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

Barry A. Warsaw barry@zope.com
Mon, 10 Sep 2001 22:54:04 -0400


>>>>> "BG" == Ben Gertzfield <che@debian.org> writes:

    BG> However, when mail comes in to a list, we need to convert it
    BG> from the 7-bit ISO-2022-JP encoding format to EUC-JP before we
    BG> archive it.  Otherwise, the archive web pages will have the
    BG> static bits in EUC-JP, and the email contents in ISO-2022-JP!
    BG> This is a mess.

I really want to keep the archiver separate from the list exploder,
although I may be missing something which prevents this.  The reason I
say this is because sites can drop in their own archivers (and
probably should/do for high volume sites, since Pipermail has its own
share of problems).  So any solution for Mailman has to take into
account a custom archiver that Does The Right Thing and completely
bypasses Pipermail.

In fact, I can imagine a site implementing a custom ArchRunner, where
our Handler/ToArchive.py drops the message into qfiles/arch and from
there Mailman might never touch that copy of the message again.

Here's the scenario: APerson sends a message to a Japanese list.  It
shows up as ISO-2022-JP.  We do the normal IncomingRunner processing
on it and drop a copy into qfiles/out.  OutgoingRunner picks up that
copy and sends it to the list membership.  That message will also be
in ISO-2022-JP, right?  Or will it need to be converted to EUC-JP?  I
think/hope the answer is that it will go out to the list membership in
ISO-2022-JP.  What I meant by the "tough luck" comment earlier was
that if someone posts a message to a Japanese list, say in
Russian/koi8-r, well, we just send it on to the list membership in
that same encoding.

So for messages traveling the poster->Mailman->list membership route,
we don't need to convert encodings, right?

I may be missing something, but it seems that nothing you've said so
far contradicts this.  If I'm wrong, please correct me here before we
talk about the archiver.  I think I understand the issues involved
there, but I want to make sure I understand this part first.  For
right now, let's keep the archiver out of the picture.

Thanks!
-Barry