|
Challenge:
Home
News Archive
Forums
CPMA
Maps
Smackdown2
People
Coverage
Features
Interviews
Links
Help Wanted
Powered by:
Columns:
Hosted
3.A.C
Interfaced
QWF
Affiliates:
Cached
Methos
Killer Instincts
CPL Europe
|
Monday, 02 July 2001 -
Bug costs Harlsom DM2 (?) ? $comment_count ?>
When I saw the end of the (pretty boring) demo of the first game of Harlsom against Fatality on DM2, I immediately got the impression 'something' was wrong. Harlsoms comment at the games end seems to indicate he too thought something was strange. After looking at the demo again, I couldn't shake the feeling that Fatality should have died in the blast that lead to Harlsoms demise. Unfortunately Fatalitys point of view is not available.(full demo on CH-TV or just the last few seconds).
KTeam history log only says:
Added "Sudden death" overtime mode. If the time is over and results are tied, game won't end while frags are equal (that is, until one player dies).
But what happens in case both players should die at the same time, by the blast of the same grenade or rocket explosion or even in a lightning gun discharge? Shouldn't the overtime just continue until another frag occurs?
I just now tested a bit a with grenades (POV1, fires; POV2) and rockets in DM4 (POV1, 5 health; POV2, 1 health, fires and POV1, 1 health, fires; POV2, 5 health). In the grenade demo, the one firing the grenade dies, with the rockets, the opponent gets killed. I guess there is a logical explanation for this behaviour in the KTeam code, but I can't find a reasoning to go along with it.
It seems like the conditions for sudden death in KTeam at least need to be better defined. Anyway, as we know now, it didn't cost Harlsom anything in the end.
Comments
|
|
|
|
| I guess q3 has the blue bug, so maybe we should call this one the sudden fatality overtime bug? |
|
|
|
|
Do rocket/grenade explosions cause an instant sphere where anything inside of that sphere gets damage?
I thought the explosion acts like an actual explosion where it starts at a point and expands.. therefore harlsom may have been closer to the point of impact (it appeared to hit the floor, not fatality) and the explosion hit him first.
Then again I'm probably completely wrong :) |
|
|
|
|
yes, i think harlsom hit the floor. but i dont think explosions cause an expanding shockwave; just an instant effect. it would need more testing to be sure. but being on a server that allows games of 5 minutes minimum is pretty boring when you're just interested in overtime :)
|
|
|
|
|
Well i haven't watched the demos yet, but i think i know what this is about. The damage function that reduces the target's health is called sequentially to every player affected by a radius damage like a rocket splash or discharge. The part that checks if the guy should die is included in the damage function by Zoid, and KTeams haven't changed that (yet). Of course it is impossible to avoid the sequentiality since the CPU can only handle one task at the same time. The only possible way to fix it is to move the death-check part out of its current place and collect the list of players that should die in the same server frame, then specially treat them if sudden death is in progress.
Note that this could happen even in a 4on4 when for example a quaded player kills 2 enemies in front of him with a rocket shot that kills the attacker himself as well, yet it's possible that the quadguy is first on the call-list in the frame causing the first death to be him, -1 frag for his team and the match is over with them losing.
If there ever will be another KTeam update, this issue is on the todo list from now. |
|
|
|
|
Hey sturm, you're supposed to be one of KTeam coders! Just fix it (and a few other things, remember our talk on #sd-client?) and ask Ceno to release the new version.
|
|
|
|
|
Zoid's code is ok, CheckRules() which checks for fraglimit hit is called in PlayerPreThink(), and PlayerPreThink() will never be called inside of damage loop.
It's a KTeams bug - here's a piece of T_Damage() code:
if (targ.health <= 0)
{
Killed (targ, attacker);
// check if sudden death is the case
if(k_sudden_death && targ.classname == "player" ) EndMatch(0);
return;
}
A better idea i guess would be to put the check either into PlayerPreThink or (still better) into StartFrame: compare the players' frags and if they are different, the one who has more wins the match.
|
|
|
|
|
| http://www.inside3d.com/qip/q1/qcwa.htm#t_radiusdamage1 |
|
|
|
|
with grenades, the player who fires the grenade always gets the damage first in my experience. below i have linked a demo where i score a direct quad grenade hit on slash yet he only takes 1x damage, since the damage was first calculated on me (4x), i died meaning that quad was over, and then the damage was calculated on him (1x).
www.daddied.com/hangtime/quadgren.qwd
this demo also shows that it cant be an expanding shockwave over time, since the person hit by the grenade receives the damage last. as u can imagine, i was rather bemused at the fact he could survive a direct quad grenade (minimum 400 damage) hit on a map with max total health and armour of 350. FYI quad wasnt about to run out at that time either; and i still took 4x damage |
|
|
|
|
| also in this specific case, can u be sure that fatality should have been killed by the last rocket? if there is no demo from his pov, we dont know how much health/armour etc he had.... |
|
|
|
|
no; i am not sure that fatality should have been killed. it was just an incident that triggered me to look into it a little bit further.
i was thinking along the lines of what sturm said, with the sequential processing of the players (there is/was a similar (?) bug with shotgun where players stand in between). however these demos i made were all recorded one after the other, without either player leaving the server (although a spectator joined somewhere). i would think the sequence of the players should then remain the same, but the second and third demo indicate the opposite. |
|
|
|
|
| This is quite common in qwctf where you hook other players, if you then fire a pointblank quadrocket at the hookie (can't use hooker can I :-)) you will often die due to splash, and if you do, the opponent will often survive. |
|
|
|
|
| believe me it happens to me ALL the time :P |
Add A Comment
|