bobd@bobd.hf.intel.com (Bob Dietrich) (04/09/91)
I recently brought up 6.7.1 on SysV 3.2.2 (386 cpu) and decided to be more efficient by having slocal filter some of my mail as it arrived. However, there seems to be a problem with appending mail to my default mailbox. It appears that slocal locks the mailbox (/usr/mail/bobd in my case) when the program starts, and when it then tries to append to this file, it fails because the lock is set. This behavior occurs with and without LOCKF versions of slocal and inc. And yes, inc and rcvstore are owned by group "mail" and group SUID. Did I build MH wrong? Is my .maildelivery file wrong? Any suggestions (other than using a second mailbox)? Thanks. Bob Dietrich Intel Corporation, Hillsboro, Oregon (503) 696-4270 (FAX: 696-4210) usenet: bobd@bobd.intel.com or uunet!intelhf!ijf1!bobd or tektronix!ogccse!omepd!ijf1!bobd or tektronix!psu-cs!omepd!ijf1!bobd ========================== Configuration: version: MH 6.7.1 #4[UCI] (bobd) of Wed Mar 27 11:38:13 PST 1991 options: [SYS5] [SYS5DIR] [TYPESIG=void] [SHADOW] [SOCKETS] [BIND] [ATZ] [TZNAME] [MORE='"/usr/bin/more"'] [FOLDPROT='"0700"'] [DUMB] [MHRC] [WHATNOW] [NOMHSEQ] [MHE] [RPATHS] [SBACKUP='"#"'] [SENDMTS] [SMTP] [SPRINTFTYPE=int] ========================== My .maildelivery file: Subject Omega pipe A "/usr/local/lib/mh/rcvstore +omega" Subject "Stock Prices" pipe A "/usr/local/lib/mh/rcvstore +stock" Subject "Stock Prices" pipe A "$HOME/bin/xshowstock" default - file ? /usr/mail/bobd ========================== Test message (file testmail): To: bobd@bobd Subject: test mail for slocal -------- foo bar. ========================== Result of executing "cat testmail | /usr/local/lib/mh/slocal -verbose -debug": temporary file "/tmp/slocala07250" selected addr="bobd" user="bobd" info="" file="/tmp/slocala07250" sender="L^A^D" mbox="/usr/mail/bobd" home="/usr2/bobd" from="" ddate="Delivery-Date: Thu, 11 Apr 91 11:48:38 PDT " now=11:48 vec[0]: "Subject" vec[1]: "Omega" vec[2]: "pipe" vec[3]: "A" vec[4]: "/usr/local/lib/mh/rcvstore +omega" vars[3]: name="reply-to" value="" hdrs[0]: name="source" value="" hdrs[1]: name="addr" value="bobd" hdrs[2]: name="Return-Path" value="L^A^D" hdrs[3]: name="Reply-To" value="L^A^D" hdrs[4]: name="From" value="L^A^D" hdrs[5]: name="Sender" value="L^A^D" hdrs[6]: name="To" value=" bobd@bobd " hdrs[7]: name="cc" value="L^A^D" hdrs[8]: name="Resent-Reply-To" value="L^A^D" hdrs[9]: name="Resent-From" value="L^A^D" hdrs[10]: name="Resent-Sender" value="L^A^D" hdrs[11]: name="Resent-To" value="L^A^D" hdrs[12]: name="Resent-cc" value="L^A^D" hdrs[13]: name="Subject" value=" test mail for slocal " vec[0]: "Subject" vec[1]: "Stock Prices" vec[2]: "pipe" vec[3]: "A" vec[4]: "/usr/local/lib/mh/rcvstore +stock" vec[0]: "Subject" vec[1]: "Stock Prices" vec[2]: "pipe" vec[3]: "A" vec[4]: "$HOME/bin/xshowstock" vec[0]: "default" vec[1]: "-" vec[2]: "file" vec[3]: "?" vec[4]: "/usr/mail/bobd" delivering to file "/usr/mail/bobd", unable to open:Error 0 (trying normal delivery) delivering to file "/usr/mail/bobd", unable to open:Error 0
truesdel@nas.nasa.gov (David A. Truesdell) (04/12/91)
bobd@bobd.hf.intel.com (Bob Dietrich) writes: >I recently brought up 6.7.1 on SysV 3.2.2 (386 cpu) and decided to be more >efficient by having slocal filter some of my mail as it arrived. However, >there seems to be a problem with appending mail to my default mailbox. It >appears that slocal locks the mailbox (/usr/mail/bobd in my case) when the >program starts, and when it then tries to append to this file, it fails >because the lock is set. This behavior occurs with and without LOCKF >versions of slocal and inc. And yes, inc and rcvstore are owned by group >"mail" and group SUID. Did I build MH wrong? Is my .maildelivery file >wrong? Any suggestions (other than using a second mailbox)? Thanks. Try simply letting the message "drop out" of the .maildelivery file. It work's fine for me. I expect that slocal would give up any special permissions before attempting to append to a file, or execute a user specified program. If /usr/mail/bobd is not writeable by bobd, it would explain your troubles. -- T.T.F.N., dave truesdell (truesdel@nas.nasa.gov) Ita erat quando hic adventi.