Work Log

Utter Confusion Top-down game:     So as I said yesterday, I worked on making the Plague Doctor's poison attack better. I wanted a physical poison vial of some kind that would splash an area instead of just immediately poisoning everyone in range. However, I was not able to get this done today.     For some reason this was way harder than I thought. My initial plan was for the Plague Doctor to spawn this vial actor which would then launch itself, and upon hitting something would incur the poison effect upon any players within range. I got all the way through to the part where it would launch before I found the problem that I would spend the rest of my time on without solving. The vial wasn't showing up. It was working, I think, just not rendering. When the AI would choose the AOE attack it would spawn the actor successfully and some time later (presumably when the thing landed) the player would get poisoned but I saw nothing appear on the client or the serve...

Juggling projects     Today I went back to working on the top down game. I was assigned to finish 2 of the enemies so that I could work with our animator and get everything  everything working; animations lining up with ai tasks and that kind of thing. I chose to finish the Plague Doctor and the skeleton archer (mostly because I had already done most of the work on them, but don't tell the team that). These enemies had had some attacks added to them in the new design doc, most of which were variations on already existing attacks.      For example, the archer was given a triple shot which would just be calling the shoot function three times. I did this but then quickly ran into issues. The first being that, as the three arrows would fire, they would all try to attach to each other like they would to a player. I fixed this simply by checking that the actor that the arrow collided with wasn't another arrow.     The next problem was that the ...

Success with multiplayer ft. Steam     I spent all of today looking more into how to get internet multiplayer working. The first attempt was the port-forwarding from yesterday. Either I did it wrong or it didn't help. I thought I opened port 7777 but after I did nothing changed and I wasn't able to connect to anyone. It's possible that the other person also needs to have a firewall rule for the same port? But I don't know.      The next thing I tried was the advanced sessions plugin that I originally tried with for the creative jam. This time I went through the setup step-by-step to see if I could get it to work. Very long story short... it (and I cannot stress enough)  FINALLY  works. I followed along with this  tutorial series up until part 7 and over 3 hours later it was all worth it.      There were some hurdles but I am overjoyed now that it works. The initial setup required adding lots of #includes to source files (on...

Why is online so hard?     So I continued my adventures in networking and I have had one minor success. That success being that I figured out how to test an online session over LAN. That is to just not "use LAN" when finding sessions. It sounds simple when I say it like that but it's actually really great because now I can test all I want from home.      That was really the only god thing that I found today, everything else wasn't all too helpful. I tried going back to the IP based connection but that didn't work still. I tried using public IP's and no luck. I manually opened port 7777 (unreal's default port) in my firewall and still no luck. After I opened the port I tried again with the sessions system but that failed as well.      I have decided for the time being that I am going to try to use the sessions system just because that gives me more options readily without any extra effort. After I decided this I did lots of reading into th...

Figuring out networking     Today I continue my search to find a way to make this game work over the internet. I started out trying to find a way to test the game when I was ready. I wanted to be able to do it all at my house, not having to rely on anyone else to test. But after looking around I couldn't find a way to do this. I wanted to be able to simulate an internet connected game over my own LAN but the only solutions I found were sketchy programs and complex router/modem crap that I didn't understand. For now I have to rely on the availability of my team to be able to test new prototypes.      Once I had that behind me I did some research into ports, public IP addresses, game sessions, and how exactly unreal handles all of it. I didn't find much that was helpful, but I did find this video from 2014 when the blueprint networking nodes were introduced. This made me give the blueprint approach one more look. It seems to work in every environment that ...

First play-tests and edited Skeleton mesh Play-testing:     So far today I have done two play-tests. One with my brother's computer on the same network and one with a team member on a different network. The one on the same network worked seamlessly. I was able to join the games, choose different classes, and end the game with an improvised trigger volume. The test across different networks didn't work though. When you would hut join nothing would happen. I should do more tests to make sure ti is a LAN issue and not an issue with the tester's computer. Although if the problem is that the current system won't work if it's not LAN, then I have a lot of work cut out for me. Skeleton:     Also I received two separate models today for the skeleton and his bow. This has made me realize one thing though; for items that need to animate (e.g the bow string) there will need to be a separate anim BP for that item. This will be a problem for future me though as af...

Git Commit:  9004b33 "Added architecture and props, Added skeleton mesh and anim" Added architecture and props:     Literally just imported any finished models from our modeler and animator. I gave feedback on scale and now everything is perfect at normal scale. Added skeleton mesh and anim:     Part of the imported props was the skeleton mesh. I applied this to the archer character and began to give it a very simple animation blueprint. This AnimBp only has three states: Idle, Walking, or Shooting. The idle and walking worked just fine but the attacking gave me some problems.      I had not planned well for this so now I had to figure out a way to know if the BT is attacking. I did this with a simple boolean that the AnimBp would check. I figure there's a better way to do this in the future, but this is all I had for now. After I had the boolean I had to allow the clients to be able to see it as the host (server) is the only th...

Git Commit: 3d3319b "Fixed video settings, Added sharpshooter collateral hits, Changed AI MoveTo task" Fixed video settings:     This was just because I realized the reverting video settings function was not working. I found that if you wanted to revert the fullscreen mode or whatever in a packaged game it wouldn't work. This ended up being because you have to "Confirm video settings" in order to revert to them. By doing this I also figured out that these settings work differently in a packaged game versus the in engine simulation. Added sharpshooter collateral hits:     I thought of this when I was showing friends the progres of the game. that it would be cool if you could shoot through multiple enemies at once. I implemented this by just doing a multi hit line trace and damaging each enemy with decreasing damage with each layer. Changed AI MoveTo task:     I was wanting to do this for quite awhile but I finally got...

Git Commit: ???? "Added video settings, Changed Sharpshooter sprint" Added video settings:     This part was my major work of today. It wasn't difficult at all, just time consuming to do all of it. I expanded the settings menu to include: Fullscreen modes Resolution Frame-rate limiting V-sync     Luckily for me, all of these settings were already built into unreal so all I had to do was put them together. The only thing I really had to do by myself was the safety measure of popping up the "Confirm Resolution" menu with a timeout that will reset the settings if they aren't confirmed. This is what all that looks like:     I also had to set the values when the player first loads up the menu. Here's that:       The current menu states look like this: Very good design, I know     I did all this for three reasons. One: I wasn't assigned anything else. Two: It would have to be done at some point. ...

Git Commit: ???? "Added Sharpshooter stamina, Fixed PlagueDoctor targeting" Added Sharpshooter stamina:     I haven't been given much to do so I decided to go back to something that I left from earlier. I was told earlier that the sharpshooter would have limited stamina. Before today though stamina was unlimited. It wasn't too hard though, I used this video as a starting point and went from there. The video talked about the timers and that was all I was missing from before. The final system will let you run to empty and stop sprinting there or deplete the stamina some fraction and charge up back to full. This is the final stamina system:     After that was done I had to have some way of conveying this to the player. I went into the first person HUD and just repurposed the cleric's mana bar as a stamina bar for the sharpshooter. But instead of making a duplicate I just had the same bar display mana if it was the cleric and stamina if it was the sh...

Git Commit: d8cb5eb "Plague Doctor working prototype"     At long last... he is complete. (For now)     Today I made the changes I wanted to make from yesterday as well as added some extra features. The first and most important thing I did today was to make the enemy attack in a cycle of random attacks. I said yesterday that I thought I would do this with custom decorators but this turned out to be unnecessary. All I had to do was use what I already had along with conditional loops. Basically, the problem was that when the doctor needed to walk somewhere it would do It for a single tick then rerun the entire tree. I added a conditional loop to the MoveTo and that fixed it. Because I could now get him to go places reliably, I added MoveTo's to the other attacks to make sure he was within range and the attack would have a higher chance of hitting. (Picture 1)     Once I had a reliable system that would not lock up like yesterday I went on to testin...

Git Commit: None     Still nothing finished yet for the Plague Doctor. It is proving to be much harder than I originally thought it would be. The only difference between this enemy and the others I have worked on is that this one has multiple attacks. That one difference has taken up two full days of work and it has yet to be fully figured out.      From the last time I worked on it I changed a few things. The first being the random attack generator. I had this in a service before but now it is a task. I realized that a service is not what I would want because it ticks and that would change the attack type rapidly, confusing the hell out of the AI. But with that figured out, I went back to figuring out how to have the attacks cycle. This was the part I was stuck on for the majority of my time today.      The problem was, once an attack was chosen it would do that attack (as expected) but then it wouldn't return to the state where it would...

Git Commit: None     Today was cut short by other responsibilities but I was still able to start on a new enemy. This was the Plague Doctor for the medieval area. I wanted to do this enemy because it would be the first (and most likely easiest) enemy that would have multiple attacks. According to the current design doc these were an AOE and cone attack that would poison a player for some amount of time and a slash that would do immediate damage.      The base of the enemy was pretty simple with most of the same code being used from the basic melee enemy until the part where I had to implement the multiple attacks. The idea was that when the enemy goes to attack it would execute a service that would set a random attack from an enum. This randomness was supposed to be weighted according to the design doc with some attacks being more common than others. I decided that this feature can wait until I know it is possible for the rest of the system to work correc...

Git Commit: b6e0596 "Changed first person UI, Fixed enemy pathfinding" Changed first person UI:     The simple change of today had to do with the first person UI, namely the health bar. Our UI artist finished the first edition of the health bar and wanted to see how it looked in-game as to adjust anything necessary before he went on to make more. The first thing we noticed was that the .png file that was exported had really bug padding around the edges that made the scale wonky which really screwed with positioning and getting clean values. Right now, it looks like this: Fixed enemy pathfinding:     When I loaded up the project today I was prepared to work on the archer enemy. Before starting that though I wanted to make sure everything else worked up to that point. I added a standard melee character to the level to make sure they play nice. They did, but I noticed that the melee character would jump up to the statue's pedestal, and, given enough t...

Git Commit: cd7dff7 "Fixed enemy health bar,  Added statue enemy" Fixed enemy health bar:     Just fixed this small issue I noticed when I was testing the new statue enemy. The issue being that the health bars above enemies heads would appear and disappear at the wrong time. I intended for the health bar to appear when the player is looking more or less directly at the enemy. What actually happened was the health bar would appear from the far left of the screen to the correct spot on the right. the problem was that I was just using the wrong conversion node from screen to world. I found the correct node from this thread.  Added statue enemy:     From the design doc I was told to make an enemy that: "These do not move on patrols they stay in one location and wait till attack come within range". Additionally during a meeting I was told that they would be stationary on a pedestal until they jump on and off to attack the player. This made me learn...

Git Commit: afb9a0e "Added ScarabSwarm enemy" Added ScarabSwarm enemy:     What I thought would be a simple enemy turned out to take up all my time today. The basic idea was that it would be an enemy that would just be a hitbox that once overlapped would do damage every interval of time. That part wasn't so bad, it was the movement that took the longest.     I started off thinking that the actor itself would be static and the actual swarm would be a particle effect that would give the illusion of movement. Turns out I was wrong and the designers said that the entire thing "moves in erratic patterns". So I had to convert the actor to a pawn then add an AI controller, blackboard, behavior tree, etc. then tell it to pick a random point close to around it and move there. The problem was, it didn't move. I struggled for a while until I noticed that I had forgotten the most important thing that I had always taken for granted when making human type enemi...

Git Commit: 976a232 "Added first person HUD" Added first person HUD:     Still no updates on the main design doc, so today I have decided to finally tackle UI. For the main reason that it was getting really annoying having to do print strings and 'watch this value' to see if my code was working the way I thought it was.     I started out with health. Simple enough as it was consistent across all players. My goal was to have a single widget blueprint that would work across all players regardless of their class. Just hiding and showing certain elements that were relevant to the class that the player is. After health though things got a whole lot more complicated. I wanted to have: Mana bar (cleric only) Ammo (sharpshooter only) Main attack cooldown indicator (different timings per class) Special attack cooldown (same story) Revive indicator (while writing this I realized I forgot to add this one) The mana bar was similar to the health bar just clas...

Git Commit: 6646905 "Added revive system, Minor fixes and changes" Added revive system:      Building off yesterday I figured out what was the problem. All of the characters do  have owning connections but only the server can access them. Once I figured that out it seemed like a no-brainer but I eventually figured out how to make the heal aura work again. The only change I had to make was to have the actual actor be spawned from the server. This is what that looks like: After that I got on to doing the thing I initially set out to do; the revive system. The hardest part of this was, without a doubt, figuring out the architecture. Trying to wrap my head around what scope was needed for everything, where code will be executed, and at the same time trying to make it as general as possible was quite difficult. I started out thinking that I wanted to do it with an interface that would do the same thing on both the reviver and the revivee player. I eventuall...

Git Commit: None     Today I made a realization; I still don't understand some fundamental ideas of networking in UE4.     I started out the day with little direction as the designers haven't updated any of the design documents or given me a task to work on after our meeting on monday. This turned out to be a blessing in disguise as it meant I found a huge bug with the core of the game. So since I don't necessarily have a deadline to meet I decided I would try to implement the reviving system I thought about yesterday. I started this out by making sure the healing from the cleric worked so that later I could use that as a reliable point of gaining health in case something went wrong with the new system.      But as I tested the heal aura I found it wasn't doing anything. I checked the blueprint and didn't see any logical error so I went to the output log to dig up errors or warnings. And indeed I found some. The first one that directly told ...

Git Commit: 8e3012c "Fixed players damaging enemies, Added the ability for enemies to damage players, Added SmoothSync to players and enemies" Fixed players damaging enemies:     This was a very odd issue that I found today. Yesterday I had only tested the damage system in the DungeonTest level which I did not realize had a different gamemode and therefore an entirely different server-client system. So today when I tested in the Island level the damage system didn't work. I was able to eventually figure out that the problem had to do with the client trying to set the enemy's health which should only be done by the server. To remedy this I added a custom event that would execute essentially the same code just on the server instead.  Added the ability for enemies to damage players:     Just a simple recreation of what we did for the creative jam game. When the enemy is close to the player it will wait an amount of time, attack with a given damage ...