Re: "send_charset" and saved record file's charset
On Wed, Jul 26, 2006 at 11:51:21AM +0200, Alain Bench wrote:
> -1) I believe that "iso-2022-jp" should be first, because any 2022-jp
Yes it should. I had euc-jp first since my (now defunct) filters
changed everything to euc-jp, so barring failure by nkf or iconv that
was the assumed charset.
> -2) It's useless to set a perfect subset together with a more complete
> charset: Set only the complete one. So "euc-jp-ms:euc-jp" could be
Noted. Forgot to clean things up after adding euc-jp-ms.
> very different, nearing incompatibility. For westerners it prohibits
> any usefull multiple $assumed_charset in practice: We have to use only
> one well choosen charset there.
I'm not sure that multiple $assumed_charset is good for CJK either.
> > send all Japanese mails in iso-2022-jp, [...] and save the record in
> > euc-jp or euc-jp-ms. [or] use utf-8 for sending, but save the record
> > in euc-jp or euc-jp-ms
>
> Impossible. The goal of FCC is to record *verbatim* what you sent.
Thinking about it, that IS what I want a *verbatim* record (BUT one
that I can read). Thanks for pointing that out.
> Would "iso-2022-jp-2" be acceptable? It is IANA approved, and has
> the fullwidth tilde (and circled numbers). Theoretically usable, but I
> ignore practical aspects. Would all or most your recipients be able to
> read it, including mailx/nkf users? If at all, it should be inserted as
Yes! Again YOU HAVE SAVED THE DAY! With iso-2022-jp-2 I get a record
that I can read with nvi and native OS tools. It will take some months
to really know if all recipients are able to read the mails. Of course
mail to myself read in mutt is okay. I also tested on a web-mail account,
and it was okay, too. If people have problems, they will be Mac and/or
Windows users. Fingers crossed because it does solve the last remaining
hitch that I know of.
THANKS!
--
henry nelson
WWW_HOME=http://yuba(dot)ne(dot)jp/(tilde)home/