spamd does not write to mailboxes. The -P option is for "spamassassin" which works somewhat differently. spamc will take the message from procmail, and pass it to spamd, which will process it, and return it back to spamc again, which will pass it on back to procmail. Without further instructions in its recipe list, procmail should then just deliver the message to the mail spool, and I would have thought it was smart enough to do locking in that default case. But what you're describing still sounds like a locking issue...
C Barry L. Kline wrote: BLK> Date: Mon, 15 Apr 2002 15:42:13 -0400 BLK> From: Barry L. Kline <[EMAIL PROTECTED]> BLK> To: [EMAIL PROTECTED] BLK> Cc: [EMAIL PROTECTED] BLK> Subject: Re: [SAtalk] Messages being munged [was: some random topic] BLK> BLK> Craig R Hughes wrote: BLK> > BLK> > Barry L. Kline wrote: BLK> > BLK> > BLK> Any ideas what I need to do to fix this? BLK> > BLK> > Quite likely a mailbox locking bug. Could you post your whole procmail recipe? BLK> > BLK> > C BLK> > BLK> BLK> Your comments and further investigation gives me an idea of the BLK> problem. Based on the docs from spamd it appears that spamd spawns a BLK> child for each message passed to it. The size of this attachment is BLK> causing a delay of in one child while other messages are being BLK> processed. Also in the docs is the tidbit that unless -P (pipe to BLK> STDOUT) is used with spamd, the output is written back to the mail BLK> spool. So Craig, I think you're dead on. I've got a couple of these BLK> child processes writing to the mailbox at one time. BLK> BLK> Now the big question -- how to fix this? Is spamd not locking mbox BLK> properly when it's writing out the cleansed message? BLK> BLK> Thanks very much for you help. BLK> BLK> Barry BLK> BLK> BLK> _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
