utf-16.el 9.66 KB
Newer Older
Dave Love's avatar
Dave Love committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
;;; utf-16.el --- UTF-16 encoding/decoding

;; Copyright (C) 2001, 2002 Free Software Foundation, Inc.

;; Author: Dave Love <fx@gnu.org>
;; Keywords: Unicode, UTF-16, i18n

;; This file is part of GNU Emacs.

;; GNU Emacs is free software; you can redistribute it and/or modify
;; it under the terms of the GNU General Public License as published by
;; the Free Software Foundation; either version 2, or (at your option)
;; any later version.

;; 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
;; along with GNU Emacs; see the file COPYING.  If not, write to the
;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
;; Boston, MA 02111-1307, USA.

;;; Commentary:

;; Support for UTF-16, which is a two-byte encoding (modulo
;; surrogates) of Unicode, written either in little or big endian
;; order: coding-systems `mule-utf-16-le' and `mule-utf-16-be'.
;; (utf-16-le is used by the DozeN'T clipboard, for instance.)  The
;; data are preceeded by a two-byte signature which identifies their
;; byte sex.  These are used by the coding-category-utf-16-{b,l}e code
;; to identify the coding, but ignored on decoding.

;; Note that un-decodable sequences aren't (yet?) preserved as raw
;; bytes, as they are with utf-8, so reading and writing as utf-16 can
;; corrupt data.

;;; Code:

;; We end up with trivially different -le and -be versions of most
;; things below, sometimes with commonality abstracted into a let
;; binding for maintenance convenience.

;; We'd need new charsets distinct from ascii and eight-bit-control to
;; deal with untranslated sequences, since we can't otherwise
;; distinguish the bytes, as we can with utf-8.

;; ;; Do a multibyte write for bytes in r3 and r4.
;; ;; Intended for untranslatable utf-16 sequences.
;; (define-ccl-program ccl-mule-utf-16-untrans
;;   `(0
;;      (if (r3 < 128)
;; 	 (r0 = ,(charset-id 'ascii))
;;        (if (r3 < 160)
;; 	   (r0 = ,(charset-id 'eight-bit-control))
;; 	 (r0 = ,(charset-id 'eight-bit-graphic))))
;;      (if (r4 < 128)
;; 	 (r0 = ,(charset-id 'ascii))
;;        (if (r4 < 160)
;; 	   (r0 = ,(charset-id 'eight-bit-control))
;; 	 (r0 = ,(charset-id 'eight-bit-graphic))))
;;      (r1 = r4)))
;;   "Do a multibyte write for bytes in r3 and r4.
;; First swap them if we're big endian, indicated by r5==0.
;; Intended for untranslatable utf-16 sequences.")

;; Needed in macro expansion, so can't be let-bound.  Zapped after use.
(eval-and-compile
(defconst utf-16-decode-ucs
Kenichi Handa's avatar
Kenichi Handa committed
71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
  ;; We have the unicode in r1.  Output is charset ID in r0, code point
  ;; in r1.
  `((lookup-integer utf-subst-table-for-decode r1 r3)
    (if r7				; got a translation
	((r0 = r1) (r1 = r3))
      (if (r1 < 128)
	  (r0 = ,(charset-id 'ascii))
	(if (r1 < 160)
	    (r0 = ,(charset-id 'eight-bit-control))
	  (if (r1 < 256)
	      ((r0 = ,(charset-id 'latin-iso8859-1))
	       (r1 -= 128))
	    (if (r1 < #x2500)
		((r0 = ,(charset-id 'mule-unicode-0100-24ff))
		 (r1 -= #x100)
		 (r2 = (((r1 / 96) + 32) << 7))
		 (r1 %= 96)
		 (r1 += (r2 + 32)))
	      (if (r1 < #x3400)
		  ((r0 = ,(charset-id 'mule-unicode-2500-33ff))
		   (r1 -= #x2500)
		   (r2 = (((r1 / 96) + 32) << 7))
		   (r1 %= 96)
		   (r1 += (r2 + 32)))
		(if (r1 < #xd800)	; 2 untranslated bytes
		    ;;		 ;; Assume this is rare, so don't worry about the
		    ;; 		 ;; overhead of the call.
		    ;; 		 (call mule-utf-16-untrans)
		    ((r0 = ,(charset-id 'mule-unicode-e000-ffff))
		     (r1 = 15037))	; U+fffd
		  (if (r1 < #xe000)	; surrogate
		      ;; 			((call mule-utf-16-untrans)
		      ;; 			 (write-multibyte-character r0 r1)
		      ;; 			 (read r3 r4)
		      ;; 			 (call mule-utf-16-untrans))
		      ((read r3 r4)
		       (r0 = ,(charset-id 'mule-unicode-e000-ffff))
		       (r1 = 15037))
		    ((r0 = ,(charset-id 'mule-unicode-e000-ffff))
		     (r1 -= #xe000)
		     (r2 = (((r1 / 96) + 32) << 7))
		     (r1 %= 96)
		     (r1 += (r2 + 32))))))))))))))
Dave Love's avatar
Dave Love committed
114 115 116 117 118 119 120 121

(define-ccl-program ccl-decode-mule-utf-16-le
  `(2					; 2 bytes -> 1 to 4 bytes
    ((read r0 r1)			; signature
     (loop
      (read r3 r4)
      (r1 = (r4 <8 r3))
      ,utf-16-decode-ucs
Kenichi Handa's avatar
Kenichi Handa committed
122
      (translate-character utf-translation-table-for-decode r0 r1)
Dave Love's avatar
Dave Love committed
123 124 125 126
      (write-multibyte-character r0 r1)
      (repeat))))
  "Decode little endian UTF-16 (ignoring signature bytes).
Basic decoding is done into the charsets ascii, latin-iso8859-1 and
Kenichi Handa's avatar
Kenichi Handa committed
127 128 129
mule-unicode-*.  Un-representable Unicode characters are decoded as
U+fffd.  The result is run through the translation-table named
`utf-translation-table-for-decode'.")
Dave Love's avatar
Dave Love committed
130 131 132 133 134 135 136 137

(define-ccl-program ccl-decode-mule-utf-16-be
  `(2					; 2 bytes -> 1 to 4 bytes
    ((read r0 r1)			; signature
     (loop
      (read r3 r4)
      (r1 = (r3 <8 r4))
      ,utf-16-decode-ucs
Kenichi Handa's avatar
Kenichi Handa committed
138
      (translate-character utf-translation-table-for-decode r0 r1)
Dave Love's avatar
Dave Love committed
139 140 141 142 143
      (write-multibyte-character r0 r1)
      (repeat))))
  "Decode big endian UTF-16 (ignoring signature bytes).
Basic decoding is done into the charsets ascii, latin-iso8859-1 and
mule-unicode-*.  Un-representable Unicode characters are
Kenichi Handa's avatar
Kenichi Handa committed
144 145
decoded as U+fffd.  The result is run through the translation-table of
name `utf-translation-table-for-decode'.")
Dave Love's avatar
Dave Love committed
146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179

(makunbound 'utf-16-decode-ucs)		; done with it

(eval-and-compile
(defconst utf-16-decode-to-ucs
  ;; CCL which, given the result of a multibyte read in r0 and r1,
  ;; sets r0 to the character's Unicode if the charset is one of the
  ;; basic utf-8 coding system ones.  Otherwise set to U+fffd.
  `(if (r0 == ,(charset-id 'ascii))
       (r0 = r1)
     (if (r0 == ,(charset-id 'latin-iso8859-1))
	 (r0 = (r1 + 128))
       (if (r0 == ,(charset-id 'eight-bit-control))
	   (r0 = r1)
	 (if (r0 == ,(charset-id 'eight-bit-graphic))
	     (r0 = r1)
	   ((r2 = (r1 & #x7f))
	    (r1 >>= 7)
	    (r3 = ((r1 - 32) * 96))
	    (r3 += (r2 - 32))
	    (if (r0 == ,(charset-id 'mule-unicode-0100-24ff))
		(r0 = (r3 + #x100))
	      (if (r0 == ,(charset-id 'mule-unicode-2500-33ff))
		  (r0 = (r3 + #x2500))
		(if (r0 == ,(charset-id 'mule-unicode-e000-ffff))
		    (r0 = (r3 + #xe000))
		  (r0 = #xfffd)))))))))))

(define-ccl-program ccl-encode-mule-utf-16-le
  `(1
    ((write #xff)
     (write #xfe)
     (loop
      (read-multibyte-character r0 r1)
Kenichi Handa's avatar
Kenichi Handa committed
180 181 182 183
      (lookup-character utf-subst-table-for-encode r0 r1)
      (if (r7 == 0)
	  ((translate-character utf-translation-table-for-encode r0 r1)
	   ,utf-16-decode-to-ucs))
Dave Love's avatar
Dave Love committed
184 185 186 187 188 189
      (write (r0 & 255))
      (write (r0 >> 8))
      (repeat))))
  "Encode to little endian UTF-16 with signature.
Characters from the charsets ascii, eight-bit-control,
eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
Kenichi Handa's avatar
Kenichi Handa committed
190 191
after translation through the translation-table of name
`utf-translation-table-for-encode'.
Dave Love's avatar
Dave Love committed
192 193 194 195 196 197 198 199
Others are encoded as U+FFFD.")

(define-ccl-program ccl-encode-mule-utf-16-be
  `(1
    ((write #xfe)
     (write #xff)
     (loop
      (read-multibyte-character r0 r1)
Kenichi Handa's avatar
Kenichi Handa committed
200 201 202 203
      (lookup-character utf-subst-table-for-encode r0 r1)
      (if (r7 == 0)
	  ((translate-character utf-translation-table-for-encode r0 r1)
	   ,utf-16-decode-to-ucs))
Dave Love's avatar
Dave Love committed
204 205 206 207 208 209
      (write (r0 >> 8))
      (write (r0 & 255))
      (repeat))))
  "Encode to big endian UTF-16 with signature.
Characters from the charsets ascii, eight-bit-control,
eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
Kenichi Handa's avatar
Kenichi Handa committed
210 211
after translation through the translation-table named
`utf-translation-table-for-encode'.
Dave Love's avatar
Dave Love committed
212 213 214 215 216 217 218 219
Others are encoded as U+FFFD.")

(makunbound 'utf-16-decode-to-ucs)

(let ((doc "

Assumes and ignores the leading two-byte signature.

Kenichi Handa's avatar
Kenichi Handa committed
220 221 222 223 224 225 226 227 228 229 230 231 232
It supports Unicode characters of these ranges:
    U+0000..U+33FF, U+E000..U+FFFF.
They correspond to these Emacs character sets:
    ascii, latin-iso8859-1, mule-unicode-0100-24ff,
    mule-unicode-2500-33ff, mule-unicode-e000-ffff

On decoding (e.g. reading a file), Unicode characters not in the above
ranges are decoded as U+FFFD, effectively corrupting the data
if they are re-encoded.  

On encoding (e.g. writing a file), Emacs characters not belonging to
any of the character sets listed above are encoded into the byte
sequence representing U+FFFD (REPLACEMENT CHARACTER)."))
Dave Love's avatar
Dave Love committed
233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250
  (make-coding-system
   'mule-utf-16-le 4
   ?u	      ; Mule-UCS uses ?U, but code-pages uses that for koi8-u.
   (concat
    "Little endian UTF-16 encoding for Emacs-supported Unicode characters."
    doc)

   '(ccl-decode-mule-utf-16-le . ccl-encode-mule-utf-16-le)
   '((safe-charsets
      ascii
      eight-bit-control
      latin-iso8859-1
      mule-unicode-0100-24ff
      mule-unicode-2500-33ff
      mule-unicode-e000-ffff)
     (mime-charset . utf-16le)
     (coding-category . coding-category-utf-16-le)
     (valid-codes (0 . 255))
Kenichi Handa's avatar
Kenichi Handa committed
251 252 253 254 255
     (pre-write-conversion . utf-16-le-pre-write-conversion)
     (dependency unify-8859-on-encoding-mode
		 unify-8859-on-decoding-mode
		 utf-fragment-on-decoding
		 utf-translate-cjk)))
Dave Love's avatar
Dave Love committed
256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273

  (make-coding-system
   'mule-utf-16-be 4 ?u
   (concat
    "Big endian UTF-16 encoding for Emacs-supported Unicode characters."
    doc)

   '(ccl-decode-mule-utf-16-be . ccl-encode-mule-utf-16-be)
   '((safe-charsets
      ascii
      eight-bit-control
      latin-iso8859-1
      mule-unicode-0100-24ff
      mule-unicode-2500-33ff
      mule-unicode-e000-ffff)
     (mime-charset . utf-16be)
     (coding-category . coding-category-utf-16-be)
     (valid-codes (0 . 255))
Kenichi Handa's avatar
Kenichi Handa committed
274 275 276 277 278
     (pre-write-conversion . utf-16-be-pre-write-conversion)
     (dependency unify-8859-on-encoding-mode
		 unify-8859-on-decoding-mode
		 utf-fragment-on-decoding
		 utf-translate-cjk))))
Dave Love's avatar
Dave Love committed
279 280 281 282 283 284 285

(define-coding-system-alias 'utf-16-le 'mule-utf-16-le)
(define-coding-system-alias 'utf-16-be 'mule-utf-16-be)

(provide 'utf-16)

;;; utf-16.el ends here