Re: A proposal

Lloyd Zusman wrote:
> After a more detailed investigation consisting of perusing the tmda
> source code, I realize that there isn't currently any way to do all of
> the things I outlined in my recent post on the tmda-user list (quoted
> below).

Hi Lloyd,

I am all for the action hooks you kindly offered to work into the code 
for everybody. ;-) But in the mean-time there are, depending on your 
setup, quite simple ways to do at least half of what you are trying to 
achieve, i.e. the part that does not involve tmda-cgi. (And provided 
learning with crm14 works in a way similar to spamassassin...)

For automatically learning delivered messages as ham: 
http://mla.libertine.org/tmda-users/2004-01/msg00526.html
In my case I have the sa-learn command in a .qmail file...

For learning messages that are about to expire from the queue as spam: 
Simply run the following bash scrip via cron once a day:
(temp=$(export PAGER='sa-learn --spam --no-sync'; tmda-pending -bS -O 
13d; sa-learn --sync);)

Of course you can also extend your queue lifetime by a day and use 14d 
in the script if you want to stick to TMDA standards.

That leaves your tmda-cgi problems. I have not really looked for a 
solution to that as I am using a maildir as pending queue and rarely 
release mail - and if then by hand via my mail reader. In any case, in 
this situation even action hooks will be of no use...

Hope this helps.

Kind regards,
Arik


PS: A word of caution. Do set the working directory of the bash script 
to an empty directory. I am not sure what sa-learn (or crm14) does when 
the pipe input is empty because no messages are expiring. In the case of 
sa-learn I suspect it will learn the contents of the current working 
directory... - Not good if it is a user's home directory. ;-)