Added an RRV screen showing what items can be acquired from the Compact Mob Farm.
Shows the entity, the items it drops and the required tier upgrade level. Hoping this also helps with confusion on how stuff works
Added an RRV screen showing what items can be acquired from the Compact Mob Farm.
Shows the entity, the items it drops and the required tier upgrade level. Hoping this also helps with confusion on how stuff works
Added a progress bar to the Compact Mob Farm GUI that hopefully makes it much easier to understand when it’ll damage the entity and it doesn’t look like it’s just doing nothing lol, especially with the 30 second default time
Waow that’s quite a bit of stardust 😅 Quite happy with this project and the ratings it got on the first ship :D (except the one that accused my devlogs of being written by ai… My writing is not that bad, right?
)
I noticed in the shipwrights’ review video that how upgrades work is not very obvious, they were trying to put a tier upgrade into a soul fan, which only accepts width, height and depth upgrades. It was mentioned in ratings that stuff is a bit hard to figure out in the beginning as well.
So to fix that, I added some silhouettes of the supported upgrades shown in the slots and also a hover tooltip saying what’s supported.
Oh and i’ve also added 26.3 support now
School starting with september really hit my productivity on this mod like a truck 
I am realizing the deadline for stardance is getting real close, so I think this is probably the point where I’ll have to ship it. There are still some TODOs in the code and bugs that need to be squashed and porting to 26.3 (which released like this week, thanks mojang
) and so I haven’t been comfortable marking the latest release as 1.0.0 yet or making announcements elsewhere, but the mod is now perfectly usable and I like where it is.
Changes since last devblog:
26.2 wasn’t actually as painful as I thought it’d be, which was a nice surprise. Adding advancements was a bit more of a pain with 26.2 support though as I needed custom criteria for some of the advancements (one for putting a soul gem in the compact mob farm and another for using a soul gem on a spawner) and Mojang, in their infinite wisdom, decided to move a bunch of advancement stuff around in 26.2.
Getting closer and closer to the actual release… could this be the last devblog before that?
Still not quite 100% about if the required kills for each mob are balanced, but it’s probably good enough. Definitely better than everything being 10 kills tho. Felt very off-balance when withers, which are hard to spawn and hard to kill, and pigs, which are just everywhere, required the same amount of kills to capture.
Notable changes:
Running in a dedicated server would crash because there are quite a few places where I referenced client-only classes in places that the jvm will load on the dedicated server as well. Mostly in the different block entity classes, where I had a clientTick method that had ClientLevel as one of it’s parameters. Even though that method would never get called on the dedicated server, java still wants it to exist when loading the class because it’s part of a method’s definition.
My current solution is very very awful… I moved those things into methods in the client initializer and… well… yeah… it just looks and is awful: https://git.offsetmonkey538.top/OffsetMods538/Ultimate-Mob-Farms/src/commit/0.0.7/modded/src/main/java/top/offsetmonkey538/ultimatemobfarms/modded/UltimateMobFarmsModdedClient.java#L65
The real, foolproof solution would be using the split sources feature of loom, which creates a second source set and makes the main source set unable to access client classes. Only problem is that it’d be a bit annoying to fully get set up in my multiloader-version gradle system and would most definitely need many changes on MonkeyLib538 as well and I just am not up for doing that atm.
Fixed compact mob farm inventory not working correctly and being incompatible with other mods’ item movement stuff. The whole point of the transfer API on both fabric and neoforge is the transaction system, where each change must store a snapshot that can be returned to if needed; Turns out ignoring the TransactionContext and not saving any snapshots is a bad idea 😅 I’ve tested and it’s compatible with create fly (see screenshot where a compact mob farm with a blaze is being used to feed a blaze burner for a steam engine) and most probably other mods as well.
Added Reliable Recipe Viewer integration that explains soul capturing
Bunch of refactoring, bug fixes, balancing, etc. Release notes can be seen on the forgejo release or on modrinth, I’m not gonna copy paste em here (this is version 0.0.5 right now)
It is now actually reviewed on Modrinth, took two weeks for that to happen (Gosh I remember when review times were like less than 24 hours
). I’m still keeping it unlisted until I get a 1.0.0 release out, but if anyone wants to check it out and share feedback or anything, then it’s right over here: https://modrinth.com/mod/ultimate-mob-farms
Before a v1, I still need better rrv and item descriptions integration, some more balancing changes, advancements, quite a few bugfixes and probably some other stuff I’m forgetting at the moment.
Made soul fans placed at different locations use a different starting frame offset, so they don’t all look the same (see before and after videos).
The way this is implemented is a bit interesting…
There’s no easy “use this model but offset any texture’s animation by 1 frame” as afaik the frame index is global per texture. What Minecraft does allow is each block state variant actually being an array of variants that the game will choose between using afaik a hash derived from the block position. This is used in vanilla for things like stone and deepslate, which have mirrored and rotated variants so caves look less repetitive or whatever.
As each texture has a global frame index, I am duplicating the soul fan texture and just changing the order of the frames in the accompanying animation metadata file and creating separate models which point to the different textures and then those models are used as variants in the blockstate.
What’s even more fun is I am using datagen for duplicating the texture and creating the animation metadata files, which definitely isn’t concerning at all. This also required me to make a pull request to fabric api because the FabricCodecDataProvider assumes all files it generates are .json files, but animation metadata is .mcmeta. (The PR isn’t yet merged so I resorted to…. injecting a mixin into fabric api, definitely not cursed at all I swear)
Added particle effects for soul fan. Figuring out where to spawn em was quite fun actually, though I did have to add a custom particle (reuses the vanilla “small gust” texture) because turns out most particles won’t obey the velocity you give them and just do their own thing, which was a bit annoying.
Added range upgrades for the fan and collector blocks. It was a bit challenging figuring out actually extending the bounding box in the correct directions based on the block is facing, but I did get it working
Implemented functionality of the soul fan, getting the strength to scale based on distance was a bit of a pain cause i needed to get the position of the center of the face that the fan’s pointing in and only calculate the distance to the entity based on the correct axis cause otherwise giving it a larger range would make the farther to the side entities get pushed less and uhh yeah, took quite a bit of experimenting with stuff to get it to work correctly
Spent the last hour and a half faffing about with my loader abstraction in monkeylib (stardance only recorded an hour cause that’s technically a separate project) for creative tabs. All that just cause I really wanted the RRV stack group for the filled soul gems to show up next to the empty soul gem, instead of at the end of the list where it did previously
Spent a bunch of time on cleaning up some code and abstracting some things out of collector block classes to make it easier to add functionality of the soul fan later on, speaking of which! the soul fan is a block that exists in game now
Spent some time thinking about what I’m gonna do for the block that moves mobs around. I think I’ve decided on going with a basic “fan” type block and now started working on what it could look like. I’m imagining the souls kinda like escaping from the soul sand and also particles spawning and moving in the direction it’s pushing.
As always I’m pretty bad at drawing textures, so the beginning here looks a bit crude, but it’s definitely a start.
It’d be really nice if aseprite had a wakatime plugin…
Added the experience collector block, which obviously picks up xp orbs and turns em into bottles of enchanting. Kind of got so focused I forgot about making more detailed devlogs 😅
Made the soul spikes actually deal extra fall damage when fell on. Current fall damage multiplier is at 5, not sure if that is too much or too little. The damage is done by a FakePlayer, so the mob will also drop loot that only drops from player kills, including experience orbs. (Seems like I need to add an XP Collector block in the future to deal with those). Oh and also did some boilerplate-ish things to soul spikes like adding a recipe, translation, etc
Turns out minecraft decides that the block you fall on is the one that’s 0.2 blocks below you…. The model of the soul spikes (which i don’t really wanna change anymore) has the base at 3 pixels high, which makes it… 0.1875 block high 😮💨
Nothing that can’t be solved with a mixin to Entity#getOnPosLegacy that changes the BlockPos it returns if the the block above the “correct” block is soul spikes
Made an okay-ish texture for the soul spikes block (many of em at once looks kinda noisy but oh well, it’ll do. Hopefully it’ll be locked up in a farm killing chamber room anyway).
Had to change the UV mapping again because for whatever reasons, minecraft decides to render a sliver of the neighboring pixel as well which looks super bad with two very contrasty colors next to each other. Solution was to move the blue parts to the bottom of the texture where there was room to have it separated by a row of blank pixels.
Finished UV mapping the spikes…. this literally took like over an hour haha, sadly time spent in blockbench doesn’t count as it doesn’t have a plugin for wakatime or whatever :/ (the 43min this devlog logs was actually setting up oxipng to automatically run on every commit to optimize the png files)
Next i guess will be to make an actual texture for it, yay. Here’s the super very interesting texture that currently just has all the different mapped sides in different colors