2008-08-29 08:01:03 +00:00
|
|
|
|
;;; thai-util.el --- utilities for Thai -*- coding: utf-8; -*-
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2018-01-01 08:21:42 +00:00
|
|
|
|
;; Copyright (C) 2000-2018 Free Software Foundation, Inc.
|
2006-12-13 01:13:58 +00:00
|
|
|
|
;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
|
2011-01-02 23:50:46 +00:00
|
|
|
|
;; 2005, 2006, 2007, 2008, 2009, 2010, 2011
|
2003-09-08 12:53:41 +00:00
|
|
|
|
;; National Institute of Advanced Industrial Science and Technology (AIST)
|
2005-03-18 06:51:41 +00:00
|
|
|
|
;; Registration Number H14PRO021
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2002-06-12 16:46:15 +00:00
|
|
|
|
;; Keywords: mule, multilingual, Thai, i18n
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
;; This file is part of GNU Emacs.
|
|
|
|
|
|
2008-05-06 04:29:13 +00:00
|
|
|
|
;; GNU Emacs is free software: you can redistribute it and/or modify
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;; it under the terms of the GNU General Public License as published by
|
2008-05-06 04:29:13 +00:00
|
|
|
|
;; the Free Software Foundation, either version 3 of the License, or
|
|
|
|
|
;; (at your option) any later version.
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
;; GNU Emacs is distributed in the hope that it will be useful,
|
|
|
|
|
;; but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
|
;; GNU General Public License for more details.
|
|
|
|
|
|
|
|
|
|
;; You should have received a copy of the GNU General Public License
|
2017-09-13 22:52:52 +00:00
|
|
|
|
;; along with GNU Emacs. If not, see <https://www.gnu.org/licenses/>.
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2001-07-15 19:53:53 +00:00
|
|
|
|
;;; Commentary:
|
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;;; Code:
|
|
|
|
|
|
2005-08-29 14:06:44 +00:00
|
|
|
|
(defvar thai-auto-composition-mode)
|
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;; Setting information of Thai characters.
|
|
|
|
|
|
1999-12-15 00:46:54 +00:00
|
|
|
|
(defconst thai-category-table (make-category-table))
|
|
|
|
|
(define-category ?c "Thai consonant" thai-category-table)
|
|
|
|
|
(define-category ?v "Thai upper/lower vowel" thai-category-table)
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(define-category ?t "Thai tone mark" thai-category-table)
|
|
|
|
|
(define-category ?u "Thai tone mark and upper sign" thai-category-table)
|
|
|
|
|
(define-category ?I "THAI CHARACTER SARA I" thai-category-table)
|
|
|
|
|
(define-category ?U "THAI CHARACTER THANTHAKHAT" thai-category-table)
|
1999-12-15 00:46:54 +00:00
|
|
|
|
|
|
|
|
|
;; The general composing rules are as follows:
|
|
|
|
|
;;
|
|
|
|
|
;; T
|
2005-03-18 06:51:41 +00:00
|
|
|
|
;; V U V U
|
|
|
|
|
;; CV -> C, CU -> C, CVT -> C, Cv -> C, CvU -> C
|
1999-12-15 00:46:54 +00:00
|
|
|
|
;; v v
|
|
|
|
|
;;
|
2005-03-18 06:51:41 +00:00
|
|
|
|
;; where C: consonant, V: vowel upper, v: vowel lower,
|
|
|
|
|
;; T: tone mark, U: tone mark and upper sign.
|
2008-08-29 08:01:03 +00:00
|
|
|
|
;; Special rule: The sign `์' can be put on the vowel `ิ'.
|
1999-12-15 00:46:54 +00:00
|
|
|
|
|
2005-03-18 06:51:41 +00:00
|
|
|
|
|
|
|
|
|
(defvar thai-composition-pattern
|
|
|
|
|
"\\cc\\(\\cu\\|\\cI\\cU\\|\\cv\\ct?\\)\\|\\cv\\ct\\|\\cI\\cU"
|
1999-12-15 00:46:54 +00:00
|
|
|
|
"Regular expression matching a Thai composite sequence.")
|
1998-10-26 08:00:11 +00:00
|
|
|
|
|
2016-06-06 17:43:10 +00:00
|
|
|
|
(let ((l '((?ก consonant) ; 0xA1
|
|
|
|
|
(?ข consonant) ; 0xA2
|
|
|
|
|
(?ฃ consonant) ; 0xA3
|
|
|
|
|
(?ค consonant) ; 0xA4
|
|
|
|
|
(?ฅ consonant) ; 0xA5
|
|
|
|
|
(?ฆ consonant) ; 0xA6
|
|
|
|
|
(?ง consonant) ; 0xA7
|
|
|
|
|
(?จ consonant) ; 0xA8
|
|
|
|
|
(?ฉ consonant) ; 0xA9
|
|
|
|
|
(?ช consonant) ; 0xAA
|
|
|
|
|
(?ซ consonant) ; 0xAB
|
|
|
|
|
(?ฌ consonant) ; 0xAC
|
|
|
|
|
(?ญ consonant) ; 0xAD
|
|
|
|
|
(?ฎ consonant) ; 0xAE
|
|
|
|
|
(?ฏ consonant) ; 0xAF
|
|
|
|
|
(?ฐ consonant) ; 0xB0
|
|
|
|
|
(?ฑ consonant) ; 0xB1
|
|
|
|
|
(?ฒ consonant) ; 0xB2
|
|
|
|
|
(?ณ consonant) ; 0xB3
|
|
|
|
|
(?ด consonant) ; 0xB4
|
|
|
|
|
(?ต consonant) ; 0xB5
|
|
|
|
|
(?ถ consonant) ; 0xB6
|
|
|
|
|
(?ท consonant) ; 0xB7
|
|
|
|
|
(?ธ consonant) ; 0xB8
|
|
|
|
|
(?น consonant) ; 0xB9
|
|
|
|
|
(?บ consonant) ; 0xBA
|
|
|
|
|
(?ป consonant) ; 0xBB
|
|
|
|
|
(?ผ consonant) ; 0xBC
|
|
|
|
|
(?ฝ consonant) ; 0xBD
|
|
|
|
|
(?พ consonant) ; 0xBE
|
|
|
|
|
(?ฟ consonant) ; 0xBF
|
|
|
|
|
(?ภ consonant) ; 0xC0
|
|
|
|
|
(?ม consonant) ; 0xC1
|
|
|
|
|
(?ย consonant) ; 0xC2
|
|
|
|
|
(?ร consonant) ; 0xC3
|
|
|
|
|
(?ฤ vowel-base) ; 0xC4
|
|
|
|
|
(?ล consonant) ; 0xC5
|
|
|
|
|
(?ฦ vowel-base) ; 0xC6
|
|
|
|
|
(?ว consonant) ; 0xC7
|
|
|
|
|
(?ศ consonant) ; 0xC8
|
|
|
|
|
(?ษ consonant) ; 0xC9
|
|
|
|
|
(?ส consonant) ; 0xCA
|
|
|
|
|
(?ห consonant) ; 0xCB
|
|
|
|
|
(?ฬ consonant) ; 0xCC
|
|
|
|
|
(?อ consonant) ; 0xCD
|
|
|
|
|
(?ฮ consonant) ; 0xCE
|
|
|
|
|
(?ฯ special) ; 0xCF
|
|
|
|
|
(?ะ vowel-base) ; 0xD0
|
|
|
|
|
(?ั vowel-upper) ; 0xD1
|
|
|
|
|
(?า vowel-base) ; 0xD2
|
|
|
|
|
(?ำ vowel-base) ; 0xD3
|
|
|
|
|
(?ิ vowel-upper) ; 0xD4
|
|
|
|
|
(?ี vowel-upper) ; 0xD5
|
|
|
|
|
(?ึ vowel-upper) ; 0xD6
|
|
|
|
|
(?ื vowel-upper) ; 0xD7
|
|
|
|
|
(?ุ vowel-lower) ; 0xD8
|
|
|
|
|
(?ู vowel-lower) ; 0xD9
|
|
|
|
|
(?ฺ vowel-lower) ; 0xDA
|
|
|
|
|
(? invalid) ; 0xDA
|
|
|
|
|
(? invalid) ; 0xDC
|
|
|
|
|
(? invalid) ; 0xDC
|
|
|
|
|
(? invalid) ; 0xDC
|
|
|
|
|
(?฿ special) ; 0xDF
|
|
|
|
|
(?เ vowel-base) ; 0xE0
|
|
|
|
|
(?แ vowel-base) ; 0xE1
|
|
|
|
|
(?โ vowel-base) ; 0xE2
|
|
|
|
|
(?ใ vowel-base) ; 0xE3
|
|
|
|
|
(?ไ vowel-base) ; 0xE4
|
|
|
|
|
(?ๅ vowel-base) ; 0xE5
|
|
|
|
|
(?ๆ special) ; 0xE6
|
|
|
|
|
(?็ sign-upper) ; 0xE7
|
|
|
|
|
(?่ tone) ; 0xE8
|
|
|
|
|
(?้ tone) ; 0xE9
|
|
|
|
|
(?๊ tone) ; 0xEA
|
|
|
|
|
(?๋ tone) ; 0xEB
|
|
|
|
|
(?์ sign-upper) ; 0xEC
|
|
|
|
|
(?ํ sign-upper) ; 0xED
|
|
|
|
|
(?๎ sign-upper) ; 0xEE
|
|
|
|
|
(?๏ special) ; 0xEF
|
|
|
|
|
(?๐ special) ; 0xF0
|
|
|
|
|
(?๑ special) ; 0xF1
|
|
|
|
|
(?๒ special) ; 0xF2
|
|
|
|
|
(?๓ special) ; 0xF3
|
|
|
|
|
(?๔ special) ; 0xF4
|
|
|
|
|
(?๕ special) ; 0xF5
|
|
|
|
|
(?๖ special) ; 0xF6
|
|
|
|
|
(?๗ special) ; 0xF7
|
|
|
|
|
(?๘ special) ; 0xF8
|
|
|
|
|
(?๙ special) ; 0xF9
|
|
|
|
|
(?๚ special) ; 0xFA
|
|
|
|
|
(?๛ special) ; 0xFB
|
|
|
|
|
(? invalid) ; 0xFC
|
|
|
|
|
(? invalid) ; 0xFD
|
|
|
|
|
(? invalid) ; 0xFE
|
1997-02-20 07:02:49 +00:00
|
|
|
|
))
|
|
|
|
|
elm)
|
|
|
|
|
(while l
|
1999-12-15 00:46:54 +00:00
|
|
|
|
(setq elm (car l) l (cdr l))
|
|
|
|
|
(let ((char (car elm))
|
|
|
|
|
(ptype (nth 1 elm)))
|
|
|
|
|
(put-char-code-property char 'phonetic-type ptype)
|
|
|
|
|
(cond ((eq ptype 'consonant)
|
2005-04-09 01:42:04 +00:00
|
|
|
|
(modify-category-entry char ?c thai-category-table))
|
1999-12-15 00:46:54 +00:00
|
|
|
|
((memq ptype '(vowel-upper vowel-lower))
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(modify-category-entry char ?v thai-category-table)
|
2008-08-29 08:01:03 +00:00
|
|
|
|
(if (= char ?ิ)
|
2005-03-18 06:51:41 +00:00
|
|
|
|
;; Give category `I' to "SARA I".
|
2005-03-28 07:31:49 +00:00
|
|
|
|
(modify-category-entry char ?I thai-category-table)))
|
1999-12-15 00:46:54 +00:00
|
|
|
|
((eq ptype 'tone)
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(modify-category-entry char ?t thai-category-table)
|
2005-03-28 07:31:49 +00:00
|
|
|
|
(modify-category-entry char ?u thai-category-table))
|
2005-03-18 06:51:41 +00:00
|
|
|
|
((eq ptype 'sign-upper)
|
|
|
|
|
(modify-category-entry char ?u thai-category-table)
|
2008-08-29 08:01:03 +00:00
|
|
|
|
(if (= char ?์)
|
2005-03-18 06:51:41 +00:00
|
|
|
|
;; Give category `U' to "THANTHAKHAT".
|
2016-06-06 17:43:10 +00:00
|
|
|
|
(modify-category-entry char ?U thai-category-table)))))))
|
1998-10-26 08:00:11 +00:00
|
|
|
|
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(defun thai-compose-syllable (beg end &optional category-set string)
|
|
|
|
|
(or category-set
|
2005-04-04 09:17:41 +00:00
|
|
|
|
(setq category-set
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(char-category-set (if string (aref string beg) (char-after beg)))))
|
|
|
|
|
(if (aref category-set ?c)
|
|
|
|
|
;; Starting with a consonant. We do relative composition.
|
|
|
|
|
(if string
|
|
|
|
|
(compose-string string beg end)
|
|
|
|
|
(compose-region beg end))
|
|
|
|
|
;; Vowel tone sequence.
|
|
|
|
|
(if string
|
2005-04-04 09:17:41 +00:00
|
|
|
|
(compose-string string beg end (list (aref string beg) '(Bc . Bc)
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(aref string (1+ beg))))
|
2005-04-04 09:17:41 +00:00
|
|
|
|
(compose-region beg end (list (char-after beg) '(Bc . Bc)
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(char-after (1+ beg))))))
|
|
|
|
|
(- end beg))
|
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;;;###autoload
|
|
|
|
|
(defun thai-compose-region (beg end)
|
|
|
|
|
"Compose Thai characters in the region.
|
|
|
|
|
When called from a program, expects two arguments,
|
|
|
|
|
positions (integers or markers) specifying the region."
|
|
|
|
|
(interactive "r")
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(let ((pos (point)))
|
|
|
|
|
(save-restriction
|
|
|
|
|
(narrow-to-region beg end)
|
|
|
|
|
(goto-char (point-min))
|
|
|
|
|
(with-category-table thai-category-table
|
|
|
|
|
(while (re-search-forward thai-composition-pattern nil t)
|
|
|
|
|
(setq beg (match-beginning 0) end (match-end 0))
|
|
|
|
|
(if (and (> pos beg) (< pos end))
|
|
|
|
|
(setq pos end))
|
|
|
|
|
(thai-compose-syllable beg end
|
|
|
|
|
(char-category-set (char-after beg))))))
|
|
|
|
|
(goto-char pos)))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
1999-01-05 04:07:10 +00:00
|
|
|
|
;;;###autoload
|
|
|
|
|
(defun thai-compose-string (string)
|
|
|
|
|
"Compose Thai characters in STRING and return the resulting string."
|
1999-12-15 00:46:54 +00:00
|
|
|
|
(with-category-table thai-category-table
|
|
|
|
|
(let ((idx 0))
|
|
|
|
|
(while (setq idx (string-match thai-composition-pattern string idx))
|
2005-03-18 06:51:41 +00:00
|
|
|
|
(thai-compose-syllable idx (match-end 0) nil string)
|
1999-12-15 00:46:54 +00:00
|
|
|
|
(setq idx (match-end 0)))))
|
|
|
|
|
string)
|
2003-02-04 13:24:35 +00:00
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;;;###autoload
|
|
|
|
|
(defun thai-compose-buffer ()
|
|
|
|
|
"Compose Thai characters in the current buffer."
|
|
|
|
|
(interactive)
|
|
|
|
|
(thai-compose-region (point-min) (point-max)))
|
|
|
|
|
|
|
|
|
|
;;;###autoload
|
2008-08-29 08:01:03 +00:00
|
|
|
|
(defun thai-composition-function (gstring)
|
|
|
|
|
(if (= (lgstring-char-len gstring) 1)
|
|
|
|
|
(compose-gstring-for-graphic gstring)
|
|
|
|
|
(or (font-shape-gstring gstring)
|
|
|
|
|
(let ((glyph-len (lgstring-glyph-len gstring))
|
|
|
|
|
(last-char (lgstring-char gstring
|
|
|
|
|
(1- (lgstring-char-len gstring))))
|
|
|
|
|
(i 0)
|
|
|
|
|
glyph)
|
|
|
|
|
(while (and (< i glyph-len)
|
|
|
|
|
(setq glyph (lgstring-glyph gstring i)))
|
|
|
|
|
(setq i (1+ i)))
|
|
|
|
|
(if (= last-char ?ำ)
|
|
|
|
|
(setq i (1- i)))
|
|
|
|
|
(compose-glyph-string-relative gstring 0 i 0.1)))))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2005-03-29 07:50:59 +00:00
|
|
|
|
;; Thai-word-mode requires functions in the feature `thai-word'.
|
|
|
|
|
(require 'thai-word)
|
|
|
|
|
|
|
|
|
|
(defvar thai-word-mode-map
|
|
|
|
|
(let ((map (make-sparse-keymap)))
|
|
|
|
|
(define-key map [remap forward-word] 'thai-forward-word)
|
|
|
|
|
(define-key map [remap backward-word] 'thai-backward-word)
|
|
|
|
|
(define-key map [remap kill-word] 'thai-kill-word)
|
|
|
|
|
(define-key map [remap backward-kill-word] 'thai-backward-kill-word)
|
|
|
|
|
(define-key map [remap transpose-words] 'thai-transpose-words)
|
|
|
|
|
map)
|
|
|
|
|
"Keymap for `thai-word-mode'.")
|
|
|
|
|
|
|
|
|
|
(define-minor-mode thai-word-mode
|
|
|
|
|
"Minor mode to make word-oriented commands aware of Thai words.
|
2018-07-02 03:34:53 +00:00
|
|
|
|
|
|
|
|
|
The commands affected are \\[forward-word], \\[backward-word],
|
|
|
|
|
\\[kill-word], \\[backward-kill-word], \\[transpose-words], and
|
|
|
|
|
\\[fill-paragraph]."
|
2005-04-04 09:17:41 +00:00
|
|
|
|
:global t :group 'mule
|
2005-03-29 07:50:59 +00:00
|
|
|
|
(cond (thai-word-mode
|
|
|
|
|
;; This enables linebreak between Thai characters.
|
|
|
|
|
(modify-category-entry (make-char 'thai-tis620) ?|)
|
|
|
|
|
;; This enables linebreak at a Thai word boundary.
|
|
|
|
|
(put-charset-property 'thai-tis620 'fill-find-break-point-function
|
|
|
|
|
'thai-fill-find-break-point))
|
|
|
|
|
(t
|
|
|
|
|
(modify-category-entry (make-char 'thai-tis620) ?| nil t)
|
|
|
|
|
(put-charset-property 'thai-tis620 'fill-find-break-point-function
|
|
|
|
|
nil))))
|
|
|
|
|
|
|
|
|
|
;; Function to call on entering the Thai language environment.
|
|
|
|
|
(defun setup-thai-language-environment-internal ()
|
|
|
|
|
(thai-word-mode 1))
|
|
|
|
|
|
|
|
|
|
;; Function to call on exiting the Thai language environment.
|
|
|
|
|
(defun exit-thai-language-environment-internal ()
|
|
|
|
|
(thai-word-mode -1))
|
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;;
|
1997-06-18 13:06:17 +00:00
|
|
|
|
(provide 'thai-util)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
;;; thai-util.el ends here
|