771
+ − 1 ;;; china-util.el --- utilities for Chinese -*- coding: iso-2022-7bit; -*-
+ − 2
+ − 3 ;; Copyright (C) 1995 Electrotechnical Laboratory, JAPAN.
+ − 4 ;; Licensed to the Free Software Foundation.
+ − 5
+ − 6 ;; Keywords: mule, multilingual, Chinese
+ − 7
+ − 8 ;; This file is part of XEmacs.
+ − 9
+ − 10 ;; XEmacs is free software; you can redistribute it and/or modify it
+ − 11 ;; under the terms of the GNU General Public License as published by
+ − 12 ;; the Free Software Foundation; either version 2, or (at your option)
+ − 13 ;; any later version.
+ − 14
+ − 15 ;; XEmacs is distributed in the hope that it will be useful, but
+ − 16 ;; WITHOUT ANY WARRANTY; without even the implied warranty of
+ − 17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ − 18 ;; General Public License for more details.
+ − 19
+ − 20 ;; You should have received a copy of the GNU General Public License
+ − 21 ;; along with XEmacs; see the file COPYING. If not, write to the Free
+ − 22 ;; Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
+ − 23 ;; 02111-1307, USA.
+ − 24
778
+ − 25 ;;; Synched up with: Emacs 21.1 (language/china-util.el).
+ − 26
+ − 27 ;;; Commentary:
771
+ − 28
+ − 29 ;;; Code:
+ − 30
+ − 31 ;; Hz/ZW encoding stuffs
+ − 32
+ − 33 ;; HZ is an encoding method for Chinese character set GB2312 used
+ − 34 ;; widely in Internet. It is very similar to 7-bit environment of
+ − 35 ;; ISO-2022. The difference is that HZ uses the sequence "~{" and
+ − 36 ;; "~}" for designating GB2312 and ASCII respectively, hence, it
+ − 37 ;; doesn't uses ESC (0x1B) code.
+ − 38
+ − 39 ;; ZW is another encoding method for Chinese character set GB2312. It
+ − 40 ;; encodes Chinese characters line by line by starting each line with
+ − 41 ;; the sequence "zW". It also uses only 7-bit as HZ.
+ − 42
+ − 43 ;; ISO-2022 escape sequence to designate GB2312.
+ − 44 (defvar iso2022-gb-designation "\e$A")
+ − 45 ;; HZ escape sequence to designate GB2312.
+ − 46 (defvar hz-gb-designnation "~{")
+ − 47 ;; ISO-2022 escape sequence to designate ASCII.
+ − 48 (defvar iso2022-ascii-designation "\e(B")
+ − 49 ;; HZ escape sequence to designate ASCII.
+ − 50 (defvar hz-ascii-designnation "~}")
+ − 51 ;; Regexp of ZW sequence to start GB2312.
+ − 52 (defvar zw-start-gb "^zW")
+ − 53 ;; Regexp for start of GB2312 in an encoding mixture of HZ and ZW.
+ − 54 (defvar hz/zw-start-gb
+ − 55 (concat hz-gb-designnation "\\|" zw-start-gb "\\|[^\0-\177]"))
+ − 56
+ − 57 (defvar decode-hz-line-continuation nil
+ − 58 "Flag to tell if we should care line continuation convention of Hz.")
+ − 59
+ − 60 (defconst hz-set-msb-table
+ − 61 (let ((str (make-string 127 0))
+ − 62 (i 0))
+ − 63 (while (< i 33)
+ − 64 (aset str i i)
+ − 65 (setq i (1+ i)))
+ − 66 (while (< i 127)
+ − 67 (aset str i (+ i 128))
+ − 68 (setq i (1+ i)))
+ − 69 str))
+ − 70
+ − 71 ;;;###autoload
+ − 72 (defun decode-hz-region (beg end)
+ − 73 "Decode HZ/ZW encoded text in the current region.
+ − 74 Return the length of resulting text."
+ − 75 (interactive "r")
+ − 76 (save-excursion
+ − 77 (save-restriction
+ − 78 (let (pos ch)
+ − 79 (narrow-to-region beg end)
+ − 80
+ − 81 ;; We, at first, convert HZ/ZW to `euc-china',
+ − 82 ;; then decode it.
+ − 83
+ − 84 ;; "~\n" -> "\n", "~~" -> "~"
+ − 85 (goto-char (point-min))
+ − 86 (while (search-forward "~" nil t)
+ − 87 (setq ch (following-char))
+ − 88 (if (or (= ch ?\n) (= ch ?~)) (delete-char -1)))
+ − 89
+ − 90 ;; "^zW...\n" -> Chinese GB2312
+ − 91 ;; "~{...~}" -> Chinese GB2312
+ − 92 (goto-char (point-min))
+ − 93 (setq beg nil)
+ − 94 (while (re-search-forward hz/zw-start-gb nil t)
+ − 95 (setq pos (match-beginning 0)
+ − 96 ch (char-after pos))
+ − 97 ;; Record the first position to start conversion.
+ − 98 (or beg (setq beg pos))
+ − 99 (end-of-line)
+ − 100 (setq end (point))
+ − 101 (if (>= ch 128) ; 8bit GB2312
+ − 102 nil
+ − 103 (goto-char pos)
+ − 104 (delete-char 2)
+ − 105 (setq end (- end 2))
+ − 106 (if (= ch ?z) ; ZW -> euc-china
+ − 107 (progn
+ − 108 (translate-region (point) end hz-set-msb-table)
+ − 109 (goto-char end))
+ − 110 (if (search-forward hz-ascii-designnation
+ − 111 (if decode-hz-line-continuation nil end)
+ − 112 t)
+ − 113 (delete-char -2))
+ − 114 (setq end (point))
+ − 115 (translate-region pos (point) hz-set-msb-table))))
+ − 116 (if beg
+ − 117 (decode-coding-region beg end 'euc-china)))
+ − 118 (- (point-max) (point-min)))))
+ − 119
+ − 120 ;;;###autoload
+ − 121 (defun decode-hz-buffer ()
+ − 122 "Decode HZ/ZW encoded text in the current buffer."
+ − 123 (interactive)
+ − 124 (decode-hz-region (point-min) (point-max)))
+ − 125
+ − 126 ;;;###autoload
+ − 127 (defun encode-hz-region (beg end)
+ − 128 "Encode the text in the current region to HZ.
+ − 129 Return the length of resulting text."
+ − 130 (interactive "r")
+ − 131 (save-excursion
+ − 132 (save-restriction
+ − 133 (narrow-to-region beg end)
+ − 134
+ − 135 ;; "~" -> "~~"
+ − 136 (goto-char (point-min))
+ − 137 (while (search-forward "~" nil t) (insert ?~))
+ − 138
+ − 139 ;; Chinese GB2312 -> "~{...~}"
+ − 140 (goto-char (point-min))
+ − 141 (if (re-search-forward "\\cc" nil t)
+ − 142 (let (pos)
+ − 143 (goto-char (setq pos (match-beginning 0)))
+ − 144 (encode-coding-region pos (point-max) 'iso-2022-7bit)
+ − 145 (goto-char pos)
+ − 146 (while (search-forward iso2022-gb-designation nil t)
+ − 147 (delete-char -3)
+ − 148 (insert hz-gb-designnation))
+ − 149 (goto-char pos)
+ − 150 (while (search-forward iso2022-ascii-designation nil t)
+ − 151 (delete-char -3)
+ − 152 (insert hz-ascii-designnation))))
+ − 153 (- (point-max) (point-min)))))
+ − 154
+ − 155 ;;;###autoload
+ − 156 (defun encode-hz-buffer ()
+ − 157 "Encode the text in the current buffer to HZ."
+ − 158 (interactive)
+ − 159 (encode-hz-region (point-min) (point-max)))
+ − 160
+ − 161 ;;
+ − 162 (provide 'china-util)
+ − 163
+ − 164 ;;; china-util.el ends here