Re: [exim-dev] Exim sieve capabilities and pysieved

Top Page
Delete this message
Reply to this message
Author: Neale Pickett
Date:  
To: pdp, levan
CC: exim-dev@exim.org, Michael Haardt
Subject: Re: [exim-dev] Exim sieve capabilities and pysieved
I haven't worked with pysieved for over a year now. I think Philippe is
still active in development, but if he's not, ping me again and I'll
make the necessary changes to support this useful addition to exim.

(Yes, I think it's a good idea :)

Are people actually using pysieved with exim now? Cool!


On 6/1/2012, "Phil Pennock" <pdp@???> wrote:

>Gentlemen,
>
>I've just committed to Exim git a new command-line option, to
>interrogate the binary for a supported list of Sieve extensions.
>
>I anticipate that for daemons speaking ManageSieve (RFC 5804) this will
>be useful for providing an accurate SIEVE capability line.
>
>At present, it emits a hard-coded list of capabilities, taking into
>account any build-time options. It does *not* handle any configuration,
>which would vary on a per-Router basis!
>
>% ./build-test/exim -bI:sieve
>comparator-i;ascii-numeric
>copy
>encoded-character
>enotify
>envelope
>fileinto
>subaddress
>vacation
>%
>
>Does this look useful for you, or am I in need of a reality check?
>
>The precise option used is still unsettled; I've only just committed it,
>other devs might point out a better way of doing this. So far, I've
>grabbed -bI:$keyword and implemented "help" and "sieve" keywords.
>
>Without knowing how many different Routers might support sieve scripts,
>the best I can think of is something like -bI:sieve-list-routers to
>iterate all Routers and emit the names of those which are redirect
>routers with allow_filter set and forbid_sieve_filter unset. Then
>permit something like "exim -bI:sieve:userforward" to interrogate the
>list of extensions supported for that Router; however, since options
>such as sieve_vacation_directory are expanded strings which can depend
>upon properties of the mail being routed, we can't guarantee useful or
>accurate results.
>
>So getting a completely correct list still needs some thought, but I
>think this gets us closer to something helpful.
>
>Feedback very much appreciated! We have until the next Exim release to
>sort out the interface details. :)
>
>-Phil