codepage.el: Codepage Mapping, Latin-7

codepage.el: Codepage Mapping, Latin-7

[Reiner Steib]

Hi,

from lisp/international/codepage.el:

--8<---------------cut here---------------start------------->8---
;; Support for the Windows 12xx series of codepages that MS has
;; butchered from the ISO-8859 specs. This does not add support for
;; the extended characters that MS has added in the 128 - 159 coding
;; range, only translates those characters that can be expressed in
;; the corresponding iso-8859 charset.

;; Codepage Mapping:
;;
;; Windows-1250: ISO-8859-2 (Central Europe) - differs in some positions
;; Windows-1251: ISO-8859-5 (Cyrillic)       - differs wildly
;; Windows-1252: ISO-8859-1 (West Europe)    - exact match
;; Windows-1253: ISO-8859-7 (Greek)          - differs in some positions
;; Windows-1254: ISO-8859-9 (Turkish)        - exact match
;; Windows-1255: ISO-8859-8 (Hebrew)         - exact match
;; Windows-1256: ISO-8859-6 (Arabic)         - half match
;; Windows-1257: ISO-8859-4 (Baltic)         - differs, future Latin-7
;; Windows-1258: VISCII (Vietnamese)         - Completely different
--8<---------------cut here---------------end--------------->8---

I think "exact match" is a little bit misleading, because
e.g. Windows-1252 is a superset of ISO-8859-1.  When reading the
comment on the "128 - 159 coding range" carefully it becomes more
clear the mapping table only refers to ISO-8859 code points.


BTW: The comment in `codepage.el' says that Latin-7 isn't official and
the it's not supported by Emacs.  This doesn't seem to be correct
anymore.

,----[ codepage.el ]
| ;; Since Latin-7 is not yet official, and Emacs does not support it,
| ;; provide translation between Windows-1257 and Latin-4 the best we
| ;; can.
`----

,----[ code-pages.el ]
| ;; used by lt_LT, lv_LV, mi_NZ
| ;;;###autoload(autoload-coding-system 'iso-8859-13 '(require 'code-pages))
| (cp-make-coding-system
|  iso-latin-7
| [...]
|  "Unicode-based encoding for Latin-7 (MIME: ISO-8859-13)"
|  ?l) ;; Lithuanian/Latvian
| (coding-system-put 'iso-latin-7 'mime-charset 'iso-8859-13)
| (define-coding-system-alias 'iso-8859-13 'iso-latin-7)
| (define-coding-system-alias 'latin-7 'iso-latin-7)
`----

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: codepage.el: Codepage Mapping, Latin-7

[Eli Zaretskii]

> From: Reiner Steib <reinersteib+gmane@imap.cc>
> Cc: Eli Zaretskii <eliz@gnu.org>
> Date: Thu, 16 Mar 2006 15:35:57 +0100
> 
> from lisp/international/codepage.el:
> 
> --8<---------------cut here---------------start------------->8---
> ;; Support for the Windows 12xx series of codepages that MS has
> ;; butchered from the ISO-8859 specs. This does not add support for
> ;; the extended characters that MS has added in the 128 - 159 coding
> ;; range, only translates those characters that can be expressed in
> ;; the corresponding iso-8859 charset.
> 
> ;; Codepage Mapping:
> ;;
> ;; Windows-1250: ISO-8859-2 (Central Europe) - differs in some positions
> ;; Windows-1251: ISO-8859-5 (Cyrillic)       - differs wildly
> ;; Windows-1252: ISO-8859-1 (West Europe)    - exact match
> ;; Windows-1253: ISO-8859-7 (Greek)          - differs in some positions
> ;; Windows-1254: ISO-8859-9 (Turkish)        - exact match
> ;; Windows-1255: ISO-8859-8 (Hebrew)         - exact match
> ;; Windows-1256: ISO-8859-6 (Arabic)         - half match
> ;; Windows-1257: ISO-8859-4 (Baltic)         - differs, future Latin-7
> ;; Windows-1258: VISCII (Vietnamese)         - Completely different
> --8<---------------cut here---------------end--------------->8---
> 
> I think "exact match" is a little bit misleading, because
> e.g. Windows-1252 is a superset of ISO-8859-1.  When reading the
> comment on the "128 - 159 coding range" carefully it becomes more
> clear the mapping table only refers to ISO-8859 code points.

Yes, the match is "exact" only in one direction: from ISO-8859
encoding to the codepage, but not the other way around.

It is meaningless to say that Windows-1252 is a ``superset'' of
ISO-8859-1, since these are encodings, not character sets.  The
underlying Emacs character set does not include characters which can
be encoded by Windows-1252, but codepage.el is only intersted in
encoding the ISO-8859-1 charset, not anything beyond that.

> BTW: The comment in `codepage.el' says that Latin-7 isn't official and
> the it's not supported by Emacs.  This doesn't seem to be correct
> anymore.

codepage.el is obsolete and is not developed anymore.  It is only used
by the MS-DOS port of Emacs; everything else uses code-pages.el.
codepage.el will die a natural death when the MS-DOS port dies,
probably when v23.1 is released.