> I guess this is time now for a 1.0.3 release.

Pretty near, I'd think.

I have a "greatly improved" watchdog that I'd like to get tested and put
in, on the "altperl" side of things.

I have had a process that periodically restarts all the slons a bit more
than hourly whether they need it or not.  In retrospect, that seems way
too heavyhanded, as it makes numbers of log files grow, it leaves around
spurious back ends, and it's just happening because some of the network
connections are OCCASIONALLY a little bit flakey.

The new version is much lighter-handed about things, zapping things only
if it's not seeing an event flow.  I need to validate that I'm looking at
the event flow correctly, but I'm periodically watching output as I get
started on Neal Stephenson's new book, and it's looking like a BIG
improvement already.

The one oddity I'm seeing is that there occasionally appear to be "future"
events, as sl_event.ev_timestamp is sometimes LATER than now().  Or
perhaps that's evidence that there are a few seconds worth of time
difference between servers, and that the NTP gods need to be appeased...