Yes, there were parts of the old code for camping waypoints remaining. I had to get rid of them and convert them to take ANGLES instead of vectors. The new camp waypoints work well now, and they finally work the way they are meant to (which has NEVER been the case so far with any version of POD-Bot), meaning that the bots now use the camp directions correctly.
The problem is that the POD-Bot waypointers have to redo ALL their camp points to make them compatible now, because all their camp directions will be messed up.
It's not necessary when using another flavor of POD-bot, because in none of them the camp directions are working correctly. But if you expect them to work right with this one, you have to redo (well, "do finally" would be more correct) all the camp waypoints.
I've also rewritten the FL_NOHOSTAGE waypoint handling in a cleaner way. I don't know if it will help to fix anycrap because honestly I have spectated these bots for 2 HOURS full in cs_assault, and I have NEVER seen one use a no-hostage waypoint incorrectly, I mean on purpose. The only case that can happen is when the bot is believing no hostage is following it.
I notice the bots still throw grenades very stupidly. I don't know what to do about this. It must be the algorithm Count Floyd made them use which must be intrinsically wrong. Perhaps some day I'll try to bang my head on this one.
I'm uploading the new bot.
http://racc.bots-united.com/releases/podbot.zip
*edit* I've also added Austin's new menu for the waypoint editor (autopath maxdistance)...
btw, the bot is finished uploading - go for it.