Bug report: Recon mission does not work for AI

FreekS · Post by **FreekS** » Sun Sep 03, 2006 9:18 am

I found in a scen under construction that my AI controlled Recon Ground missions did not launch.

When viewing in the SE, from the SAME side as the unit on recon mission the mission launches normally, BUT when viewing it (toggle all) from another side the mission does not launch.

Tried several things to repair:
- build scen in 3.7.1 SE or in 3.6 SE -- no difference
- delete and rebuild mission (so mission is created after ready time is defined) -- no effect
- delete and rebuild aircraft on mission -- no effect.

I have a small test scen: (let me know who to send it to)
To reproduce:

1. Open test scen in 3.7.1. SE
2. Run game
You will see your own Spy aircraft observing nothing
3. Open test scen in 3.7.1. SE
4. Switch sides to side '1'
5. run game
You will see plane rise from land base on a recon mission

Conclusion is that I believe in 3.7 the Recon ground mission NEVER works for the AI. It can be used by the player to create and operate (but almost no player will ever do this).

I suggest High Priority.

Freek

hermanhum · Post by **hermanhum** » Sun Sep 03, 2006 10:26 am

Sample file for the PlayersDB v6.4.9 at StrategyZoneOnline

jkruny · Post by **jkruny** » Sun Sep 03, 2006 11:44 am

Thanks for the report Freek, and thanks for the file Hermann.

I'll give this one a whirl this weekend. I don't often use Recon Ground Missions in scenes that I'm personally working on, so I never have stumbled on this bug.

Freek, did you say this bug existed in the 3.6.3 SE version of the scene also?

hermanhum · Post by **hermanhum** » Sun Sep 03, 2006 12:04 pm

The mission works fine in H3.6.3

Only when the situation arises in ANW 3.7 is there a problem. The test file for H3 has been attached to verify that the mission does work in the older version.

Bucks · Post by **Bucks** » Mon Sep 04, 2006 11:00 am

FreekS,

Have you run the scen with the "Game" exe and not just "Scenedit" exe? Also to test properly add a third neutral side to watch AI controlled missions if using scenedit.

Btw I believe that if you have the capability you should attempt to test running a server/client setup on the same machine. This is especially useful for observing AI controlled behaviour from spectator mode.

There have always been issues with testing with scenedit, long ago since the days of H2. I don't have the latest PDB so I can't test the scen. (Pls send if you'd like me too) but I opened it and there are only 2 sides which immediately sets alarm bells ringing here. Like I said I always create a third side to use to observe from ( show all units ON).

Cheers

Darren

(PS delete third side b4 scen is published)

hermanhum · Post by **hermanhum** » Mon Sep 04, 2006 11:38 am

This factor was taken into consideration. The test scenarios were made so that the behaviour was observable from a side that was not in control of the missions planned for the AI.

jkruny · Post by **jkruny** » Mon Sep 04, 2006 12:24 pm

Hey Guys,

Yes, at some point in the past testing in the SE was not reliable.

But IIRC, just post-Jesse, the game codes for the GE and SE were merged or blended or some such. My memory of this is slightly foggy, but it was a pretty big deal for scene builders at the time. It meant you now longer had to build in the SE, and then test in the GE.

Since then, all GE code changes are automatically built into the SE also, so as far as I can tell, testing in the SE will result in the same behaviors as in the GE, with the added benefit that the tester can see what all sides and units are doing.

I do still test from the GE occasionally tho, this gives you the proper perspective from the players standpoint.

And I always play Mike Mykytyn's scenes several times from the GE before succumming to the "how did he do that?" syndrome.

jkruny · Post by **jkruny** » Mon Sep 04, 2006 1:24 pm

Hmmm,

Some other oddities here as well.

When the Shadow does go on station for the recon mission, fuel goes to 185 hrs and there is no usage, it will stay on station forever. It does show proper fuel levels and usage tansiting to and from the mission station however.

Still working the problem guys.

Thanks

hermanhum · Post by **hermanhum** » Mon Sep 04, 2006 3:57 pm

The fuel oddity is a database problem and not a game engine problem.

The fuel consumption setting for the plane is inadvertently set at zero. This is a carryover from the HUD-II and will be fixed in the next release.

jkruny · Post by **jkruny** » Mon Sep 04, 2006 4:12 pm

yeah, sorted that out.

Recon Ground missions are indeed broken.

Reported bug to AGSI.

Thanks guys.

jkruny · Post by **jkruny** » Mon Sep 04, 2006 4:16 pm

As a side note, it does appear however that Recon Ship Missions work properly.

FreekS · Post by **FreekS** » Mon Sep 04, 2006 4:48 pm

Thanks Jeff,

Darren, I discovered the problem in GE on a large new scen I'm building and then tracked it down exactly as you suggested by observing in SE from a neutral side so as to not influence AI behaviour. In fact if you run in SE from the same side as the mission, the mission will launch!

The test scen was built to demo the probelm is the smallest and simplest scen possible.

I hope this one gets priority; it may sound that other mission types (ASW of recce ship) sould be used instead of recce ground, but my experience is planes on those units are more easily distracted by far away contacts than on recce ground.

Freek

jkruny · Post by **jkruny** » Mon Sep 04, 2006 4:59 pm

The test scene I built has a neutral observation side.

But even though the mission will launch in your test scene if you have that side selected, there is still a serious problem, because it will not repeat if you order the Shadow to land, refuel/rearm. There is just something wrong with Recon Ground Missions.

Thanks Freek

Bucks · Post by **Bucks** » Tue Sep 05, 2006 12:56 am

The fuel oddity is a database problem and not a game engine problem.

The fuel consumption setting for the plane is inadvertently set at zero. This is a carryover from the HUD-II and will be fixed in the next release.

You are kidding right?

You're saying it didn't have fuel from the HUD-II???

Ok check this:-

ID#258 ES-3A Shadow v1
Type: RCN (was AEW)
Flag(+): Terrain Following
Fuel(-): #0226 Aviation Fuel x 5753 v0 x 1
Fuel(+): #0305 Aviation Fuel x 5888 v1 x 1
Comm(-): #0108 HF Radio (Air) v0 x 1
Comm(-): #0110 UHF/VHF Radio (Air) v0 x 1
Comm(-): #0111 NATO Sonobuoy Reciever.99 v0 x 1
Comm(+): #0161 VHF Radio (Air) v0 x 1
Comm(+): #0163 AN/ARC-184 Have Quick v2 x 1
Comm(+): #0225 UHF (air, secure) v1 x 1
Sensor(-): #0075 ESM Level 3 v0 0-360 x 1
Sensor(-): #0617 AN/ASQ-81(V)2 MAD v0 0-360 x 1
Sensor(-): #0618 AN/APS-137(V)1 ISAR v0 Fore 90 x 1
Sensor(-): #0722 OR-89 v0 0-360 x 1
Sensor(+): #0618 AN/APS-137(V)1 ISAR v1 Fore 180 x 1
Sensor(+): #1319 OR-263 IIR v0 Fore 270 x 1
Loadout(-): #0120 Ferry #02 (2 x 600 Litre) v1
Loadout(+): #0083 Tanker #003 (S/ES-3A/B Viking) v0
Loadout(+): #0084 Ferry #01(2 300 USG) DT v1
Mount(-): #0222 Viking Sonobuoy Dispenser v0 Rear 180 x 1
Mount(+): #0600 AN/ALE-44 Decoy Dispenser v0 PMF+PB+PS+PMA x 1
Mount(+): #0600 AN/ALE-44 Decoy Dispenser v0 SB+SMF+SMA+SS x 1

The above is a comparison between the ES-3A available in the HUD-II at Harpgamer Harpoon 2 and 3 Databases Page and the version in the proposed HUD3. As anyone can see there was fuel in it. I dare you to download and check the version in the available HUD-II.

I have to ask why you'd blame me for a bug that I didn't create? Why you want to publically say yes there's an issue with the PDB, but it's not my fault it was there for 4 years and we just never got around to fixing it? When really the HUD-II ES-3A-Shadow has 5753 aviation fuel.

If you check above you'll see I've EDITED the fuel load as follows:

Fuel(-): #0226 Aviation Fuel x 5753 v0 x 1 -removed
Fuel(+): #0305 Aviation Fuel x 5888 v1 x 1 - added will appear in next release. ( edit made to correct range issue)

Herman your statement is incorrect, check your PDB and then check the web available HUD-II. You will actually see that the error was made in the PDB, possibly when the platform was COPIED from the HUD without the relevant fuel record being available.

I have been back to harpoon for about 6 weeks, and now I have to say that after a 4 year hiatus I can't have missed much while I was away. Would people stop blaming others for things that well they simply aren't responsible for. It's situations like this that cause flame wars etc THINK B4 U POST

I am in my 40th year not my 14th!!! I've been through high school and I'm not coming here for a return visit if I can help it. Would we all please mature a little and I'm sure we'll start making progress with the future growth of the game. Rather than SHC - Spontaneous Harpoon (not Human) Combustion.

Thanks in Advance

Darren Buckley

hermanhum · Post by **hermanhum** » Tue Sep 05, 2006 1:57 am

No one is attempting to pass blame onto anyone else. On the contrary, the users of the PlayersDB are indebted to the editors of HUD-II, H3DB, and HCDB for being allowed to use their work in the futherance of the PlayerDB.

The data was mis-identified as coming from the HUD-II. Instead, it was derived from the H3DB entry as can be seen from the attached image. This was done in the hope of both databases being fixed.

At issue is not the fuel allocated to the aircraft, but the consumption rate assigned to the engines for that aircraft. As anyone can see by the image, the Boost Consumption Rate is Zero. Therefore, when the plane went to maximum speed, the consumption rate actually dropped to zero giving the plane unlimited endurance.

Apologies to all who have taken exception to the reporting of this database error.

Bucks · Post by **Bucks** » Tue Sep 05, 2006 5:02 am

Fair enough mate, in future help me out with good identification of these errors and I'll fix them accordingly. You had me looking for a 0 fuel record. You were right it's boost consumption rates and I have adjusted them accordingly, thanks Herman.

I looked at PlayersDB v6.4.9 and you were right to say there was no fuel there. So looks like a combination of a couple of DB errors/omissions.

Thanks

Darren

Bucks · Post by **Bucks** » Tue Sep 05, 2006 5:36 am

ORIGINAL: FreekS

Thanks Jeff,

Darren, I discovered the problem in GE on a large new scen I'm building and then tracked it down exactly as you suggested by observing in SE from a neutral side so as to not influence AI behaviour. In fact if you run in SE from the same side as the mission, the mission will launch!

The test scen was built to demo the probelm is the smallest and simplest scen possible.

I hope this one gets priority; it may sound that other mission types (ASW of recce ship) sould be used instead of recce ground, but my experience is planes on those units are more easily distracted by far away contacts than on recce ground.

Freek

Freek,

I ran some tests of my own using a scen I have for sonar testing (Just added a set of S-3 series aircraft to the airbase in the scen. ( Using HUD3 development database - Not Publically available )

I looked at a number of ways of viewing the scen and you're right to say it will work when you switch to the ground recon mission's controlling side. I only noticed it launch once, automatically other times it was really lazy.

I did discover a way to make it work. Transfer to the mission controlling side and if the mission doesn't start or even continue after the first aircraft launches and completes the mission. Open the "edit mission" window. Highlight nothing then click "cancel" to get out of the mission editor immediately. You'll wait the appropriate amount of time for launch (2 mins) and suddenly the mission will commence.

I believe doing this refreshes the mission controller. Here's a screenshot and if you look at the "message window" you'll see the list of "ready" aircraft. This must have "refreshed" the "mission controller" and the aircraft you can see to the north east of the base is one of the ES-3As assigned to the mission which has launched after I followed the procedure with the mission editor above.

Scen Consisted of 2 of each of the S-3 (S-3A/B & ES-3A) series aircraft added to my sonar test scen (also allowed me to test new fuel consumption rates - Thanks again Herman). I tried it with and without weapons loaded and that made no difference in the AI undertaking the mission. They were assigned to a "Ground Recon" mission and I tried passive/intermittent/active radar settings again no difference. the area they had to search contained both enemy and neutral ground units and I tried them (the ground units) with the 3 radar settings as well. The Ground units were also tried with and without "auto-detect" enabled "home" key function in scenedit. (I was hoping that by using no auto-detect that the AI would have something to look for).

There is a bug appearing in "Mission Control" for the AI. See also this thread:

Bug: Old or destroyed targets not cleared from q

Confirmation of a similar test scenario setup using other databases would be appreciated. Freek, Jeff you might want to try to recreate this behavior?

Thanks in advance

Darren Buckley

FreekS · Post by **FreekS** » Sat Sep 09, 2006 7:17 pm

Darren,
Sounds like you confirm the Recon Ground mission does not work when controlled by AI. Most players do not use missions during gameplay but the AI needs them. As Jeff indicated Recon Ship works and I now use recon ship for finding ground targets. Not sure if the Recon ship mission would move outside of the refpoints when it investigates ships - in my scen that is not a problem.
Freek