Bug-NUM_FOR_EDICT Error: Bad Pointer

[stefanhendriks]

I also thought there was a problem when you are trying to get the ent pointer from index 0. I thought you had to start from 1 till gpGLobals->maxclients...

so try and find any for loop which starts at 0.. Well, i can't bet on it. But i remembered in some cases it goes wrong here.

Another thing could be you have a <= instead of a <...

I am not sure, but perhaps its
0 till 31
instead of
1 till 32

you can figure this out yourself to be sure, gtg.

[[NvT]_KaszpiR_]

i have been testing podbots versions for a long time now
and i have noticed this problem form time to time
few notes

1. more bots means more probable the error will occur - i have tested that 20 botts will crash serverr very fast

2. it may be something with dying and maybe item dropping

3. i'm not sure if it is possible but i have noticed fw times that server crashed when i was under fire and was sure i gonna die and then seeing the bot that aimed at me was jumping

gonna make more detailed checks if you need
i hope the notices will not be misleading, so just think of them and not use them for 100%

one more thing - with the latest official release of podbot 2.6mm from ehre (that is around 2004-04-28 ) the bot crashes really much rarely
also i was wondering if some optimization methohds during compiling can have dramatic effect on this (like -02 or other switches that sometimes are not good)

i'm testing podbot on amd k6-II 500 on cs 1.6 linux box with amx
the bot crashes even if running pure hlds without plugins (except metamod)

[biohazerd87]

if you notice this error
Error:Failed to load dialog file Resource/UI/ClassMenu.res

Function trace:
CGameUI::Start
CGameUI::Initialize
CBaseUI::Start
CBaseUI::Initialize
Host_Init
CFileSystem_Steam::Mount()
it is trying to load a file "ClassMenu.res" which apparently doesn't exsist i don't know if this has a bearing on it but i hope it helps

[Huntkillaz]

as i said b4 it's a steam error and has nothing to do with bot's if u check the steam forums ....most ppl get that error and there's no proper fix for it...
and it doesn't really affect the game play n stuffs

[biohazerd87]

true but why do i get this error log everytime CS crashes with the NUM_FOR_EDICT Error

[Huntkillaz]

i'm not sure but i think it get generated at any hl.exe crash as i haven't had many crashes of "NUM_FOR_EDICT Error: Bad Pointer" yet i got plenty of error logs of that steam classres error

i think u'll have to system.out.print....umm c++ alternative to that to find out whats going on.....it'z a long way to go that way but it'z the best way...

[sPlOrYgOn]

okay i've did a few changes and I'd like for the people who get this error a lot to test it...
http://www.mapzap.org/files/podbot/podbotBETA.zip

[biohazerd87]

nope still getting the error

[sPlOrYgOn]

Then there really must be something else that can cause that error that I don't know about...

[edit]
finally been able to crash...
I seems to only get it on cs_office...
I've tried de_dust, awp_map, as_oilrig, cs_office, and cs_assault so far...
but only cs_office has crashed...
leaving me with log files.
http://www.mapzap.org/L0512006.log.txt
http://www.mapzap.org/L0512007.log.txt
both happened on cs_office
they still don't help me much
[/edit]

[[NvT]_KaszpiR_]

i got crashes on any map :/

one more thing
first i was using hlds with default binary, that is hlds_486
but i got amd k6-II cpu
in the result the hlds was freezing very often - not crash but freezed.

then i started to expeoriment with binary hlds_amd
not only the hlds works now much smoother but the bots are crashing the hlds and not freesing
i got crashes very very rarely now - myabe once on few map changes, but always...
i am using the podbot recompiled on my own machine with the fixes for the 'unnamed' bots, using gcc 2.95.3