T2,T3,T4 Artillery Glitches

[Swaygin]

Please see replay #3642366 for a demonstration and commentary on these issues. I attempted to attach a game log, but it seems my replay.log file is for the last non-sandbox, multiple-player game I played instead of the sandbox game I just completed.

Note: All mods were indeed disabled.

First please let me say that the artillery changes are awesome. Personally, I've found that there were some problems with not firing, and even with these fixed, artillery required far too much attention.

The particular change I'm referring to is queued attack orders at points on the map (but not units). If I've queued up attacks on 4 different points, for example, all artilleries will now fire 3 shots at a target and then move on to the next.

All of the issues I'll refer to occur only for the 3rd, final shot. There may be issues in general, but I haven't noticed any in my limited experiments.

For T2 artillery, there seems to be a problem only with Seraphim, the Zthuthaam. The third shot is consistently skewed in the direction of the next target. I believe the turret is moving too quickly, causing it to fire when already beginning to move on to the next target.

For the UEF T3 artillery, the Duke, and the Cybran T3 artillery, the Disruptor, the third shot is consistently at maximum range, irrespective of the distance of the targets.

For the Mavor and Salvation, the third shot travels a ridiculous distance, and likely well beyond the edge of even the largest maps. I believe this problem is the same as that for the Duke, i.e. the third shot travels at maximum possible range (which, for the Mavor, is technically limited to 4000)

The Aeon T3 artillery, the Emissary, appears to have the same problem as the Zthuthaam, i.e. the third shot is skewed. Additionally, the third shot is often far, far short of the target (i.e., well beyond the designed variation in accuracy). The Seraphim T3 artillery suffers from this range glitch as well, although it sometimes significantly exceeds the target.

For all artillery, however, I haven't noticed any problems when it is attacking a single point.

[nine2]

game.log is the log of the last game you played
replay.log is the log of the last replay you watched

so post game.log

[Swaygin]

Oops, I will do so in the future.

I have commented throughout this replay, which I made specifically to illustrate these issues. I would be surprised if they could not be replicated.

If the game.log file is required I can make another replay, but I'd prefer not to if possible.

[ckitching]

The log probably won't be interesting, and there're handy steps to reproduce anyway.
Perhaps a regression from https://github.com/FAForever/fa/commit/ ... 067d2aa0e6

[CodingSquirrel]

This is something I have noticed as well. Another good example is the t1 arty. The first two shots fire properly, and the 3rd goes in weird places. In the past I believe I also had an issue when testing this on an Asswasher, the third bomb would sometimes be way off. Last night I also tested the Gunther and it would pause a long time before firing and would only fire one shot per ground fire location.

I'll probably take a look later tonight at what's driving that system to see what might be going on.

[Swaygin]

I'm confident it is related to the Emissary problem; I was checking artillery because I'm the one who pointed out this problem, and IceDreamer was nice enough to take a look and fix it.

There is another problem unfortunately. T3 Artillery will consume power after it has been constructed but without having been shot (4k for UEF, 4250 for Aeon, which I believe are the energy costs for firing).

A 10-second way to verify this is to play a sandbox game and use Alt+F2 to create a heavy artillery. You will see that there is a significant energy deficit approximate to the amounts above.

The only way to remove this cost, that I've found, is by ceasing to run an energy deficit.

Afaik, this problem only occurs when there is a power stall at the time of completion, but this alone isn't sufficient.

It is very peculiar behavior; the actual resource cost (in the upper left) continuously decreases from ~4k to almost nothing, but then shoots back up to 4k. I have only "fixed" this by building power. I can't narrow down the exact conditions, but in experimenting this has occurred "naturally" with T2 artillery (Aeon's, the Miasma) as well.

I've attached a replay demonstrating some of these things and the game.log. I immediately demonstrate the effect using Alt+F2, but my comments are disorganized as I don't fully understand the exact conditions for which the other cases occur when construction is "natural."

Also, I'd like to say thanks to all you guys for improving upon this already-amazing game, which for some reason I've only recently discovered. I wish I could help with coding, but I have no idea how to do anything besides VBA for Excel and Maple

[Zock]

Sometimes spawning units with alt f2 can cause wired behaviour itself, i think to make sure the problem is real it is better to spawn some engies and a paragon and build it on +10 gamespeed. I don't know the exact problems with alt f2, so it is probably not the case here, but it is generally better to avoid it.

[Swaygin]

Please check out Replay #3647005. This concisely demonstrates the problem with T2 artillery, with commentary. No mods were used. The game.log is also attached.

The problem seems to occur only when power stalling (so I can't use a paragon to generate the effect). Once begun, it lasts for a little while (I clocked it once at 12 seconds, don't know if this is consistent though), apparently irrespective of the power stall being fixed. I.e., power could afterward be maxed and with a surplus, while the artillery will continue to drain energy.

It is a pain to demonstrate in Sandbox mode; I'd recommend looking at the replay. If you'd like to see it first hand:

(1) Alt+F2 some mass storages, mass extractors, T3 pgens and a single energy storage.
(2) Upgrade your ACU to T3 and wait until your mass exceeds that necessary for constructing the artillery.
(3) Destroy the extractors and pgens. Your power income should now be only the 20 from your ACU.
(4) Construct the T2 artillery using your engineering-upgraded ACU. You can put the speed up very high and/or reclaim with your engineers, but to be certain it will occur do not build pgens.
(5) When it nears completion, stop all reclaiming and slow down your game speed.
(6) Look at your power cost after completion and for a short while after; it should cycle from ~140 to ~120, ~100, ~80, ~60, ~40, down to around 20-30, at which point it jumps back to ~140.

Although demonstrating this glitch is very complicated, this does not mean the problem is obscure or minor. I don't know the exact conditions for it to occur, but it is certainly far broader in scope than the detailed procedure above would suggest. The procedure is just a reliable way to generate the error and get some kind of idea as to its possible cause.

[IceDreamer]

The power cycle is normal. Once completed, Artillery which require E to fire drain the cost for their first salvo. If stalled, they can't drain at the intended rate, so they pull a bit, then a bit less, spread it out. If you're heavy stall, it takes a long time. Once the alloted E has been drained, it should cease until it fires, at which point it will drain again.