That did the trick. I was going through the code and I suspected that was the problem in cfg-grammar.y but I am no code monkey so I let it be ;-) Thanks again for your help. Now I hope that this parameter will help bring down the CPU load on one of my loghosts (it has a whole bunch of parallel connections). Thanks, -----Original Message----- From: syslog-ng-bounces@lists.balabit.hu [mailto:syslog-ng-bounces@lists.balabit.hu] On Behalf Of Roberto Nibali Sent: April 26, 2006 2:19 PM To: Syslog-ng users' and developers' mailing list Subject: Re: [PATCH] Re: [syslog-ng] time_sleep() ADDENDUM:

The grammar was missing an entry, the attached patch should fix it. Please try and report back if this solves the problem for you.

And my patch was missing the documentation update, which I've obtained by fraud from the marvellous syslog-ng-2 tree.

------------------------------------------------------------------------

--- syslog-ng-1.6.10/src/cfg-grammar.y 2006-03-14

00:34:13.000000000 +0100

+++ syslog-ng-1.6.10-ratz/src/cfg-grammar.y 2006-04-26 22:58:20.000000000 +0200 @@ -631,6 +631,7 @@ | KW_USE_DNS '(' yesno ')' { configuration->use_dns = $3; } | KW_TIME_REOPEN '(' NUMBER ')' { configuration->time_reopen = $3; } | KW_TIME_REAP '(' NUMBER ')' { configuration->time_reap = $3; } + | KW_TIME_SLEEP '(' NUMBER ')' { configuration->time_sleep = $3; } | KW_LOG_FIFO_SIZE '(' NUMBER ')' { configuration->log_fifo_size = $3; } | KW_GC_BUSY_THRESHOLD '(' NUMBER ')' { configuration->gc_busy_threshold = $3; } | KW_GC_IDLE_THRESHOLD '(' NUMBER ')' { configuration->gc_idle_threshold = $3; }

And attached is the better version for Bazsi to apply if he's inclined to do so. This is about the third time I've chased down a parser-based bug, so I wonder if using (f)lex/yacc for this really pays off ;)? Considering my state of mind, I reckon I could be really off the track with this patch as well ... Cheers, Roberto Nibali, ratz -- echo '[q]sa[ln0=aln256%Pln256/snlbx]sb3135071790101768542287578439snlbxq' | dc