Hey all, This is not a formal bug report as such, as I've mentioned it numerous times in the past but don't want to resurrect those older (and prior version) threads. However, I did want to add some information from my own testing... In an attempt to snag a save where I can demonstrate this issue for the devs to look at, I've come across one common problem: No matter what issues I've having with unresponsive turrets when I SAVE my game, said turrets respond instantly the moment that save is loaded. Basically, turrets aren't inherently broken - though they sometimes seem this way during play, especially prolonged play - but they do start to fail over time. In my latest example, I'd approached a POI in my CV. I could see Drones were nearby on the Radar, but they weren't rendered in yet. What was rendered in was them firing - but that's a separate, previously reported bug. None of my Six Cannon turrets did anything, and at the very least three had a perfect LoS on the attacking Drone. I save my game - turrets still 100% unresponsive - and exit to the menu. Upon reloading, ALL my turrets instantly open fire, destroying the Drone in a couple of seconds. Perfect. Note: I'm always running the latest Experimental version, I just play Single Player currently and my games are always unmodified. I hope this rather informal report is of some value, I just wanted to clarify how this issue seems impossible to recreate from a save, as the act of loading causes all turrets to instantly wake up, and respond perfectly. Scoob.
I'm not quite sure at the moment Pantera. I'd been playing for a couple of hours before I moved near the POI in question. Upon reloading my save, the Turrets on my CV remained responsive for some time, easily taking out any incoming Drones before they could cause damage. Some time later, after I'd captured the POI and placed my own Cannon Turret on it a Base attack triggered. I noticed that the newly built Base Turret was very slow to respond once more, yet it couldn't have been more than half an hour or so since I'd loaded this save. Incoming Drones got off multiple Shots - both Rocket and Minigun - before they were defeated. I'll try to be mindful of time played while observing turret responsiveness - sometimes I lose track of time. Scoob.
Hi Scoob (long time, hope all is well), Hi Pantera; I can confirm/back up Scoob's report - I haven't submitted on this in at least a year or two, but the issue remains. I haven't had a clock/stopwatch on it either, but I have noticed that sometimes, turrets don't engage enemies... I've got an HV build with ventrally mounted turrets specifically for taking out pesky spiders (haaaaate spiders... lol), raptors, etc... I have sat and watched as they approach, and no firing... I then thrust away (so they aren't directly underneath me) and watch again, same problem... rinse and repeat 2 or 3 more times... then for no apparent reason 1 or more of the 4 turrets will start firing. Same thing applies on my dorsal turrets - whether drones (I have 15mm for them) or base turrets (I have rockets for those)... sometimes I can be wheeling around the base for minutes waiting for the turrets to pick a target and fire... sometimes they fire immediately and mop up the enemy. As perhaps a "connected" issue/observation, I have noticed that on my rocket turrets in particular, when they DO start firing at turrets they don't always register the target's destruction - they will fire 2-3 more rounds of missile at the same point on the POI, before either finding another target or just going "dormant". Like Scoob, IIRC this issue has been around for a LONG time in some form or another. If you can think of any logging, savegame, or other debug info we could collect to track it down, I'd be happy to gather it for you.
I was playing SP today and had the devil of a time trying to get my hover turret to fire at drones. It wouldn't even load rounds automatically first time until I manually loaded, then it only fired sporadically. I ended up just relying on gatlings. I have noticed this problem previously.
If other can confirm the not firing > save > reload > instantly start firing - obviously Single Player only - that I experience that'd be a start. Then we can try to time how long it is from loading a game that turrets start to become less responsive. Not that easy to do while just playing normally - I can go hours without engaging a target at times - but I'll certainly be making a note of time played etc. for the next time this occurs for me. Hey @Jacob Adjani - it's all good thanks, hope the same is true for you. Scoob.
An interesting note from some playing today - earlier I submitted what I thought was a completely unrelated bug report (about the map in space not functioning correctly... not showing resource icons, discovery notifications, etc.). Due to another issue I'm experiencing, I have to completely exit the game to type anything up - so I did so, submitted the bug, then loaded the same game back up... and presto... all the resource and POI icons are present on the map! I know they are completely different code components - so this leads me to think... overall memory leak? In my experience with software development (business stuff, not gaming), a problem that develops over time is often a memory leak/memory lock not clearing. Just thinking out loud here... 2 totally different methods, but displaying similar behaviours - functional on load (clean memory addresses), function ceasing/decreasing over time (as memory registers are claimed/assigned)? @Pantera - Apart from my layman's guess above, I've been thinking this one through, and I'm not sure how we could conclusively determine cause here - I think this needs to be tested with a full debug window/logging setup at Eleon... not sure how else one could gain insight as to what the targeting algorithms are doing at any given time.
I agree this is a tricky one. Is it just playtime that causes turret performance to degrade, or some specific action taken by the player that slows their response time? Usually, by the time I experience this bug, I've just been playing not logging my exact activities or time spent, though I will try to be more aware of these in future. This does sound like some sort of memory leak, but it's leaking in such a way that's preventing the turret code from working / introducing a potentially huge delay in response. Scoob.
im having nothing but problems with POI's reacting on my game. They work sometimes and other times they dont its very bad sense the updates.
Ok... we may be on a "wrong track" here - just had a "use case" that runs contrary to some of our earlier assumptions... I loaded up a savegame, jumped in my HV to go mine a nearby ore deposit, and on approach saw a rocket drone... When I got in range, dorsal turrets fired and took it out, and I'm thinking "yup, just like @Scoob said... fresh load, works fine"... ... ... then... One of my ventral turrets took a single shot off to the left of my view at the time... I didn't see them, but a group of 3 arachnids were just out of frame to the west of me. I waited... nothing... spinning the camera around, they were in plain view... but other than the first single round shot that alerted me to their presence... nothing. No firing. They approached and as before, I'm thrusting away to keep them from being directly below me (and therefore outside the turrets elevation)... this repeats multiple times and now I'm getting sporadic fire from different turrets at different times... it takes out 2 of the 3, and now I'm far enough away from the 3rd that he's not aggro'd on me anymore, but he's still WELL within range... again, no firing. I jumped out of my cockpit, exited the HV, with a mind to seeing if I could "line up" my view with the turret - thinking maybe it had something to do with line of sight... best I was able to tell their was a clear path to target, or at least some of the target... So, since the above occurred (or at least started) within 60 seconds of load... this is one case that refutes the "memory leak" theory. Could it actually be "line of sight"? Is the target code not finding clear pathing to the hit boxes on the targets maybe? Are the hit boxes too small?
Not sure it helps, as I can't find any *cause* info, but: just had another occurrence - this time with Zirax turrets; I had been playing for a while, landed on a new planet, jumped out of my CV to do some tweaking/re-design on it's attached Scout SV... as I've got her torn halfway apart, the heretofore undiscovered Epsilon Patrol Vessel decides to mosey on over to where I'm parked and my CV, docked HV, and I start taking some light fire from it's plasma turrets. I set a land speed record for slapping 2 Pulse Lasers on my SV (had them and some ammo laying around my cargo holds), jump in my SV (naked as she was), and engage the Patrol Vessel. It has 2 ventral plasma cannons, and a single dorsal one... I fly up to the ventral aft turret, and at first it's ignoring me, and now the CV and HV on the ground as well... the ventral forward and dorsal turrets, not sure, they didn't have line-of-sight on me at the time. After I've punched a few shots at the ventral aft turret... maybe 20 seconds... it "wakes up" turns and engages me. Only got 2 rounds off before I blow it up. With the ventral aft turret gone, I'm now looking directly at the ventral forward one... and it's looking directly at me... BUT not firing. I fire away at it until it's destroyed, it never having fired at me. Same story when I maneuvered around topside to engage the dorsal turret - it didn't even swivel to "look" at me... just sat inert until I killed it. Again, I have no idea what info beyond the above account of events I can provide. @Pantera, 1) is there any "combat logging" or anything else available or that can be turned on to help track this? - and - 2) in the course of the above adventure, I used SV Pulse Laser guns for the first time in a long while... I noticed that I had to click to fire each round... as in, click-and-hold did NOT fire more than one round... they used to fire VERY rapidly, which was what I liked about them; not sure if this is a bug, or a change made that I missed in release notes? Let me know if you want me to submit a formal bug rep on it please.
This issue can be noticed all the time in general PvP, but it's sporadic and not easily replicated. It seems to be independent of playfield load, and ping.
My read on everyone's observations here is that this is an intermittent, and yet consistent and persistent concern - I'm thinking the only way to track/trigger it would be in a full debugger environment (which players don't have the ability to do). You guys (the QA Team) have the tools required for that? Can you throw up a debug window and go bomb around in a CV or HV and pick a fight with something... see what's happening in the code in real time?
This is my take on it as well. This has been such a long standing issue now and there is absolutely zero possibility of us providing a save setup where this is reproducible (since the act of leaving the game resets the bug so to speak). I see no other way to progress forward with it. This is why so many of us experience it from time to time but don't bother reporting it over and over. It's simply never going to be reproducible by providing a game save I'm afraid.
@Pantera Ok, now I'm becoming more convinced it's line of sight (or perhaps it could be a dreaded multiple cause issue, and line of sight is just one component). Forgive the novel that follows, but I want to lay out why I'm thinking this way. I just completed an all-nighter running tests on this on the starter planet. I built my favourite HV (the one you have the BP for), and went about exploring looking for opportunities to test while just playing the game in general. This was about an 8 hour stretch, so lots of consecutive time with playfields loaded etc. I focused on my ventral turrets - the HV's lower/ventral weapons are 4 x 15mm retractable turrets, mounted inverted at the four outward corners of the ship's underside. In the version of the BP you have (which has been my "standard" one for some time), those turrets are "recessed" by 1 block... meaning that when extended the tops of the turret layer are level with the lowest block on the build. I carved out niches for them, with nothing obstructing them in the outer-forward and outer-side arcs relative to their mount position. Put another way, at zero elevation, they should each have a field of fire in the 90 degree range, perhaps a bit more. Of course, we are rarely engaging enemies that crawl/slither across the ground - from a 3 meter high hover - at zero elevation. So in theory, their practical firing arcs should be (and I've seen demonstrated many times) much wider. All that said, after experiencing the same thing (spiders/raptors approach, no firing, thrust away, maybe a shot or two) over and over again... I decided to test a theory. I redesigned the underside of the HV and dropped the BASE part of the turrets level with the undercarriage... meaning, with the exception of a 1x1 "strut" placed behind each turret (as "landing gear"), the TURRET section of each cannon has an near 360 degree field of fire even at zero elevation. I ran this for a few hours, and while unfortunately the problem over all persists... I did notice an immediate, consistent, and marked improvement in responsiveness of the turrets in situations, as close to the ones earlier in the night as I could make them, versus the "recessed" mounting. It's not conclusive to be sure, and as I said, I also think this could be the bane of all QA - a problem with multiple causes - BUT the change was VERY noticeable. Is there anything you could look into code wise as to how the angle-of-attack / pathing-to-target algorithms are handling engage/no-engage calculations? How strict are they? Again, thinking out loud, but the testing this past night was significant enough in degree that I thought it warranted mention. ALSO - how does the path to target handle "full" vs "partial" blocks - I have made some observations with my drill turret that could relate to this... despite have zero VISUAL obstructions, their are some positions that will not drill the ground... I suspect it's because their may actually be a "block" in the way, even though it's not visually there (as in ramp blocks, thin wall blocks, half wall, etc.)
I have to agree with pretty much everything here, especially Jacob's most recent report. I have a HV with four turrets, placed at each corner of the vehicle, with >270 degree firing arcs (the only obstruction being the vehicle itself, at one corner of the turret placement), and being placed directly above the hover pads so they are very close to the ground for firing at spiders and other crawling enemies. I have seen many times, where at least two of the turrets should have a completely open firing view of an enemy, yet they do nothing for several moments, forcing me to move around to keep the enemy from getting too close, and trying to make sure the turrets maintain an open firing view of the enemy, but they'd still only fire sporadically, if at all. So improving LoS for the turrets may help, but it is definitely not the primary cause of this problem. Perhaps on a related issue, I've also noticed that base cannon turrets have "insta-hit" projectiles (since I can see the impact before the projectiles actually reach the target), and yet they miss almost every single shot, until the target gets worryingly close to the base. How can their targeting algorithms be off so badly? If all they have to do is line up a shot, and then fire while the target is in the LoS, how are they missing? They do not need to "lead" the target, because the shot will hit instantly if lined up correctly, and this can't even be the problem anyway because I noticed that the shots were being aimed below the target as it was flying in. Why, in all the universe, would they be aiming below the target if the target is maintaining the same altitude while flying in?? Even if they were trying to lead the target, then they should be aiming slightly above the center of the target! So what is going on with that targeting behavior??
Hey @Pantera , another thought/idea just struck me... do you know/can you look into what precise point on the turret block is the start point for a path to target calculation? I've been using retractable turrets exclusively since you guys released them, and it occurred to me - they are 2 blocks high... IF the start point calculation is actually made from a node placement at the end of the barrel of the extended turret... well never-mind, thought/idea isn't relevant to the investigation. IF the start point is determined through an averaging of the x/y/z dimensions of the total block, the block center point, or the "first" cube within the block - which would likely put it in the "base" NOT on the turret... then this could have a HUGE impact on targeting. None of my last couple threads address the issue of "generally non-responsive" turrets - which I'm still seeing (meaning turrets that don't engage at all, or for extended periods of time - particularly those on the dorsal side that have an undeniable path to target to flying enemies)... BUT I'm thinking with this info we could potentially either confirm or eliminate one possibility (i.e. targeting code).
I don't have retractable turrets. I have tried changing targetting path from the turret by lowering/raising front/back of HV and changing position, elevation to see if this was a consideration during the failures - I believe most players would try this instinctively (once the shock of impotent turrets has worn off. ) - but it doesn't seem to have an impact. Drones in particular have a higher elevation so it is less likely pathing was obstructed by either the turret or parts of the HV or terrain. It seems to me that the turret not recognizing an enemy or a predator at times or is not completing the task after recognition.
Hey Russell; I agree overall it does seem like recognition of target, or speed of recognition (or lack thereof), would explain everything people are witnessing... my thinking on the "pathing theory" was to have Eleon at least give it a quick look - more with an eye to eliminating it as a potential cause. To quote my favourite Vulcan - "if we eliminate whatever is impossible, whatever remains - no matter how improbable - must be the truth". Unfortunately, @Scoob and I have been reporting issues with turrets for 2.5 years... this issue or group of issues is pretty "long in the tooth" at this point (to say the least)... I figured perhaps if the community helped Eleon "step through" 1st tier diagnosis, we might actually get a fix for it after all! Here's hoping!
