Autonomous CAP and GCI AI fighter script

[Quax456]

Bump

 

Gesendet von meinem K00E mit Tapatalk

[Stonehouse]

Sorry, news is still that it's getting there. Testing continues around some issues trying to sort them out before I pass it on to others. Hoping to avoid other people getting headaches and stress from using it.

[Justificus]

Sorry, news is still that it's getting there. Testing continues around some issues trying to sort them out before I pass it on to others. Hoping to avoid other people getting headaches and stress from using it.

 

Awfully considerate of you mate! Looking forward to its' release.

[Stonehouse]

lol don't know about considerate, more selfish really as if I miss things by being deliberately slack then it only comes around again :music_whistling:

 

More helping you guys helps myself.:D

[Rivvern]

Yea Stonehouse is working really hard with this awesome and gamechanging script. Hang in there guys. A functioning version will come!

[Pikey]

Dunno if this is useful, but i am working with your last script OK for spawning, the only issues seems more DCS related, the spawned AI are as dumb as you can get. I seem to remember this being a problem before also. Please when testing consider the AI's intelligence and behaviour.

[ESAc_matador]

I am planning a mission.

 

There are few triggers, but the important ones.

 

MIST + CAPGCI does not work. I think I put everything right!! but no CAP... nor CGI.

The last Bridge352CAP.miz

[Stonehouse]

Hi matador,

That's a 1.2.16 mission but the mist version you are using is the one Grimes built for 1.5beta - so I am not sure if it is compatible. Maybe ask over in the mist thread if what you are doing is going to work or perhaps first go to Grimes Github site and grab Mist 3.7.51 or whatever was the latest prior to 1.5 beta being released. Give it a try and see if that sorts it out.

 

Cheers,

Stonehouse

[Stonehouse]

Thanks Pikey, yeah the AI has been raised as an issue in the new version. Trying to see what can be done but effectively all we can do is set ROEs, evasion response and waypoints and the rest is up to DCS.

[ESAc_matador]

Cheers!

[ESAc_matador]

still does not work... the file must be corrupted. I have used this mission thousand times...

 

Good excuse to start a new one.

 

by the way... the 4.0 Mist works fine as I can see. I used the template and it works.

 

REgards anyway!

Edited October 31, 2015 by ESAc_matador

[Stonehouse]

Another bit of an update. Hit some setbacks with a new method of checking the status of CAPs and it's impact on how often they spawn. There is also a bug that is proving difficult to find where multiple groups are being assigned to a intruder group where it should only be one. This also affects CAP spawning - as in too many are spawned.

 

So some more rewriting is taking place and hopefully will be back to testing again towards the end of the week. It inches closer to being ok to release but it is a bit of a slow grind and I'm also due to co-ordinate a major software release at work around the weekend of the 21st so I predict my spare time will not exist pretty soon except for sleeping and trying to de-stress after trying to get people to sign off various things and chasing people up so we pass each phase exit criteria in good shape. Long story short is I'm still working on it but it's taking longer than I hoped, some of that is outside my control at present.

[Justificus]

Thanks for the update! No worries, real life is the priority. We all appreciate you working on getting this out to us. Take your time, just not too much. ;)

[nolove]

Thanks for the updates! been keeping my eye on this for a bit now!

 

thanks stonehouse!

[Stonehouse]

Been ill with the flu for the last week so that plus dealing with work has meant I've really not progressed I'm afraid. Hoping to get back onto it later this coming week.

[Pikey]

Hey Stonehouse, this is an idea for an enhancement.

There are a couple of methods of persisting through server restarts now and I'm building a proper campaign. One thing I noticed, you warehouse your CAP "pool" in the script itself and thats great for one session but can't handle multiple sessions.

 

Is there anyway to output the total aircraft remaining as they fluctuate to a lua file for inclusion at runtime? Preferably by each airport, but a net total would also be handy. I had to manually count the destroyed aircraft during the mission and subtract it the next time the mission was launched. Yes this would require changes to the missionscripting.lua sanitisation but iuts a net leap forward in capability.

[Stonehouse]

Yeah it's a good idea and can go on the list but I don't think I will attempt it this time around. Now work is starting to calm down a bit I've started looking at the script again in the last couple of days and fixed up a few things but the over targeting of intruders bug is still doing my head in and that is the show stopper for this version. I am starting to think I might have to rewrite the allocation of interceptors to targets section of the script but need to be careful. Already my changes to the way the table recording active and available CAP groups is handled has caused knock on issues that then require investigation and more fixes. I don't want to get myself to the point of major rewrites on a 5000 line script or you won't see anything for months.

[Pikey]

theres a couple of things in the SE that need to be reverted first like the group ID issue with the introduction to multi cockpits. probably worth watching for those fixes before spending any time else yeah, its a waste.

[lukrop]

I forked your repo and almost completly rewrote this gem. You might want to check it out. If you want me to, I could even send you a pull request but I still got some plans with it so maybe I'll wait with the pull request.

 

https://github.com/lukrop/GCICAP

 

Here are the biggest differences at a glance: https://github.com/lukrop/GCICAP#fork-changes

[Stonehouse]

Hi lukrop,

Probably would have been nice to have been contacted first but from my viewpoint I believe you've acted within the scope of the copyright notice so I wish you luck and success with your version. I do suggest that you contact Snafu before going much further as he is the original owner of the intellectual content and may have a different view of matters entirely. I cannot speak for him or give you permission to proceed.

 

Just as an fyi the version you forked was a failed release and should really have been removed from github and I would suggest looking at the prior one.

 

Only other thought at the moment is that your script internally is vastly different to the one I've worked on and therefore I can't really help people who use your version. It would make sense for you to begin a new thread for yours to avoid confusion when people need help and support.

 

In the end I guess if your version supersedes this one I'll retire and get a lot more flying in lol and use it myself.

[lukrop]

Probably would have been nice to have been contacted first but from my viewpoint I believe you've acted within the scope of the copyright notice

Since the code, in the github repo, is licensed under the MIT license everything should be ok. ;)

 

Just as an fyi the version you forked was a failed release and should really have been removed from github and I would suggest looking at the prior one.

Whoops, didn't even notice that.

 

Only other thought at the moment is that your script internally is vastly different to the one I've worked on and therefore I can't really help people who use your version. It would make sense for you to begin a new thread for yours to avoid confusion when people need help and support.

That was the part I didn't want to do, without checking in on this thread first, since my stuff could be merged back into the original repo. Thats how open source should work. So we don't have masses of forks, outdated main projects and users confused which version to use. But this is entirely up to the original author(s). I would send a pull request if anyone wants me too.

 

In the end I guess if your version supersedes this one I'll retire and get a lot more flying in lol and use it myself.

Woha, nobody retires. :D That's the good thing about version control systems, we can collaborate. I feel you though, didn't get much flying this weekend either.

Edited December 1, 2015 by lukrop

[tovivan]

Any idea when will we get a working, 1.5-compatible version? The latest (14 Sept) doesn't work with 1.5.

[Stonehouse]

Like I said previously there is a problem with interceptions and targets getting multiple interceptors allocated that slowly results in CAPs spamming the mission. At inspection the code Snafu has written looks ok and should work which means that the only way to find the issue is slow and tedious debugging essentially using the old print line method to tell you the trace of lines executed and display variable values. Unfortunately as does everyone I have higher priorities like work and family so working on the script falls down to when there is time and - being honest - energy. I design and build software for a living and sometimes after a day at work the last thing I want to do is do more coding especially when there are no real tools to assist you.

 

If someone very clever with lua and the windows environment (FSF Ian are you there?) wanted to help us script builders my wish would be for them to build a real time interactive debug tool where I can see the code of a chosen script from those in a mission in a window and step execution through each line one at a time or choose break points to stop execution as well as displaying the list of variables and tables and their values and allow me to choose which data items I want to see. I have the equivalent to that at work for a different language and environment and it really helps productivity and bug squashing.

 

So short answer to your question it is in progress and it will come out as soon as I can get it done. It would be great to have more people work on this script as it has been basically just me for near 2 years for quite a complicated piece of software. Especially as I had to learn lua along the way. In my opinion though it really needs not to have people rewrite completely different versions unless they want to take over working on and supporting the script completely (which is fine by me). Rather I would have them enhance what is there using the script as it is now in a co-ordinated fashion. The problem (as I see it) with things like Lukrop has done is since his version is so radically different, retrofitting anything he does back into the "real" version will take more work to analyse and rework his code so it fits the existing script. You cannot just cut and paste from his version to the main one. ie it actually increases the effort required.

 

If it is any comfort my own squadron mates are hounding me for the next version too. Of course they want all of our group to fly too so there is another priority problem. I haven't been able to fly for a long time as it is a choice to work on this script or fly a mission as there is only so much free time. My flights tend to be 5 mins here and there to work on scripts or check if a mod works under DCS 1.5/2.0.

 

Perhaps I should just do the bare minimum to get the January 2015 version working under 1.5 even though it has other issues instead of the version I am working on now? What do the users say? It will still take time to do that of course if that is what the majority wants.

Edited December 3, 2015 by Stonehouse

[tovivan]

Thanks for the reply!

[SNAFU]

Hey Stonehouse, I am amazed that you still have the energy to keep my piece of junk-script living. :)

 

For me it was an experiment, how far I could get with my idea and at some point,I just had to stop, where the effect to effort ratio approached the zero point. At the end of the day I realized, you can get fairly randomized fighter intercepts with triggers and ingame solutions in far less time. And since I only play closed cooperative campaings, which are highly controlled by ATOs, SOPs, ROEs, SPINS and other strange words, I myself found little need for the script at all.

 

For someone professional or with a background in scripting, I guess it would be best to start from the scratch. But then again, thinking about the hundreds of hours you spent (and I already spent about maybe 200hrs work and more testing into this) I feel a little guilty since I started this and so I honestly have to say to you... The time spent on this is not worth the result, better go out, enjoy the sun, have a quality time with your friends and loved ones, is time far better spent. You never know, what happens tomorrow. :cry::smilewink: