rob@pbhyf.PacBell.COM (Rob Bernardo) (01/28/89)
Just ran into an interesting situation at one site where ELM has just been installed. The default configuration of ELM has it check 6 times for a lock on mail spool files (i.e. /usr/mail/*) and then giving up (rather than assuming the lock is defunct and removing it). One site that just installed ELM, had the following line in the /etc/profile: mailx -e to inform a user who is logging in whether s/he has mail. Unfortunately 'mailx -e' leaves a mail lock file lying around for the user, and since ELM (in the defualt configuration) won't remove the lock, ELM would report the file locked and give up. They never noticed the problem before, because they were using mailx, and mailx would remove the lock file after the requisite number of tries. Two possible solutions. Don't use mailx -e, but rather mail -e, or configure ELM to remove the lock file after the requisite number of tries. I recommend the former and not the latter. -- Rob Bernardo, Pacific Bell UNIX/C Reusable Code Library Email: ...![backbone]!pacbell!pbhyf!rob OR rob@pbhyf.PacBell.COM Office: (415) 823-2417 Room 4E750A, San Ramon Valley Administrative Center Residence: (415) 827-4301 R Bar JB, Concord, California
david@pacbell.PacBell.COM (David St.Pierre) (01/30/89)
In article <4602@pbhyf.PacBell.COM> rob@pbhyf.PacBell.COM (Rob Bernardo) writes: > [mailx -e bug noticed] > >They never noticed the problem before, because they were using mailx, >and mailx would remove the lock file after the requisite number of tries. /bin/mail also removes it if the contents are non-numeric or (numeric and the process no longer exists). We have experienced problems with mail .lock files hanging around and preventing people from logging in again. If some mail programs put a numeric like "1" in the lockfile, that would cause a lot of problems. >Two possible solutions. Don't use mailx -e, but rather mail -e, or >configure ELM to remove the lock file after the requisite number of >tries. I recommend the former and not the latter. We should have mailx fixed on the 3b20s by Monday (the problem is in setfile()) but other systems should heed Rob's recommendation - the mailx -e option is broken in SV_R3.0. -- David St. Pierre 415/823-6800 {att,bellcore,sun,ames,pyramid}!pacbell!david