1.21.11 Snapshots

[devout vault]

yup yup

[rare lynx]

I'll have to deal with that

[devout vault]

we also need to change Neoform

[limpid coral]

But good news, you have a whole update cycle to make the changes

[rare lynx]

need to use jst to rewrite the sources to the new mappings merged into the old classes

it's automatable, just painful

[limpid coral]

how many patches are there? is it worth automating?

[hallow chasm]

I'm waiting for a vineflower release before releasing neoform for 25w44a because I am not patching more than 200 annotations

[rare lynx]

yes lol, there are hundreds

[limpid coral]

Oh well yeah go ham lol

[gloomy axle]

yeah, ok, locals are confirmed (see #789950127774105602 message), which will make many parts sooo much less pain

looking at stuff like FogRenderer#setupColor

[rare lynx]

https://mk.asie.pl/notes/aefqgch0mn

asie (@asie)

www.minecraft.net/en-us/article/removing-obfuscation-in-java-edition

Congratulations to who I assume are Mojang's programmers for finally pushing this through. When I started work on alternate community-sourced Minecraft mappings in 2016, I did not imagine that it would start a domino chain culminating in Mojang's legal team finally conceding that the charade of binary obfuscation for one of the most reverse engineered games in the world is no longer beneficial to maintain.

Now let's hope Fabric finally lets Yarn go independent, as they should have done long ago.

mk.asie.pl · Oct 29, 2025, 3:46 PM

[limber osprey]

idk where that is linking but it ain't accessible for me

[gloomy axle]

did you try clicking on it

[limber osprey]

now it works

[gloomy axle]

it's a message in fabricord

[limber osprey]

it turned into unknown and now it works

[gloomy axle]

[gloomy axle]

yeah, that's what happens when you're not in the server it links to

[split oak]

day one rendering changes mapped yay

[limber osprey]

no, I mean it was literally giving me an error message. I did try it like normal first.

it doesn't matter though, it works now! 😅

[gloomy axle]

discord jank ¯_(ツ)_/¯

[dreamy frost]

There are 790 patch files in Neo itself. Not sure how many NeoForm patches there are

[gritty moat]

🥳

[limpid coral]

Great time to move to mixins instead of patches

[split oak]

how much will this speed up gradle refresh

[limber osprey]

not unless we somehow get source visible mixins

[limpid coral]

that's theoretically possible right?

[spring junco]

OBFUSCATION IS DEAD

[limpid coral]

just a bit of a pain

[dreamy frost]

And even if we had source-visible mixins, there are some things you simply cannot do with mixins while patches can

[limber osprey]

there was talk about moving patches to mixins when they were working on runtime mojmaps, I'm pretty sure, so there has been a bit of thought that has gone into this already

[gloomy axle]

yeah, there was at least talk about doing that

[lone cairn]

https://tenor.com/view/atg-capa-stucapa-studiocapa-pandu-gif-24944581

^ mojang right now

[fallow dagger]

That's no use

[devout vault]

this actually has the incredible benefit of making our patches super readable lol

[fallow dagger]

Yes that's a very good point

[devout vault]

(and reducing rejects, once we're over the initial hump)

[next wolf]

im late to everything... whats this about mojang no longer obf mc? thats crazy

are they like releasing sources (would love to have comments/javadocs in vanilla code mojang please)
or just the compiled jar skipping the obf proguard step they currently do

[lethal ice]

starting with next snapshot, they will provide alternative files that will be deobfuscated, so we can observe the differences
and starting with the next snapshot cycle after the current drop, the main files will be deobfuscated

no sources

only jars with real class/method/field names and LVTs

[gloomy axle]

no docs

[echo pendant]

no capes!
-# wait, wrong reference

[gloomy axle]

I mean, that's also not wrong

when deobf cape

[lethal ice]

deobf capes implies the current capes are obfuscated

which means they contain some meaningful information and aren't just pixel art

[gloomy axle]

I meant capes because java edition is now no longer obfuscated

[worthy ember]

Well, we can't add more of those either.

[split oak]

capes for everyone who had to deal with obfuscated jars and stuck around

[next wolf]

how different will vanilla code look now that we wont have to deobf everything
will we see all the magic numbers and stuff turn into static final references and such?

[split oak]

no

those are inlined

the only real difference is the names of things

local variables and parameter names

its still the compiled code, not the source files (for now)

[next wolf]

local var names will be real sweet, could get param names via parchment but i guess aside from javadocs thats now dead

[split oak]

yes parchment and yarn are reportedly EOL

[next wolf]

hehe yarn that should have died when offical mappings released

also i hope parchment doesnt full die

i like the javadocs it provides

[split oak]

well

I hope parchment dies

in the way forge did

replaced

I want the yarn people and the parchment people to merge into a big javadoc group

and make the best javadocs

[next wolf]

i mean if mojang drops some kind javadoc jar that we can merge during setup
that would be great, having mojang comments and such would be real nice

would love to see what kind of cursed comments mojang makes

[split oak]

I suppose thatd be similar to release deobf mappings

its not a full sources jar its just the docs

[gloomy axle]

Parchment won't die

It will just be docs only

most likely

[devout vault]

• unpick + annotations

hopefully

maybe

[split oak]

annotations?

I thought we have those

did they not ship annotations before

unpick will be nice tho

[gloomy axle]

If they're only Retention(SOURCE), then no

[gloomy axle]

Oh, is unpick part of parchment?

[split oak]

I thought unpick was fabric

[next wolf]

built by and used by fabric, but that doesnt mean we cant use it

[turbid dune]

There was already talks of including the unpick data in parchment, the news has just moved up the timeline on that

[split oak]

now that they dont have to worry about silly things like the reason they began the project

[onyx rivet]

can we deal with the lambda problem please

this feels like a good time 😔

i don't know how well it's dealt with right now

[echo pendant]

the lambda problem?

[onyx rivet]

patches offset the name of lambdas, making the same injection wrong between fabric and neoforge

such as lambda$addMainPass$8 having different index on Neo because all of the LevelRenderer patches

however, I haven't checked in a bit and I don't know if this has somehow been fixed

[hidden forum]

My understanding is that it's kinda impossible to fix. It'd require a lot of manual work to to write mappings for fixing it

[onyx rivet]

yeah, makes sense

or just forbid lambdas in patches

[worthy ember]

That would also work, I guess. Just define functions?

[devout vault]

no that's not the root cause

Vineflower also decompiles in a different order from bytecode order

you get reordered lambdas just because of that alone

patches just add to it

[elfin kelp]

doesn't neo not change classes that are unpatched in prod?

[shrewd violet]

We patch all classes these days

[devout vault]

yes, since if we did not, you'd have a delta between neo moddev environments and prod for mixins

[echo pendant]

looks at https://github.com/neoforged/NeoForge/issues/2117

[devout vault]

yup

[hardy perch]

Ok im reading this thread

So in a nutshell we get the jars unobfuscated but there's no documentation so we gotta figure it out

[jade plume]

we get the jars unobfuscated

👌

we do get param names now which should be a huge help

[echo pendant]

what we'll getting next snapshot compared to this snapshot is all the parameter names and local variable names

[hardy perch]

The 1.12.11 modding scene iis going to be crazy

however...

[gloomy axle]

and especially locals

[jade plume]

lotta times I'd override some method and have no idea what the param names were so I just made shit up

[hardy perch]

we're gonna stay in 1.21.1 until the new popular version drops

1.22.1 is gonna be the next one, i just know it

[gloomy axle]

we already had params via parchment, but not locals, and having those is gonna make stuff nicer

[gloomy axle]

but yeah, parchment obviously didn't cover everything

jellybiscuits

[lofty wedge]

Egg on your face when it turns out that's also what Mojang called it :P

[gloomy axle]

maybe they did ¯_(ツ)_/¯

[hardy perch]

still i wonder with these new changes is neoforge is going to be rewritten again? Kinda like what happened post 1.20.1 after branching off forge

[gloomy axle]

we'll know next week

[hardy perch]

yeah

[split oak]

mojang we love u please make this new era of unobfuscation 1.22 so I can stop being pressured to lag behind 🙏

[plush mortar]

you can also make your own decision to use whatever version you want

[jade plume]

it's going in 1.21.11 so 1.22 will have unobfuscation, yes

but yeah you can just ignore the pressure

when people ask me to backport the new stuff I did in the minor versions to 1.21.1 I just throw bricks at them

reason I'm doing new stuff in the minor versions is so that when we finally get 1.22, all of my stuff will be nice and settled down and stable too and I won't have to worry about breaking my own apis or save compat mid-version

like four out of five dentists in user-support are crashing because some mod's dependency made a breaking change in the middle of 1.21.1

[hybrid perch]

who is pressuring who ?

[split oak]

my players r pressuring me

cuz 1.21.1 has create and all the big mods that have decided theyre staying on 1.21.1 for the forseeable future

[plush mortar]

well, its your project, do with it what you want

[split oak]

no point if there's nobody to play it

i did eventually decide to port

but it is unfortunate that I lose that mod availability

[fallow dagger]

well that's on you for deciding to port 🤷

it's clear that 1.21.2+ versions are not gonna stabilize atm

I will start porting my bigger mod (MI) but I expect very few players on new versions for now

so I am first taking the time to add a few more features to 1.21.1

[split oak]

there are fewer players on newer versions because there are fewer mods on newer versions

its a cycle

[fallow dagger]

given the version churn it's obvious that it just cannot be stable

[split oak]

when will it be stable?

the new drops system is very successful i doubt theyll stop it

there will always be a new update every 3 months

[hybrid perch]

drops are staying. the versioning will change

[split oak]

if its not rendering they have plenty of other things to refactor

[fallow dagger]

whenever 1.22 gets released and people stick to it (instead of moving to 1.22.1+/1.22.2+), is my best guess

[plush mortar]

it'll probably be more rendering too

[split oak]

the new update cycle does not allow for "stable" releases

we dont even know if 1.22 is a thing

ur looking forward to a hypothetical

[devout vault]

what does stable mean anyway

if there's enough incentive in terms of accumulated features, ppl will port

[split oak]

1.22 could be 3 years out or not coming at all

[fallow dagger]

stable = enough "big modders" agree on sticking to a version 😛

[devout vault]

i.e. java 25 is an incentive

proper parameters and local variables throughout the entire vanilla codebase?

[hybrid perch]

problem with that approach is some will never move

[devout vault]

DEFINITELY an incentive

😄

[fallow dagger]

that means no stable version for 3 years

[split oak]

I think (hope and pray) if this release ends up being Java 25 + deobf jars, it will become the next major modding version

[fallow dagger]

the problem is that 1.21.11 just doesn't sound catchy

1.22 is a different story 😄

[hybrid perch]

is ae2 updated? generally it's one of the mods that starts porting saga

[lethal ice]

there's a certain group of modders that are refusing to participate in the "drops" system altogether and are waiting for the next hiccup to cause minecraft updates to have a gap of at least 1 year

[split oak]

cuz they also said the rendering stuff doesnt have much left to go

[devout vault]

it is

we dont release though

[plush mortar]

source?

[devout vault]

it was also updated to .5, .6, .8

[split oak]

their article

hold on lemme get it

[hybrid perch]

oh why so?

[devout vault]

main is still 1.21.1 for AE2

ppl PR to 1.21.1

[split oak]

This work is mostly complete for entities, block entities, particles and the UI. Next, we plan to tackle chunks and small leftovers.

[hybrid perch]

oh don't want additional burden of maintaining 2 branches ig

[split oak]

chunks and small leftovers are what's left according to this

[plush mortar]

can you link the article?

[split oak]

https://www.minecraft.net/en-us/article/the-road-to-vibrant-visuals-on-java

its when they moved from the VV discord

[plush mortar]

afaict, that specifically talking about the state extraction for a render thread

not renderer work in general

[devout vault]

yup that's also how i took that

given that currently that work doesnt even have much benefit I think given the submission still happens on the same thread

[plush mortar]

it can help batch up the submit better

[devout vault]

true, yes

but in terms of parallelism

i gathered that they'd move submission to another thread?

once it's actually all using that approach (extract/submit)

[plush mortar]

doing that means moving OpenGL (assuming they stick with GL) to another thread

[devout vault]

or the other way around 😄

[split oak]

so its gonna be like 3 years before modders decide on a stable version?

hoping wishing praying that 1.21.12 (inevitable hotfix) has enough appeal that a good amount of people move

[plush mortar]

for mac reasons, technically yes other way around

but still GL getting its own thread

[devout vault]

wait mac was only affected by first thread message pumping

not rendering

(I thought)

yes definitely, given that we render from different threads too

[plush mortar]

thats correct, so really its whatever thread processes the event queue that matters and has to stay as the first thread

idk how timing with the two threads would be handled there though, so it could go either way really

[solar stone]

https://www.minecraft.net/en-us/article/removing-obfuscation-in-java-edition

So does this mean we get eventually get access to sources jar too?

[twin viper]

likely no

[solar stone]

Ah rip

[twin viper]

theres a big difference between 'heres the compiled code' and 'heres the source code', especially for saying that you actively try make sure people dont steal it

[solar stone]

I do wonder if this changes everything tho, for future server software like Bukkit...

[limpid swan]

Method-coded param names and inlined magic constants are still going to be an annoyance

[trail mesa]

I think Paper has already been running with mojmap in prod, or at least has done all the enablement work for it, so probably not a huge deal for them

[solar stone]

Yeah

[limpid swan]

I think that'd only happen if MS and Mojang completely write off JE as a substantial revenue source

[solar stone]

Wdym?

[limpid swan]

As long as selling licenses to JE still makes money I don't think legal and/or corporate will let them publish sources

[stark crane]

Crash logs will be easier yay

[random gust]

So if I understand, its not as huge as I thought ? I thought this change would make it so much easier for the modding community and allow to literally just diff each versions to see what changed & if it affects their code

[gloomy axle]

We already had that

We had half the information already ever since mojmaps exist

Now we have the other half

[sweet prism]

100%!

[lone cairn]

it makes some parts easier
it doesn't change a whole lot, tbh
its nice that they do this though, but I feel like this has been in the works for at least a year, as many mojang employees working on JE were modders in the past

[vestal tundra]

would it take a year to remove a couple lines of gradle related to obfuscation ?

[twin viper]

then get it through the legal department

[vestal tundra]

indeed 😅

[lone cairn]

would be funny if they used ANt to build all of it

[devout vault]

Makefiles

[lethal ice]

I feel like it's more like 15 years to get it through legal

[deep iris]

Microsoft hasn't even owned Minecraft for that long

[lethal ice]

however long ith as been :P

[deep iris]

There's a good chance it's been in the works for a decent time since they provided mojmap originally though

[deep iris]

September 2014, over a decade now

[lethal ice]

10 years then :P

[sweet prism]

Fascinating

[fallow dagger]

MDG-Fabric when?

[devout vault]

(I was thinking the same thing lol)

One Cute Elephant to rule them all

[viral nacelle]

https://xkcd.com/927/

xkcd: Standards

Fortunately, the charging one has been solved now that we've all standardized on mini-USB. Or is it micro-USB? Shit.

[fallow dagger]

I mean, don't we all obviously agree that MDG is the nicest Gradle plugin? 😁

[rare lynx]

totally unbiased opinion

[devout vault]

How can you not with such a cute mascot.

[dreamy frost]

I would love that, simply because Loom has been pissing me off every time I've updated AtlasViewer to a new MC version. Update FAPI? Sorry, need new loader. Update loader? Sorry, need new Loom. Update Loom? Sorry, need new Gradle. And, oh, btw, can't update Gradle without reverting absolutely fucking everything...

[fallow dagger]

https://tenor.com/view/baby-elephant-running-away-cute-gif-24680637

[fallow dagger]

We could really consider it

The biggest problem atm would be the recomp cycle messing up the lambda names

[rare lynx]

damn, emotional manipulation! I approve, !

[fallow dagger]

But otherwise running fabric loader shouldn't be too difficult without obf

[rare lynx]

it should be possible to remap the lambdas post comp, not sure if luke has already managed to finish doing it, but technically it should be possible to figure out the correct name of the lambda by cross-referencing what method is calling it

[edgy river]

I always update gradle manually via the distribution url

Much easier

[twin viper]

ugh i get this too when i update my one multiloader mod i maintain.
and when i complain in the fabric discord they ensure me that everything is backwards compatible.. (this was not a one off occurence)
how is that even relevant

[obtuse igloo]

Uhh... no actually you can't, if there's recomp involved. Because the descriptors also shift

It's not just the names

(i tried this to be clear 😛)

[forest stag]

Old Fabric API and old loader do not work on newer versions. That is like expecting old Neo to work on a newer version.

Newer loader requiring new Loom and new Loom requiring new Gradle is a reasonable point

[split oak]

fabric loader is supposed to be version independent

as is MDG, idk about loom

[forest stag]

It is, but old versions still don't work on new versions

[twin viper]

this is what everyone says but what they mean to say is that its backwards compatible

[forest stag]

That only means new versions work on old versions

[twin viper]

and no one seems to understand the difference

[split oak]

oh

backwards compatible is very different

is MDG backwards compatible or version independent?

[twin viper]

MDG is version independent provided mojang doesnt change some stuff, like suddently deciding to remove obfuscation

[split oak]

makes sense

[twin viper]

(in case anyone is wondering why - https://github.com/FabricMC/fabric-loader/blob/9efe3fdef26f171125dc20e090fd9e5dd19fc615/minecraft/src/main/java/net/fabricmc/loader/impl/game/minecraft/McVersionLookup.java#L358-L447)

(thats probably not the only reason, but its an interesting piece of code regardless)

[split oak]

ah right

loader has to update every snapshot cycle

[untold coral]

But split sources! :p.

[devout vault]

Oh we'd love to

Have to do it in Neo first, though

Or rather, Neo is split

We have to publish it split, though

[fallow dagger]

Not really no

[devout vault]

yeah we do

[fallow dagger]

No 😛

We can split in NFRT

[devout vault]

no we can't

we don't actually want to process neoforge-universal jar

we just wanna put it onto the classpath using gradle

[fallow dagger]

It's not particularly slow to process if it's just splitting two jars up

[devout vault]

I do not want to process the NF jar at all

(and yes, yes, you are right, we are right now, but that's really due to legacy)

[fallow dagger]

It also depends if we want to have NFRT just pass it through unchanged or if we also want to add special logic to MDG

[devout vault]

pass through what?

[fallow dagger]

NFRT needs the universal jar to compile Minecraft

[devout vault]

yes, as a classpath item, but merging it into a merged jar is shit

[fallow dagger]

Ah right we are still doing that

Heh yeah that does suck

[devout vault]

I also apologize for completely derailing this thread lol

[next wolf]

https://www.youtube.com/watch?v=YNvWGMO2LKs

0:35
huh game rule command does not accept the namespace
that feels wrong, what if 2 game rules have the same path but diff namespace

[twin viper]

neo patch?

[next wolf]

mayhaps

if not using it already should be just swapping the arg out to use the RL arg type

[twin viper]

ideally it would support the short form w/o namespace too if it doesnt conflict

[next wolf]

thats built into the suggestion provider

someone recently PRd it so RLs match substr from both parts + full

and i think vanilla yoinked it too iirc

[latent ivy]

The thing with the game rule command is that because you can either have integer or boolean game rules, you can't know which argument type to expect if you just use the resource argument. So game rules are currently just literals with the corresponding type argument coming after. Literals do not support partial matches (unless what you write starts the same way), so if namespaces were allowed, you couldn't type /gamerule advance_time to get autocomplete suggestions because the argument is a literal minecraft:advance_time.

[twin viper]

can you not resolve the previous dynamic (non-literal) parameter to provide the suggestions for later ones

[solar stone]

Brigadier limitation 🤔

Each suggestion is an individual command (sub command)

[latent ivy]

It's that the argument that you use for setting the game rule value is fixed when you register the command. But which argument it is (int or boolean) depends on which game rule you select while writing the command at which point the argument types are already fixed.

[solar stone]

Yea

[twin viper]

Ah, because you specify the argument type in the command registration

[latent ivy]

Yes

[solar stone]

If not done already, prob could do a search bar for the game rule menu

Atleast

[twin viper]

the hack then would be double register each gamerule, once namespaced and once just as the path (if no conflicts)

[latent ivy]

And then your suggestions go
agamerule
gamerule1
gamerule2
minecraft:agamerule
minecraft:gamerule1
minecraft:gamerule2
minecraft:zgamerule
zgamerule

Ain't that fun

[solar stone]

The real solution is to design brigadier to use codecs! XD

[twin viper]

youd need to use a different sort, assuming brigadier allows that

[solar stone]

A key/value sub cmd system

[twin viper]

ah but they are literals

yeah that is a problem

entire command is a greedy string that gets manually parsed: problem solved

[solar stone]

xD

[twin viper]

dispatch commands

[solar stone]

Hmm

[edgy river]

I worked at that for a bit, but it's definitely awkward and requires jank mixin on fabric where I was doing it

[thorn hatch]

why not PR it to brigadier?

[edgy river]

Because it was awkward and jank, and I only really did it as a proof of concept gist for someone on the fabricord. The code existed for about 20 hours on my computer

It was just, it works? Cool. Throw away

[next wolf]

tomorrow y’all know what that means first unobf’d jar

Oh and the shenanigans that comes with a new snapshot but that should be expected

[solemn furnace]

Java 25 tomorrow, maybe?

[twin viper]

nah java 26 with value preview obviously

[lethal ice]

"we have decided to go back to java 5"

[next wolf]

Oh gods no

[solemn furnace]

Is it Java 5 or Java 1.5?

[lethal ice]

I don't know when their public naming changed from 1.x to x

https://www.oracle.com/java/technologies/java-archive-javase-v14-downloads.html
https://www.oracle.com/es/java/technologies/java-archive-javase5-downloads.html

1.4 -> 5

the internal version remained 1.x until 1.8, then java 9 was 9.0

[devout vault]

"We migrated everything to Kotlin"

[lone cairn]

pffff

at least could have done scala

[lime oyster]

You know, I like how they shuffled around gamerules and its still only really limited to ints and bools

At least if you want to use other features

[gleaming ore]

Who's to say that isn't prep work?

[lime oyster]

Maybe? But I'm not confident on that

They merged the type and key into one unit, which makes me think it's just some simplification work

[atomic fern]

Scala was fun. What was not fun for me was unwinding 15 yr old ChickenBones scala magic

also, iirc, scala support in forge was originally ChickenBones's suggestion. So blame him for scala.Arrays :) (for those who know what that pain was)

[lone cairn]

Scorge was the second iteration, AFAIK
made by, back then Illy

[atomic fern]

yeah, i'm meaning when it was included in forge, circa 1.4/1.5

maybe 1.6, not sure on the timeline

famously, uses a lot (or used to) of bandwidth for forge maven

[lone cairn]

yes
eh, I might revive it at some point for forge (and perhaps neo, who knows)
or just a general gradle plugin, dunno

[atomic fern]

mm, there was another language loader for it

[viral nacelle]

scalable cats force right?

[atomic fern]

I don't remember the name

thats the one

but, scala is very academic, its not particularly friendly

nor does it promote good memory/gc patterns

[lone cairn]

can still be fun 😛

[atomic fern]

oh absolutely, traits were based af

i still also miss implicit functions

i remember a lot of sensitive code in fmp explicitly using java collections everywhere to avoid the immutable collections lib

[gleaming ore]

still waiting for someone to make a C# lang loader

Don't care if it gets transpiled 4 times, I want one lol

[viral nacelle]

wasn't there somebody working on a WASM loader

[atomic fern]

(and if you want them at runtime for java, with java, compatible with scala, https://github.com/TheCBProject/TraitMixinCompiler )

[viral nacelle]

that'd basically allow everything

[thorn hatch]

I mean a wrapper that wraps neoforge + minecraft via ikvm in a c# project could provide a bridge

[lime oyster]

Here's your primer: https://github.com/neoforged/.github/pull/37/commits/a7ec132ade6b983aa6aa1de91783818c596e0b3e

[next wolf]

Heh the snapshots not dropped yet, champ must be on the ball

[lone cairn]

yes, but that is java IN C#
not C# in java 😛

[lime oyster]

Some sampler additions, block/terrain split in pipelines, game rule changes, slot sources, and jspecify

[next wolf]

Unless he’s only just now posting last weeks?

[lone cairn]

orrrrr

he works for mojang, and gives us secret insights

[next wolf]

If that’s the case tell us champ, fucked are we with the unobf’d jar tomorrow xD how broken are our tools gonna be

[jade plume]

RenderPipelines that were used by both a standalone block and the terrain has been split into separate pipelines: one with prefix _BLOCK and _TERRAIN, respectively. This includes the solid, cutout, translucent, and tripwire pipelines. No block variant exists for the translucent pipeline.
how do I render translucent blocks in renderers then

[gleaming ore]

Comm if you figure that out lemme know, Gander 21.10 is still broken on that aspect

[jade plume]

oh uhh

in 21.10 if I have a SubmitNodeCollector and I'm trying to render blocks that are naturally transparent just do submitBlock and they work

[gleaming ore]

Yes but submitBlock does not tesselate (try rendering a bunch of glass)

[jade plume]

hmm

[fallow dagger]

It does not what?

[jade plume]

to render solid blocks transparently I extend ModelBlockRenderer (so I can unhardcode the alpha) and call tesselateBlock on it

[gleaming ore]

submitBlock:

[jade plume]

woah

[thorn hatch]

nano

[gleaming ore]

WRONG copy, sorry

[jade plume]

please this is a christian minecraft server

[thorn hatch]

I am at the office...

[shrewd violet]

Well not anymore......

https://tenor.com/8yQW.gif

[gleaming ore]

I firmly blame Discord for its copy paste there, and can't take that back. Deleted as fast as possible.

[deep wadi]

Where message log

[jade plume]

I checked message logs (in case I needed to delete it from there too) but didn't see it, maybe the log bot doesn't do threads?

[deep wadi]

BIG oversight

[atomic fern]

a man of culture i hear

[gleaming ore]

ANYWAY MOVING ON

Rendering glass with submitBlock does this

[jade plume]

is that wrong?

[gleaming ore]

But if you look at more correct rendering:

[atomic fern]

looks fine to me, just no culling

[gleaming ore]

Notice the lost translucency

You should see the red concrete through the glass

[atomic fern]

it is translucent, just obscured faces are not culled

your second one is incorrect, you should see the red, and the sides of the command block

[jade plume]

I'm confused, is the picture labeled "more correct rendering" the incorrect rendering?

[gleaming ore]

It's more correct in that it removes additional geometry but incorrect on translucency

[atomic fern]

the first is closer to correct than the second imo

[jade plume]

yeah when I saw the first I was like "yes this is what I would expect it to render like"

[gleaming ore]

While submitBlock does translucency correctly but does not tesselate and remove the additional geometry (since it's one block upload per call)

[thorn hatch]

nano is on a cracked german windows install?

[gleaming ore]

No, that was from someone in FC, not mine

But Gander is mine so shareable

[atomic fern]

Seems like we should provide an overlaod to submitBlock which takes a side mask for occluded faces, seems quite easy

it boils down into ModelBlockRenderer.renderModel, line 452, that for loop

[thorn hatch]

well they did say that they are not done with static block model rendering

[atomic fern]

https://ss.ln-k.net/39b10.png

there specifically

[gleaming ore]

What would be more useful is a renderBatchedBlock or renderBatched submit call

Which is a combination of tesselate and submitBlock

[atomic fern]

Submitting them individually is fine, they will use caced buffers

it should really not be any different

[dreamy frost]

If you want to call a Windows install that isn't activated because MS are dicks and won't re-activate a valid license after changing the hardware three times a cracked install, then sure, that's what I'm using 😛

[gleaming ore]

Switch to Zorin, team Linux is nice :>

[dreamy frost]

Multiplayer games and Intel Quartus would like a word with you

[gleaming ore]

I can play plenty of games just fine

And Intel is too busy shooting themselves in the feet to make me care lol

[dreamy frost]

You certainly won't be playing BF6 on Linux any time soon

[gleaming ore]

M'kay

I've got Rocket Lab, Portal, Minecraft, Factorio, Shapez, over half of Steam... think I'm good. And the more things that work with Steamdeck and Linux, the more pressure is put on all the AAA companies to give a shit

[atomic fern]

they will literally never give a shit

they only want profits

[gleaming ore]

'kay. Hope companies making stuff like Silksong eat them alive

[atomic fern]

3% of the userbase is not profit

once the market share is like 70/30, we can probably expect them to start caring a tiny bit

[gleaming ore]

Anyway. Topic. Rendering sucks. What do y'all think tomorrow will bring?

[atomic fern]

MORE RENDERING!

[gleaming ore]

...besides that, we KNOW that'll happen

[atomic fern]

very excited to diff asm dumps of obf remapped an unobf

i have been bashing my head against RT errors in CG for a while, and i want to know if i'm insane or not

[deep wadi]

Minecraft will drop Linux support!

[atomic fern]

(my 3% was from the latest steam hardware survey)

[gleaming ore]

TG, now update that for the year and a half since

[dreamy frost]

They won't give a shit until the problem of non-invasive ACs is solved in a way that's simpler than a Windows-only kernel-level AC (which I know is a controversial topic)

[lime oyster]

Is there a larger view of that?

Curious how linux changed once steam decks became a thing

[deep wadi]

That was Google images

[lime oyster]

oh

[deep wadi]

https://store.steampowered.com/hwsurvey/steam-hardware-software-survey-welcome-to-steam

What covers was saying

[gleaming ore]

So

[atomic fern]

3% is fucking huge, fyi, linux was at sub 1% for yeeeaars

[gleaming ore]

Win 10 is dead. That's a LOT of market that might suddenly move

[jade plume]

linux is less than "none of the above"

[gleaming ore]

And I know a ton of people reaaaaal mad at MS right now, and willing to try Linux...

[atomic fern]

yeah, come to linux, we have good customizable de's, and cracked translation layers :)

[next wolf]

oh did i tell you nano i ditched zorin a while back
been on fedora now for quite a while

[deep wadi]

“Imma try linux!”
Sees 100+ different flavors of Linux and gets overwhelmed

Goes back to windows

[atomic fern]

just use ubuntu, its fine

[gleaming ore]

Ubuntu, Mint, or Zorin

[atomic fern]

if your starting, it doesn't matter

[gleaming ore]

All are very very friendly

[atomic fern]

i used ubu for like a year then moved to arch

[lime oyster]

The only windows computer I have that's not decomissioned at this point is my main work computer, though that's only because the apps I need only exist for windows/mac and wine isn't good enough

The rest is like 5 different distros of linux because of testing things

[lunar storm]

Wsg everyone what did I miss

Is the snapshot in 5 mins?

[next wolf]

your a day early

[lunar storm]

Oh wait I'm an idiot 😭

[next wolf]

and it would have been ~50mins ago

if it was today

[lunar storm]

😅😅

[next wolf]

for reference snapshot drops in ~<t:1762268400:R>

[solar stone]

Frog day tomorrow?

🤞

[lethal ice]

likely

[hybrid perch]

neofrog?

[solemn furnace]

FXAA in Minecraft pls 🤞

[low kraken]

Don't worry, we've got you
https://github.com/MrGlockenspiel/activate-linux

[hybrid perch]

lets not suggest zorin now. please. thank you

[split oak]

what's wrong with zorin?

im looking at it rn it actually looks really nice

and it supports windows apps with is my main issue with switching to linux i might actually try this one out

[hybrid perch]

everything supports windows apps. zorin is not special in that way

[split oak]

none of the linux things i tried supported windows and they felt gross to use, very basic looking no polished look

[hybrid perch]

1. uses heavily modded shell of a different project and is prone to breakage
2. uses and ships outdated software
3. tech advertised on their website is not even their own

[split oak]

i want windows without the microsoft of it all that still supports windows games and apps

(probably not for this thread but plz enlighten me in #squirrels-🦊)

[hybrid perch]

use any modern distro in that case. zorin is not special as I said

[lethal ice]

zorin isn't special in tech, they just package things to be less annoying to people that have used windows all their life

that's their whole selling point

above all, it's one of the few distros that supports nvidia proprietary drivers out of the box without having to go through 50 hoops to get them installed and working

almost everything else whines about them not being foss

I wonder if there's any distro that includes the best choice for any purpose even if that choice is paid software and it adds to the price of a license 🤔

[twin viper]

(this moved to #squirrels-🦊)

[solemn furnace]

Frog day!

[devout vault]

The anticipation is killing me!

[rare lynx]

wrong answer, you were already dead!

[split oak]

mojang is killing me with the snapshots actually

but i love it

[atomic fern]

Saaaame! Gimmie unobf!

[split oak]

and J25!

and Vulkan!

[solemn furnace]

and DirectX 12!

oh wait no

[next wolf]

todays the day

[split oak]

6 hours

[next wolf]

~<t:1762268400:R> to go

[split oak]

is it not an hour after that

[next wolf]

always been 3pm my time

[next wolf]

which is in ~5hrs

[split oak]

ig its between 10 and 11 est

[split oak]

this one is in 6 hours

guess it was late

[next wolf]

i mean they try to release around that time sometimes they are later some times earlier

[next wolf]

but always around this time

[solemn furnace]

If Mojang adds Vulkan, I think it will be via bgfx.

[next wolf]

ngl im not rly expecting much this week since they are also shipping the unobf jar

[split oak]

i am

shipping the unobf jar probably doesnt require much effort

all they have to do is disable the proguard thing and upload a new jar

[next wolf]

not saying it does, im just expecting that to be the focus of this week

[split oak]

yeah probably nothing super big like a new backend

but i think some more rendering changes and maybe J25

since J25 is also trivial

and would make a nice break in the versions

so the newer versions could simply start with J25

[polar kernel]

Maybe also new content of the drop for non technical audience

[low kraken]

Dst switch, iirc

[split oak]

i forgor

[low kraken]

Both show up as 15:00ish for me in the uk

[split oak]

does sweden not do DST

[low kraken]

It does
Cet(UTC+1) -> cest(UTC+2) in the summer

Apparently

But since the UK changes by the same amount in the same direction at the same time, the relative change is 0

[next wolf]

oh yeah that happened recently didnt it
ngl i rarly ever pay attention to when that happens

but should still be around the time i posted

[thorn hatch]

it's Tuesday 16:00 mojang time when they have the release scheduled, it just takes a few minutes to show up

which is in ~3h 15min

[low kraken]

<t:1762268400:R>
<t:1762268400:f>

[next wolf]

~< 25mins to go

[lone cairn]

oh no

https://tenor.com/view/knife-cat-cat-knife-cat-knife-cat-angry-gif-9404759922388476738

[atomic fern]

~T- 5 mins

[next wolf]

inb4 mojang is late

[atomic fern]

[thorn hatch]

T-1 min

[lethal ice]

I'm watching an exam

[echo pendant]

as in, proctoring/invigilating?

[thorn hatch]

better than writing it

[atomic fern]

what kind of exam?

[thorn hatch]

T-0

https://www.minecraft.net/en-us/article/minecraft-snapshot-25w45a

[echo pendant]

gimme

[solemn furnace]

T+1

[next wolf]

they late 🙁 /j

[atomic fern]

afjklhsdf\

[solemn furnace]

25w45a

[next wolf]

Snapshot 25w45a

• Primer: https://github.com/ChampionAsh5357/neoforged-github/blob/primer/2111-or-22/primers/1.22/index.md
• Article: https://www.minecraft.net/en-us/article/minecraft-snapshot-25w45a
• Changelog: https://misode.github.io/versions/?id=25w45a
• Notion: https://apexmodder.notion.site/2524f070f88880648482e3cbdcba566a?v=2524f070f88880b29f60000c52785ecb&p=2a14f070f888805bb0dee2658269e0ab&pm=c
• SnowMan: https://github.com/neoforged/Snowman/commit/3ea660bd0b9481306ce5bc99c3673f02361baba2

SlicedLimes Videos:

• Main: https://www.youtube.com/watch?v=JBhZriLnIFs
• Pack: N/A

||<@&1067092163520909374>||

[crystal plume]

[Reference to](#mojira message) #mojira [➤ ](#mojira message)Version 25w45a has been released.

lolothepro30 referenced · Nov 4, 2025, 3:01 PM

25w45a

Affected

3 issues

Fixed

45 issues

Released

2025-11-04

Quoted

[thorn hatch]

where da unobf at

[lone cairn]

WJOOOO

time to

[shrewd zealot]

Our final set of features for Mounts of Mayhem

[lone cairn]

With this week's snapshot, we're also introducing a new concept for data packs called timelines, that can modify game behavior based on the in-game day time – more about that in the full changelog below.

[gloomy axle]

there it is

[lone cairn]

jesus christ mojang

[jade plume]

[lone cairn]

timelines

[dark nacelle]

New version detected: 25w45a.

[lone cairn]

what is this, Doctor Who - Minecraft edition??

[atomic fern]

As we are preparing to remove obfuscation from Java Edition, you can also get a non-obfuscated experimental version of this snapshot using the Minecraft Launcher:

Download this zip file
Unpack the folder into your "versions" folder of your local Minecraft application data folder
Start (or restart) the Launcher
Create a new launch installation and select the "unobfuscated 25w45a_unobfuscated" version
Start the game and the remaining files will be downloaded

https://piston-data.mojang.com/v1/objects/de334d80d9ddc5abb94c611b8ad10f9125c4c421/25w45a_unobfuscated.zip

[echo pendant]

so not in the manifest

[atomic fern]

its not

zip contains another version manifest

[echo pendant]

good, we don't have to worry about that pain

for now

alright, time to change toolings to accept custom version manifests

[gloomy axle]

yeah, that's the first thing I checked

[echo pendant]

me too

[lone cairn]

Sci, can I task YOU to work on the timeline stuff?

sounds perfect for you /s

[gloomy axle]

items atlas

[thorn hatch]

Item textures were split out of the blocks atlas into a separate new items atlas

[lime oyster]

Does the non-obfuscated not come with the source?

[echo pendant]

more sledgehammers to rendering!

[echo pendant]

they never will

they're just unobfuscated

no name mangling

[lone cairn]

why would they give the source

[gloomy axle]

full LVT

[lime oyster]

I mean, at that point, the only thing they're not giving us is any javadocs they have

Assuming the LVT is not mangled

[gloomy axle]

I assume snowblower can't access those since they're not in the manifest

[gloomy axle]

they said we would be able to see locals

[thorn hatch]

if they have any

[echo pendant]

nope

[turbid wedge]

We got a data-driven animation system for the server-side, now can we get one for the client?

And yes Ik NeoForge has one, but that's only for modded entities

[gloomy axle]

I thought vanilla has one

ah way, data driven

[gloomy axle]

ok, so how do I view the source then

I wanna see locals

[atomic fern]

its real

the LVT is alive

[turbid wedge]

The NeoForge one basically just defines a codec for the Vanilla one lol

[atomic fern]

its all there

https://ss.ln-k.net/940f2.png

[gloomy axle]

I guess I could run the server jar through a standard decompiler

[turbid wedge]

Hallelujah

[next wolf]

https://github.com/neoforged/Snowman/commit/3ea660bd0b9481306ce5bc99c3673f02361baba2

-# Note that SnowMan is still Obf->Obf

so remember how i said i wasnt expecting much this week....
wtf are timelines

[gloomy axle]

oh, the server jar has an interesting structure

[lime oyster]

Oh hey, item atlas

[next wolf]

oh.. thats concerning

[worthy ember]

What’s in it?

[next wolf]

contained so many help utilities that are just gone

[jade plume]

1170 lines of helper methods

was it renamed?

[next wolf]

or moved eslewhere?

[gloomy axle]

but hey, I see locals, I'm happy

[lime oyster]

Did ResourceLocation really get renamed to Identifier?

[dark nacelle]

Minecraft Snapshot 25w45a

Greetings from the Tuesday snapshot shipping room! Our final set of features for Mounts of Mayhem brings the game's strongest armor tier to your favorite mount – with Netherite horse armor! As well as making your steed more stylish, Netherite horse armor w...

Nov 4, 2025, 3:02 PM

[next wolf]

please no

we are not yarn

[echo pendant]

they did

the mad lads actually renamed it

[turbid wedge]

THEY WHAT

[next wolf]

well time to retire i suppose /j

[lime oyster]

And Util still exists

[gloomy axle]

this decompiler also can't handle records

[echo pendant]

[lime oyster]

In net.minecraft.util

[turbid wedge]

No way

[next wolf]

ah it was just moved

[turbid wedge]

I preferred ResourceLocation

[worthy ember]

Wait, what?

[next wolf]

this is a neat new class tho

[turbid wedge]

That's not new

[next wolf]

its not?

new to me atleast

[turbid wedge]

I don't think it's new

[echo pendant]

that's not new

can you guess what it was previously called

[lime oyster]

Was originally ResourceLocationPattern

[turbid dune]

Did someone at Mojang deciding just to apply yarn naming to the whole project

[gloomy axle]

not my minecraft

[wise seal]

I like the new timeline stuff...
It looks to be more about time of day based events... but I could see it being used for longer running story stuff....
Imagine a slowly crumbling unstable dimension....

[echo pendant]

alright boys, I'm announcing my retirement from Minecraft modding /j

[next wolf]

inb4 "unobf" is just "mojang->apply yarn->ship"

[gloomy axle]

official were a ruse, they used yarn all this time

[echo pendant]

Beardifier, iirc a specifically not Yarn name, is still in

[lime oyster]

Welp, I'll start primer work tomorrow

Definitely not because I'm dreading all the renames

[turbid wedge]

Noo the skybox field of a dimension is an enum

[echo pendant]

the sheer amount of changes alone the ResourceLocation to Identifier rename did

[gloomy axle]

[lime oyster]

It's going to be the jspecify hell all over again

Except worse

[echo pendant]

[turbid wedge]

Ah yes the blocks array

[lime oyster]

Wait, they fixed spelling???

[chrome shore]

haha that was me on fabricord pointing that out, slicedlime 👀 reacted my message

[echo pendant]

perfect, thank you for that

[gloomy axle]

should have named it citroën

[chrome shore]

(also popping in to say Identifier forever! hope you guys don't mind it...)

[lone cairn]

doctor who in minecraft

[turbid dune]

Smh, each snapshot they'll progressively apply more and more yarn naming

[turbid wedge]

This is going to be absolutely so painful for multiversion mods

[gloomy axle]

time to reimplement remapping to just remap it back to ResourceLocation

[lime oyster]

It's less I don't care and more why is there so many changes

[lunar storm]

Are you shitting me?😭

[lone cairn]

in secret, Mojang is going to keep Yarn alive, just because

[echo pendant]

all the Yarn maintainers were in fact alts of—

[lunar storm]

What do mojangstas say about this, like what's the reason

[gloomy axle]

My best guess is that this a tribute to yarn

[lunar storm]

Identifier forever I guess

[turbid dune]

Could also be several Mojangsters prefering the name as well

[lunar storm]

Yh but the timing is suspicious

[lone cairn]

Modmuss being a secret Mojang emplyee, and this is just a stab at all the haters

[lunar storm]

Like why now and not before

[lone cairn]

tbf, some people were extremely toxic that Yarn was sunsetted..... so I can see Mojang just stabbing those who were like that

[gloomy axle]

well sure, but still

[lunar storm]

I want them to reply, if they're online rn

[thorn hatch]

was just RL renamed to the yarn name or anything else too?

[turbid wedge]

Everything that was named after RL

[next wolf]

some classes moved locations too

[turbid wedge]

Actually where is the class anyway?

I don't remember where it is

[hallow chasm]

has anyone else checked the "preparing to remove obfuscation from Java Edition" link at the bottom of the blog post

[lunar storm]

Did Identifier move locations btw?

[turbid wedge]

I could've sworn it was n.m.data

[next wolf]

mainly net.minecraft.Util -> net.minecraft.util.Util

[lone cairn]

makes you wonder if Cpw is also just a Mojang emplyee, and that is why he is never around. He is cursed to never speak!

[gloomy axle]

net.minecraft.resources

[turbid wedge]

Ah

[next wolf]

yeah Identifier is still where RL was

[lunar storm]

But RL just made sense

[next wolf]

smh they renamed RL to identifier but didnt do IdentifierKey or some shit

[lunar storm]

They did RLPattern to identifierpattern

[lone cairn]

yes, as I said, I think Moajng just did this, to spite all the toxic shit heads who were toxic towards the blogpost of Yarn being sunsetted

that is my theory

[gloomy axle]

they didn't fix the error messages tho

[lone cairn]

or that Modmuss is just Jeb /s

[hallow chasm]

not the best snapshot in terms of patches that need to be fixed

[next wolf]

warnings dont exist, you can ignore them... right

[gloomy axle]

depends, if it's yellow, sure, but if it's orange, no

[next wolf]

i see the word warn i ignore, simple as that

[lunar storm]

Right in time for Spooky season

[fallow dagger]

Is Identifier the only rename?

[lunar storm]

RLpattern turned into IdentifierPattern as well

[thorn hatch]

and anything named after RL

[echo pendant]

so about 75%-80% of the codebase got affected

[gloomy axle]

except ResourceKey apparently, what is/was the yarn name for that one?

[fallow dagger]

RegistryKey?

[lunar storm]

Registry key yeah

[next wolf]

see that one makes sense

since the keys are mostly for registires

[fallow dagger]

Yeah

[lunar storm]

Mostly being the operative keyword

[spare hearth]

time to add a bunch more vars to my codebase or finally figure out stonecutter's string replacement

[next wolf]

var wont help you when you need to call one if the static methods to construct a RL Identifier

[spare hearth]

i have a helper method for that anyway since my multiversion extends to before 1.21

[hallow chasm]

on the bright side

[echo pendant]

huzzah!

[next wolf]

now who wants take to take the unobf jar and spit out a new parchment version with all the new param names

gimme those early unobf param names in my 21.10 workspace

uhh is this just another class being moved or do dimensions not have special effects anymore?

[gloomy axle]

I mean, that shouldn't be too hard to setup

[proper rover]

how is the unobf decomp?

[next wolf]

i just know its something alot of modded dimensions used so kind concercning if its gone

[gloomy axle]

bad, it's Identifier now

[proper rover]

that's good though

but i was meaning about the decompile quality :p

I'm still working on getting my tools changed to the new stuff, but I'd love to know if there is any bugs that needs addressing

i want to release vf 1.11.2 if not

[latent ivy]

Replaced with Timelines

[hallow chasm]

I will do the neoform update later since there are many patch failures due to mojang renaming stuff

[rare lynx]

the fuck you mean they renamed it to identifier

[proper rover]

alright

[rare lynx]

who do I stab

[proper rover]

it's fewer characters!

[latent ivy]

Maybe we don't single out devs 😄

[devout vault]

Identifier is so much better

[turbid dune]

enough batter to make several cakes?

[next wolf]

fine, we stab all of mojang then /j

[hallow chasm]

For the amount of patches that broke

[north summit]

so you are suggesting.... tnt?

[rare lynx]

[north summit]

https://tenor.com/view/madagscar-penguins-kaboom-gif-9833865

[rare lynx]

this is such a moment

[ocean flower]

What makes a name like ResourceLocation or Identifier good or bad outside of being used to them?

[proper rover]

is the obfuscated jar part of the version manifest?

[gloomy axle]

principle

[devout vault]

There's a separate link at the bottom to a special manifest

[proper rover]

alright

[gloomy axle]

no

[devout vault]

So, not part of the "production" manifest, no

[proper rover]

thanks!

[devout vault]

https://piston-data.mojang.com/v1/objects/de334d80d9ddc5abb94c611b8ad10f9125c4c421/25w45a_unobfuscated.zip

contains the version manifest for the unobfuscated version

[wild junco]

Primarily how easy it is to tell what they do, aka how ResourceLocation represents the location of a resource.

[rare lynx]

I mean, since we have s here anyway, could someone clarify whether 1.21.11/1.22 or whatever the release version is named will be the first one that's unobfuscated or whether the first snapshot in the new cycle will be the first

[gloomy axle]

they said the latter

Starting with the first snapshot following the complete Mounts of Mayhem launch, we will no longer obfuscate Minecraft: Java Edition.

[deep wadi]

Mojang should’ve used Id instead, even fewer characters

[latent ivy]

missed opportunity

[gloomy axle]

They should have just used the obfuscated names, saves so many keystrokes

[wild junco]

Depending on the definition it's not even an identifier, since the resource that is located there can change at any time

[gloomy axle]

Why write ResourceLocation.fromNamespaceAndPath(…) if you can just write like akd.h(…)

[gloomy axle]

Should have named it NamespacedLocationOrIdentifierOrSomething

[solemn furnace]

Is it possible to pin this?

[proper rover]

by the way, are all params final now?

[solemn furnace]

Ty

[next wolf]

orion sniped me pinning that

[proper rover]

i heard that somewhere so I was interested in whether development of the unfinaling vf plugin was going to happen

[gloomy axle]

From what I know mojang's codebase is final by default

[echo pendant]

based

[lone cairn]

wait, people read pins?
I have never opened them in this channel

[next wolf]

i mean thats fair for this channel, the pins can get spammy during ports

-# which again is why im asking for snapshot dedi channels/threads/forum or something, but i dont think that will happen

[vestal tundra]

ngl i feel like 1.22 is a new 1.12.2 to 1.13 situation

[jade plume]

or 1.16.5 to 1.17

[vocal ivy]

Is it possible that they will change back to ResourceLocation in the next snapshot

[next wolf]

https://tenor.com/view/futurama-professor-farnsworth-a-man-can-dream-gif-12491746995726919913

[vocal ivy]

They also keep three location references in Identifier, which seem to have been not checked at all

[next wolf]

some exceptions also still use RL

[atomic fern]

wiat, the fuck happened to ResourceLocation

[edgy river]

it got a little stringy

cought up in the yarn

[vocal ivy]

yarn hacked mojang(JK)

[atomic fern]

this is completely unnaceptable

[shrewd violet]

Somebdoy needs to run git revert real quick

[atomic fern]

but the mojang gods do as they choose i guess

[next wolf]

covers your late to the party

[deep wadi]

Enjoying your little block game eh

[atomic fern]

bro, i was fixing toolchains and my bot for the snpashot

i can only do so many things

[vocal ivy]

ResourceLocation has been around for longer than I've been in the game, do they really understand what they're doing

[vestal tundra]

LOL then make it remap mixins aswell...

[solar stone]

Add back ResourceLocation 😭

[vocal ivy]

mcp again

[vestal tundra]

about to make a lib to bring back RLs 🤣

[solar stone]

🤣

[jade plume]

time is now data-driven: *crickets*
ResourceLocation is now Identifier: *confused screaming*

[vocal ivy]

not enough locations

[vestal tundra]

shame on them

[edgy river]

done

[worthy ember]

You can see our priorities are correct.

[vocal ivy]

replay mod have a remap tool that support remap kotlin and mixin

but loom seem only support java

[edgy river]

yeah. if I decide I want to die in gradle and gpl I may package that into a loom extension that doesn't actually have anything to do with loom

[edgy river]

correct

[lone cairn]

you will get what mojang serves you, and you will enjoy it!
/s

[next wolf]

pshh il just sneak in a simple RL patch into kits port

[vestal tundra]

done ?

[lone cairn]

are you asking for mojang to change everything to Yarn, to spite neo?

[next wolf]

make Identifier sealed
add new class ResourceLocation
extend Identifier
add static bouncers

ez patch

[ember galleon]

migrate mappings supports mixins now

[vestal tundra]

yoo finally

[ember galleon]

all of them, including some third party annotations

[vestal tundra]

i had to fix 70 mixins by hand last time

[edgy river]

last time Mojang released a fully deobfuscated copy of the game?

did I sleep in on a day?

[next wolf]

that will be when 21.11/22.0 drops (well mounts of mayhem)

we get obf and unobf jars for each snapshot until then

[vestal tundra]

what will Parchment become

[next wolf]

dead or just javadocs

[vestal tundra]

:(

[next wolf]

it has no purpose shipping param names if mojang is giving us them

which was its main purpose

[solar stone]

Variable names?

[atomic fern]

and var names, yes

[next wolf]

those are included too afaik in unof jar

[atomic fern]

full lv table is now available

[next wolf]

and i think method params are also final everywehre now

[proper rover]

In my internal testing, it looks like the new (unobf) version has ~20 fewer compile errors when compared to the last snapshot

[vestal tundra]

it's this snapshot ?

[next wolf]

yes 45a is todays snapshot

[vocal ivy]

full local variable ,but not good as parchment

[next wolf]

what does parchment give that the unobf doesnt?
other than javadocs

[lethal ice]

nothing

parchment has a few javadocs, zero local variables
unobf has zero javadocs, all the local variables

solution: parchment becomes a javadoc provider

[next wolf]

yeah i see no reason for parchment to stick around
i mean the javadocs are nice but i could understand it dieing out too

[atomic fern]

can intellij staple javadoc jars onto binary jars? may be able to just literally publish a javadoc jar on maven

[lethal ice]

technically it has another potential feature: manage unpick definitions

[lethal ice]

(unpick as in de-inline)

[next wolf]

oh please, i wanna see all those constants in vanilla code get used

so many of them

no more magic numbers

like why do this, just AT the constant public and use it
so much more readable imo and if mojang changes it for what ever reason
we dont have to remember to update this random 1

[stark crane]

In 1.22 do we think we will have a lts version that all mods will settle on?

[proper rover]

for what it's worth we're working on a decompiler-level solution for unpick, based on unpick v4

[next wolf]

doubt it

with mojag pushing more frequent updates now

[gloomy axle]

well, from what I gather the idea is to have parchment as javadocs + unpick + annotations

fabric discord said something about merging, but idk

[unreal jacinth]

oh identifier wasnt a joke

[worthy ember]

Unfortunately not.

[spare hearth]

there is no 1.22 and there is no new lts

[echo pendant]

"there is no queen of england" /j

[lunar storm]

net.minecraft.resources.Identifier just sounds too wrong

[atomic fern]

this is the world we live in now

or rather, level

world went away many years ago

[split oak]

what the hell no please

resource location makes so much more sense they couldn't have paid homage with Cat#meow or something why this

this is going to require so much refactoring for everyone

[devout vault]

They use these for many things that are not resources.

[split oak]

everything is a resource so that's simply not true

whether it's a file resource or a registry resource

[edgy river]

Just use migrate mappings /j

It still works from mojmaps -> mojmaps

As long as the intermediary name doesn't change

[split oak]

all my names revolve around rl

modLoc and things

now they have to be converted to modId

[hybrid perch]

thats gonna cause some confusion

[next wolf]

modId(<modId>) ->modId:<modId> ah totally not confusing at all

[split oak]

actually

modId(path) -> modid:path

but modLoc is just such a nicer name

idk if this change will grow on me I genuinely really don't like Identifier

and now the code base is a mess of both forms bc they didn't change everything

Identifier location just kill me

[lone cairn]

don't give them new ideas

[lunar storm]

I wouldnt mind World tbh

[lone cairn]

you say that
but soon they will come with something stupid like SPWorld, MPWorld, ThreadedLevel

[lunar storm]

what are those supposed to be

[split oak]

I would prefer world to identifier

[lunar storm]

Yea identifier is just weird

[lone cairn]

Single Player World
Multi Player World
and then a common Level, where the 2 can talk to each other, in a multi-threaded way

[lunar tapir]

not sure why, but part of me is kinda surprised they would name a parameter "thing", I always feel unprofessional when I do it lol

[gloomy axle]

I mean, it's fairly correct in this context
but I get what you mean

[lunar tapir]

a few lines down they use element instead

[gloomy axle]

Well, that's not T, that's ResourceKey<T>

They could have called it value, but I guess that's not that much better than thing

[split oak]

hello people with src access

what the blits looking like can i see

[split oak]

day 1 param names i prayed for times like this

[next wolf]

https://www.youtube.com/watch?v=JBhZriLnIFs

[gloomy axle]

for me locals are the more interesting part, no more f3, flag1, i2

[twin viper]

given rl is called id now, renaming modid anyone?

[split oak]

modNamespace

or just namespace

[twin viper]

yeah namespace was my thought

[sour torrent]

Dude Im so happy that its Identifier I have maintained its my least favorite part of Mojmap

[twin viper]

Identifier.fromNamespaceAndPath

love to see it

[split oak]

RL is better fight me

[twin viper]

i dont like identifier as much as the next guy, but this is one hell of a method name that i kinda wish they changed with the class rename

[split oak]

i support the of overloads

[sour torrent]

It is literally an id I will stand on that hill

[split oak]

no it is not

a texture location is not an ID

all IDs are RLs but not all RLs are IDs

[jade plume]

it's the id of a texture

[twin viper]

its the location of resources such as textures and sounds.
Its also the id of registry onjects like items

[split oak]

it is the location of the resource in the registry

[sour torrent]

The way I see it, the blocks arent registered in the resource part of my code they are registered in the code part of my code

I will ragebait on this I dont care

[split oak]

...what

resources are not just assets

anything can be a resource even things in code

[devout vault]

did anyone bother doing a snowblower with the deobf?

[devout vault]

BTW here's a VF plugin to remove the final from the parameters: https://github.com/shartte/VineflowerUnfinal

[next wolf]

asking as the smooth brain that i am, what does final on method params do exactly?
i mean me the dev creating the method, its like final on class fields, cant modify the param
but for a caller, does it rly change or do anything?

[devout vault]

no, it has no effect on callers

[next wolf]

so whats the benefit of removing final?

like its vanilla code we are removing them from
so we arnt inside the method bodies, we are the callers
whats the benefit?

[devout vault]

we are removing them now

generally they make the signatures less readable if you're just trying to understand what is going on

[next wolf]

im just trying to understand why/what the benefit is?
i mean in neodev i suppose its helpful, where you can modify the bodies
but as a general modder whos mostly calling the methods
im strugling the see a benefit/reason behind removing final

[next wolf]

readability i suppose makes sense, some methods are already long enough in param length let alone all the final keywords now
-# looks at guiGraphics.blit

[jade plume]

as a modder it makes the vanilla source file easier to read because the list of parameters has less boilerplate words in it

[devout vault]

and yes, if remove them from neodev, we have to remove them from moddev

but generally it's both, I'd say, so we have been removing them as soon as they showed up in the obfuscated jars 1+y ago

[twin viper]

does final params affect mixin?

[devout vault]

no

[dreamy frost]

Do final params even have any runtime checks?

[devout vault]

no

[dreamy frost]

That's what I thought

[devout vault]

The flag that makes them final is in the same block as the parameter names (which is optional, anyway)

The tables you get when you compile with -parameters

That started showing up in 2024 in the obf'd jars

we're currently stripping those access flags in ART during the remapping from obf->named

but assuming we'd want to continue doing that when we directly decompile the new jars without any prior bytecode processing

then we need a vineflower plugin (or do it afterwards in JST)

Well, JST would be bad, since we patch the sources before JST, if I remember correctly. And the patches should be based on the non-final parameters.

[gleaming ore]

screms in late to the party, and also Identifier

meanwhile, in Mojland.. I imagine there's a bit of amusement happening

https://tenor.com/view/helluva-boss-gif-20187203

^ Yarn rn

also Moj already announcing they aren't going back smh /s

[edgy river]

👀 ResourceKey

-# You're next

[gleaming ore]

oh lawd

buffalo buffalo buffalo buffalo..

[echo pendant]

don't worry, once we go unobf all the pain will go away

[gleaming ore]

...that sounds threatening

sci do you have a gun