtable (make-hashtable 50))
39
40 (defun define-category (designator doc-string)
41 "Make a new category whose designator is DESIGNATOR.
42 DESIGNATOR should be a visible letter of ' ' thru '~'.
43 STRING is a doc string for the category.
44 Letters of 'a' thru 'z' are already used or kept for the system."
45 (check-argument-type 'category-designator-p designator)
46 (check-argument-type 'stringp doc-string)
47 (puthash designator doc-string defined-category-hashtable))
48
49 (defun undefine-category (designator)
50 "Undefine DESIGNATOR as a designator for a category."
51 (check-argument-type 'category-designator-p designator)
52 (remhash designator defined-category-hashtable))
53
54 (defun defined-category-p (designator)
55 "Return non-nil if DESIGNATOR is a designator for a defined category."
56 (and (category-designator-p designator)
57 (gethash designator defined-category-hashtable)))
58
59 (defun defined-category-list ()
60 "Return a list of the currently defined categories.
61 Categories are given by their designators."
62 (let (list)
63 (maphash #'(lambda (key value)
64 (setq list (cons key list)))
65 defined-category-hashtable)
66 (nreverse list)))
67
68 (defun undefined-category-designator ()
69 "Return an undefined category designator, or nil if there are none."
70 (let ((a 32) found)
71 (while (and (< a 127) (not found))
72 (if (gethash a defined-category-hashtable)
73 (setq found a))
74 (setq a (1+ a)))
75 found))
76
77 (defun category-doc-string (designator)
78 "Return the doc-string for the category denoted by DESIGNATOR."
79 (check-argument-type 'defined-category-p designator)
80 (gethash designator defined-category-hashtable))
81
82 (defun modify-category-entry (char-range designator &optional table reset)
83 "Add a category to the categories associated with CHAR-RANGE.
84 CHAR-RANGE is a single character or a range of characters,
85 as per `put-char-table'.
86 The category is given by a designator character.
87 The changes are made in TABLE, which defaults to the current buffer's
88 category table.
89 If optional fourth argument RESET is non-nil, previous categories associated
90 with CHAR-RANGE are removed before adding the specified category."
91 (or table (setq table (category-table)))
92 (check-argument-type 'category-table-p table)
93 (check-argument-type 'defined-category-p designator)
94 (if reset
95 ;; clear all existing stuff.
96 (put-char-table char-range nil table))
97 (map-char-table
98 #'(lambda (key value)
99 ;; make sure that this range has a bit-vector assigned to it
100 (if (not (bit-vector-p value))
101 (setq value (make-bit-vector 95 0))
102 (setq value (copy-sequence value)))
103 ;; set the appropriate bit in that vector.
104 (aset value (- designator 32) 1)
105 ;; put the vector back, thus assuring we have a unique setting for this range
106 (put-char-table key value table))
107 table char-range))
108
109 (defun char-category-list (char &optional table)
110 "Return a list of the categories that CHAR is in.
111 TABLE defaults to the current buffer's category table.
112 The categories are given by their designators."
113 (or table (setq table (category-table)))
114 (check-argument-type 'category-table-p table)
115 (let ((vec (get-char-table char table)))
116 (if (null vec) nil
117 (let ((a 32) list)
118 (while (< a 127)
119 (if (= 1 (aref vec (- a 32)))
120 (setq list (cons a list)))
121 (setq a (1+ a)))
122 (nreverse list)))))
123
124 ;; implimented in c, file chartab.c (97/3/14 jhod@po.iijnet.or.jp)
125 ;(defun char-in-category-p (char category &optional table)
126 ; "Return non-nil if CHAR is in CATEGORY.
127 ;TABLE defaults to the current buffer's category table.
128 ;Categories are specified by their designators."
129 ; (or table (setq table (category-table)))
130 ; (check-argument-type 'category-table-p table)
131 ; (check-argument-type 'category-designator-p category)
132 ; (let ((vec (get-char-table char table)))
133 ; (if (null vec) nil
134 ; (= 1 (aref vec (- category 32))))))
135
136 (defun describe-category ()
137 "Describe the category specifications in the category table.
138 The descriptions are inserted in a buffer, which is then displayed."
139 (interactive)
140 (with-output-to-temp-buffer "*Help*"
141 (describe-category-table (category-table) standard-output)))
142
143 (defun describe-category-table (table stream)
144 (let (first-char
145 last-char
146 prev-val
147 (describe-one
148 (lambda (first last value stream)
149 (if (and (bit-vector-p value)
150 (> (reduce '+ value) 0))
151 (progn
152 (if (equal first last)
153 (cond ((vectorp first)
154 (princ (format "%s, row %d"
155 (charset-name
156 (aref first 0))
157 (aref first 1))
158 stream))
159 ((charsetp first)
160 (princ (charset-name first) stream))
161 (t (princ first stream)))
162 (cond ((vectorp first)
163 (princ (format "%s, rows %d .. %d"
164 (charset-name
165 (aref first 0))
166 (aref first 1)
167 (aref last 1))
168 stream))
169 (t
170 (princ (format "%s .. %s" first last)
171 stream))))
172 (describe-category-code value stream))))))
173 (map-char-table
174 (lambda (range value)
175 (if (and (or
176 (and (characterp range)
177 (characterp first-char)
178 (eq (char-charset range) (char-charset first-char))
179 (= (char-to-int last-char) (1- (char-to-int range))))
180 (and (vectorp range)
181 (vectorp first-char)
182 (eq (aref range 0) (aref first-char 0))
183 (= (aref last-char 1) (1- (aref range 1))))
184 (equal value prev-val)))
185 (setq last-char range)
186 (if first-char
187 (progn
188 (funcall describe-one first-char last-char prev-val stream)
189 (setq first-char nil)))
190 (funcall describe-one range range value stream))
191 nil)
192 table)
193 (if first-char
194 (funcall describe-one first-char last-char prev-val stream))))
195
196 (defun describe-category-code (code stream)
197 (let ((standard-output (or stream standard-output)))
198 (princ "\tin categories: ")
199 (if (not (bit-vector-p code))
200 (princ "(none)")
201 (let ((i 0)
202 already-matched)
203 (while (< i 95)
204 (if (= 1 (aref code i))
205 (progn
206 (if (not already-matched)
207 (setq already-matched t)
208 (princ " "))
209 (princ (int-to-char (+ 32 i)))))
210 (setq i (1+ i)))
211 (if (not already-matched)
212 (princ "(none)")))
213 (let ((i 0))
214 (while (< i 95)
215 (if (= 1 (aref code i))
216 (princ (format "\n\t\tmeaning: %s"
217 (category-doc-string (int-to-char (+ 32 i))))))
218 (setq i (1+ i)))))
219 (terpri)))
220
221 (defconst predefined-category-list
222 '((latin-iso8859-1 ?l "Latin-1 through Latin-5 character set")
223 (latin-iso8859-2 ?l)
224 (latin-iso8859-3 ?l)
225 (latin-iso8859-4 ?l)
226 (latin-iso8859-9 ?l)
227 (cyrillic-iso8859-5 ?y "Cyrillic character set")
228 (arabic-iso8859-6 ?b "Arabic character set")
229 (greek-iso8859-7 ?g "Greek character set")
230 (hebrew-iso8859-8 ?w "Hebrew character set")
231 (katakana-jisx0201 ?k "Japanese 1-byte Katakana character set")
232 (latin-jisx0201 ?r "Japanese 1-byte Roman character set")
233 (japanese-jisx0208-1978 ?j "Japanese 2-byte character set (old)")
234 (japanese-jisx0208 ?j "Japanese 2-byte character set")
235 (japanese-jisx0212 ?j)
236 (chinese-gb2312 ?c "Chinese GB (China, PRC) 2-byte character set")
237 (chinese-cns11643-1 ?t "Chinese Taiwan (CNS or Big5) 2-byte character set")
238 (chinese-cns11643-2 ?t)
239 (chinese-big5-1 ?t)
240 (chinese-big5-2 ?t)
241 (korean-ksc5601 ?h "Hangul (Korean) 2-byte character set")
242 )
243 "List of predefined categories.
244 Each element is a list of a charset, a designator, and maybe a doc string.")
245
246 (let (i l)
247 (define-category ?a "ASCII character set.")
248 (define-category ?l "Latin-1 through Latin-5 character set")
249 (setq i 32)
250 (while (< i 127)
251 (modify-category-entry i ?a)
252 (modify-category-entry i ?l)
253 (setq i (1+ i)))
254 (setq l predefined-category-list)
255 (while l
256 (if (and (nth 2 (car l))
257 (not (defined-category-p (nth 2 (car l)))))
258 (define-category (nth 1 (car l)) (nth 2 (car l))))
259 (modify-category-entry (car (car l)) (nth 1 (car l)))
260 (setq l (cdr l))))
261
262 ;;; Setting word boundary.
263
264 (setq word-combining-categories
265 '((?l . ?l)))
266
267 (setq word-separating-categories ; (2-byte character sets)
268 '((?A . ?K) ; Alpha numeric - Katakana
269 (?A . ?C) ; Alpha numeric - Chinese
270 (?H . ?A) ; Hiragana - Alpha numeric
271 (?H . ?K) ; Hiragana - Katakana
272 (?H . ?C) ; Hiragana - Chinese
273 (?K . ?A) ; Katakana - Alpha numeric
274 (?K . ?C) ; Katakana - Chinese
275 (?C . ?A) ; Chinese - Alpha numeric
276 (?C . ?K) ; Chinese - Katakana
277 ))
278
279 ;;; At the present, I know Japanese and Chinese text can
280 ;;; break line at any point under a restriction of 'kinsoku'.
281 (defvar word-across-newline "\\(\\cj\\|\\cc\\|\\ct\\)"
282 "Regular expression of such characters which can be a word across newline.")
283
284 (defvar ascii-char "[\40-\176]")
285 (defvar ascii-space "[ \t]")
286 (defvar ascii-symbols "[\40-\57\72-\100\133-\140\173-\176]")
287 (defvar ascii-numeric "[\60-\71]")
288 (defvar ascii-English-Upper "[\101-\132]")
289 (defvar ascii-English-Lower "[\141-\172]")
290 (defvar ascii-alphanumeric "[\60-\71\101-\132\141-\172]")
291
292 (defvar kanji-char "\\cj")
293 (defvar kanji-space "$B!!(B")
294 (defvar kanji-symbols "\\cS")
295 (defvar kanji-numeric "[$B#0(B-$B#9(B]")
296 (defvar kanji-English-Upper "[$B#A(B-$B#Z(B]")
297 (defvar kanji-English-Lower "[$B#a(B-$B#z(B]")
298 (defvar kanji-hiragana "\\cH")
299 (defvar kanji-katakana "\\cK")
300 (defvar kanji-Greek-Upper "[$B&!(B-$B&8(B]")
301 (defvar kanji-Greek-Lower "[$B&A(B-$B&X(B]")
302 (defvar kanji-Russian-Upper "[$B'!(B-$B'A(B]")
303 (defvar kanji-Russian-Lower "[$B'Q(B-$B'q(B]")
304 (defvar kanji-Kanji-1st-Level "[$B0!(B-$BOS(B]")
305 (defvar kanji-Kanji-2nd-Level "[$BP!(B-$Bt$(B]")
306
307 (defvar kanji-kanji-char "\\(\\cH\\|\\cK\\|\\cC\\)")