Some things

[Dave Ahn <ahn@vec.wfubmc.edu>]

Here are a bunch of replies to several messages from Vanilla-list over the
last few weeks:

----
First, are we sure that Tom's (or whoever it was) patch to ntserv/detonate.c
is needed?  The dogfight robot already has its own custom detonate code in
robots/detonate.c:

	    /* Teammates' torps are always friendly except in dogfight,
	       when only own are */
	#ifdef DOGFIGHT
	    if (j == me)
	      continue;
	#else
	    if (Teammates(j, me))
	      continue;
	#endif

----
Second, Carlos, your patch to call ltd_update_deaths() actually breaks the
stats because it incorrectly adds a death count to every player at endgame.
What is needed is something like:

	for (i=0; i<MAXPLAYERS; i++) {
	  saveplayer(i);
	}

The sleep(2) line doesn't do anything, because player stats are saved only
when a player dies.  In the INL robot mode, players aren't killed off at
end game; their states (not stats) are reset using obliterate() in inl.c
(which doesn't save any stats).

The PW stats seemed to work fine, because they are log-file based and not
player file based.  I'll work on this in the next couple of days and get
a fix commited.

----
Tom, I do not have any examples of ltd_dump output other than the one at
http://pickled.fox.cs.cmu.edu/netrek/results/beta-server/ltd_dump.txt

However, I noticed that there are a couple of games/scrimages scheduled
on that server this week.  You could use the output from those.

>First, SB stats don't appear in the stats I have; are those working at all?
>What's the schedule on them?

Yes, ltd_dump.c version 1.2 and higher will report out LTD starbase stats
summary at the top of the file.  The starbase stats are also included in the
SB column for each player.  If you found some missing SB stats (for example,
there was one game where Kevin's base lasted until end-game, but ltd_dump
reports him as having had 0 stats in SB), it is due to the more
general saveplayer@endgame problem I reported before I went on vacation.
This is something not directly related to the stats package.  I will commit
a fix to this soon, maybe tomorrow if I get over this jet lag.

>Kill/death types are not comprehesive; there are deaths by phaser, torp,
>and smack, but not det, explosion or planet.  This makes for significant
>discrepancies between total deaths and the death types, and it also leaves
>out some important information.

They are not comprehensive, because they were not listed in the original
stats documents you sent me over a year ago.  I had provided the full list
of the LTD stats to inlcouncil and vanilla-list and called for requests for
additional stats, but nobody asked for them.

Having said this, it is fairly easy to add kills/deaths by det/explosion/
planet.  For now, kills by det/planet happen rarely incomparison to
explosions, so you could approximate the number by (kt - kbp - kbt - kbs).

I will go ahead and add these additional stats.  If there are other stats
that are needed, please tell me as soon as possible.  I want to put the
LTD stuff behind me and start working on some other things needed in the
server.

>Do we have any provision for reporting pop/plague and other army statistics
>on the server side?

No, I don't believe so.  At one point, I did have army pop stats incorporated
into the robot, but I will need to track down that chunk of code.

-- 
Dave Ahn <ahn@vec.wfubmc.edu>        |  "When you were born, you cried and the
                                     |  world rejoiced.  Try to live your life
Virtual Endoscopy Center             |  so that when you die, you will rejoice
Wake Forest Univ. School of Medicine |  and the world will cry."  -1/2 jj^2