On Wed, Apr 17, 2002 at 10:42:46PM +0200, Nicolas Rachinsky wrote: > * Edmund GRIMLEY EVANS <[EMAIL PROTECTED]> [2002-04-17 21:31:00 +0100]: > > Nicolas Rachinsky <[EMAIL PROTECTED]>: > > > > > > I think getmail does better than fetchmail if the network goes down > > > > while it's polling the server: fetchmail has an annoying habit of > > > > losing its fetchids when this happens, resulting in the delivery of > > > > several hundred duplicate messages in my case. However, I don't care > > > > about that so much any more as I have a procmail script to filter out > > > > duplicates (using an MD5 sum of the message contents rather than > > > > trusting Message-IDs). > > > > > > Would you share it? I would like such a thing. > > > > The start of my .procmailrc is below. The in-line Perl script removes > > all header fields except Date, From, Subject, To and Cc. MD5 sums are > > appended to $MAILDIR/MD5, so you should remove the beginning of that > > file from time to time. Duplicates are sent to =dupes. It's not > > efficient or elegant, but it seems to work. > > Thanks. Looks quite nice.
Somehow I've missed the start of Edmund's .procmailrc. But why wouldn't one do it this way rather than MD5 all that stuff manually? Have you found the message-id header to be that unreliable? :0 Wh: msgid.lock | formail -D 8192 .msgid.cache -- David Kelly N4HHE, [EMAIL PROTECTED] ===================================================================== The human mind ordinarily operates at only ten percent of its capacity -- the rest is overhead for the operating system.
