.:: Bots United ::. - PodBotmm on 32 Bit AMD?

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

- Bug Reports (http://forums.bots-united.com/forumdisplay.php?f=49)

- - PodBotmm on 32 Bit AMD? (http://forums.bots-united.com/showthread.php?t=7562)

Re: PodBotmm on 32 Bit AMD?

The crash can also happen at the start via hlds_run - directly at the beginning:

Code:

CRASH: Wed Apr 14 21:41:41 CEST 2010

Start Line: ./hlds_i686 -debug -nojoy -noipx -pidfile hlds.pid -debuglog debuglog.txt -port 27050 -tos -heapsize 131072 -zone 8192 -game cstrike +sv_lan 1 +sv_proxies 0 +maxplayers 20 +ip 178.63.3.149 +map de_dust

Core was generated by `./hlds_i686 -debug -nojoy -noipx -pidfile hlds.pid -debuglog debuglog.txt -port'.

Program terminated with signal 11, Segmentation fault.

[New process 10221]

#0  0xf75aa6dc in feof () from /lib32/libc.so.6

#0  0xf75aa6dc in feof () from /lib32/libc.so.6

#1  0xeec18e1b in MPluginList::ini_startup ()

   from ./cstrike/addons/metamod/dlls/metamod_i386.so

No symbol table info available.

From        To          Syms Read   Shared Object Library

0xf76e6320  0xf76e6428  Yes         /home/cs/hlds_nowon.so

0xf76e4400  0xf76e4638  Yes         /home/cs/hlds_fixes.so

0xf76e1a20  0xf76e27d8  Yes         /home/cs/hlds_muid6.so

0xf76d2a60  0xf76d3958  Yes         /lib32/libdl.so.2

0xf76bf170  0xf76ca3f8  Yes         /lib32/libpthread.so.0

0xf769a410  0xf76b3908  Yes         /lib32/libm.so.6

0xf755b230  0xf7662af8  Yes         /lib32/libc.so.6

0xf76ec7f0  0xf770427f  Yes         /lib/ld-linux.so.2

0xf6ff1630  0xf707e86a  Yes         /home/cs/ORIG3111e/hlds_l/engine_i686.so

0xf6f49460  0xf6f87f51  Yes         ./libhlwon.so

0xf6dd4060  0xf6e56749  Yes         ./libSteamValidateUserIDTickets_i386.so

0xf6cf12b0  0xf6d00b49  Yes         /home/cs/ORIG3111e/hlds_l/filesystem_stdio_i386.so

0xeeb9b2a0  0xeec21ab7  Yes         ./cstrike/addons/metamod/dlls/metamod_i386.so

0xee9d5e00  0xeea7a044  Yes         /home/cs/ORIG3111e/hlds_l/cstrike/dlls/cs_i386.so

Stack level 0, frame at 0xffdb6a50:

 eip = 0xf75aa6dc in feof; saved eip 0xeec18e1b

 called by frame at 0x0

 Arglist at 0xffdb6a48, args:

 Locals at 0xffdb6a48, Previous frame's sp is 0xffdb6a50

 Saved registers:

  ebp at 0xffdb6a48, esi at 0xffdb6a40, edi at 0xffdb6a44, eip at 0xffdb6a4c

End of crash report

----------------------------------------------

But after the auto-restart via hlds_run, its running now ...

EDIT:

I nearly forgot: This is the debug-version of 3B20r - and it RUNS at least!
But this is something, the latest development version also could ...
The non-debug version of 3B20q was the library, that crashed the servers directly and constantly.

But we can stay at 3B20r of course :)
Already much better for me ...

EDIT:

So far, it seems that i can run both the downloadable development version 3B20q and the debug version 3B20r, which is not an official download.

On our UWC3 Server, i *constantly* get the following messages - with both above versions:

SZ_GetSpace: overflow on SVC_RuleInfo
SZ_GetSpace: overflow on SVC_RuleInfo
SZ_GetSpace: overflow on SVC_RuleInfo
SZ_GetSpace: overflow on SVC_RuleInfo
SZ_GetSpace: overflow on SVC_RuleInfo
SZ_GetSpace: overflow on SVC_RuleInfo
SZ_GetSpace: overflow on SVC_RuleInfo
SZ_GetSpace: overflow on SVC_RuleInfo

About 1 per second ...

And also, the server crashes in the beginning one to several times, till it finally runs.

Here the CRASH of one of those starts:

Code:

----------------------------------------------

CRASH: Thu Apr 15 00:50:02 CEST 2010

Start Line: ./hlds_i686 -debug -nojoy -noipx -pidfile hlds.pid -debuglog debuglog.txt -port 27020 -tos -heapsize 131072 -zone 8192 -game cstrike +sv_lan 1 +sv_proxies 0 +maxplayers 32 +ip 178.63.3.149 +map de_dust2

Core was generated by `./hlds_i686 -debug -nojoy -noipx -pidfile hlds.pid -debuglog debuglog.txt -port'.

Program terminated with signal 11, Segmentation fault.

[New process 11386]

#0  0xf75e66dc in feof () from /lib32/libc.so.6

#0  0xf75e66dc in feof () from /lib32/libc.so.6

#1  0xeec54e1b in MPluginList::ini_startup ()

   from ./cstrike/addons/metamod/dlls/metamod_i386.so

No symbol table info available.

From        To          Syms Read   Shared Object Library

0xf7722320  0xf7722428  Yes         /home/cs/hlds_nowon.so

0xf7720400  0xf7720638  Yes         /home/cs/hlds_fixes.so

0xf771da20  0xf771e7d8  Yes         /home/cs/hlds_muid6.so

0xf770ea60  0xf770f958  Yes         /lib32/libdl.so.2

0xf76fb170  0xf77063f8  Yes         /lib32/libpthread.so.0

0xf76d6410  0xf76ef908  Yes         /lib32/libm.so.6

0xf7597230  0xf769eaf8  Yes         /lib32/libc.so.6

0xf77287f0  0xf774027f  Yes         /lib/ld-linux.so.2

0xf702d630  0xf70ba86a  Yes         /home/cs/NEW3111e/hlds_l/engine_i686.so

0xf6f85460  0xf6fc3f51  Yes         ./libhlwon.so

0xf6e10060  0xf6e92749  Yes         ./libSteamValidateUserIDTickets_i386.so

0xf6d2d2b0  0xf6d3cb49  Yes         /home/cs/NEW3111e/hlds_l/filesystem_stdio_i386.so

0xeebd72a0  0xeec5dab7  Yes         ./cstrike/addons/metamod/dlls/metamod_i386.so

0xeea11e00  0xeeab6044  Yes         /home/cs/NEW3111e/hlds_l/cstrike/dlls/cs_i386.so

Stack level 0, frame at 0xfff393a0:

 eip = 0xf75e66dc in feof; saved eip 0xeec54e1b

 called by frame at 0x0

 Arglist at 0xfff39398, args:

 Locals at 0xfff39398, Previous frame's sp is 0xfff393a0

 Saved registers:

  ebp at 0xfff39398, esi at 0xfff39390, edi at 0xfff39394, eip at 0xfff3939c

End of crash report

----------------------------------------------

What else can i do?

Re: PodBotmm on 32 Bit AMD?

The crash doesn't point at all to podbot_mm_i386.so file...

But - in Your log I have seen such file - hlds_nowon.so. I don't think this is any officially file provided in the latest supported by VALVE CS1.5 update...
I can still run CS1.5, but only on LAN network. If You are using any addons they are letting You to play with Your friends via internet old CS1.5 (WON servers are off so it is impossible to play so), I'm assuming You are using some illegal and unsupported stuff. If You had the original license for CS1.5 - with the same license You could install steam and CS1.6 (without CZERO) and You wouldn't have any problems like that. Since You decided to stick with no-won CS1.5, I'm assuming Your CS1.5 is illegal. Otherwise what is the reason to not update Your CS to supported by VALVE and other mod developers CS1.6?

Re: PodBotmm on 32 Bit AMD?

Oeh.

There are lots and lots of players using CS 1.5 either on LAN or in the internet - with completely legal software!

Yes, overloading a library function is required - but this is legal too!

About the why: Because we love HL1 and CS 1.5, of course! CS 1.6 is crap already and everything on Steam is just ..... <deleted by myself> :)

But to ensure you, that this the above has nothing to do with those libraries, i started testing without them, of course!

The behavior of PodBotmm (compared to other bots who do run) is not changed when using the above libraries or other MetaMod plugins like soundcheck, amxx or hlguard!

If you prever, i can of course run all tests without *any* other stuff - but as i said, it is the same behavior:

The official stable version completely crashes all servers i have - regardless of plugins or library overloads.
The development version or the debug version only show crashes at the "quit" command or sometimes at the start of CS.

But as you said, the last crashes i posted do not contain hints about podbot - to be sure, i am going to test "quit" on a fresh install without anything: Neither Metamod nor any other library and report back!

Re: PodBotmm on 32 Bit AMD?

I can use both the development and the latest debug version ...

But both have a serious problem:

I get the following message *every* second or so, when i watch the server with HLSW:

SZ_GetSpace: overflow on SVC_RuleInfo

The messages go away, when i stop HLSW to watch the server.

Also, i get message overflow messages when connecting to the server, may get droppped and need to issue "retry" to get connected at all!

All of the above stops, when i have no bots or JoeBot or EBODbot instead if PodBotmm running!

This is true even when NO AMXX ist loaded ... PodBotmm alone as MetaMod plugin is enough to get this behavior!

:-|

*sob*

Re: PodBotmm on 32 Bit AMD?

What is the value of pb_latencybot in Your podbot.cfg file?
SVC_RuleInfo message iis spamming Your console because of too many cvars loaded (old HL1 on WON/NOWON servers supports 2048 bytes only). To avoid that You need to hide some cvars like is described in this topic:
http://forums.alliedmods.net/showthr...p?t=492&page=2

http://forums.alliedmods.net/archive...php/t-482.html

Re: PodBotmm on 32 Bit AMD?

pb_latencybot is at 2 ...
Shall i change that?

I am confused that the bots can have such an effect - the other bots don´t have this effect.
Maybe changing latency to 0 to 1 could help?

I will check that CVAR stuff - thank you very much!

:)

Re: PodBotmm on 32 Bit AMD?

yeah - please put it to 0 i.e.
That fluctuating ping wasn't tested under CS1.5 and it may cause some side effects...

Re: PodBotmm on 32 Bit AMD?

With latency set to 0, i still get "SZ_GetSpace: overflow on SVC_RuleInfo" ...

But i did not hide any CVARs so far - will report back :)

Just tell me, if you are interested in testing something yourself on one of the servers :)

Re: PodBotmm on 32 Bit AMD?

Quote:

Originally Posted by TJA (Post 61485)

Just tell me, if you are interested in testing something yourself on one of the servers :)

No. I'm not interrested in no-won servers. If You are still getting the same error, try with standard installation and create a LAN server (without that no-won stuff installed). If You wiil not getting that message - that will mean podbot mm is incompatible with no-won binaries and I'm not going to provide any support for it. When I create my HLDS with CS1.5 (windows), I'm not getting that message (even with 30 bots placed there). I don't think under linux there can be a difference. The only difference for me is that no-won addon You have installed (which changes some VALVE binaries - which is illegal BTW).