Some strange stuff

[>BKA< T Wrecks]

Hello all; I've recently mailed the new dll (Apr 26th) to my LAN mates and tried it out myself, and the result so far is the following (in no particular order):

• The slow-motion bug still exists on some maps! Occurs on CS 1.5 Win 32 listenserver, CS 1.6 Win32 listenserver and CS1.6 Licux DediServer. What are we doing wrong? I mean, several ppl said it was fixed... ???
  
• A command to kick a specific bot / player is missing, one of my friends reported. Some friends of him got angry when he had to use "kick random bot", and a human got kicked instead. (1.6 W32 LS)
  
• The min_bot and max_bot values established in podbot.cfg get read correctly, but have no effect - bots only enter the game via bot menu command or "addbot" command in cfg. (1.6 linux DS)
  
• The message "NUM_FOR_EDICT: bad pointer" appears on 1.6 linux DS... ??? This one totally beats me. I know it from a VERY old dll, and it was fixed. Didn't see it myself, either. Odd...
  
• Bots spinning around in circles when throwing grenades (related to the other grenade bug?) Occurs very rarely, my friend Robert has observed it 2-3 times, he says. (1.6 Linux DS)
  
• The message "The bomb has been planted" does not appear in the console. (1.6 Linux DS)
  
• Although speech is turned on, bots don't use team messages like "follow me" any more, although they sometimes seem to follow one bot, but nothing can be heard (1.6 Linux DS)
  
• When waypointing, on some maps (saw it on cs_rio and cs_industrywest) in one small area of the map waypoints start blinking or disappearing from sight (although they are still there and work perfectly), plus they cause the same behaviour in other WPs and connections around. (1.5 W32 LS)

Our bad or some dll quirks?

[sPlOrYgOn]

yes slow mo still exists but i think it is because of the changelevel command.

the min/max bot seems to be not implemented (the joining part at least)...

NUM_FOR_EDICT i bet PMB can fix this

bots spinning while throwing nade... been here for quite a while..

the rest i'll take a look...

[Pierre-Marie Baty]

Thanks a lot for the feedback T Wrecks

Do you guys use the "changelevel" command, or anything (AMX, AdminMod) that may make use of it ?
For example, if AMX or AdminMod have a "next map" command, they may call changelevel and you wouldn't know it...

The changelevel command is very unsafe in CS 1.6. Lots of weird errors happen when it's used a few times. Try running with only metamod and the bot, to see if it gets better...

*edit*
presently I have no idea where to start looking for this NUM_FOR_EDICT error. If you could manage to find a procedure that would reproduce the bug 100% of time, I'd enjoy that a lot
The same for all the other bugs would not be superfluous at all either, provided it's doable

[sPlOrYgOn]

Originally Posted by Dethpod

sPlOrYgOn,

I'm running your version of podbot2.6-mm and it seems to run fine except when the map changes. So far maps de_luxor & de_prodigy are crashing the server.

Here's the error:
L 04/29/2004 - 11:02:24: FATAL ERROR (shutting down): NUM_FOR_EDICT: bad pointer
FATAL ERROR (shutting down): NUM_FOR_EDICT: bad pointer


I hope this helps in development,

-Dethpod

from thread http://forums.bots-united.com/showthread.php?t=1510
hope that helps..

[>BKA< T Wrecks]

Phew... and I thought we all were just too stupid. So there ARE a few bugs left.
First a little general info: Robert, who told me most of the bugs, is running MetaMod and PODBot only, no further plugins. Changelevel is no longer used, only map (thx for your concise explanation of this issue, btw!).

Robert will send me some logfiles as soon as he manages to do so; I hope they'll help. Btw, he got the NUM_FOR_EDICT on cs_office mainly.

@ The grenade spinning round thingy already known: Sorry for repeating, I didn't know that. My bad!

Oh, and one more comment: Besides these bugs, they all LOVE this bot! The bots' behaviour in general is awesome! Just had to say that... cuz it's true and the bot (and all ppl involved in coding it) damn well deserve it!

[eCartman]

I've seen slow-motion too when playing 1.6 DS window server. But it goes away after 10 minutes of play.

[sPlOrYgOn]

eCartman i believe your slow mo is different.
your slow mo must be when it is calculating Vistables.

[Pierre-Marie Baty]

I've been trying to get this num_for_edict crash, but to no avail... the bots can load de_prodigy without problems here it seems (CS 1.5).

I think I'll let this to sPlOrYgOn or somebody else...

[>BKA< T Wrecks]

I'm only using the latest dll on my 1.5 listenserver, too (my waypointing installation). Apart from the Slow-motion issue, I don't have problems. And I'm not even sure about this slomo thing, because when I change between maps that have an up-to-date pvi file saved and ready, I hardly experience this effect. However, yesterday I had the first CS crash in weeks when mapchanging between 3 maps I knew I had good pvi's for.
That's why I was very surprised when Robert told me all that stuff that went wrong - I didn't have a single "bad pointer" error! Not on the maps already manetioned, not anywhere else.
And last night, I joined the W32 1.6 listenserver of one of my friends, and we playtested a bit - we only changed between standard maps that had up-to-date pvi's (either regular mapcycle changes or "map" command), and the slow-motion effect was definitely there. However, the bot ran 100% rock solid and didn't crash once.

[Pierre-Marie Baty]

Ah, I smell the slowmo problem might be an incorrect reset of the internal parameters of the EstimateNextFrameDuration() function after a map change...

hmm, yes... the more I think about it the more I'm convinced of it.

msecnum, msecdel and msecval must not be reset after a map change.

They should

Better do that in ServerActivate()

tell this to sPlOrYgOn