[Mailman-Developers] Mailman CVS sends out Japanese templatemails in EUC-JP

[Tokio Kikuchi]

Barry, Ben,

Sorry I am late in responding your discussion. 
(I am busy since 'they' are back to school. ;-)

Japanese (and may be other language) encoding glitch arise in

1. Archiving and displaying messages in Web Pages.
2. Moderation interfacing
and
3. Message body munging like adding header and footer.

All in all, my experience with good old Majordomo and Mailman 2.0
tells me that it is safe to convert the mail message when it comes
in and use only one code (EUC-JP) to process it. And, convert to
the conventional code (ISO-2022-JP) when the message is going out.

There may be many occasions when you want to add new feature in 
the internal processing but it is no good if you always have to 
think aout the code. EUC-JP is safe in using HTML when you must
escape < and > because all the double-byte characters are 8th bit set.

Perhaps we should collect more discussion from other language users
for converting other languages. 
I think we can put i18n of the message handling off to the 2.2 or later.
Until then, separate contributing patches from several language user 
may be sufficent.

> Here's a patch to add a new function to mimelib.Message.  It returns
> the charsets of each of the text/* parts in a message, and None for
> each part that is not text or does not have a charset.  This will make
> converting the iso-2022-jp parts much easier.

Ben, my Japanese patch worked at least for my environment. Internal
crafted messages are converted to ISO-2022-JP using the last-week CVS.
We should look at it more carefully, I think.
And, thanks for your suggestion to use mimelib for code-finding.
I will look at it when I have time.

--
Tokio