Re: problem with cvs snapshot?

[mea@nic.funet.fi]

> Hello,
> 
> just wondering if anyone has hit the following error using the latest
> snapshot (50s9 on Aug 1) in the CVS tree:
> 
> zmailer: interpret: recursed more than 40 levels deep on invocation!  script termination condition error ?
> 
> I got this error while the router is processing a message with 173
> recipients (after list expansion using a non-standard program). The
> error occurred after it has processed the 79th address.  After this
> error occurs, the router stops processing messages and mailq will
> return "no daemon" for router status (though the router daemon is
> still running).
> 
> At first I thought that I have a cf script problem, but the problem
> persisted even after I replaced my cf scripts with fresh copies from
> the cvs tree. I was running 50s2 before (which successfully processed
> the same message).

	Ok, so if the problem isn't in the scripts, then something
	which has been done to the router in the mean time is the
	problem...


	Say, when you did 'cvs update', did you just do 'make'
	to get new binaries, or 'make clean;make' ?
	(after doing reconfigure ?)

	The Makefiles do not do sufficient dependency directed
	compilation without command 'make depend' right after
	the (re)configuration phase where the fresh set of
	Makefiles are generated.


	If the problem occurs with clean compile too, then I would
	like to see a full transcript of a processing session with
	'rtrace' mode active.  The transcript will be long, no doubt..

	# $MAILBIN/router -i
	...
	z# cd $POSTOFFICE/router
	z# rtrace
	z# process 12345-678		<<-- whatever the file spoolid is

> -- 
> Ambrose Li > home > acli@ada.ddns.org (fast but probably quite unreliable)

/Matti Aarnio <mea@nic.funet.fi>