+01Jan00=======================================================================
+1. Fixed Julian routines to get rid of the millenium bug! (do as I say, not
+as I do!).
+2. updated gdx.pl so that external programs can do a sh/dx just as a normal
+user can. On reflection this could be used as a model for all sorts of
+"external" interface routines.
+11Dec99=======================================================================
+1. added missing pc48 so that it calls the DXDb::process routine
+10Dec99=======================================================================
+1. allow msgs to marked as read or unread
+02Dec99=======================================================================
+1. DX command doesn't recognise '6M' as '6m' in SH/DX ON 6M.
+2. set/node now sets the homenode as the callsign.
+3. changed the label for the 18Mhz band from '18m' to '17m' in the bands
+table.
+01Dec99=======================================================================
+1. started the callbot which lives at n2tly.dxcluster.net 9000.
+2. added www.qrz.com
+3. added www.buck.com
+4. added k4ute at no4j.com
+28Nov99=======================================================================
+1. fixed all the stupid errors introduced in 1.36 confusing DXUser->get with
+DXCluster->get_exact.
+24Nov99=======================================================================
+1. extended badmsg.pl stuff so that it works like the forward file and allows
+one to reject on to, from, origin and subject fields.
+2. added a swop.pl again in a forward.pl like format that allows simple swop
+for a field or test on a field and set another.
+3. Fixed the problem that read private mail follows you around as well as
+unread private mail does.
+4. Change the logic of connections so that the node will accept callsigns with
+SSIDs to -8 as different entities from each other. SSIDs > 8 are subtracted
+from 15 and are assumed to be coming in via netrom diddled connections. This
+affects lots of the commands as well. I may not have got them all.
+5. Hangup clients on HUP signals. I originally ignored them, for reasons that
+are now lost in the mists of time. HUP signals now cause clients to die
+(cleanly) and disconnect everything correctly.
+6. Added show/time command
+21Nov99=======================================================================
+1. lose extraneous rubbish after the callsign as sent by BBSes.
+2. Change end label on sh/moon to RGain dB rather than Loss dB.
+3. Remove spurious call to DXCluster->get_exact in pc50.
+4. Allow rcmd to locally connected nodes that are not (for some reason) in the
+nodes table.
+5. Try and send a PC39 in more cases to a cluster that I want to disconnect
+from in the hope that it notices and disconnects its end.
+20Nov99=======================================================================
+1. Added set/unset logininfo which will tell anybody that has this set when
+someone has either logged in or out of this node.
+2. Added set/pinginterval command which alters the length of time between
+pings sent to nodes, the value is minutes < 120 and seconds above that, 0
+means disable pinging (not recommended).
+3. If someone does a sysop whilst connected via ax25, don't show privilege
+level or passwd in stat/user.
+4. Fixed the problem with the dx command where if the callsign ended in a ?
+then the call and the freq would be replicated in the remarks.
+5. Set/unset ann, dx, talk, beep, wx, wwv all are now permanent (ie stored in
+the user file)