Re: progress
 

Odd - I guess that's just an annoying reality of different distros doing different things.
Anyway, with https://github.com/tschumann/sandbot...a7287ffafbc558 you should be able to compile without changing things manually (hopefully). I think I'll have to stick with this approach as compiling for multiple different distros isn't something I have the time to do.

What was the problem with team picks that you mentioned? I didn't quite understand what was happening.

Re: progress
 

Well in the Metamod loaded version of Sandbot the bots are joined but won't select any team nor spawn. It's kinda like it was depending on HPB's bot_start.cpp for the bots to work for Op4CTF maps.

Re: progress
 

Hm, I'll check it out and see if I can replicate it.

Re: progress
 

I was unable to replicate but I've added some logging - see https://github.com/tschumann/sandbot...eeec1355a34684 (and the previous patch) - could you build and then tell me what you see in the console when it runs?
What version of MetaMod are you using too?

Re: progress
 

Well this may surprise you, because it appears that this Sandbot hates AMXX v1.8.2 for some odd reason. However it does seem to work well on Spawn Chat Protection and my modified version of AdminMod:-


There must be some AMXX plugin or maybe the AMXX Core Engine that is blocking those bots from selecting a team and spawning. However despite AMXX ruining Sandbot the teams are a bit uneven and it does block some bots from selecting a team.

http://steamcommunity.com/id/SCO_Rob...01045018411047

Re: progress
 

Ah right. I think someone else has had a similar issue before (some other plugin breaking something). Annoying.

Yeah I'm aware that the joining code isn't great - it's on my todo list to make teams even etc.

Anyway, looks like I should be good to tag a release soon. Thanks for the help.

Re: progress
 

It's a pleasure and happy xmas!

Re: progress
 

Merry Christmas to you too!
I've released v0.4.1.1 too: https://github.com/tschumann/sandbot...s/tag/v0.4.1.1

I now realise I've forgotten to add some more notes to the readme. I guess I'll fix that up after I merge it back to master.

Re: progress
 

Looks like that newer version still gives me that error from above on my Xubuntu 14.04 server for OP4CTF. But when I recompiled the src code on my Linux Mint with g++-4.8 it appears to work better. I don't know if your compiler from your Linux Distro or GCC have some missing dependencies.

Re: progress
 

Yeah I'm aware that's still a problem - some mismatch in libstdc++ versions which I haven't figured out yet (need to go back through this thread and try and nail it all down).
I wanted to get the main fix of waypoints not working in Linux out - in the meantime yeah it should compile if all else fails.

Re: progress
 

Sure feel free to check and test those recompiled Sandbot in my OP4CTF server that I've mentioned via PM and you can use "meta list" ingame to check Metamod addons.

Re: progress
 

Thanks - I'll do a bit of reading and see just how gcc and libstdc++ linking is actually meant to work.

I don't think you've sent me a PM.

Re: progress
 

The PM was sent by Steam Chat.

Re: progress
 

Alright I'll check it out.

Re: progress
 

I've added some Op4CTF waypoints for op4ctf_blackmesa and op4ctf_orange.

Re: progress
 

Thanks, I'll add them to the next release (if that's okay with you).

Re: progress
 

Wooow, I don't know if there is still a CTF server for Opposing Force (I've already googling it but never find any)

I hope we can play it together

BTW, how about sandbot progress? Have you solved a bot_count and bot_names changing problem? :)

Re: progress
 

I'm still working on the bot_count thing - I've tidied up the bot addition code and it seems to work better (still might have some quirks) but it should mean but subtraction should be doable. Still need to do some more tidy-up before I can get customisable bot names (probably not next release).

Re: progress
 

For tschumann, thanks for working on that part. I will waiting for it.

For RoboCop, I've already tested your waypoint and this is my review

op4ctf_blackmesa : This map is so huge. Even with waypoints, bot is still hard to capture the flag. Sometimes they are just roaming around without any purpose

op4ctf_orange : This map is not large, but there is only one way to get the flag. Therefore, instead of "finishing objective", they are only killing each other forever (to make it worse, because there are too many bots, we can't even touch flag at all, and it will just causing a chaotic gameplay unless you are smart enough to running and ignoring everything which bot seems impossible to doing that)

Well, I didn't mean to say if your work is terrible. Thanks for your try, and as reward for you, I will give you this.

This is a custom map called "op4ctf_omen". I've included a waypoints for that map. Believe it or not, this map is easier to take and capturing a flag. Of course the bots will get in your way to the flag, but it still possible even in large numbers. Trust me. In my first try, both I and my opponent got 1 point each.

Oh, one more thing. If you have hpb_bot installed, I made a waypoint for hpb_bot too. With metamod, you can combine sandbot and hpb_bot together so you can play with these bots on that map.

I will waiting for your review.

Re: progress
 

Well those new waypoints I've created aren't fully tested. Here is another new waypoint, only this time its for op4cp_park for Capture Point maps.

Re: progress
 

Are you happy for me to distribute them with Sandbot anyway?
The bots don't properly play op4cp_park do they? The waypoints are just to make them go to the capture points?

Re: progress
 

Well mainly your bots don't seem to get unstuck by wrench hitting their enemies at very close proximity, nor strafe very much to make them more evasive and professional.

I'm not sure on how well they capture properly, but in op4ctf_omen they don't capture as they supposed to be as they tend to randomly come back and forth.

Re: progress
 

What a robocop said is right. Sometimes after BOT takes a flag, and the enemy team gets in their way, they will turn back around chasing them instead of going back to their base to capture a flag

However, compared to previous map, at least this map is better right? Taking a flag is easier although capturing it needs some skill

I will test op4cp_park later. Very interesting map

Re: progress
 

Yeah I think the bots might prioritise enemies over capturing the flag - that probably needs some tweaking (even on op4ctf_hairball they just get stuck killing each other in the middle of the ice).

Re: progress
 

Well, I've already tested op4cp_park, and I don't know why BOT just roaming around in capture points after they have finished capture it. I was hope if they will looking for another capture points, but in fact they didn't.

As a result, stand by in the same place together, it will make my penguins go happy :clap:

Re: progress
 

Okay so in BotThink the bot will look for an enemy 40% of the time, but BotThink gets called every frame I think, so that's no good. I might change FindEnemy to only look for a new enemy if the bot doesn't have the flag (or if it has the flag and has very low health or something like that).

Yeah I still need to add support for capture point maps.

Re: progress
 

Well this may not be relevant but thanks to your sandbots it appears that [APG] clan gained a dozen fans for noticing on making our HL1 servers look less dead and have bots with more common sense.

http://steamcommunity.com/sharedfile...?id=1284620366

Re: progress
 

Glad to hear it - it would be a shame for GoldSource engine games to die out completely. I still enjoy Half-Life modding but obviously the community is much smaller these days.

Anyway, are you happy for your waypoints to be bundled from the next release?

Re: progress
 

It seems OK to me, even though the way I've recompiled SandBot Linux builds on my Linux seems to work better. I don't mind being your Linux Programmer even though I got a lot to learn C++.

Also I think we need a bit more communication to help our efforts and replies go quicker, may I suggest we use Bots-United Discord?

https://discord.gg/zCSvyyC

Re: progress
 

I intend to keep supporting the Linux version (better hopefully - I'm also working on Linux versions of https://github.com/tschumann/basis and http://www.moddb.com/mods/sandpit and I'm using what I learn from Sandbot).

I've mainly been posting here to keep it all more visible (and to keep traffic coming to Bots United - the more bot developers the better).

Re: progress
 

Also I think I found a useful solution that I've borrowed from Cheeseh's RCBot. What it is, is a Crashlog.txt script that is used for RCBot to take a record of causes of the crashes, like missing wad required for the server as well like "FATAL ERROR" crashes.

Also someone may need to update this site as vBulletin 3 won't be compatible for PHP 7.2 sadly.

Re: progress
 

Is it something in the Sys_Error callback?
The logging in Sandbot is a bit of a mess - I didn't like the write to file that HPB_Bot did (I wanted to be able to see stuff in the console). I'm moving towards a standard logging function which should allow a more unified writing to file option.

Re: progress
 

That could that yes. Also I've now got up to nearly over 200 op4 maps as well as added a few more CTF and CP maps which you find out here:-

http://apg-clan.org/fastdl/gearbox/maps/

Re: progress
 

Nice - good conservation.

Hopefully capture point support won't be too hard.

Re: progress
 

@RoboCop so I'm trying to get to the bottom of versions of libstdc++ etc - so you have gcc 4.8.5 and libstdc++.so.6.0.24 - do you have any other instances of libstdc++ installed? Do all of your libstdc++.so.6 symlinks point to the same file or are there multiple libstdc++.so.6.*.*?

Re: progress
 

Well as far I can tell the sandbot.so that I've recompiled appears to be linked with only '/usr/lib/i386-linux-gnu/libstdc++.so.6', if my amateur C++ skilled mind doesn't deceive me:-

It doesn't appear to be linked to 'usr/lib/i386-linux-gnu/libstdc++.so.6.0.24' from my server. I am not sure but I think that most Metamod Addons, as well as HL1 Bots seem to adapt better with GCC/G++ 4.8.

Also I got bad news my old Linux Laptop died a couple weeks ago. So I had to install the Debian Jessie i386 Netboot into VMWare and install DropBox where I've kept the src files.

Re: progress
 

If you are really sure that the issue is with libstdc++ you can link it statically to the bot .so. Just copy the libstdc++.a file(from your distro) to the same dir where the makefile is and add -L. to the linking command line.

Re: progress
 

Is /usr/lib/i386-linux-gnu/libstdc++.so.6 a symlink?
And yeah I think I was using gcc 4 before maybe but I don't want to get stuck on some old version.

Well I can't say I'm sure but it seems to be part of the issue. I might give static linking a go.

Re: progress
 

Also hopefully I'll be able to get Day of Defeat bots to use knives and pistols. I haven't been able to work out which weapon id corresponds to what weapon but I've realised I can block the spawning of some weapons so I should be able to figure it out now.

Re: progress
 

Well its a shame ShrikeBot isn't open source and it doesn't seem likely that SandBot will work with ShrikeBot waypoints.