[mew-int 01217] Re: iso-8859-15 (aka Latin-0 or Latin-9)

[Kazu Yamamoto ( 山本和彦 )]

Hello, 

> I want to support a mechanism which allows you to live with Latin-0
> and Latin-1 in more convenient manner. The following questions were
> asked to a guy by me. To hear comments from other guys, I would like
> to disclose them here:

I have implemented a translation mechanism between Latin 0 and Latin 1
in Draft mode.

Please test this on Emacs 21 (not Emacs 20) with anon CVS version of
Mew.

You can set a default latin charset to mew-charset-latin. Its default
value is "iso-8859-15" (Latin 0). This means if there are both Latin 0
and Latin 1 characters in a draft, Latin 1 characters are converted to
Latin 0 if possible.

If you want to choose Latin 1, set mew-charset-latin to "iso-8859-1":
	(setq mew-charset-latin "iso-8859-1")

Concrete algorithm:

If there are both Latin 0 and Latin 1 characters in a draft, Mew
checks user's default value, say "iso-8859-15" (Latin 0). Mew then
checks whether or not conflicted characters of Latin 1 exist.

If not, Latin 1 can be converted to Latin 0 without loss. So, Mew
converts Latin 1 to Latin 0 and tells you through the mini-buffer.
The "charset" parameter becomes "iso-8859-15".

If exist, Latin 1 cannot be converted to Latin 0 without loss. So, Mew
does NOT convert Latin 1 to Latin 0. Eventually, mew-charset-m17n, its
default value is "utf-8", is used to encode the draft. The "charset"
parameter becomes "utf-8".

This algorithm works for both writing a new message and replying a
received message. The Mule-UCS package is not necessary unless "utf-8"
is used.

Please give a try and send me comments. Thanks in advance.

--Kazu