I think it's a combination of primarily two things -- users setting up long-ish vacation messages and doing (arguably inelegant) things to black/whitelist addresses.
In the cases I've seen so far, a user will either have an explicit list of a few hundred email addresses from which they want all mail to be discarded or a list of 40-50 globbed domains they want to discard mail from coupled with a list of a few hundred "whitelisted" specific addresses. That's invariably in addition to "if (match) {discard;}" blocks matching on the spam analysis headers we add to inbound messages -- I haven't had a chance to analyze how much actual *effect* the black/whitelisting may be having, but I suspect a lot of it is actually superfluous, since most of what they're blacklisting is probably also being flagged by our spam detectors.
It's made worse, I think, by the way that our Autosieve user interface adds comments to the script when it's editied, and compounded by users who do a lot of:
if (a) {
blah
}
if (b) {
blah
}rather than:
if (a || b) {
blah
}but at this point, weaning the folks who use it a lot off the former model is more of a long-term solution -- lifting the size limit a bit is a shorter-term stopgap. Fortunately, 99% of our users who use sieve have scripts down in the 1-2K range -- it's just that the handful with large scripts are...vocal... :-) For the most part, the large scripts aren't complex -- just overly verbose and heavily commented...
--Thanx, --Rob--
Ken Murchison wrote:
I don't see where a 64k limit would be a problem, but >32k is a pretty big script. What exactly is causing the size to be so large?
--- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
