Spring RTS Engine

BA running with XE10.1B using IKs config

crash on startup bug in latest build: unreproducible



Now I know you would rather spent 20 minutes arguing against the bug tracker, rather than the 2 minutes it takes to signup, but if you are going to insist on forum posting then you really need to provide:

• Map
• Mod
• OS
• Version of NTai

So far all I have is:

It crashes on startup and there are no logs.

• When you launch spring?
• When it starts the countdown?
• When the countdown finishes?
• When the commander has been spawned?
• Several seconds in?
• In the lobby?
• During loading sequence?

I know your trying to help but so far you've been lucky in that I had an idea of what it is, but when I've utterly squashed that bug in release builds what will you do then? I don't even know what version of NTai your running! Or even what version of spring, I have to infer it, make assumptions, which is very very very bad.

All this leaves AF confused, and a confused AF cannot fix bugs.

So please, I would rather you reeled off the version numbers of your components and exactly what you saw from a 3 year olds perspective, rather than speculation on what you think might be the issue. Be crystal clear, and leave nothing to chance, no ambiguousness, no inference, no assumptions.

Is there a way to setup NTai so it doesn't write logs, btw? Or maybe always overwrite the same or something, basically I'm worried that if used a lot by unwary people, NTai would make the Spring folder grow to humongous size just from its log.

Sorry for being a pain and thankyou umrug for signing up.

Given the nature of XE11, Im tempted to pencil in a 10.5 with some of the things that didnt make it into 10 or 10.1 with some of the refactors of XE11 ( incomplete though), mainly Im thinking attacking. I wouldnt count on that though as I may plough through to XE11, which is mainly a completion fo ym re-architecting effort before the goodies of XE12

XE10.1B?
I only found a on your website
problem solved
got link to B \?|

look again, a was replaced with b not long after

XE10.1b is almost completely comatose for me on XTA 9.59 (their commanders only react to being shot at) with default settings, console kept on getting spammed with errors like: arm_u0commander is missing from the efficiency array. Played on Kolmogorov with 3 NTAIs, 2 of which being Core, none of which being allied with anyone. Running on Windows Vista SP1, Springlobby 0.35, Spring 0.80.5.1

EDIT: Tried it on BA, same behaviour, only error I got was something about it not being able to save BA.tdf in the learn folder.

After two long years, NTai is back again into Kernel Panic installer.

From now on, and for as long as it works for me.

^_^ My supervisor will be pleased!

ding dong

UI Coursework: 10 days
Main Coursework 1: 10 days
Final year project presentation: December 8th

If you want to check what has been done, you dont need to prod me, you can see in meticulous line for line detail on github and on redmine.

Hi,

here the screenshot with a lone depthcharger on a 100% Land-map.
(Astro Turf V1)
And he have build later a second ...

Regards
R-TEAM

hmm, well theres code to check for water weapons, but then it would appear it doesnt catch all water weapons =/

Any ideas how I would do the check properly then?

Hi,

have to Add that is the Mod Rev-V60.
Maybe not full supportet by your AI ?
But the other AIs doing no waterweapon build on silly positions.
So it musst a general usable way in checking here ...

Regards
R-TEAM

Okay it seems NTai is broken atm, and nobody told me. Ive been busy with courseworks! =(

I'm going to bring forwards certain plans.

I'm not going to re-implement attacking in C++, I'm bringing forward the plans with Lua, and I'm re-architecting slightly the behaviour management system.

To reflect the changes I'm going to start a branch of the AI using the NTai LE1 version-ing system to indicate this is not quite the same AI.

Also, Im going to be implementing behaviours and tasks in lua. Behaviours will also start to resemble components of a state/rule machine built using blocks branches and chains. These behaviours then operate by adding tasks to a queue that the unit follows.

Of course the whole tree thing will be replaceable should people disagree with it.

gg ntai

IK, obvious leap here is not that there will be no re-implemented attacking you want so badly.

The obvious conclusion is it will be implemented in lua

Please tell me you'll still be maintaining current NTai! I don't mean continuing to improve it, but merely to fix NTai whenever a new Spring version breaks it.

its intended that there should be a lua configuration layer which resembles closely the current config format, and remember currently my only tried and tested config is the kernel panic config anyway

NTAI has spent the last 3 years constantly being re-developed, with stable useable versions approximately 1% of time

You'll have to forgive my tired despair that rather than fix the version you have your doing it again for the 1 millionth time :p

The whole point of me doing any of the work I've been doing was in order to make the changes for lua