|
333
|
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 Amdahl Corporation.
|
|
|
5 ;; Copyright (C) 1995 Sun Microsystems.
|
|
|
6
|
|
|
7 ;; This file is part of XEmacs.
|
|
|
8
|
|
|
9 ;; XEmacs is free software; you can redistribute it and/or modify it
|
|
|
10 ;; under the terms of the GNU General Public License as published by
|
|
|
11 ;; the Free Software Foundation; either version 2, or (at your option)
|
|
|
12 ;; any later version.
|
|
|
13
|
|
|
14 ;; XEmacs is distributed in the hope that it will be useful, but
|
|
|
15 ;; WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
16 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
17 ;; General Public License for more details.
|
|
|
18
|
|
|
19 ;; You should have received a copy of the GNU General Public License
|
|
|
20 ;; along with XEmacs; see the file COPYING. If not, write to the
|
|
|
21 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
|
22 ;; Boston, MA 02111-1307, USA.
|
|
|
23
|
|
|
24 ;;; Commentary:
|
|
|
25
|
|
|
26 ;; Functions for working with category tables, which are a particular
|
|
|
27 ;; type of char table. Some function names / arguments should be
|
|
|
28 ;; parallel with syntax tables.
|
|
|
29
|
|
412
|
30 ;; Written by Ben Wing <wing@666.com>. The initialization code
|
|
333
|
31 ;; at the end of this file comes from Mule.
|
|
|
32 ;; Some bugfixes by Jareth Hein <jhod@po.iijnet.or.jp>
|
|
|
33
|
|
|
34 ;;; Code:
|
|
|
35
|
|
412
|
36 (defvar defined-category-hashtable (make-hashtable 50))
|
|
333
|
37
|
|
|
38 (defun define-category (designator doc-string)
|
|
|
39 "Make a new category whose designator is DESIGNATOR.
|
|
|
40 DESIGNATOR should be a visible letter of ' ' thru '~'.
|
|
|
41 STRING is a doc string for the category.
|
|
|
42 Letters of 'a' thru 'z' are already used or kept for the system."
|
|
|
43 (check-argument-type 'category-designator-p designator)
|
|
|
44 (check-argument-type 'stringp doc-string)
|
|
|
45 (puthash designator doc-string defined-category-hashtable))
|
|
|
46
|
|
|
47 (defun undefine-category (designator)
|
|
|
48 "Undefine DESIGNATOR as a designator for a category."
|
|
|
49 (check-argument-type 'category-designator-p designator)
|
|
|
50 (remhash designator defined-category-hashtable))
|
|
|
51
|
|
|
52 (defun defined-category-p (designator)
|
|
|
53 "Return non-nil if DESIGNATOR is a designator for a defined category."
|
|
|
54 (and (category-designator-p designator)
|
|
|
55 (gethash designator defined-category-hashtable)))
|
|
|
56
|
|
|
57 (defun defined-category-list ()
|
|
|
58 "Return a list of the currently defined categories.
|
|
|
59 Categories are given by their designators."
|
|
|
60 (let (list)
|
|
|
61 (maphash #'(lambda (key value)
|
|
|
62 (setq list (cons key list)))
|
|
|
63 defined-category-hashtable)
|
|
|
64 (nreverse list)))
|
|
|
65
|
|
|
66 (defun undefined-category-designator ()
|
|
|
67 "Return an undefined category designator, or nil if there are none."
|
|
|
68 (let ((a 32) found)
|
|
|
69 (while (and (< a 127) (not found))
|
|
412
|
70 (if (gethash a defined-category-hashtable)
|
|
|
71 (setq found a))
|
|
333
|
72 (setq a (1+ a)))
|
|
|
73 found))
|
|
|
74
|
|
|
75 (defun category-doc-string (designator)
|
|
|
76 "Return the doc-string for the category denoted by DESIGNATOR."
|
|
|
77 (check-argument-type 'defined-category-p designator)
|
|
|
78 (gethash designator defined-category-hashtable))
|
|
|
79
|
|
|
80 (defun modify-category-entry (char-range designator &optional table reset)
|
|
|
81 "Add a category to the categories associated with CHAR-RANGE.
|
|
|
82 CHAR-RANGE is a single character or a range of characters,
|
|
|
83 as per `put-char-table'.
|
|
|
84 The category is given by a designator character.
|
|
|
85 The changes are made in TABLE, which defaults to the current buffer's
|
|
|
86 category table.
|
|
|
87 If optional fourth argument RESET is non-nil, previous categories associated
|
|
|
88 with CHAR-RANGE are removed before adding the specified category."
|
|
|
89 (or table (setq table (category-table)))
|
|
|
90 (check-argument-type 'category-table-p table)
|
|
|
91 (check-argument-type 'defined-category-p designator)
|
|
|
92 (if reset
|
|
|
93 ;; clear all existing stuff.
|
|
|
94 (put-char-table char-range nil table))
|
|
|
95 (map-char-table
|
|
|
96 #'(lambda (key value)
|
|
|
97 ;; make sure that this range has a bit-vector assigned to it
|
|
|
98 (if (not (bit-vector-p value))
|
|
|
99 (setq value (make-bit-vector 95 0))
|
|
|
100 (setq value (copy-sequence value)))
|
|
|
101 ;; set the appropriate bit in that vector.
|
|
|
102 (aset value (- designator 32) 1)
|
|
|
103 ;; put the vector back, thus assuring we have a unique setting for this range
|
|
|
104 (put-char-table key value table))
|
|
|
105 table char-range))
|
|
|
106
|
|
|
107 (defun char-category-list (char &optional table)
|
|
|
108 "Return a list of the categories that CHAR is in.
|
|
|
109 TABLE defaults to the current buffer's category table.
|
|
|
110 The categories are given by their designators."
|
|
|
111 (or table (setq table (category-table)))
|
|
|
112 (check-argument-type 'category-table-p table)
|
|
|
113 (let ((vec (get-char-table char table)))
|
|
|
114 (if (null vec) nil
|
|
|
115 (let ((a 32) list)
|
|
|
116 (while (< a 127)
|
|
|
117 (if (= 1 (aref vec (- a 32)))
|
|
412
|
118 (setq list (cons a list)))
|
|
333
|
119 (setq a (1+ a)))
|
|
|
120 (nreverse list)))))
|
|
|
121
|
|
412
|
122 ;; implimented in c, file chartab.c (97/3/14 jhod@po.iijnet.or.jp)
|
|
333
|
123 ;(defun char-in-category-p (char category &optional table)
|
|
|
124 ; "Return non-nil if CHAR is in CATEGORY.
|
|
|
125 ;TABLE defaults to the current buffer's category table.
|
|
|
126 ;Categories are specified by their designators."
|
|
|
127 ; (or table (setq table (category-table)))
|
|
|
128 ; (check-argument-type 'category-table-p table)
|
|
|
129 ; (check-argument-type 'category-designator-p category)
|
|
|
130 ; (let ((vec (get-char-table char table)))
|
|
|
131 ; (if (null vec) nil
|
|
|
132 ; (= 1 (aref vec (- category 32))))))
|
|
|
133
|
|
|
134 (defun describe-category ()
|
|
|
135 "Describe the category specifications in the category table.
|
|
|
136 The descriptions are inserted in a buffer, which is then displayed."
|
|
|
137 (interactive)
|
|
412
|
138 (with-output-to-temp-buffer "*Help*"
|
|
|
139 (describe-category-table (category-table) standard-output)))
|
|
333
|
140
|
|
|
141 (defun describe-category-table (table stream)
|
|
|
142 (let (first-char
|
|
|
143 last-char
|
|
|
144 prev-val
|
|
|
145 (describe-one
|
|
|
146 (lambda (first last value stream)
|
|
|
147 (if (and (bit-vector-p value)
|
|
|
148 (> (reduce '+ value) 0))
|
|
|
149 (progn
|
|
|
150 (if (equal first last)
|
|
|
151 (cond ((vectorp first)
|
|
|
152 (princ (format "%s, row %d"
|
|
|
153 (charset-name
|
|
|
154 (aref first 0))
|
|
|
155 (aref first 1))
|
|
|
156 stream))
|
|
|
157 ((charsetp first)
|
|
|
158 (princ (charset-name first) stream))
|
|
|
159 (t (princ first stream)))
|
|
|
160 (cond ((vectorp first)
|
|
|
161 (princ (format "%s, rows %d .. %d"
|
|
|
162 (charset-name
|
|
|
163 (aref first 0))
|
|
|
164 (aref first 1)
|
|
|
165 (aref last 1))
|
|
|
166 stream))
|
|
|
167 (t
|
|
|
168 (princ (format "%s .. %s" first last)
|
|
|
169 stream))))
|
|
|
170 (describe-category-code value stream))))))
|
|
|
171 (map-char-table
|
|
|
172 (lambda (range value)
|
|
|
173 (if (and (or
|
|
|
174 (and (characterp range)
|
|
|
175 (characterp first-char)
|
|
|
176 (eq (char-charset range) (char-charset first-char))
|
|
|
177 (= (char-to-int last-char) (1- (char-to-int range))))
|
|
|
178 (and (vectorp range)
|
|
|
179 (vectorp first-char)
|
|
|
180 (eq (aref range 0) (aref first-char 0))
|
|
|
181 (= (aref last-char 1) (1- (aref range 1))))
|
|
|
182 (equal value prev-val)))
|
|
|
183 (setq last-char range)
|
|
|
184 (if first-char
|
|
|
185 (progn
|
|
|
186 (funcall describe-one first-char last-char prev-val stream)
|
|
|
187 (setq first-char nil)))
|
|
|
188 (funcall describe-one range range value stream))
|
|
|
189 nil)
|
|
|
190 table)
|
|
|
191 (if first-char
|
|
|
192 (funcall describe-one first-char last-char prev-val stream))))
|
|
|
193
|
|
|
194 (defun describe-category-code (code stream)
|
|
|
195 (let ((standard-output (or stream standard-output)))
|
|
|
196 (princ "\tin categories: ")
|
|
|
197 (if (not (bit-vector-p code))
|
|
|
198 (princ "(none)")
|
|
|
199 (let ((i 0)
|
|
|
200 already-matched)
|
|
|
201 (while (< i 95)
|
|
|
202 (if (= 1 (aref code i))
|
|
|
203 (progn
|
|
|
204 (if (not already-matched)
|
|
|
205 (setq already-matched t)
|
|
|
206 (princ " "))
|
|
|
207 (princ (int-to-char (+ 32 i)))))
|
|
|
208 (setq i (1+ i)))
|
|
|
209 (if (not already-matched)
|
|
|
210 (princ "(none)")))
|
|
|
211 (let ((i 0))
|
|
|
212 (while (< i 95)
|
|
|
213 (if (= 1 (aref code i))
|
|
|
214 (princ (format "\n\t\tmeaning: %s"
|
|
|
215 (category-doc-string (int-to-char (+ 32 i))))))
|
|
|
216 (setq i (1+ i)))))
|
|
|
217 (terpri)))
|
|
|
218
|
|
|
219 (defconst predefined-category-list
|
|
|
220 '((latin-iso8859-1 ?l "Latin-1 through Latin-5 character set")
|
|
|
221 (latin-iso8859-2 ?l)
|
|
|
222 (latin-iso8859-3 ?l)
|
|
|
223 (latin-iso8859-4 ?l)
|
|
|
224 (latin-iso8859-9 ?l)
|
|
|
225 (cyrillic-iso8859-5 ?y "Cyrillic character set")
|
|
|
226 (arabic-iso8859-6 ?b "Arabic character set")
|
|
|
227 (greek-iso8859-7 ?g "Greek character set")
|
|
|
228 (hebrew-iso8859-8 ?w "Hebrew character set")
|
|
|
229 (katakana-jisx0201 ?k "Japanese 1-byte Katakana character set")
|
|
|
230 (latin-jisx0201 ?r "Japanese 1-byte Roman character set")
|
|
|
231 (japanese-jisx0208-1978 ?j "Japanese 2-byte character set (old)")
|
|
|
232 (japanese-jisx0208 ?j "Japanese 2-byte character set")
|
|
|
233 (japanese-jisx0212 ?j)
|
|
|
234 (chinese-gb2312 ?c "Chinese GB (China, PRC) 2-byte character set")
|
|
|
235 (chinese-cns11643-1 ?t "Chinese Taiwan (CNS or Big5) 2-byte character set")
|
|
|
236 (chinese-cns11643-2 ?t)
|
|
|
237 (chinese-big5-1 ?t)
|
|
|
238 (chinese-big5-2 ?t)
|
|
|
239 (korean-ksc5601 ?h "Hangul (Korean) 2-byte character set")
|
|
|
240 )
|
|
|
241 "List of predefined categories.
|
|
|
242 Each element is a list of a charset, a designator, and maybe a doc string.")
|
|
|
243
|
|
|
244 (let (i l)
|
|
|
245 (define-category ?a "ASCII character set.")
|
|
|
246 (setq i 32)
|
|
|
247 (while (< i 127)
|
|
|
248 (modify-category-entry i ?a)
|
|
|
249 (setq i (1+ i)))
|
|
|
250 (setq l predefined-category-list)
|
|
|
251 (while l
|
|
|
252 (if (and (nth 2 (car l))
|
|
|
253 (not (defined-category-p (nth 2 (car l)))))
|
|
|
254 (define-category (nth 1 (car l)) (nth 2 (car l))))
|
|
|
255 (modify-category-entry (car (car l)) (nth 1 (car l)))
|
|
|
256 (setq l (cdr l))))
|
|
|
257
|
|
|
258 ;;; At the present, I know Japanese and Chinese text can
|
|
|
259 ;;; break line at any point under a restriction of 'kinsoku'.
|
|
|
260 (defvar word-across-newline "\\(\\cj\\|\\cc\\|\\ct\\)"
|
|
|
261 "Regular expression of such characters which can be a word across newline.")
|
|
|
262
|
|
|
263 (defvar ascii-char "[\40-\176]")
|
|
|
264 (defvar ascii-space "[ \t]")
|
|
|
265 (defvar ascii-symbols "[\40-\57\72-\100\133-\140\173-\176]")
|
|
|
266 (defvar ascii-numeric "[\60-\71]")
|
|
|
267 (defvar ascii-English-Upper "[\101-\132]")
|
|
|
268 (defvar ascii-English-Lower "[\141-\172]")
|
|
|
269 (defvar ascii-alphanumeric "[\60-\71\101-\132\141-\172]")
|
|
|
270
|
|
|
271 (defvar kanji-char "\\cj")
|
|
|
272 (defvar kanji-space "$B!!(B")
|
|
|
273 (defvar kanji-symbols "\\cS")
|
|
|
274 (defvar kanji-numeric "[$B#0(B-$B#9(B]")
|
|
|
275 (defvar kanji-English-Upper "[$B#A(B-$B#Z(B]")
|
|
|
276 (defvar kanji-English-Lower "[$B#a(B-$B#z(B]")
|
|
|
277 (defvar kanji-hiragana "\\cH")
|
|
|
278 (defvar kanji-katakana "\\cK")
|
|
|
279 (defvar kanji-Greek-Upper "[$B&!(B-$B&8(B]")
|
|
|
280 (defvar kanji-Greek-Lower "[$B&A(B-$B&X(B]")
|
|
|
281 (defvar kanji-Russian-Upper "[$B'!(B-$B'A(B]")
|
|
|
282 (defvar kanji-Russian-Lower "[$B'Q(B-$B'q(B]")
|
|
|
283 (defvar kanji-Kanji-1st-Level "[$B0!(B-$BOS(B]")
|
|
|
284 (defvar kanji-Kanji-2nd-Level "[$BP!(B-$Bt$(B]")
|
|
|
285
|
|
|
286 (defvar kanji-kanji-char "\\(\\cH\\|\\cK\\|\\cC\\)")
|
