Re: nnimap, agent and expiration

[Simon Josefsson <jas@extundo.com>]

kai.grossjohann@uni-duisburg.de (Kai Großjohann) writes:

> Simon, can you say something to clear this up?  Maybe the IMAP node
> should mention this?  Or the expiry node should mention it?

Indeed.  I added a sub-node to the IMAP node, similar to the node
about splitting in IMAP, on expiring in IMAP:

Expiring in IMAP
----------------

   Even though NNIMAP is not a proper NNMAIL derived backend, it
supports most features in regular expiring (see *note Expiring Mail::).
Unlike splitting in IMAP (see *note Splitting in IMAP::) it do not clone the
NNMAIL variables (i.e., creating NNIMAP-EXPIRY-WAIT) but reuse the
NNMAIL variables.  What follows below are the variables used by the
NNIMAP expiry process and any related notes on their usage.

   A note on how the expire mark is stored on the IMAP server is
appropriate here as well.  The expire mark is translated into a IMAP
client specific mark `gnus-expire' and store on the message.  This
means that only likely only Gnus will understand and treat the
`gnus-expire' mark properly, although other clients may allow to view
client specific flags on a message.  It also means that your server
must support the permanent storage of client specific flags on
messages.  Most though, fortunately.

`nnmail-expiry-wait'

`nnmail-expiry-wait-function'
     These variables are fully supported.  The expire value can be a
     number, the symbol IMMEDIATE or NEVER.

`nnmail-expiry-target'
     This variable is supported, and internally implemented by calling
     the NNMAIL functions that handle this.  It contains an optimization
     that if the destination is a IMAP group on the same server, the
     article is copied instead of appended (that is, uploaded again).