Page 54 of 103

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 02:08
by alfred007
Berserk Cyborg wrote:Could you upload a save on Gamma 6? Yeah, two hours is a lot for this mission. What would you think if I made it 1 hour and 20/30 minutes for Normal (thus about an hour on insane)?
Though I would be able to finish the level within 15 minutes I spent a lot of time getting prepared for Gamma 08. At the moment I would say that 2 hours depending on difficulty is a good solution. I needed the whole time of the level to get prepared for Gamma 08 because of the power limit for insane difficulty. We can later test it out if we can change the timer to 1:30 for normal difficulty because power management is a bit tricky in Gamma mission.
Bethrezen wrote:How did you manage that one ? because the best i have been able to do is get my units to veteran yeah i can get commanders and scanners up that high but not regular units, although that might be because I typically play with more units then you because by alpha 12 I have 3 squads of 18 combat units, 3 commanders, 2 mobile repair units and like 4 trucks
Yes, one point is that I just use 30 units during the Alpha campaign. And to get my units to this experience I let every enemy factory as long alive as possible. For example, at Alpha 01 I use a suicide squad to recover the engineering artifact. Then I build 4 repair tanks and place my tanks south of the first scavenger factory. The first and second scavenger factory produces units and I destroy them until there are 2 minutes left on the timer. In this way, I destroy more than 600 scavengers only in Alpha 01. Spread over the 20 tanks I have at the end of Alpha 01 means that at the end of Alpha 01 all my units are minimum regular, some professionals. And in Alpha 02 I place a mg tank in front of every scavenger factory backed with two repair tanks. Every time one tank became a professional I change the tank to make another one professional rank. In this way, I destroy over 300 scavengers in Alpha 02 with 20 units makes all minimum professional and 2 or 3 veteran. From Alpha 04 on I try to destroy as many units as I can by mortars attached to a sensor until I get commanders at Alpha 06. And for one more example in Alpha 09 I don't destroy the scavenger factories but place a commander with 3-4 lancer tanks in front of the factories with some repair facilities nearby and kill them one after the other. That brings more than 400 killed enemies. And doing that consequent every level during the Alpha campaign and recycling every commander as soon as he has an experience of 256 I have so many specialist units at the end of the Alpha campaign.

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 03:48
by Bethrezen
Berserk Cyborg wrote:
Bethrezen wrote: Not sure if it would be possible to have the free units of beta 1 and 2 have different ranks depending on difficulty though Berserk Cyborg would need to answer if that's possible or not.
Yes. Though at that point they would be mostly used for recycling or cannon fodder anyway since a good player would have professional units at this time.
what I tend to do with the free units is recycle them and then turn them in to AAA so that at least some of my AAA start with experience and then later when i gain access to vtols I convert them to vtols and then I usually just send a huge armada of vtols and just bomb them back to the stone age

you have to watch out for the AAA but when using heavy bodies and set to retreat ant medium damage its usually ok ya need to watch doing that on Gamma though coz them SAM sites are kind of nasty but on beta they don't normally have nearly enough AAA to stop 36 heavy body bombers, so all ya need to do is set them to patrol over an enemy base and then them do there thing

wahhhaaaa how do you like getting bombed into submission :twisted:

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 03:58
by Berserk Cyborg
Pushed b86f23a7459de639ee7e3b61de13e0a0ce1c21ff.

VTOL group attacks shortly after mission start, the LasSat strike every ten seconds and is limited to the upper half of the map (takes about an hour), and mission timer changes with difficulty. Also, that is a lot sensor towers you built. O_o

The Gamma 8 mission time completion will be faster I think because of the six extra research upgrades since the research squashing. Think it takes about 50 minutes now. And no idea what the last Gamma mission will be like. I was lenient on the factory timers, but there are more factories than ever so it could be like pre-nerf Alpha 12 on steroids.

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 04:11
by alfred007
Berserk Cyborg wrote:Also, that is a lot sensor towers you built. O_o
They are stronger than walls and hold the Nexus reinforcements back so that they can get destroyed by my artillery.

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 04:17
by alfred007
I looked into the code of Gamma 07 and my question is what happens when the level lasts longer than one hour. Does the LasSat go further south? When you play with normal difficulty you have 2 hours. And when he needs one hour for the upper half that means for me that he reaches the southern edge in two hours. But he has to stay in the upper half during Gamma 07 no matter how long the level last.

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 04:37
by alfred007
And now I looked into the code of Gamma 09 to get an idea of what to expect. For the NX-NEFactory you set a throttle of 10 seconds. Is that really intended? Only scavenger factories had such a short throttle until now.

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 05:19
by Berserk Cyborg
alfred007 wrote:I looked into the code of Gamma 07 and my question is what happens when the level lasts longer than one hour. Does the LasSat go further south?
No. It will randomly fire at a tile within the upper half of the map unless something tries to go back up there. Then it will lock onto whatever it is and blast it to bits.
alfred007 wrote: And now I looked into the code of Gamma 09 to get an idea of what to expect. For the NX-NEFactory you set a throttle of 10 seconds. Is that really intended?
Uhh... that might be tough. I think 30-40 seconds would do ok in that case.

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 16:40
by Bethrezen
ok so updated to the latest master
warzone2100-master-20180120-023743-b86f23a.exe (20-Jan-2018)

Alpha 06

Ok so i have been going back over alpha because I wanted to see how the levels where affected by the change that makes units come after you if you attack them i also wanted to see how well this actually works and this is easier to do on alpha since the units are weaker and you don't have to worry about getting bombed, anyway I think this still needs a little work, because it doesn't seem to work consistently, I noted a few instances on alpha 06 where I would bomb units from beyond there site range with my mortars and as before they would just sit there doing nothing.

For example when you bomb this group of scavs or mortars from beyond site range with your mortars then the rest will come after you as expected.

Image

however this doesn't seem to apply properly when you get further into the level as you can see in this screen shot.

Image

I'm bombing the enemy tank and you can see its smoking yet as as before its just sitting there like a lemming getting shelled to death, shouldn't that tank react to getting hit and come and attack you?

While making the rest of a group of units come and get you if you attack one of its members is an improvement I think this should also work on individual units as well so if as in this case there is only a lone unit that unit will react to getting attacked and move in the general direction the fire is coming from and come and attack you.

here is a save for alpha 06 I made so that you can see the issue your self, all ya need do is move the scanners foreword a bit and let your mortars start bombing the tank and you will see that it does nothing.

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 18:30
by Berserk Cyborg
It is a different problem and has been with the library forever. You will notice that as a mission progresses that group reactions slow down more and more to the point that it takes seconds for them to do anything.

Edit:

Code: Select all

function __camTacticsTick()
{
	var dt = CAM_TICKS_PER_FRAME;
	for (var group in __camGroupInfo)
	{
		queue("__camTacticsTickForGroup", dt, group);
		dt += CAM_TICKS_PER_FRAME;
	}
	queue("__camTacticsTick", dt);
}
As more and more groups are made it will take longer to execute commands. Thing is, groups are not removed even if they are destroyed so the newer groups will gradually take longer to do stuff. I think removing dead groups would be ok since camStopManagingGroup(group) already exists and can be put to use finally.

The below change seemed to help. You can either start the mission over, or load your current save, create a new save from it, and then load the new one. Might take a couple seconds since there are 80+ useless enemy groups on your save.
updated-campaign.wz

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 22:35
by Bethrezen
It is a different problem and has been with the library forever. You will notice that as a mission progresses that group reactions slow down more and more to the point that it takes seconds for them to do anything.

As more and more groups are made it will take longer to execute commands. Thing is, groups are not removed even if they are destroyed so the newer groups will gradually take longer to do stuff.
I see well that doesn't seem like a very smart move, who ever coded the obviously didn't take into account the performance issues that would result from the game having to iterate through more and more dead groups because they are not removed when destroyed.

One would assume that once a unit, structure, object or group is destroyed then all reference to said unit, structure, object or group would be removed to avoid issues.
The below change seemed to help. You can either start the mission over, or load your current save, create a new save from it, and then load the new one. Might take a couple seconds since there are 80+ useless enemy groups on your save.
does that work for all levels or just alpha 06 ? because if so what i might do is just start from alpha 1 again to make sure it does cause any issues, because alpha isn't to difficult so should take to long

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 22:50
by alfred007
Berserk Cyborg changed libcampaign.js so the changes will work for all levels.

Re: Help needed testing 3.2.x Campaign games!

Posted: 20 Jan 2018, 23:22
by Berserk Cyborg
Bethrezen wrote: One would assume that once a unit, structure, object or group is destroyed then all reference to said unit, structure, object or group would be removed to avoid issues.
The actual game objects are like droids, structures, and features. The campaign library stores some basic information for the droid groups in an object called __camGroupInfo. Without removing the groups that have no droids in them the newer groups would have a delayed reaction time (in retrospect I bet that caused slow repair retreating).

Typically there are not many groups active at a time in most missions, so the timing impact would be low even on a fast unit production map like Alpha 12. I guess in a way some missions got easier as time went on.
Bethrezen wrote: does that work for all levels or just alpha 06 ? because if so what i might do is just start from alpha 1 again to make sure it does cause any issues, because alpha isn't to difficult so should take to long
All missions.

Edit:
I think I fixed the CB droids being useless in 8cad813a7ff3421370d927942d2feba4ad7e0933.

Re: Help needed testing 3.2.x Campaign games!

Posted: 21 Jan 2018, 02:52
by alfred007
I updated to master b86f23a and replayed Gamma 07. Now everything looks good.

Then I played Gamma 08 and again I have to say that the queue of 5 seconds for the LasSat is too short. Like in Gamma 07 it should be set to 10 seconds. The fact that the LasSat prefer droids over structures makes it impossible for the player to repair oil derricks the LasSat has destroyed or to rebuild some sensor towers for the artillery. That reminds me of the song of Sidney Youngblood: All we can do is to sit and wait. That's really annoyingly and also boring so I would prefer if you give the LasSat a coin flip if he targets a structure or a droid. Then you have at least a chance to do something instead of simply waiting until the end of the level.
I found one weird issue. I have two saved games, one shortly after the 2nd level missile code is cracked and one a few seconds before the 3rd level missile code is cracked. When I reload the saved game with a few seconds before the 3rd level missile code is cracked the game shows me immediately the video that the 2nd level missile code is cracked, despite the fact that I did this about 15 minutes before. A few seconds later the 3rd level missile code is cracked, the video is played where the satellites get destroyed and then the game didn't end. When I reload the saved game shortly after the 2nd level missile code is cracked and play until the end everything works as it should.
The timer of 1 hour is ok, after all necessary research are still 10 minutes left on the timer.
In the file attached are the logs from Gamma 08 and the two saved games from Gamma 08.

Re: Help needed testing 3.2.x Campaign games!

Posted: 21 Jan 2018, 03:54
by Bethrezen
ok so download the mod started again and have just played up to the same point on alpha 06 that i remarked about earlier and the mod seemed to help and it does start moving now when I shell it with my mortars, although it's reaction still seems a little slow though it is quicker then before.

Also for some reason instead of coming after my mortars like you would expect it actually drives off round the other side of the plateau and goes after my lancers which seems a little odd, so think it still needs a little work it is however a step in the right direction.

here is my alpha 06 save so you can see what I mean, just mind and remember that you will have to add the mod or it wont load.

Re: Help needed testing 3.2.x Campaign games!

Posted: 21 Jan 2018, 04:11
by Bethrezen
General Issues

Ok so i got 3 more issues to add to the general issues list.

1.) You can incorrectly build oil derricks on burning oil wells, by telling your builder to build an oil derrick on a burning well, then saving and reloading and the builder will then incorrectly commence construction even though the oil well is still burning, as you can see in this screen shot.

Image

This doesn't seem to be level dependent you can do this on any level.

2.) When you have more then 1 commander selected and you for example tell them to hold fire the hold fire state will not be applied to all the selected commanders and you will have to do each one individually, the same holds true when you for example have 3 commanders selected and you click to recycle only 1 will go, the same thing happens when you for example have say 3 commanders selected and then change the retreat state it will not be applied to all 3 commanders and you will have to click each individually.

This is incorrect behaviour when you have more then one commander selected what ever command you give should be applied to all the selected commanders you shouldn’t be needing to issue the same order multiple times to each commander individually so this is something that needs looking at because its kind of irritating and makes working with multiple commanders more difficult then it should be.

3.) When you target say a wall and order your units to fire your units will only fire the number of shots required to destroy the target, however when removing feature objects to avoid pathing issues this is not the case and all your units will fire at the target instead of just the 1 or 2 which are actually needed to destroy it, so again this is something else that should be looked at, because this is something that wastes a lot of time.