[Emacs-diffs] Changes to emacs/lisp/international/utf-16.el [EMACS_21_1_RC]

[Kenichi Handa]

Index: emacs/lisp/international/utf-16.el
diff -c emacs/lisp/international/utf-16.el:1.1.2.3 
emacs/lisp/international/utf-16.el:1.1.2.4
*** emacs/lisp/international/utf-16.el:1.1.2.3  Tue May  6 03:00:00 2003
--- emacs/lisp/international/utf-16.el  Fri Jun 20 22:06:11 2003
***************
*** 26,37 ****
  
  ;; 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.
--- 26,42 ----
  
  ;; Support for UTF-16, which is a two-byte encoding (modulo
  ;; surrogates) of Unicode, written either in little or big endian
! ;; order and either with or without the leading BOM (a two-byte
! ;; signature which identifies their byte sex)a.
! ;;
! ;; We provides these base coding systems.
! ;;    name                                    endian  BOM
! ;;    ----                                    ------  ---
! ;;    mule-utf-16le                           little  no
! ;;    mule-utf-16be                           big     no
! ;;    mule-utf-16le-with-signature            little  yes
! ;;    mule-utf-16be-with-signature            big     yes
! ;;
  ;; 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.
***************
*** 109,115 ****
                  (r1 %= 96)
                  (r1 += (r2 + 32)))))))))))
  
! (defconst utf-16-le-decode-loop
    `(loop
      (read r3 r4)
      (r1 = (r4 <8 r3))
--- 114,120 ----
                  (r1 %= 96)
                  (r1 += (r2 + 32)))))))))))
  
! (defconst utf-16le-decode-loop
    `(loop
      (read r3 r4)
      (r1 = (r4 <8 r3))
***************
*** 117,123 ****
      (write-multibyte-character r0 r1)
      (repeat)))
  
! (defconst utf-16-be-decode-loop
    `(loop
      (read r3 r4)
      (r1 = (r3 <8 r4))
--- 122,128 ----
      (write-multibyte-character r0 r1)
      (repeat)))
  
! (defconst utf-16be-decode-loop
    `(loop
      (read r3 r4)
      (r1 = (r3 <8 r4))
***************
*** 127,163 ****
  
  )
  
! (define-ccl-program ccl-decode-mule-utf-16-le
    `(2                                 ; 2 bytes -> 1 to 4 bytes
!     ,utf-16-le-decode-loop)
    "Decode UTF-16LE (little endian without signature bytes).
  Basic decoding is done into the charsets ascii, latin-iso8859-1 and
  mule-unicode-*.  Un-representable Unicode characters are decoded as
  U+fffd.")
  
! (define-ccl-program ccl-decode-mule-utf-16-be
    `(2                                 ; 2 bytes -> 1 to 4 bytes
!     ,utf-16-be-decode-loop)
    "Decode UTF-16BE (big endian without signature bytes).
  Basic decoding is done into the charsets ascii, latin-iso8859-1 and
  mule-unicode-*.  Un-representable Unicode characters are
  decoded as U+fffd.")
  
! (define-ccl-program ccl-decode-mule-utf-16-le-with-signature
    `(2
      ((read r3 r4)
!      ,utf-16-le-decode-loop))
!   "Like ccl-decode-utf-16-le but skip the first 2-byte BOM.")
  
! (define-ccl-program ccl-decode-mule-utf-16-be-with-signature
    `(2
      ((read r3 r4)
!      ,utf-16-be-decode-loop))
!   "Like ccl-decode-utf-16-be but skip the first 2-byte BOM.")
  
  (makunbound 'utf-16-decode-ucs)               ; done with it
! (makunbound 'utf-16-le-decode-loop)
! (makunbound 'utf-16-be-decode-loop)
  
  (eval-and-compile
  (defconst utf-16-decode-to-ucs
--- 132,168 ----
  
  )
  
! (define-ccl-program ccl-decode-mule-utf-16le
    `(2                                 ; 2 bytes -> 1 to 4 bytes
!     ,utf-16le-decode-loop)
    "Decode UTF-16LE (little endian without signature bytes).
  Basic decoding is done into the charsets ascii, latin-iso8859-1 and
  mule-unicode-*.  Un-representable Unicode characters are decoded as
  U+fffd.")
  
! (define-ccl-program ccl-decode-mule-utf-16be
    `(2                                 ; 2 bytes -> 1 to 4 bytes
!     ,utf-16be-decode-loop)
    "Decode UTF-16BE (big endian without signature bytes).
  Basic decoding is done into the charsets ascii, latin-iso8859-1 and
  mule-unicode-*.  Un-representable Unicode characters are
  decoded as U+fffd.")
  
! (define-ccl-program ccl-decode-mule-utf-16le-with-signature
    `(2
      ((read r3 r4)
!      ,utf-16le-decode-loop))
!   "Like ccl-decode-utf-16le but skip the first 2-byte BOM.")
  
! (define-ccl-program ccl-decode-mule-utf-16be-with-signature
    `(2
      ((read r3 r4)
!      ,utf-16be-decode-loop))
!   "Like ccl-decode-utf-16be but skip the first 2-byte BOM.")
  
  (makunbound 'utf-16-decode-ucs)               ; done with it
! (makunbound 'utf-16le-decode-loop)
! (makunbound 'utf-16be-decode-loop)
  
  (eval-and-compile
  (defconst utf-16-decode-to-ucs
***************
*** 184,190 ****
                    (r0 = (r3 + #xe000))
                  (r0 = #xfffd))))))))))
  
! (defconst utf-16-le-encode-loop
    `(loop
      (read-multibyte-character r0 r1)
      (translate-character utf-translation-table-for-encode r0 r1)
--- 189,195 ----
                    (r0 = (r3 + #xe000))
                  (r0 = #xfffd))))))))))
  
! (defconst utf-16le-encode-loop
    `(loop
      (read-multibyte-character r0 r1)
      (translate-character utf-translation-table-for-encode r0 r1)
***************
*** 193,199 ****
      (write (r0 >> 8))
      (repeat)))
  
! (defconst utf-16-be-encode-loop
    `(loop
      (read-multibyte-character r0 r1)
      (translate-character utf-translation-table-for-encode r0 r1)
--- 198,204 ----
      (write (r0 >> 8))
      (repeat)))
  
! (defconst utf-16be-encode-loop
    `(loop
      (read-multibyte-character r0 r1)
      (translate-character utf-translation-table-for-encode r0 r1)
***************
*** 204,212 ****
  )
  
  
! (define-ccl-program ccl-encode-mule-utf-16-le
    `(1
!     ,utf-16-le-encode-loop)
    "Encode to UTF-16LE (little endian without signature).
  Characters from the charsets ascii, eight-bit-control,
  eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
--- 209,217 ----
  )
  
  
! (define-ccl-program ccl-encode-mule-utf-16le
    `(1
!     ,utf-16le-encode-loop)
    "Encode to UTF-16LE (little endian without signature).
  Characters from the charsets ascii, eight-bit-control,
  eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
***************
*** 214,222 ****
  `utf-translation-table-for-encode'.
  Others are encoded as U+FFFD.")
  
! (define-ccl-program ccl-encode-mule-utf-16-be
    `(1
!     ,utf-16-be-encode-loop)
    "Encode to UTF-16BE (big endian without signature).
  Characters from the charsets ascii, eight-bit-control,
  eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
--- 219,227 ----
  `utf-translation-table-for-encode'.
  Others are encoded as U+FFFD.")
  
! (define-ccl-program ccl-encode-mule-utf-16be
    `(1
!     ,utf-16be-encode-loop)
    "Encode to UTF-16BE (big endian without signature).
  Characters from the charsets ascii, eight-bit-control,
  eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
***************
*** 224,234 ****
  `utf-translation-table-for-encode'.
  Others are encoded as U+FFFD.")
  
! (define-ccl-program ccl-encode-mule-utf-16-le-with-signature
    `(1
      ((write #xFF)
       (write #xFE)
!      ,utf-16-le-encode-loop))
    "Encode to UTF-16 (little endian with signature).
  Characters from the charsets ascii, eight-bit-control,
  eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
--- 229,239 ----
  `utf-translation-table-for-encode'.
  Others are encoded as U+FFFD.")
  
! (define-ccl-program ccl-encode-mule-utf-16le-with-signature
    `(1
      ((write #xFF)
       (write #xFE)
!      ,utf-16le-encode-loop))
    "Encode to UTF-16 (little endian with signature).
  Characters from the charsets ascii, eight-bit-control,
  eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
***************
*** 236,246 ****
  `utf-translation-table-for-encode'.
  Others are encoded as U+FFFD.")
  
! (define-ccl-program ccl-encode-mule-utf-16-be-with-signature
    `(1
      ((write #xFE)
       (write #xFF)
!      ,utf-16-be-encode-loop))
    "Encode to UTF-16 (big endian with signature).
  Characters from the charsets ascii, eight-bit-control,
  eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
--- 241,251 ----
  `utf-translation-table-for-encode'.
  Others are encoded as U+FFFD.")
  
! (define-ccl-program ccl-encode-mule-utf-16be-with-signature
    `(1
      ((write #xFE)
       (write #xFF)
!      ,utf-16be-encode-loop))
    "Encode to UTF-16 (big endian with signature).
  Characters from the charsets ascii, eight-bit-control,
  eight-bit-graphic, latin-iso8859-1 and mule-unicode-* are encoded
***************
*** 249,256 ****
  Others are encoded as U+FFFD.")
  
  (makunbound 'utf-16-decode-to-ucs)
! (makunbound 'utf-16-le-encode-loop)
! (makunbound 'utf-16-be-encode-loop)
  
  (let ((doc "
  
--- 254,261 ----
  Others are encoded as U+FFFD.")
  
  (makunbound 'utf-16-decode-to-ucs)
! (makunbound 'utf-16le-encode-loop)
! (makunbound 'utf-16be-encode-loop)
  
  (let ((doc "
  
***************
*** 268,280 ****
  any of the character sets listed above are encoded into the byte
  sequence representing U+FFFD (REPLACEMENT CHARACTER)."))
    (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
--- 273,285 ----
  any of the character sets listed above are encoded into the byte
  sequence representing U+FFFD (REPLACEMENT CHARACTER)."))
    (make-coding-system
!    'mule-utf-16le 4
     ?u       ; Mule-UCS uses ?U, but code-pages uses that for koi8-u.
     (concat
!     "UTF-16LE encoding for Emacs-supported Unicode characters."
      doc)
  
!    '(ccl-decode-mule-utf-16le . ccl-encode-mule-utf-16le)
     '((safe-charsets
        ascii
        eight-bit-control
***************
*** 288,299 ****
                 unify-8859-on-decoding-mode)))
  
    (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
--- 293,304 ----
                 unify-8859-on-decoding-mode)))
  
    (make-coding-system
!    'mule-utf-16be 4 ?u
     (concat
!     "UTF-16BE encoding for Emacs-supported Unicode characters."
      doc)
  
!    '(ccl-decode-mule-utf-16be . ccl-encode-mule-utf-16be)
     '((safe-charsets
        ascii
        eight-bit-control
***************
*** 307,319 ****
                 unify-8859-on-decoding-mode)))
  
    (make-coding-system
!    'mule-utf-16-le-with-signature 4 ?u
     (concat
      "Little endian UTF-16 (with BOM) for Emacs-supported Unicode characters."
      doc)
  
!    '(ccl-decode-mule-utf-16-le-with-signature
!      . ccl-encode-mule-utf-16-le-with-signature)
     '((safe-charsets
        ascii
        eight-bit-control
--- 312,324 ----
                 unify-8859-on-decoding-mode)))
  
    (make-coding-system
!    'mule-utf-16le-with-signature 4 ?u
     (concat
      "Little endian UTF-16 (with BOM) for Emacs-supported Unicode characters."
      doc)
  
!    '(ccl-decode-mule-utf-16le-with-signature
!      . ccl-encode-mule-utf-16le-with-signature)
     '((safe-charsets
        ascii
        eight-bit-control
***************
*** 328,340 ****
                 unify-8859-on-decoding-mode)))
  
    (make-coding-system
!    'mule-utf-16-be-with-signature 4 ?u
     (concat
      "Big endian UTF-16 (with BOM) for Emacs-supported Unicode characters."
      doc)
  
!    '(ccl-decode-mule-utf-16-be-with-signature
!      . ccl-encode-mule-utf-16-be-with-signature)
     '((safe-charsets
        ascii
        eight-bit-control
--- 333,345 ----
                 unify-8859-on-decoding-mode)))
  
    (make-coding-system
!    'mule-utf-16be-with-signature 4 ?u
     (concat
      "Big endian UTF-16 (with BOM) for Emacs-supported Unicode characters."
      doc)
  
!    '(ccl-decode-mule-utf-16be-with-signature
!      . ccl-encode-mule-utf-16be-with-signature)
     '((safe-charsets
        ascii
        eight-bit-control
***************
*** 348,358 ****
       (dependency unify-8859-on-encoding-mode
                 unify-8859-on-decoding-mode))))
  
! (define-coding-system-alias 'utf-16-le 'mule-utf-16-le)
! (define-coding-system-alias 'utf-16-be 'mule-utf-16-be)
! (define-coding-system-alias 'utf-16-le-with-signature
!   'mule-utf-16-le-with-signature)
! (define-coding-system-alias 'utf-16-be-with-signature
!   'mule-utf-16-be-with-signature)
  
  ;;; utf-16.el ends here
--- 353,369 ----
       (dependency unify-8859-on-encoding-mode
                 unify-8859-on-decoding-mode))))
  
! (define-coding-system-alias 'utf-16le 'mule-utf-16le)
! (define-coding-system-alias 'utf-16be 'mule-utf-16be)
! (define-coding-system-alias 'utf-16le-with-signature
!   'mule-utf-16le-with-signature)
! (define-coding-system-alias 'utf-16be-with-signature
!   'mule-utf-16be-with-signature)
! 
! ;; For backward compatibility.
! (define-coding-system-alias 'mule-utf-16-le 'mule-utf-16le-with-signature)
! (define-coding-system-alias 'utf-16-le 'mule-utf-16le-with-signature)
! (define-coding-system-alias 'mule-utf-16-be 'mule-utf-16be-with-signature)
! (define-coding-system-alias 'utf-16-be 'mule-utf-16be-with-signature)
  
  ;;; utf-16.el ends here