Tim Legant <[EMAIL PROTECTED]> writes:

> I'd send him a request to add another number (not a flag but a set
> of flags) that programs like TMDA could use (non-MUAs).  He's
> deliberately designed that 'info' field so it's infinitely
> extensible.

This is a good idea, but still I go back to the issue of coexisting
with MUAs.

MUAs only seem to understand supported ``info starting with "2,"''
flags and don't behave nicely when confronted with anything else, such
as ``info starting with "3,"'' flags or a "2," flag it doesn't
recognize.

For example, take the file:

1037926661.362732.acl:3,C

If you read the message with Mutt, Mutt blows away the ``:3,C'', and
the file is now named

1037926661.362732.acl:2,S

Similarly:

1037926661.362732.acl:2,C

is renamed

1037926661.362732.acl:2,S

So I think for pending to be a Maildir, TMDA would have to understand
and use only the ``info starting with "2,"'' flags defined for MUAs on
djb's Maildir page to have any hope of coexisting with MUAs---and
that's the whole point. If you can't use an MUA safely on pending,
there's no point to make it a Maildir.

Does this sound right?

If so, the question becomes, is it possible for TMDA to use only the
MUA flags, and still retain its crucial functionality?

I think TMDA only needs to distinguish between a delivered and a still
pending message. The further refinement between delivered by release
and delivered by confirmation isn't that important I don't think.

See http://cr.yp.to/proto/maildir.html

Perhaps the "F" flag can be used to indicate the message has been
delivered by TMDA? If it's safe to assume the MUA won't toggle this
on, then maybe. Does anyone know when (if) the MUA toggles "F" on a
Maildir message?
_________________________________________________
tmda-workers mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-workers

Reply via email to