bug#5387: MS950 alias for CP950 charset

bug#5387: MS950 alias for CP950 charset

[jidanni]

I signed up on the Legislature of Taiwan's website, and the confirmation
mail had
From: sysop@ly.gov.tw
Subject: =?BIG5?B?pd+qa7B8pf6yebjqsFS69C2/76XBqkGwyLHSsMq9VLt7qOc=?=
Mime-Version: 1.0
Content-Type: text/html; charset=MS950
Content-Transfer-Encoding: quoted-printable

And it turns out MS950 is an alias for CP950, so perhaps emacs should
incorporate this alias, even though this is the first time I've seen it.
Perhaps make all MSxxx be aliases for CPxxx.

bug#5387: MS950 alias for CP950 charset

[Kenichi Handa]

In article <87aawfzl6w.fsf@jidanni.org>, jidanni@jidanni.org writes:

> I signed up on the Legislature of Taiwan's website, and the confirmation
> mail had
> From: sysop@ly.gov.tw
> Subject: =?BIG5?B?pd+qa7B8pf6yebjqsFS69C2/76XBqkGwyLHSsMq9VLt7qOc=?=
> Mime-Version: 1.0
> Content-Type: text/html; charset=MS950
> Content-Transfer-Encoding: quoted-printable

> And it turns out MS950 is an alias for CP950,

Where did you get that information?

> so perhaps emacs should
> incorporate this alias, even though this is the first time I've seen it.
> Perhaps make all MSxxx be aliases for CPxxx.

I checked <http://www.iana.org/assignments/character-sets>
and found that only MS936 is listed as an alias of GBK.

---
Kenichi Handa
handa@m17n.org

bug#5387: MS950 alias for CP950 charset

[jidanni]

>> And it turns out MS950 is an alias for CP950,
K> Where did you get that information?
I inferred it from my single encounter.
K> I checked <http://www.iana.org/assignments/character-sets>
K> and found that only MS936 is listed as an alias of GBK.
That makes two... or 1.5.

bug#5387: MS950 alias for CP950 charset

[Kenichi Handa]

In article <87my0fy44r.fsf@jidanni.org>, jidanni@jidanni.org writes:

>>> And it turns out MS950 is an alias for CP950,
>>> Where did you get that information?
> I inferred it from my single encounter.
>>> I checked <http://www.iana.org/assignments/character-sets>
>>> and found that only MS936 is listed as an alias of GBK.
> That makes two... or 1.5.

And,
iconv: MS932, MS936, MSCP949, MSCP1361
python: ms932, ms936, ms949, ms950, ms1361

hmmm...

I've just installed the attached change to accept msXXX.
canonicalize-coding-system-name is used by
coding-system-from-name, and, at least, rmail uses it.

If you are using rmail, please try the latest code, or the
attached patch.

---
Kenichi Handa
handa@m17n.org

=== modified file 'lisp/international/mule-cmds.el'
--- lisp/international/mule-cmds.el	2010-01-13 08:35:10 +0000
+++ lisp/international/mule-cmds.el	2010-01-15 12:33:24 +0000
@@ -226,19 +226,22 @@
 ;; and delimiter characters.  Support function of
 ;; coding-system-from-name.
 (defun canonicalize-coding-system-name (name)
-  (if (string-match "^iso[-_ ]?[0-9]" name)
-      ;; "iso-8859-1" -> "8859-1", "iso-2022-jp" ->"2022-jp"
-      (setq name (substring name (1- (match-end 0)))))
-  (let ((idx (string-match "[-_ /]" name)))
-    ;; Delete "-", "_", " ", "/" but do distinguish "16-be" and "16be".
-    (while idx
-      (if (and (>= idx 2)
-	       (eq (string-match "16-[lb]e$" name (- idx 2))
-		   (- idx 2)))
-	  (setq idx (string-match "[-_ /]" name (match-end 0)))
-	(setq name (concat (substring name 0 idx) (substring name (1+ idx)))
-	      idx (string-match "[-_ /]" name idx))))
-    name))
+  (if (string-match "^\\(ms\\|ibm\\|windows-\\)\\([0-9]+\\)$" name)
+      ;; "ms950", "ibm950", "windows-950" -> "cp950"
+      (concat "cp" (match-string 2 name))
+    (if (string-match "^iso[-_ ]?[0-9]" name)
+	;; "iso-8859-1" -> "8859-1", "iso-2022-jp" ->"2022-jp"
+	(setq name (substring name (1- (match-end 0)))))
+    (let ((idx (string-match "[-_ /]" name)))
+      ;; Delete "-", "_", " ", "/" but do distinguish "16-be" and "16be".
+      (while idx
+	(if (and (>= idx 2)
+		 (eq (string-match "16-[lb]e$" name (- idx 2))
+		     (- idx 2)))
+	    (setq idx (string-match "[-_ /]" name (match-end 0)))
+	  (setq name (concat (substring name 0 idx) (substring name (1+ idx)))
+		idx (string-match "[-_ /]" name idx))))
+      name)))
 
 (defun coding-system-from-name (name)
   "Return a coding system whose name matches with NAME (string or symbol)."

bug#5387: MS950 alias for CP950 charset

[Reiner Steib]

On Sat, Jan 16 2010, jidanni@jidanni.org wrote:

>>>>>> "K" == Kenichi Handa <handa@m17n.org> writes:
> K> I've just installed the attached change to accept msXXX.

Please add "(Bug#5387)" to the ChangeLog entry.

> K> canonicalize-coding-system-name is used by
> K> coding-system-from-name, and, at least, rmail uses it.
> I'll CC the gnus people to make sure they will use it too.

Gnus should use all coding-systems / charsets provided by Emacs.  No
change in Gnus required.

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

bug#5387: MS950 alias for CP950 charset

[Kenichi Handa]

In article <871vhqw8ss.fsf@marauder.physik.uni-ulm.de>, Reiner Steib <reinersteib+gmane@imap.cc> writes:

> On Sat, Jan 16 2010, jidanni@jidanni.org wrote:
>>>>>>> "K" == Kenichi Handa <handa@m17n.org> writes:
> > K> I've just installed the attached change to accept msXXX.

> Please add "(Bug#5387)" to the ChangeLog entry.

Ah, ok, just done.

> > K> canonicalize-coding-system-name is used by
> > K> coding-system-from-name, and, at least, rmail uses it.
> > I'll CC the gnus people to make sure they will use it too.

> Gnus should use all coding-systems / charsets provided by Emacs.  No
> change in Gnus required.

But MS950 is still not a coding-system in Emacs.
coding-system-from-name is a function to guess a coding
system from the given name.

(coding-system-from-name "MS950")
cp950
(coding-system-p 'MS950)
nil
(coding-system-p 'CP950)
nil
(coding-system-from-name "CP950")
cp950
(coding-system-p 'cp950)
t

---
Kenichi Handa
handa@m17n.org

bug#5387: MS950 alias for CP950 charset

[jidanni]

>>>>> "K" == Kenichi Handa <handa@m17n.org> writes:
K> I've just installed the attached change to accept msXXX.
K> canonicalize-coding-system-name is used by
K> coding-system-from-name, and, at least, rmail uses it.
I'll CC the gnus people to make sure they will use it too.