.:: Bots United ::. - jk_botti v1.20 release - bug fixes, new feature: lift support

.:: Bots United ::. (http://forums.bots-united.com/index.php)

- JKBotti (http://forums.bots-united.com/forumdisplay.php?f=83)

- - jk_botti v1.20 release - bug fixes, new feature: lift support (http://forums.bots-united.com/showthread.php?t=6377)

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

Ok installed, min_bots 0, max_bots 6. Now in the server there is 9 players and still one bot, it works better than before but still is not perfect. ;)

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

here's 1.22. I added 'jk_botti debug_minmax 1/0' option. Turn it on when min/max start going wrong and copy console debug output here.

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

Maybe something is wrong in your client struct...

Code:

status

#      name userid uniqueid frag time ping loss adr

# 1 "TimeJumpper" 20 UNKNOWN 101  1:20:15    0    0

# 2 "Digimon-man" 17 UNKNOWN 127  1:20:15    0    0

# 3 "Player_R0x[BMOD-lovers]" 15 UNKNOWN 135  1:20:15    0    0

# 4 "Medium_Sausage" 16 UNKNOWN 151  1:20:15    0    0

# 5 "Bitch_Slapper" 5 UNKNOWN  85  1:20:15    0    0

# 6 "mister_million_names" 8 UNKNOWN  89  1:20:15    0    0

# 7   "DuWi" 18 *********  80 16:51   72    0 *********

6 users

[jk_botti] debug_minmax mode ENABLED

[jk_botti] client count: 6, bot count: 5, max_bots: 6, min_bots: 0

[jk_botti] client_count < max_bots: FALSE

[jk_botti] client_count > max_bots: FALSE

[jk_botti] bot_count > min_bots: TRUE

[jk_botti] Should add bot: FALSE

[jk_botti] Should remove bot: FALSE

Before that there were 2 real users but still the bots was with one more bonus and when there is 6 users or more to play, there is always one bot in to the server.

EDIT: OK I just checked the server again and it seems that all is fine for now. (11:50 GMT+3)

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

Ok, I got new version to test with another way of check if client is valid.

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

OK I just installed it but something strange showed up in to the console:

Code:

[jk_botti] Loading waypoint file: valve/addons/jk_botti/waypoints/crossfire.wpt

[jk_botti] - loaded: 321 waypoints, 1338 paths

[jk_botti] [matrix load] - loading jk_botti waypoint path matrix

Got a NaN origin on ammo_crossbow

Got a NaN origin on ammo_crossbow

Got a NaN origin on ammo_crossbow

[jk_botti] Executing valve/addons/jk_botti/jk_botti.cfg

The map is crossfire and I don't have any other mm plugins exept JKBotti. But anyway I'm not 100% sure that "Got a NaN origin on ammo_crossbow" come from JKBotti.

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

Quote:

Originally Posted by The Storm (Post 55523)

"Got a NaN origin on ammo_crossbow" come from JKBotti.

I did get same warning on my linux server, so I removed jk_botti. Same error game up => I removed metamod. Same error => I developed win32 gamedll loader for linux-hlds (based on xine win32 codec loader). Same error game up.

I guess it's in engine bug (currently running engine_amd.so). I don't remember this appearing few years back so it's quite new. I'll check other engine_ix86.sos to see if it's compiler bug. Reverse-engineering shows that "Got a NaN origin on %s" is located in engine_xxxx.so/swds.dll ...

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

"Got a NaN origin on %s", SV_CheckVelocity in QuakeWorld. It seems that only one entity causes triple error output:
http://cvs.icculus.org/cvs/twilight/...phys.c?rev=1.6

I ran engine_i686 and engine_amd, both show error. engine_i486.so hangs on same place.

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

More debuging engine_i686.so shows that this isn't caused by one entity but once for every three dimensions of checked vector memory. Bug: Use of uninitialized vector memory?

I think I read about this sort of problem on some mm plugin.. gcc and msvc initialize array/vector memory differently, gcc making NaNs and msvc zero. HL-engine probably assumes msvc behavior.

Anyways it's valve bug.

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

Ok running JKBotti 1.23, here output.

Code:

#      name userid uniqueid frag time ping loss adr

# 1 "d.f.j.rocky" 469 ********   3 02:03   22    0 ************

# 2 "Freshman Neuman" 430 ******** 175  1:23:26  177    0 *********

# 3    "Neo" 440 ********* 248  1:09:25   33    0 ***********

# 4 "master_gollum" 465 *********   3 05:59   31    0 *********

# 5 "D@RKNE$$" 462 ***********  33 10:17   73    1 ***********

# 6 "Big_Van" 6 UNKNOWN 717 18:32:19    0    0

# 7 "deniz91" 454 ******** 110 30:59   22    0 **********

# 8 "master hacker" 447 *********  83 55:22  359    0 **********

# 9 "The cocrissiumKil" 429 **********  34  1:24:55   79    0 **********

#10 "unnamed" 460 **********  9 15:45   74    0 **********

#11  "DrFlo" 461 **********  41 11:04   19    0 **********

#12 "Sir Sui <Jan>" 463 **********  38 10:14   18    0 **********

#13 "albertmendez1955" 470 **********   0 01:23  164    0 **********

#14  "Wofka" 467 **********  10 04:58   23    0 **********

#15   "PyRo" 471 **********   0 00:53   70    0 9**********

#16     "bu" 472 **********   0 00:18  163    0 **********

16 users





[jk_botti] client count: 16, bot count: 1, max_bots: 6, min_bots: 0

[jk_botti] client_count < max_bots: FALSE

[jk_botti] client_count > max_bots: TRUE

[jk_botti] bot_count > min_bots: TRUE

[jk_botti] Should add bot: FALSE

[jk_botti] Should remove bot: TRUE

[jk_botti] Test UTIL_PickRandomBot(), return value: 0

"[jk_botti] Should remove bot: TRUE" but the bot is not removed...

EDIT: I will hide the STEAM_ID's and IP's for security reasons.

Re: jk_botti v1.20 release - bug fixes, new feature: lift support

Ah, logic error in PickRandomBot.