JG4_Deciman Posted November 24, 2020 Posted November 24, 2020 Hi there, this is for experts... The trains are still bugged And even if I tried everything to check I have no clue what is wrong. The mission inside contains a very complex logic with different layers Every layer serves the sub layer and is served by the sub layers It contains a very complex train system (still unfinished but working until the bug happens) When spawning at 'Rotterdam' you can simply move to external view and watch what happens.... The entire trains (max 10 the same time active) handle their spawning and movement completely themselves. There should be NEVER 2 trains running on the same track Whenever different tracks meet the trains react on possible ways to move (or wait or despawn) Trains passing by is enabled for different positions on the map (groups named as checkpoint) The entire logic IS working at nearly all positions Simply watch the area at Schiedam (1007-7) where a very high concentration of trains using the logic is happening It IS working But on a different track (Gouda -> Den Haag) same logic has massive problems. Train spawning at Gouda proceeds to 0808-5, waits (because other train wants to drive from Den Haag to Gouda) lets the train pass by (despawning and spawning again) spawns in the right orientation is ordered to move towards Den Haag but decides to turn arround, move a while and derail And after implementing some debugging items (smoke) I am 100% sure that ALL waypoint from his starting position (and so all waypoints into the direction the crazy train moves) are disabled and were NEVER triggered again. And to say it to the devs directly: There is NO way to reduce the (working!) logic for debugging. This bug happened at several different positions The same logic was used on working AND bugged positions Even reducing the used waypoints didn't change it So now it's up to somebody who knows more than I do to tell me WHY a train ignores to move where it was told to move to, WHY a train decides where to move even if it was told to move somewhere else and WHY the same logic in most cases is working, and in some cases fails Deci Rotterdam.zip
Gambit21 Posted November 24, 2020 Posted November 24, 2020 Simple test mission - isolate your problem logic.
JG4_Deciman Posted November 24, 2020 Author Posted November 24, 2020 As there is no problem logic there is nothing to isolate.
JG7_X-Man Posted November 24, 2020 Posted November 24, 2020 2 hours ago, JG4_Deciman said: Hi there, this is for experts... I was going to chime in as I have had that issue and solved it when @Sketch gave me an idea. However, as you asked for an expert. I will let someone else help you. Good luck.
Gambit21 Posted November 24, 2020 Posted November 24, 2020 36 minutes ago, JG4_Deciman said: As there is no problem logic there is nothing to isolate. Perfect - problem solved.
JG4_Deciman Posted November 24, 2020 Author Posted November 24, 2020 1 minute ago, Gambit21 said: Perfect - problem solved. No, not solved But to confirm there is no problem inside the logic, here is what IS programmed... ANY waypoints is only triggered by an additional timer (0 delay) which is enabled/disabled by the logic and all these timers are triggered by a special timer (150ms delay) whenever a setting is made. Nothing inside the mission is triggering a waypoint directly Default setting after Mission Start: -> SETTING Offline for each of the 10 trains of each location That means: ALL waypoint, waypoint triggering and triggered by waypoints actions are deactivated for the particular train of the location Once a train spawned the initial logic is activated, performing some actions, leading the train to the final waypoint and when reaching that - the next location is activated (initial setting and initial waypoint) - the previous location is set offline (so EVERYTHING is deactivated, made visible for the problem location with smoke) So I am 100% sure that no waypoint is triggered, can be triggered or will trigger any action by accident And at the point where the train behaves strange I've debugged all triggering of waypoints by subtitles. The only waypoint triggered is the one the train should move to, but it turns around and moves into the complete opposite direction. And all these things only happen at a specific point, even if the same (!) logic works as it should at different positions So I think you can imagine that I have absolutely no idea why... Deci
Sketch Posted November 24, 2020 Posted November 24, 2020 (edited) Haha me an expert! Not really! It sounds like a track/road bug and not an issues with mission design. However, Gambit is right - you should isolate the issue, and use the process of elimination. To me it sounds like this: Train X spawns in at location N Train X moves to location Y Train X deactivates at location Y Train X activates, after some time or some logic, at location Y Train X moves to location Z, but instead moves towards location N And your question is, "Why does train X moves to location N instead of location Z? My logic looks correct, but it's very complex and I need a rubber ducky right now." Is this correct? I'm making assumptions here based on what I've read so far. If it is, then we should make two tests. One where we know the train will do everything correctly - it will move to a waypoint, deactivate, activate after some time, and move to another waypoint. This will be the control group since we know that it works. Then we make another train that moves to location Y from location N - where you believe the track has an issue at. Make sure this new train deactivates, activates after some time, and attempts to move to location Z - just like the control group train. There should not be any different logic. If this new train doesn't move to location Z, like your control group train, then we can send your isolated mission and it's logic to the developers as bug feedback. However, if both the control group train and the test train work as intended, then we know there's something wrong with your complex logic in the big mission. (And that's a whole 'nother can o' beans!) This is the process of elimination and the process of isolation. Good luck, and let me know how it goes. Maybe I can help in some way, right now though; I'm not sure. Edited November 25, 2020 by Sketch Clarity 2
Gambit21 Posted November 25, 2020 Posted November 25, 2020 (edited) 9 hours ago, JG4_Deciman said: And all these things only happen at a specific point, even if the same (!) logic works as it should at different positions So I think you can imagine that I have absolutely no idea why... Deci So make a very simple test mission spawning a train at that point and see if you can duplicate the error just with that minimal logic. That's what we mean by isolate, and what I meant earlier by "isolate problem logic" it may not be your problem technically, but a bug with something else...but it's still problem logic. Capeesh? I'm curious as well about this because I've noticed train high strangeness more than once (trains coming off the rails/crashing) and I'd really like it all sorted for Normandy. Edited November 25, 2020 by Gambit21
Sketch Posted November 25, 2020 Posted November 25, 2020 8 hours ago, Gambit21 said: So make a very simple test mission spawning a train at that point and see if you can duplicate the error just with that minimal logic. That's what we mean by isolate, and what I meant earlier by "isolate problem logic" it may not be your problem technically, but a bug with something else...but it's still problem logic. Capeesh? I'm curious as well about this because I've noticed train high strangeness more than once (trains coming off the rails/crashing) and I'd really like it all sorted for Normandy. I went ahead and created this simple mission. As a mostly multiplayer designer, I actually had a lot of fun making the single player cameras work the way I wanted them to, and learned a bit about how I want them to work next time. Does the track line that @JG4_Deciman mention in his first post have a bug or is there some underlying issue with the complex code? You'll just have to watch the video... And of course, here's the mission file for you to examine and what not. Testing Trains on Rheinland.zip
JG4_Deciman Posted November 25, 2020 Author Posted November 25, 2020 Hi @Sketch The track is the same even if the waypoints are a bit different. The most differences between your mission and mine are: Singleplayer (you) vs Multiplayer (me) mode Activate/Deactivate Train (you) vs Spawn/Delete Train (me) No deactivation of waypoints (you) vs deactivation of waypoints/actions triggered by waypoints (me) Maybe somewhere there the reason for the strange behaving (me) and the performing as intended (you) train is located I tried to reduce my version step by step since a while and I could reduce the number of waypoints used by 1/3, but no change Distance between waypoints: - even logic with closer waypoints is working, problem track has about 600m distance to the next waypoint after respawning But I'll continue to create a 'scripted' version using the base parts of the logic and bypass/delete everything not needed in that case Deci
Sketch Posted November 25, 2020 Posted November 25, 2020 4 hours ago, JG4_Deciman said: Singleplayer (you) vs Multiplayer (me) mode This shouldn't matter, and when I test, I'm going to use time dilation to make things go much faster for me. 4 hours ago, JG4_Deciman said: Activate/Deactivate Train (you) vs Spawn/Delete Train (me) This shouldn't really matter, as they function nearly the same way. 4 hours ago, JG4_Deciman said: No deactivation of waypoints (you) vs deactivation of waypoints/actions triggered by waypoints (me) This is where we have a huge difference and it's most likely where your issue is. I'm again making assumptions, but if the train is going one way, then suddenly going another way, then another way again, it's because one or more of your waypoints to that particular train is getting activated or deactivated when it shouldn't be. 1
JG4_Deciman Posted November 26, 2020 Author Posted November 26, 2020 Maybe I've got an idea, but there is no way to prove it.. My logic used activate/deactivate Triggers for waypoints and actions (timers) activated by waypoints. So it deactivates unused/not needed waypoints and actions (that is a MUST) and it activates needed waypoints (but does NOT trigger them) Assuming the same I realized when using 'Modifiers' for timers set to 0 delay (they were not only reset, but also triggered!) and assuming that activating a waypoint will also trigger it that would explain (partially) why the train behaves in a way it should not. But it would also not explain why this is not happening always, but only at some locations... Checking is impossible I've tried to remove the 'enabler' for waypoints and leave the 'disabler' working. But triggering any disabled waypoint will not make the train move there. And I have not yet any idea how to confirm what I suppose to be the reason... Deci
JG4_Deciman Posted November 27, 2020 Author Posted November 27, 2020 Ok, some (or better more) tests later... Deactivating a waypoint will - NOT use it when it is triggered - NOT let the train stop after it has been told to drive there (but that is not happening inside my logic) So I can be sure, every disabled waypoint CANNOT be telling any train to drive there even if triggered by accident As sending trains to waypoints uses an additional timer (activated/deactivated by logic) there is also a second instance for eleminatig that Next level: Reduced the trains to 2 (without randomizers) and let them spawn directly Problem train left the station, station logic was disabled (made visible by smoke) and was never reenabled for that train (made visible same way) So all waypoints and triggering timers were disabled (no way to trigger them even by accident) Train can only move to the next location (checkpoint) Train arrived at checkpoint, stopped as intended and was waiting for the other train that should pass it. That part is working. Waiting train despawns, other train passes, waiting train respawns (orientation is as intended, heading towards east) Datamining the following by subtitles shows: Respawned train is told ONLY to head to he waypoint EAST of it but turns around and heads WEST following the other train And I really have no clue WHY this happens and why this happens ONLY there (because 3 different locations using the same logic are working perfectly) Next thing I've tried was to implement a delay of 100ms to all settings timers (activating/deactivation the waypoints) so giving the server time to perform any internal waypoint action before it is disabled by me... no difference So I still have a train, ignoring the given destination and moving wherever it wants. And all waypoints beeing at the direction the train moves ARE disabled... The only activated and triggered waypoint is 600m away from it in its orientation Deci __Debugging.zip
JG4_Deciman Posted November 28, 2020 Author Posted November 28, 2020 Ok, a lot of debugging actions later I can declare this as a BUG @-DED-Rapidus Now it's up to the devs to find out WHY this happens and why this happens only at SOME positions on the map AND to fix it soon! My logic (and I mean the full version) IS working but even a very small scripted version gives the bugged result. Train will spawn at 0809-2 2 minutes after mission start. So you'ld have enough time to spawn in, move to external view and move the camera to that position (right of spawning place) Some seconds after spawning the train will change nation from 'neutral' top red and 1 minute after spawning the train will move towards the first waypoint. The train will proceed to 3 waypoints at 0808-5 (one by one) And the delay for triggering the next waypoint is always 200ms Every waypoint triggered activates green smoke on its position Every waypoint reached activated red smoke above its position and deactivates the green smoke on its position. And as expected (and not surprising me at all) the train reaches the 2nd waypoint at 0808-5 (where all the problems are located) is told to drive to the final waypoint, but instead of driving there it turns around and moves where ever it decided to move to... Same strange behaviour was already noticed at other places So finally I can prove that it is not my logic, but a bug. Deci Debugging 2.zip 1 1
jollyjack Posted November 28, 2020 Posted November 28, 2020 (edited) Sidling: we need them trains .... no ifs, buts or maybes. And more ships, don't forget the ships! The more the merrier ! I would love to see that armored train moving and shooting at tanks or airplanes too, a game thing i am willing to pay for ... as with the announced flak trucks. Blowing up railways and roads is still only possible at bridges alas: to quote a well known forum-member here: This game is all about blowing up things. Food for more experimenting below, alas the viewer plane skids off the Moscow map if you don't watch out, feel free to add to this, and if you're lucky a train jumps track and crashes in to the other, i saw it happening ! TRAIN RACE.zip Edited November 29, 2020 by jollyjack 1
JG4_Deciman Posted November 28, 2020 Author Posted November 28, 2020 In addition: As it seems this bug is related to waypoints, so it may affect ALL units using them and that would include planes, too... Deci
Gambit21 Posted November 28, 2020 Posted November 28, 2020 There’s no bug with regards to waypoints and aircraft/vehicles that I’ve seen. 1
JG4_Deciman Posted November 29, 2020 Author Posted November 29, 2020 16 hours ago, Gambit21 said: There’s no bug with regards to waypoints and aircraft/vehicles that I’ve seen. IIRC there was (or is?) a bug with planes taxiing. Have not used taxiways up to now, but I think I've read that something is sometimes not working as it should. And as basic a taxiway is nothing but a row of waypoints fore planes (with very specific conditions...) Deci
jollyjack Posted November 29, 2020 Posted November 29, 2020 Spent some time on taxi route mapping. Jim's manual is quite good on explaining the hows. Especially with a flight of more than 3 aircraft it seems like 'you gotta get lucky' to make it work. Planes taxi-ing in a flock bump int into all the obstacles they can find on their route and into each other. Behave like a flock of geese before take-off whilen the run for a dog. With all that testing and reloading in game and an often crashing ME i usually give up and choose start on runways instead.
JG4_Deciman Posted November 29, 2020 Author Posted November 29, 2020 @-DED-Rapidus Testing was always the same: Using my PC (3.4gHz, 16mb RAM, SSD installation, win10) for both, hosting and playing Running a dServer (separate account, separate installation of IL2) and 'playing' (or better watching) with my main account. None of my installations has any mods installed dServer was told to save logs and host mission NOT public Deci
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now