[Vehicles] Reversed SPMA/Helios Artillery

As far as I can tell, the camera shell should not have any rendering restrictions. I wonder what might be causing this...

[edit]
Actually, I just saw that ONSMortarShell defines a CullDistance of 10000 UU. That is the base class for anything any SPMA-style vehicle will lob at you with its main cannon, including any camera shells. In other words, that's a problem you should also have with the standard SPMA and the Perses.
(But I'll fix it anyway, of course.)
In the mean time, mappers can modify the default CullDistance of the affected custom projectiles under Actor >> Projectile >> ONSMortarShell. (Remember to turn off "Placeable classes Only?" so you can see them.)

[edit2]
In a future release I'll fix the Culldistance issue and I'll also make AVRiL rockets redirect to the Helios/Hephaestus Artillery again. Should I also do that for the "standard" Reverse SPMA?

I've never noticed the same trouble seeing the camera that I have on Arctic-Junk but I don't think there are a lot of maps where a SPMA gets used much. Capn Gordans is the one that stands out the most and it's fog distance is 10000. With 4 of them in this map I definitely see more use of them than I have in any other map. Raising the cull distance on the camera to match Arctic-Junk would be great. They are already hard to see on that map for some reason, especially the blue one.

What about them drawing on the ground? I saw that again yesterday. Tried to get a screenie but the round ended.

Imo, the Avril should redirect on any SPMA based vehicle but I tend to have opinions like that that others don't agree with. People like to sit back and easily dominate because of unbalanced play, . While it's much better than D-Junk, A-Junk still suffers from being very difficult to make a comeback on and one team is often pounded back to their base with very little chance of recovering. The new Artillery vehicles make that worse and are kind of a replacement for the ridiculous cluster bomb, in that they can sit back and suppress nodes with little chance of getting killed because you have to get by so many of the Artillery's teammates to get to it, just like the Arbalest. UT is an arena shooter. Any weapon that breaks the requirement for a direct line-of-sight needs to be carefully considered.

Right. I guess I'll give a default value (Reverse SPMA being "redirection-proof", Helios and Hephaestus not) and allow mappers to relatively easily change that by editing a default value for that vehicle class.
The reason I changed that part in the first place was that the AVRiL redirection made the SPMA weak. You essentially draw in all enemy AVRiLs from around you by launching that camera shell. Literally every opponent can attach you without you having a way to defend yourself. The Helios now has the Poltergeist/Basilisk energy shock wave for the driver to take care of incoming AVRiL rockets, while the Hephaestus driver could use the flamethrower to do the same thing.

Camera visibility probably should be fixed, though. It's bad enough that you are being attacked by a vehicle you can't see. You really should at least be able to figure out how it saw you.

Originally Posted by Wormbo

[edit2]
In a future release I'll fix the Culldistance issue and I'll also make AVRiL rockets redirect to the Helios/Hephaestus Artillery again. Should I also do that for the "standard" Reverse SPMA?

Yes, I think the R-SPMA will be used a lot more than the other two.

Originally Posted by Wormbo

...The reason I changed that part in the first place was that the AVRiL redirection made the SPMA weak. You essentially draw in all enemy AVRiLs from around you by launching that camera shell. Literally every opponent can attach you without you having a way to defend yourself...

True but by removing the Avril redirect you flip that situation around so that the SPMA can attack without the victims having a way to defend themselves or attack back. And the SPMA can one-shot kill someone on foot where as it takes several Avril hits to kill a SPMA.

If a SMPA driver hears an Avril lock and then destroys their own camera the Avril lock is lost, right? So a good SPMA driver, when hearing an Avril lock, is going to destroy their camera and move slightly to avoid the drifting Avril. If a SPMA driver lets someone lock Avrils onto them by redirect of the camera and doesn't do anything about it then they deserve to die. I think even with the Avril re-direct the SPMA driver still has the advantage.

Originally Posted by Binger

True but by removing the Avril redirect you flip that situation around so that the SPMA can attack without the victims having a way to defend themselves or attack back. And the SPMA can one-shot kill someone on foot where as it takes several Avril hits to kill a SPMA.

If a SMPA driver hears an Avril lock and then destroys their own camera the Avril lock is lost, right? So a good SPMA driver, when hearing an Avril lock, is going to destroy their camera and move slightly to avoid the drifting Avril. If a SPMA driver lets someone lock Avrils onto them by redirect of the camera and doesn't do anything about it then they deserve to die. I think even with the Avril re-direct the SPMA driver still has the advantage.

I agree with this. The SPMA can nestle itself in spots unreachable by foot soldiers.

I suppose that's a reasonable argument for keeping the redirect in all cases. However, I'll still add some slight modifications there. The redirect will no longer kick in immediately if the attacker has a line of sight to the vehicle. Instead, the rocket will have to be underway for at least one second already before considering the turn towards its eventual target. Also, that message "SPMA acquired" will only be displayed when the AVRiL rocket actually switched target from the camera to the vehicle, not when the vehicle was the target from the start.

Originally Posted by Wormbo

I suppose that's a reasonable argument for keeping the redirect in all cases. However, I'll still add some slight modifications there. The redirect will no longer kick in immediately if the attacker has a line of sight to the vehicle. Instead, the rocket will have to be underway for at least one second already before considering the turn towards its eventual target. Also, that message "SPMA acquired" will only be displayed when the AVRiL rocket actually switched target from the camera to the vehicle, not when the vehicle was the target from the start.

Would all this also work with your battery launcher? I've included it on the same map as the SPMAs with varied results.

Yes, the Battery Launcher and its rocket extend the AVRiL and its rocket, respectively. The entire redirection logic is based on the AVRiL rocket's API and otherwise entirely implemented in the artillery vehicle.

New version available! (link in first post updated)

Changes:

• Added missing death message to Helios nuclear explosion and increased its network priority to make it show up more reliably under load
• AVRiL rockets are now redirected to the artillery vehicle again, but only after at least one second of flight
• Fixed camera shells (and primary artillery shells) not necessarily being visible all the time due to limited CullDistance inherited from stock SPMA
• Improved visibility of blue camera shell thrusters against sky/fog
• Adjusted bot AI criteria for picking the gunner seat over of the driver seat
• Improved time to target calculation for shell fuse timer in case an invalid target location was selected by the player
• Reduced amount of Hephaestus napalm by 45%, but increased its damage by 25%
• Reduced average fire particle count for Hephaestus napalm by 50%
• Reduced network bandwidth impact for Hephaestus napalm (I hope without significantly breaking it)

Originally Posted by Wormbo

New version available! (link in first post updated)

Changes:

• Added missing death message to Helios nuclear explosion and increased its network priority to make it show up more reliably under load
• AVRiL rockets are now redirected to the artillery vehicle again, but only after at least one second of flight
• Fixed camera shells (and primary artillery shells) not necessarily being visible all the time due to limited CullDistance inherited from stock SPMA
• Improved visibility of blue camera shell thrusters against sky/fog
• Adjusted bot AI criteria for picking the gunner seat over of the driver seat
• Improved time to target calculation for shell fuse timer in case an invalid target location was selected by the player
• Reduced amount of Hephaestus napalm by 45%, but increased its damage by 25%
• Reduced average fire particle count for Hephaestus napalm by 50%
• Reduced network bandwidth impact for Hephaestus napalm (I hope without significantly breaking it)

Thanks, I'll update these for the next version of ArcticJunkyard.

Sometimes (often) I cannot shoot down the camera for one of these (or both, not sure). No matter how many times I hit it it will not fall. Other times it comes down with one shot. Very frustrating. Also, sometimes the camera will appear to be sitting on the ground. I can NEVER destroy it when it's on the ground.

Originally Posted by Wormbo

New version available! (link in first post updated)

Changes:

• Added missing death message to Helios nuclear explosion and increased its network priority to make it show up more reliably under load
• AVRiL rockets are now redirected to the artillery vehicle again, but only after at least one second of flight
• Fixed camera shells (and primary artillery shells) not necessarily being visible all the time due to limited CullDistance inherited from stock SPMA
• Improved visibility of blue camera shell thrusters against sky/fog
• Adjusted bot AI criteria for picking the gunner seat over of the driver seat
• Improved time to target calculation for shell fuse timer in case an invalid target location was selected by the player
• Reduced amount of Hephaestus napalm by 45%, but increased its damage by 25%
• Reduced average fire particle count for Hephaestus napalm by 50%
• Reduced network bandwidth impact for Hephaestus napalm (I hope without significantly breaking it)

Awesome, I will swap in my Pyroclastic edit too.

Andrew

Wormbo please look into the camera issue with these. About half the time I can shoot at it over and over and it does nothing. Other times it comes down in one shot. It seems that the camera is often not actually where it shows. When this is happening my Avril will not lock on to it but will sometimes lock on to a place where it is not showing. Also when it is showing on the ground I believe it is actually in the air directly above that spot.

I did a demorec of this but it shows me aiming away from the camera. I will try to remember to get an actual screen capture video but just trust me that my description of the issue is accurate and that others are getting the same thing.

Oh, I see the problem. The original SPMA camera shell has a location correction feature that also still exists in the Reversed SPMA camera shells, except that it does not get the necessary correction information anymore.

THB, I don't see myself releasing anything any time soon. A fix should be relatively easy, though. If you use Reversed SPMA embedded in a map, you can edit the ReverseSPMAArtilleryCameraShell class (show non-placeable classes, then go to Actor >> Projectile >> ONSMortarShell >> ONSMortarCamera >> ReverseSPMAArtilleryCameraShell, doubleclick it to open script editor) to include the line "RealLocation = Location;" both in DeployCamera() and EndedRotation(). Just add it anywhere at the start or end of both functions, then click the "Compile changed scripts" button (or select it from the menu) to update the bytecode accordingly. Finally, save the map to persist the changes. (No further rebuilding required.)

Background: ONSMortarCamera uses simulated state Deployed to perform the logic, while ReverseSPMAArtilleryCameraShell moved it to simulated function DeployCamera() and simulated event EndedRotation(). (I don't remember why, but it was probably a cleaner solution to a different replication-related problem.)

Originally Posted by Wormbo

Oh, I see the problem. The original SPMA camera shell has a location correction feature that also still exists in the Reversed SPMA camera shells, except that it does not get the necessary correction information anymore.

THB, I don't see myself releasing anything any time soon. A fix should be relatively easy, though. If you use Reversed SPMA embedded in a map, you can edit the ReverseSPMAArtilleryCameraShell class (show non-placeable classes, then go to Actor >> Projectile >> ONSMortarShell >> ONSMortarCamera >> ReverseSPMAArtilleryCameraShell, doubleclick it to open script editor) to include the line "RealLocation = Location;" both in DeployCamera() and EndedRotation(). Just add it anywhere at the start or end of both functions, then click the "Compile changed scripts" button (or select it from the menu) to update the bytecode accordingly. Finally, save the map to persist the changes. (No further rebuilding required.)

Background: ONSMortarCamera uses simulated state Deployed to perform the logic, while ReverseSPMAArtilleryCameraShell moved it to simulated function DeployCamera() and simulated event EndedRotation(). (I don't remember why, but it was probably a cleaner solution to a different replication-related problem.)

Thanks, I'll incorporate this into my next ArcticJunkyard edit.

Thank you Wormbo. Does this apply to the Helios also?

Putting the code in the Reversed SPMA class will apply to all related vehicles.