1.21 Snapshots

[weary moon]

https://tenor.com/view/godfather-massacre-sad-gif-16810633

[odd flare]

yeah....
my childhood is ruined

[red sigil]

maybe they are adding a Builder profession

to villagers

[odd flare]

only way to know, is to wait for the snapshot....
which might be today
would be odd, thoguh

[unreal timber]

they've actually just added AI to the fishing rod bobber so it hunts fish instead of passively waits there

[oblique socket]

mojang wtf

[fallow horizon]

i noticed some argb/abgr weirdness in the changes as well, so yeah

[obsidian tide]

NativeImage?

[oblique socket]

i'd recommmend checking banners

they're... weird

[obsidian tide]

wait no, NativeImage was worse

[oblique socket]

in this snapshot, banners will sometimes show up all white in your inventory

[manic oracle]

i remember seeing that getting fixed in mojira a few days ago pretty sure

[oblique socket]

and I don't know if it's just me, but there seems to be a "dithering" effect on banners

idk if that was always there

[manic oracle]

yep

[oblique socket]

we just need official entity batching 😔

[sinful hazel]

the argument is argb, but it's storing rgba 😛

but packing that as an int on a little endian system is going to be inversed from how opengl interprets it (hence abgr)

[odd flare]

when vulkan? /s

[lone tree]

I wonder, since they changed the vertex system in 24w21a, what other changes they'll make in the next snapshot...?

more optimization?

[red sigil]

this change was a surprise

so, who knows

there could be more surprises

[fiery galleon]

inb4 they casually drop data pack items

[uncut birch]

are there any big endian systems anymore?

because this makes me wonder how many programs are out there that silently assume little-endian

[red sigil]

I would guess 90%+ of software written for PC assumes little endian, but software written for other platforms may be more ... flexible

[uncut birch]

networking code i think is usually a little better because the networking library comes with helper functions for "convert to big endian" so you're less tempted to just roll your own that unconditionally swaps bytes or just hardcode the values backwards in the first place

[TCP/IP and many older application protocols are big-endian because the committees that defined them had decided that it should be the standard]

[cosmic wharf]

iirc it has to do with the fact many graphics controllers store their info in BGRA format, and ARGB is just reversing that endianness

unless I misunderstand what the two red circles are on about

[uncut birch]

literally the other way around, it's RGBA format in memory, ABGR when treated as 32-bit integers

i.e. you have 0x11 0x22 0x33 0xff in memory for the red, green, blue, alpha, when accessing that from java as a 32-bit integer it looks like 0xFF332211, and minecraft itself wants to use ARGB where you'd use 0xFF112233

so the function accepts ARGB, converts it to ABGR, and writes that to memory using the 32 bit little endian write instruction

so the actual bytes in memory are in R G B A order

[red sigil]

it's why some people like big-endian more

because if you display memory left-to-right as you would text

[uncut birch]

the graphics hardware doesn't actually care [beyond that you have to tell it what format you're using], but the names OpenGL uses for these formats are strictly based on the byte order

[red sigil]

you get the number in the way you'd write it in code

while with little-endian, you get it backward

[uncut birch]

[which makes me wonder how it manages 16-bit formats where the green bits are split between the two bytes]

[red sigil]

with data like R10G10B10 or R5G6B5

byte-swapping doesn't work

you have to do additional shift/mask to convert endianess

so people usually don't bother

[cosmic wharf]

many GPU display controllers are BGRA format, not RGBA

[red sigil]

but really, that's only an issue if the endianess of the gpu is not a match for the endianess of the cpu

and they usually do match because gpus are made for x86 :P

[cosmic wharf]

Also, fun fact, this is the historical reason why microsoft GDI used BGRA for colors

[fallow horizon]

ok now that the event refactor is done, how do I start kits properly

do I just remove schurli's branch and run the action again?

[primal spade]

yeah that will do

[fallow horizon]

still 88 rejects

let's see how much of a grind this will be

[fallow horizon]

the answer is extremely grindy

only down to 62 rejects

pushed what I have so far

basically all the rendering changes are done as well as a few misc things I spotted along the way

[vapid gale]

Two minor things:

• https://github.com/neoforged/Kits/compare/bda06425bb06...6b91b6659356#diff-277a8ebe8d9e2bc9500d73bd1bc535c7c492aac6d078c30f78fb5562d6697d02R44-R45: there's a constant for the radius in RenderChunkRegion, would be nice to use that instead of a magic number
• https://github.com/neoforged/Kits/compare/bda06425bb06...6b91b6659356#diff-f989923af099fd13d74147576083745701de84e3839e3801c5f7c90255b41ef0L118-R135:
  please add a note regarding https://github.com/neoforged/NeoForge/issues/635, then we can finally close that issue

[fallow horizon]

surprised that issue is still unfixed, I'm pretty sure the patch dates back to 1.12

thanks for the comments, I'll fix both

[vapid gale]

Yeah, I'm very surprised they fixed neither that nor the SpriteCoordinateExpander issue despite touching both of those directly

[fallow horizon]

Can I just close the Neo issue now then?

[vapid gale]

I'd say yes

[uncut river]

why is it called "Kits" btw

[lone tree]

SpriteCoordinateExpander The bug report inside is invalid, by the way.

It doesn't cause any bugs in Vanilla.

In the bug report, there's no mention of this affecting Minecraft Vanilla.

https://bugs.mojang.com/browse/MC-201374 This bug still affects Minecraft 24w21b?

[hard ivy]

baby fox

[uncut river]

oh

[odd flare]

Does that really matter?
Snapshots are meant to be broken, so if there is a issue that was introduced a snapshot before, and still not fixed, then maybe it needs a bit more time

[lone tree]

No, this bug was introduced in Minecraft 1.16.3, but since Mojang fixed similar bugs in Minecraft 24w21a, I wonder if they didn't fix this bug.

[lone tree]

#mojira message nice

[mighty prawn]

[Jump to referenced message](#mojira message) in #mojira

lolothepro30 referenced · May 27, 2024, 4:01 PM

3 tickets have just been resolved as Fixed!

MC-169698

Blocks within igloo basements can generate waterlogged when generating in close proximity to water

MC-266471

Trial chambers can intersect aquifers, causing some blocks inside the structure to be waterlogged

MC-268185

Copper grates in trial chambers are waterlogged

Quoted

[torn notch]

👀

[torn notch]

blehh, hope they redo model loading soon

[half plaza]

apparently ender pearls call changeDimension when teleporting the owning entity now even though the ender pearl is set to lose it's owner when changing dimensions

[uncut birch]

always, or only between dimensions?

[half plaza]

always

[uncut birch]

so throwing them through a portal doesn't work, what about having an ender pearl suspension chamber in another dimension

what if you're offline when an ender pearl owned by you loads [iirc that would lose the owner before]

[half plaza]

changing dimension seems to be the only time the owner is lost although it only checks the level it is currently in for the owning entity

[lone tree]

Imagine if Mojang changes the light engine again in the near future.

[odd flare]

they might be working on that in the background

it has been known that they are using github, so it is likely on some branch

[viscid crane]

It's chuesday, innit

No snappy

[light arrow]

I thing java edition uses git... no idea what bedrock uses

🐸

[viscid crane]

Bedrock is using subversions, yes

Their developer builds have all sorts of extra settings and configurations related to it

All on a cloud azure backend

including CI

idk, it's a microsoft environment

So I imagine it's a hellscape

Of barely functioning systems, constant rewrites due to unnecessairy changes and just inconvinience

[primal spade]

I hope nobody is still using svn in 2024

[obsidian tide]

our uni is

[viscid crane]

ip over carrier pidegon

[light arrow]

mm yes patch mailing lists

[austere ocean]

[red sigil]

at least it's email. could be floppies like we used to do in the past ;P

[obsidian tide]

we do have gitlab, but the system for exercises with automatic junit tests still uses svn for some reason

[austere ocean]

is the crown glued to your head or does the hair hold it in place

[viscid crane]

I love legacy software

[primal spade]

My god you guys should find a better uni

We used git

[viscid crane]

Our unis don't use VSC

They just want zips emailed to them

[austere ocean]

I think it's the hair, which is concerning

[uncut river]

I'm yet to go to uni but that sounds psychotic 😅

[red sigil]

my friend teaches at the university I'm doing the PhD in, and they have different systems for handing over homework. main one is some website thing

but eg, my friend teaches git in his class, and expects people to commit to github

and uses the github history to track progress

[light arrow]

may I introduce you to the way my father "wrote" and transferred programs... punch cards

[jagged stag]

we'll see, after all 2021 is still not over yet

[austere ocean]

covid time

[red sigil]

I'm not that old heh. the first computer I used was a 286 with 5" floppies, and the oldest I have used was an Amstrad CPC464 I found at an uncle's basement and took home

[red sigil]

one of the big projects at my lab that has been in development for over 10 years is on svn.

and it's written like shit and my friend is desperate and wants to rewrite it all because of how horrible the code is

[light arrow]

I still have a C64 for gaming at home

[viscid crane]

I have a C64 with elite

I also still have my indy

But I haven't been able to access it yet, I need an IDE adapter to get rid of the password

[austere ocean]

:3c

[light arrow]

please stay on topic... 🐸

[austere ocean]

it's not Wednesday tho

wait is it

it's not Wednesday

[light arrow]

but tomorrow is and this is the snapshots channel

[red sigil]

Schurli is right, we have allowed the squirrels to sneak into this channel

[hard ivy]

[odd flare]

https://tenor.com/view/eye-roll-ugh-im-tired-with-this-here-we-go-again-hugh-laurie-gif-17296078

[lone tree]

Ok, it's only one bug report fixed

[lone tree]

https://fxtwitter.com/gegybeans/status/1795518076152352781 fyi

[half plaza]

There is no method in VertexConsumer for ending a vertex

[primal spade]

Hmmm?

Check how they handle it when blitting sprites in GUI maybe?

[half plaza]

They end the vertex either when a new vertex is started or the build method is called (which isn't included in the interface)

[twilit slate]

Hmmm it is that day
I really wonder what dev methodology they use? Do they have sprints?

Is today the sprint closing day?
Who is the scrum master?
What ceremonies do they have?

[sinful hazel]

nap time is a very important ceremony

[turbid hatch]

I recently heard that tree refactors is a ceremony they have

[twilit slate]

I agree, but it needs a ceremony master that reads the bed time story, about every player being satisfied with the new update, so the devs can have nice dreams

[remote void]

https://www.youtube.com/watch?v=ia8Q51ouA_s

[alpine pelican]

New version detected: 1.21-pre1.

[jagged stag]

FUCK

pre

already

<@&1067092163520909374>

[austere ocean]

WUH

I Completely forgot it was snapshot day lmao

[jagged stag]

heh, did it before the fabric bot

[austere ocean]

ok what did they change

data driven mob effects pls

[turbid hatch]

do they change a lot on pre-releases?

[austere ocean]

tell it to work faster smh

[remote void]

Quick! Merge all Neo PRs now!

[austere ocean]

the beatings will continue until decompilation speeds improve

[light arrow]

pre... 😢 no datapack brewing

[austere ocean]

aw

[fallow horizon]

Kits is down to 6 errors on 24w21b

[light arrow]

gen patches prepare neoform and update... compile errors can be done later

[austere ocean]

WHAT
LEASHABLE BOATS

https://mdcfe.dev/mc-changes?ver=1.21-pre1

[lone tree]

Note: Minecraft 1.21.1 is already planned

yeap

[obsidian tide]

it's minecraft source code, even if not the full one, so ...

[austere ocean]

With this snapshot, we are temporarily removing the functionality of lighting blocks on fire using the Fire Aspect enchantment, as we want to refine it further. This functionality will be reintroduced in a later release.

[fallow horizon]

we start with 24w21b for 1.21-pre1, right?

[lone tree]

hhahaha

[austere ocean]

MC-902 The end obsidian platform resets every time entities go through the end portal, which can cause blocks to be deleted
damn lmao

[obsidian tide]

wait, that was a thing?

[alpine pelican]

Minecraft 1.21 Pre-Release 1

We're bringing you the first Pre-Release of 1.21 with leashable boats, portal-able Ender Pearls and bug fixes!With this snapshot, we are temporarily removing the functionality of lighting blocks on fire using the Fire Aspect enchantment, as we want to refi...

May 29, 2024, 2:01 PM

[lone tree]

this

[austere ocean]

ye

[primal spade]

OMFG

[obsidian tide]

so they broke obsidian farms?

[hard ivy]

1.21-pre1

https://mdcfe.dev/mc-changes?ver=1.21-pre1
https://www.minecraft.net/en-us/article/minecraft-1-21-pre-release-1
https://github.com/neoforged/Snowman/commit/4dded6afd88d6f43d37e18085aa8cb05dd2aa1ee

[primal spade]

No way

[austere ocean]

they did afaict

surprise!

[obsidian tide]

I mean, technically, lava is infinite since dripstones exists, so you can recreate it via that

it's just slower

although, don't the towers respawn if you respawn the dragon?

So I guess there's still other ways

[austere ocean]

there's always another way

[obsidian tide]

but being able to just stand besides the spawn platform and occasionally drop an item through the portal was nice ¯_(ツ)_/¯

[austere ocean]

what

[digital scaffold]

They might still regenerate it btw, just make it only replace air

[lone tree]

what is renderFace?

[austere ocean]

I don't have the diff on hand

[digital scaffold]

Ocelot's still decompiling for me

Ok it's done

Diff time

Finally

[light arrow]

MC-902 The end obsidian platform resets every time entities go through the end portal, which can cause blocks to be deleted
welp there goes obsidian farming

[austere ocean]

supposedly it still works in this prerelease?

[austere ocean]

h

[jagged stag]

people really need to start reading the backlog

[digital scaffold]

Lol

- "0xBADFOOD",
+ "0xBADF00D",

[austere ocean]

[digital scaffold]

openUrl is no more

It is now openUri

And obviously takes URI instead of URL

openFile is also openPath now

[obsidian tide]

so I we can open files?

[jagged stag]

you we can open files

[austere ocean]

it would appear that they changed the end platform so it doesn't replace blocks, just air

[obsidian tide]

[austere ocean]

which is what the bug report was about, I guess

cool prerelease either way

[fallow horizon]

model loading got a refactor that probably won't affect many people but affects me greatly

[austere ocean]

1.21 is on the horizon :3c

[obsidian tide]

time to build an obsidian box filled with lava

[fallow horizon]

of note: ResourceLocation is now final and ModelResourceLocation is now a record that wraps ResourceLocation

[jagged stag]

    public static final ReportType CHUNK_IO_ERROR = new ReportType(
        "Minecraft Chunk IO Error Report",
        List.of(
            "I have failed you!",
            "Let's not do it again...",
            "Worst magic trick ever!",
            "Remember to backup your worlds regularly",
            "Pirates stole your chunk!",
            "Ker-chunk!",
            "Ideally, this shouldn't be here",
            "Let's hope it wasn't anything important",
            "Computers were a mistake",
            "Welp",
            "Not my proudest moment",
            "Who needs blocks in a block game, right?",
            "This chunk is no more...it has ceased to be...this is an EX-chunk",
            "loss.mca"
        )
    );

[hard ivy]

Ender Pearls that are thrown through portals can now teleport players to a different dimension than the one they are currently in
the possibilities

[red sigil]

what the... pre1 already

[jagged stag]

[hard ivy]

With this snapshot, we are temporarily removing the functionality of lighting blocks on fire using the Fire Aspect enchantment, as we want to refine it further. This functionality will be reintroduced in a later release.
wait what

I didn't know they added that

[hard ivy]

IT BEGINS

[light arrow]

ohno no more pass any one of them

[fallow horizon]

yeah I actually rely on that iirc

so this will be fun

[prisma smelt]

this sucks

[light arrow]

call monica here

[hard ivy]

@gusty flower i hope you like sequenced maps

[red sigil]

wth is a sequencedmap

[digital scaffold]

Java 21 feature

It's a map that is also like a queue

LinkedHashMap is the primary implementation

[hard ivy]

it's a map that retains the order of the key-value pairs

the sequence, mayhap

[red sigil]

A Map that has a well-defined encounter order, that supports operations at both ends, and that is reversible. The encounter order of a SequencedMap is similar to that of the elements of a SequencedCollection, but the ordering applies to mappings instead of individual elements.

The bulk operations on this map, including the forEach and the replaceAll methods, operate on this map's mappings in encounter order.

[hard ivy]

it's part of the sequenced collections feature

[digital scaffold]

Well not necessarily retains, more like just has a defined order

[red sigil]

does LinkedHashMap implement this interface now
right yes they do

All Known Implementing Classes:
ConcurrentSkipListMap, LinkedHashMap, TreeMap

[digital scaffold]

For example, SortedMap extends SequencedMap

[hard ivy]

LinkedHashSet inherits from SequencedSet

[digital scaffold]

It doesn't retain insertion order, but it does have a defined order (which is sorted order)

[hard ivy]

no more "okay, time to iterate this set just to get the last element"

[red sigil]

well-defined order is nice

it means you can tell it won't change under your nose

[austere ocean]

what difference does it make in this case?

idk enough about the rendering system

[light arrow]

what impl do they use to initialize it?

[red sigil]

i would imagine it's a linkedhashmap, but I haven't checked

[digital scaffold]

I think it was a sorted map from fastutil

[obsidian tide]

their fix is this

[austere ocean]

hm

[digital scaffold]

So they're broken instead of deleted now

Good fix 👍

[turbid hatch]

just tested and it breaks all blocks but drops them as items

[obsidian tide]

yeah

[turbid hatch]

right

[obsidian tide]

lemme test fluids

[turbid hatch]

I should I have read more

also when you kill the ender dragon the portal drops 2 torches not sure if that was always a thing lol.

[gusty flower]

Interesting 🤔

[red sigil]

immediate() uses a sorted map

[obsidian tide]

ok, water just gets yeeted still

[light arrow]

automatic concrete farms with end platform breaking... hello

[obsidian tide]

which I guess is the best they can do

[red sigil]

while RenderBuffers uses a linked hashmap

[light arrow]

they could drop the old water item

[obsidian tide]

[obsidian tide]

automated cobble farm

but for items

and without tnt duping

[gusty flower]

Did any other code change that uses that sequenced map, or was it just the type change?

[digital scaffold]

So this allows you to instabreak any block in the game?

[austere ocean]

[digital scaffold]

Does this work on unbreakables?

[obsidian tide]

[austere ocean]

in what situation can you have an unbreakable in the same position as the end platform lmao

[digital scaffold]

Oh I just saw cursed code

if (switch (c) {

See EntitySelectorParser

[hard ivy]

say hello to switch expressions

[red sigil]

nope, seems they just want .keySet() to have predictable ordering

[austere ocean]

[obsidian tide]

bedrock just gets yeeted

[austere ocean]

good stuff I guess
lmk if any other massive internal changes are hiding somewhere :p

[obsidian tide]

wait, I can do that now?

[austere ocean]

but I need to get back to what I'm supposed to be working on

[red sigil]

you could do that in java17 already

[gusty flower]

Weird. I wonder if there's another refactor coming later down the line which actually takes advantage of the sort order.

[obsidian tide]

I didn't realize, or forgot

[red sigil]

I have been using switch expressions for ages in my mods lol

[obsidian tide]

but I can do the instanceof one with J21 now, right?

[jagged stag]

or both?

[red sigil]

no idea about pattern matching switch in java

my brain goes to C#

[agile canyon]

For some reason I'm not able to pull it...

Time to clean everything!

Apparently I can't find lwjgl-freetype-3.3.3-natives-macos-patch

Agh

[half plaza]

Leashable.tickLeash((Entity & Leashable)this);

Apparently this is valid java code

[hard ivy]

ah yes, good ol intersection type

iirc the name

[austere ocean]

the bitwise and of Entity and Leashable :p

[hard ivy]

for the curious, that's JLS (21 SE) 4.9, Intersection Types

[digital scaffold]

Wait it is? I didn't realize intersection types were allowed in casts

[obsidian tide]

for type bounds?

[digital scaffold]

Huh, apparenlty they are

[hard ivy]

(from JLS 21 SE 15.16, Cast Expressions)

[obsidian tide]

yeah, so type bounds

[digital scaffold]

I thought it was only allowed in generic bounds

[hard ivy]

probably for cases when <T extends Entity & Leasable> void accept(T leashableEntity)

[obsidian tide]

yeah, that's what I'm reading too

[agile canyon]

Don't know why, but I love that

[hard ivy]

so if (e instanceof Entity && e instanceof Leashable) thing.accept((Entity & Leashable) e);

[obsidian tide]

but sadly no e instanceof Entity & Leashable 😔

[austere ocean]

which is apparently Dd`a`a

[agile canyon]

What is Mojang's maven repo?

[digital scaffold]

https://libraries.minecraft.net

[fallow horizon]

libraries.minecraft.net, I think?

[agile canyon]

Cool

[obsidian tide]

can't browse it tho

[agile canyon]

yeah, and that didn't seem to fix it...

[twilit slate]

errr this is what i get for having a lot of work

[agile canyon]

Guess I'll just download the resource manually

[lone tree]

when data-driven blocks?

[digital scaffold]

TIL CommonLinks existed

[austere ocean]

maybe the real data-drivers were the codecs we made along the way

[digital scaffold]

All the constants used to be inlined, so I never knew it existed

[agile canyon]

And that's still not working...

[digital scaffold]

The attribute mining_fatigue was spelled minining_fatigue

[agile canyon]

I think I need to update gradle's cache manifest

[primal spade]

FINALLY

This is great

[fallow horizon]

dynamic resources is in shambles

[primal spade]

I'm sure you'll find a way

[fallow horizon]

I'm probably just going to rewrite it

think it can be done cleaner now with Mojang's refactors

are there instructions on how to properly update neoform patches?

it seems to have four folders instead of one

[obsidian tide]

why isn't res loc a record yet?

[digital scaffold]

Forced public constructor

Basically because this is Java and not Kotlin

[viscid crane]

Any noteworthy changes?

[obsidian tide]

I don't see how that's a problem?

[viscid crane]

on a technical and code side

[digital scaffold]

Well ResourceLocation has a private primary constructor, you see

But records can't have that

[red sigil]

I wish they allowed records to have private constructors

java had this whole idea of "data classes"

[digital scaffold]

Me too

[red sigil]

which was designed for optimal jvm performance

and best encapsulation

[digital scaffold]

Both Kotlin and Scala allow private constructors for their data classes

Why can't Java?

[red sigil]

and the rules were like:

1. all fields must be data classes or primitive types
2. superclass must be another data class (or object)
3. no public constructor, only factory methods

and they go and make records incapable of point 3.

[obsidian tide]

unless I'm seeing something wrong, the string string one is public?

[digital scaffold]

No, it's private since last snapshot

It's all factories now

[obsidian tide]

ah well, I'm not looking at the latest snapshot

[viscid crane]

I don‘t think that first point is satisfied by records either, no?

Members can be all sorts of objects

[digital scaffold]

It's not enforced by the language, but you can enforce it on yourself

3 is literally impossible in the language though

[obsidian tide]

make the public one throw

[digital scaffold]

Then you have no constructor

[viscid crane]

You can‘t override the default public constructor

[half plaza]

Turns out not even the developer of vineflower knew that you could cast to intersection types

[digital scaffold]

In Ocelot, it decompiles as (Entity)((Tameable)entity), yeah

[primal spade]

Can you?!

[digital scaffold]

Apparently yes

We're all finding this out today

[viscid crane]

I did not know that either, wtf

[digital scaffold]

The Mojang dev was probably like "I wonder if this works"

[red sigil]

what's an intersection type?

[digital scaffold]

And it did

[mighty prawn]

coehlrich

[Reference to](#1233004919322054678 message) #1233004919322054678 [➤ ](#1233004919322054678 message)```java
Leashable.tickLeash((Entity & Leashable)this);

Apparently this is valid java code

May 29, 2024, 2:47 PM

[white pond]

Ah I was about to see if you already did sth coehlrich hehe

[digital scaffold]

MyClass & MyInterface & MyInterface2

[red sigil]

oh huh

[obsidian tide]

for <T extends X & Y>

[red sigil]

how new is that?

[fallow horizon]

@half plaza is the normal update procedure for neoform to just delete broken patches and remake them from scratch?

[obsidian tide]

J21

afaict

[viscid crane]

that is really useful to know

[red sigil]

aha, that explains it

[digital scaffold]

Well intersection types have been around since generics I think

The ability to cast to them, Idk

[viscid crane]

wait, can you instanceof intersection types too?

[red sigil]

I definitely tried in the far past and didn't work

[viscid crane]

[digital scaffold]

Wait what?

[obsidian tide]

no

[red sigil]

I think that's a question

[viscid crane]

x instanceof Entity & Leashable z

[viscid crane]

unfortunate

[digital scaffold]

No you cannot

[half plaza]

Basically yes but also modify patches that only require simple changes

[viscid crane]

tragic indeed

That would be cool

[digital scaffold]

It would be, but alas

[austere ocean]

TIL that exists

[red sigil]

how would it work, two instanceof?

[digital scaffold]

Yeah

[obsidian tide]

wait, according to the JLS, that was a thing since J8?

[red sigil]

also isn't string final? so that would be illegal regardless

[digital scaffold]

It would compile to basically x instanceof Entity z && x instanceof Leashable

[digital scaffold]

You can still do instanceof String

[viscid crane]

Then z wouldn‘t have leashable stuff accessible though

[red sigil]

yes but String & Something would never be true

[digital scaffold]

String implements Constable

[red sigil]

ah

nevermind then

[obsidian tide]

also Comparable

and char sequence

[digital scaffold]

And Serializable and ConstantDesc

[austere ocean]

Kotlin has where syntax for requiring a generic type parameter to implement an arbitrary number of interfaces, apparently

[red sigil]

I just didn't realize constable was a java interface :P

[austere ocean]

there's also & Any but I think that's special syntax

[digital scaffold]

Kotlin doesn't support intersection types, so no

Only intersection with Any is allowed

[austere ocean]

thought so

[hard ivy]

what's the problem, officer constable?

[viscid crane]

how very bri*ish

[obsidian tide]

yeah, intersection casts existed since J8

[agile canyon]

Why is PacketReport a data provider?

[digital scaffold]

What else would it be?

Wait, is there another interface for reports?

[agile canyon]

I mean, it's just BlockReport but for packets, just seems weird in my head

[obsidian tide]

admin, please ban this block, they did something bad

[obsidian tide]

Doesn't

record Foo(int i, float f) {
  public Foo(int i, float f) {
    throw new UnsupportedOperationException("Use the factories");
  }
}
```do that tho?

[torn notch]

In what ways? Just the model RL recordifying and no longer inheriting from RL? Or more?

[hard ivy]

well, given that other constructors must call the canonical record constructor IIRC

[digital scaffold]

Well now you can't instantiate the record

[obsidian tide]

ah

[hard ivy]

Gaming32 'bout to prove me wrong

ah good, my memory isn't that bad yet

[obsidian tide]

but you can override the default constructor, it just has to actually do something

[obsidian tide]

I mean, with that you're just adding to it iirc, not "overriding" it

the parameter less version that is

[vapid gale]

The parameter list on the canonical constructor is implicit, you can access the parameters even without writing out the parameter list

[obsidian tide]

yes, but you can't change them

[digital scaffold]

That's extending it, but you can replace it entirely by specifying the parameter list (as I showed in my second example)

But you can't decrease the access

[obsidian tide]

like this.a = ... would be illegal iirc

[torn notch]

I do appreciate that a small part of my wish was fulfilled :p now they just need to go all the way

[obsidian tide]

yeah, I understand now

[hard ivy]

but you can change the parameters before they get assigned, iirc

[digital scaffold]

Yes

[agile canyon]

Teleports, stop changing please

[vapid gale]

It's not, which is why this works: https://github.com/neoforged/NeoForge/pull/283/files#diff-23d3f7ba78584f5642bf1a4b317efe179fd81e40ebc784203aa6071ba970e232R11-R17

[digital scaffold]

So uh, they removed the new gamerule

But they not only didn't remove the translations, but they fixed their grammar

[hard ivy]

Mojang's going all-in on teleports

[obsidian tide]

wait, is parameter less version executed before or after assignment?

[digital scaffold]

Before

[obsidian tide]

because this doesn't work

[digital scaffold]

Yeah, that doesn't work because that would be a double assignment to the final field

Doesn't matter that it's before the generated one

[agile canyon]

Leashable is probably going to need some extensions

You can leash a boat?

[lone tree]

yeap

bedrock parity

We're bringing you the first Pre-Release of 1.21 with leashable boats, portal-able Ender Pearls and bug fixes!

[twilit slate]

What in gods name is that color scheme?
like i thought for a sec it was c#...

[agile canyon]

VoxelShape constructor now protected

And done with the primer: https://gist.github.com/ChampionAsh5357/d895a7b1a34341e19c80870720f9880f/revisions#diff-d98e96ce400ed97b94b31eb272da1aaf8e5bc302879c32f34168e708477578c6

[digital scaffold]

Custom

Well, it's based on VS Code

It's a port of my custom VS Code scheme to IntelliJ

[balmy estuary]

Where the MC article link?

[torn notch]

second line in those links

[balmy estuary]

Ah

How does one make a Resource location now?

[torn notch]

example in here

[balmy estuary]

Ah

[stray scroll]

Release date is June 13th

[odd flare]

thats like 2 weeks

before my internship even stops

[stray scroll]

https://youtu.be/qIFJ5c8laPg

[odd flare]

also, good to see you alive again XD

[hard ivy]

IT BEGINS

[true totem]

🚨 🚨

[digital scaffold]

It's pre when they leave the weekly snapshot schedule, right?

Or is that rc?

[austere ocean]

[mighty prawn]

https://hammertime.cyou/

[lone tree]

it would be nice to have a stable version of NeoForge 20.6 before then

[digital scaffold]

Still no 3D mace

[obsidian tide]

Do it yourself

[austere ocean]

<t:1718291640:R>

[red sigil]

that's way sooner than I thought. huh.
I expected first week of july at the earliest

[austere ocean]

two weeks

[red sigil]

yes, prereleases is when they start to release as soon as fixes are made

[lone tree]

The question is: hotfix or not?

[austere ocean]

first major MC release since the fork, right?

[white pond]

yes

[austere ocean]

neat

[white pond]

but really. no one cares about 1.20.6

at all

[austere ocean]

I care :(

[white pond]

will you still care when 1.21 is out?

[austere ocean]

no

[white pond]

See 😄

[austere ocean]

but I care right now :p

[white pond]

all will be well /handwave

[austere ocean]

mmm

Create is still planning to target 1.20.4 neo afaict, so I think it'll stick around for a while

[white pond]

Which one had item components? 1.20.5, right?

[austere ocean]

yeah the dev status page says they're not doing 1.20.6 (yet)

[austere ocean]

yup

[obsidian tide]

and J21, right?

[austere ocean]

yup (2)

[red sigil]

1.20.5 but 1.20.5 "doesn't exist"

it's just 1.20.6 :P

cos 1.20.6 was just a hotfix

[obsidian tide]

because llamas forgot what they were holding iirc

[austere ocean]

compromise nomenclature: 1.20.5½±½

:p

[red sigil]

and supposedly another bug that was in the security category and didn't show up in the changelogs

cos I heard it could be used for duping

[white pond]

But seriously, 1.20.5 is the big jump and while I also played in 1.20.4, it doesn't have much appeal to players over 1.20.1, right?

[red sigil]

1.20.4 never got any traction

modpack authors, and so players, used 1.20.1

[austere ocean]

I'll be interested to see how much Create changes that

[red sigil]

and the community has decided 1.20.x for x > 1 are betas for 1.21

so modpack authors probably won't care until 1.21

[austere ocean]

how will porting neo to 1.21 work? I forgor

[red sigil]

same as poting to any other update

edit build.gradle/gradle.properties to specify new numbers
rename mods.toml if you didn't already for 1.20.6
and then despair as you fix the compile issues

[austere ocean]

I meant more as in when can we expect 21 to be available :p

[red sigil]

currently the porting work is in the private repository, where the team are dealing with the patches

once it's in a state in which the patches can round-trip without issues, it will be moved to a public porting branch, where updates will be published for development purposes

an exmaple of this kind of branch is: https://github.com/neoforged/NeoForge/tree/port/24w14a

[austere ocean]

right

[red sigil]

these porting branches are NOT stable. they will break, and they will crash, because it's work in progress stuff

but sometimes they work well enough for your own mod porting efforts

[austere ocean]

will be interesting to see what happens wrt mod adoption

[digital scaffold]

Oh, you are on the Create team

Idk how I never noticed till now

[crystal umbra]

How much does create use item nbt?

[red sigil]

I know of at least one person that has ported to 1.20.6 by creating an NBT data component to hold their custom nbt stuff :P

cursed, cos it can't be mutable

[agile canyon]

CustomData is very much fun like that

[digital scaffold]

Oh right, those

I've probably spent over 6 hours porting those

[crystal umbra]

Do you not datagen recipes?

[digital scaffold]

Granted my mod is mostly based on advancements

[obsidian tide]

Hello

[austere ocean]

yeag
but whatever ends up happening will influence what versions people play on (although I'm curious to see how heavily)

[twilit slate]

Yeah but create is a mess of a code base. It is spaghetti all the way. Basically unusable

Start with the data gen classes ....

[austere ocean]

using branches for MC versions has always bugged me a bit, but I have no clue what the alternative might be

eventually it feels like maintaining multiple copies of the codebase with enough subtle differences to be annoying

[digital scaffold]

I personally use a preprocessor, but I realize that's not for everybody

But it allows me to support 14 version/loader combinations at once

[austere ocean]

the versioning scheme I use reflects this: <mc minor>.<mc patch>.<mod version>.<mc version-specific patches>
example: WAILT 20.6.1.1 is MC 1.20.6, mod version 1, with 1 version-specific patch

it's cursed, but it actually has semantic meaning

ah, the fuck-it-we-ball scheme

:p

[weary moon]

Use NoVersion™️

:P

[austere ocean]

version files with the Unix timestamp of their release

[austere ocean]

imo semver is not sem for MC mods unless the mod has an API

the website even says that a project using semver must have a public API, iirc

Software using Semantic Versioning MUST declare a public API. This API could be declared in the code itself or exist strictly in documentation. However it is done, it SHOULD be precise and comprehensive.

[obsidian tide]

dfu version schema

just a counter

but weird

[austere ocean]

such is the nature of mod versioning, I suppose

[digital scaffold]

I mean, it's just <major>.0.<ci build number>

Where major has the same meaning as SemVer

Breaking changes

[austere ocean]

the minor version is for moral support

[magic estuary]

World saves can be considered a public API -- in that case, changes that break a world or something built in a world would be breaking changes

[austere ocean]

[magic estuary]

Heck, that's basically what MC itself does

[obsidian tide]

that's if you'd actually use DFU

[steep copper]

DFU is off the table for mods, we've discussed that to death

[obsidian tide]

I mean, aiui, you can still set up your data fixers, but that's gonna be separate from the rest, sadly

[steep copper]

Many people. It has multiple issues that cannot cleanly be solved.

[obsidian tide]

yes ^

separate from the rest

[steep copper]

Issue #1: The vanilla schemas run ONCE per world.

No changing that.

Have fun only releasing one mod version ever for an MC version, that's not how it works in the real modding world

Hope it's perfect

I'm out

Not having this discussion AGAIN

[hard ivy]

what if we drag luke into this discussion

[obsidian tide]

you'd have to give everything a passthrough fixer iirc

[torn notch]

Lol. I should go compile the fabric discussion I dragged him into...

[magic estuary]

Exactly; hence why MC is still at 1.x -- DFU means that if you treat the saves as the public API, there's never been a breaking change

And yeah, you can inject your own datafixers linked to changes in Minecraft version with absolutely no issue -- hell, neoforge has a couple of these (that I wrote) to upgrade the old neo attributes to the corresponding new vanilla ones!

The part of "DFU for mods" you'll never be able to get working right is datafixers associated with mod versions within a single MC version

...and, well, any datafixers, vanilla or otherwise, working consistently inside modded containers

And the former issue -- datafixers to upgrade mod versions within a MC version -- you're not going to ever get working with DFU for a number of reasons. You end up with situations where DFU can't reasonably order schemas and fixers from different mods or vanilla, and where that leads to issues, no matter what your do -- that's the discussion on fabricord that Ratatosk mentioned; someone thought they had a cool new way to get that working, which would have been awesome, and unfortunately it didn't work for a variety of reasons, mostly tied up in the fundamentals of what makes DFU tick

[gusty flower]

Yeah, I've said a few times now that the best you're going to get really is registering schemas for the vanilla datafixers to look through when applying vanilla datafixes

[magic estuary]

Well, no. You can do better than that

You can register fixers to run on vanilla versoin upgrades

You just can't fix anything within a vanilla version

But if your mod format changes at the same time as an MC version change -- you can add a fixer for that

Takes a mixin because that all runs far before mod ctors run, of course

You can also register your stuff to the vanilla schemas (takes the same sort of early mixin entrypoint) -- which is useful if you want to support vanilla fixers running within your inventories, but is also... potentially a lot of work, and a maintenance burden as you have to keep it around and add new ones plus fixers if you change the format, etc.

[obsidian tide]

so components?

[magic estuary]

Yeah you can with some work add a fixer to move your old NBT into a component

Takes a mixin or two and some DFU knowledge

And won't work if your item is inside a modded container

[obsidian tide]

I guess I'll stick with verifyComponentsAfterLoad

because that last point is the thing I really don't like about dfu

[magic estuary]

Well I mean how else would it work? It's gotta discover that stuff is in a nested structure somehow -- and the only way that'll work is for whoever's adding the container to register that structure in a schema

Which they can do -- it's one mixin -- but there's then a maintenance burden associated with that in terms of keeping it up to date and adding fixers if you change it

[obsidian tide]

yes, I know, doesn't mean I have to like it

[magic estuary]

You can do that for a single mod's stuff, but it fails when you have more than one mod, or modded stuff and vanilla stuff, that is in the same data structure and needs fixed

Because for that to work it's all gotta be in the same schema chain

And you can't sensibly "merge" disparate schema chains

So, if you have a format that contains entirely your own data -- nothing from other mods or vanilla that'll need fixed -- you can totally use datafixers for it

Heck, I've got a system that uses datafixers to update config files this way

But if, say, you change the format that items are stored in your BE in the middle of a single vanilla version -- and another mod changes the way their item data is formatted in the middle of a single vanilla version -- you can't built a system with DFU that will sensibly apply both updates in all cases, without the two mods being explicitly aware of eachother

[primal spade]

Realistically you should be able to

Both orders should work?

[red sigil]

I assume we'll see another prerelease today

[alpine pelican]

New version detected: 1.21-pre2.

[odd flare]

@red sigil you were right!
sound the alarms!

[alpine pelican]

Minecraft 1.21 Pre-Release 2

We're rounding out the week with the second Pre-Release of 1.21, with a lot of bug fixes. In case you did not catch the news, the full release of 1.21 is planned for the 13th of June!Happy mining!ChangesEnd Crystals are now immune to fire damageAdded an at...

May 31, 2024, 1:14 PM

[twilit slate]

<@&1067092163520909374>

1.21 Pre-2

https://www.minecraft.net/en-us/article/minecraft-1-21-pre-release-2

[lone tree]

🚨

[turbid hatch]

ominous big 1.21 Pre-2

[lethal pivot]

wtf does attestation mean

oh attesting

proof

right

[steep copper]

End platforms are now features o.o

[light arrow]

Added new placement modifier type fixed_placement to place features in a set of fixed positions

hmmm

[sudden oyster]

This is to fix the bug from the Pre1 where the crystals instantly exploded once entering the End as they were damaged by the fire below them.

[steep copper]

lol, imagine that

[enters end]
[kaboom]

[digital scaffold]

I bet this is ProGuard removing the useless pop

Because this would compile to aload_0; aload_1; invokevirtual; pop; return;

[uncut river]

any big changes guys?

[obsidian tide]

[digital scaffold]

New block tag

[austere ocean]

• MC-272596 - Throwing an ender pearl into the end fountain skips the credits
  

[digital scaffold]

minecraft:mob_interactable_doors

[austere ocean]

do you reckon they finally made the portal respawn code less cursed? or just added special handling for the pearl

[lone tree]

Enderdragon death rendering?

[digital scaffold]

They already made it less cursed a few weeks ago

[austere ocean]

lovely

[digital scaffold]

So what they did to fix the ender pearl thing is simply make it so the ender pearl can't go through the end portal if its owner hasn't seen the credits

So it just did like before and teleported the player so they're inside the portal, and they go through normally

[austere ocean]

makes sense

the real question is how, while implementing the end platform changes, nobody at mojang noticed the crystals instantly exploding :p

[digital scaffold]

Probably a test world without crystals

With the pearl not going through the portal, you'll respawn at your bed/anchor

I think if you have seen the credits already, you'll respawn at world spawn

Because I think that's where the pearl will go

Finally AABB.getMinPosition and AABB.getMaxPosition are a thing

[lone tree]

Does this mean better culling of block entities?

[steep copper]

Missed that. Niiiice

[digital scaffold]

No? It's just more API

[lone tree]

sad

[digital scaffold]

Has nothing to do with BEs

[sudden oyster]

Is this for using doors (villagers) or breaking them (zombies)? I guess the first but just want to be sure.

[digital scaffold]

Anything using InteractWithDoor, lemme check

So villagers, piglins, and piglin brutes, it appears

[agile canyon]

Take your primer: https://gist.github.com/ChampionAsh5357/d895a7b1a34341e19c80870720f9880f/revisions#diff-d98e96ce400ed97b94b31eb272da1aaf8e5bc302879c32f34168e708477578c6

[hard ivy]

IT BEGINS CONTINUES

[austere ocean]

[fallow horizon]

send help, this is pain

finally in-game though

[twilit slate]

spawn an ender dragon and use a furnace...

[primal spade]

AO is fucked

[fallow horizon]

yep

[vapid gale]

In vanilla or just in Neo?

[fallow horizon]

I highly doubt Mojang would release a pre in this state but I'll check vanilla

vanilla is fine

[vapid gale]

I just checked your update commit: regarding the mutable fields we used to patch into BlockElementFace, I would just patch MutableObjects into the record, recreate the old constructor and initialize the mutable objects as empty

[fallow horizon]

Alright

I considered doing it that way but wasn't sure if it'd be considered ugly

will do

[digital scaffold]

Oh speaking of constructors

ServerStatus has an extra isModded param, but no Vanilla constructor that passes a default value

I might PR that rq

[fallow horizon]

💀

lack of parchment is not helping

[blissful nebula]

I don't really see the need for this

[digital scaffold]

Replied in #neoforge-github

[fallow horizon]

Where else would blocks rendered as items get their light from besides the packed light provided by the item renderer?

[digital scaffold]

Btw why are SRG param names still used? Why not just do like locals and let the decompiler fill in a name?

[steep copper]

Level context

[fallow horizon]

haven't been able to figure it out and need to do other things, my progress is on 1.21-pre2 on Kits if someone else wants to debug

[vapid gale]

I'll take a look in a bit

[magic estuary]

They won't because schemas, and how DFU discovers nested data structures. If its all happening within one vanilla schema update, you're fine -- that's the case where you change stuff alongside a vanilla version change

But otherwise, no, it doesn't work and there's no way to fix it with nested structures

[primal spade]

Well I am thinking of a parallel versioning system where multiple mod changes get turned into independent version bumps

[magic estuary]

Yeah you can't get that to work

Because you don't know whether an update in mod A should be applied before or after an update in mod B -- and if those are separate version bumps that ordering does matter

[primal spade]

No it doesn't matter

If the mods are independent both orders should work

Imagine "move stack to other part of NBT" and "change stack". These can be applied in any order

[magic estuary]

...No. Just no. Remember that data can be nested and that one mod's data can be nested inside anothers and that schemas, which define how that nesting works, are versioned right alongside fixes. Someone suggested exactly what you did back in fabricord a while back and went and tested it and sure enough, it had the issue I'd thought it would

Gimme a sec to write up an example versioning chain and why it causes issues

[primal spade]

Idk, it seems possible on a conceptual level but I don't care enough to investigate it or even support it in my mods should it actually work

[magic estuary]

It's impossible on a conceptual level.

Mod A:

1. add display_rack, add to schema telling DFU that its data is nested here in stuff field.
2. move display_rack item data to items field. Change NBT format of gadget item

Mod B:

1. add pedestal BE, add to schema with single item in displayed field
2. move pedestal item data to items field. Change NBT format of magic_wand item

Now, when upgrading from Mod A, Mod B versions 1, 1 to version 2, 2 -- which set of fixes do you apply first? Lets say you choose to directly layer the schemas -- so the lowest schema of one mod is parented by the highest of another. If you apply mod A first, it relocates both BEs fine -- but any gadgets inside of a pedestal are not fixed. Similarly, if you apply mod B first, magic_wands inside of display_racks are not fixed. If you don't directly layer the schemas -- say, you're looking for some way to merge/interweave them that gets rid of this problem -- then you run into the issue of where you decide to line them up. The sensible pick -- and the only one that solves this specific issue -- would be the mod versions of the data you're trying to update -- but this has a few problems; (a) huge performance issues as you have to do this separately for every chunk/world/whatever, and/or keep a lot of cached stuff around, (b) interleaving this with vanilla version updates is basically impossible, and (c) you lose out on datafixer optimizations, because you can't feasibly re-run them for the whole mod+vanilla stack for every chunk, so datafixing is now slow

Building a DFU-like system that did this is totally possible, yeah

You just lose out on all the stuff that makes DFU advantageous to begin with -- notably, the performance benefits you gain from having a single, guaranteed chain of schemas you can build and optimize early on

And, cause DFU is only performance-sensible because of that advantage, you really wouldn't want to use DFU for that

[primal spade]

I don't think I know enough DFU to understand the difference between DFU and DFU-like

[fallow horizon]

Mod versions would generally be deterministic across many chunks so couldn't an optimized DFU structure just be cached per mod version set

[magic estuary]

Yes-ish. That is to say, yes, but in a large modpack with, say, 100s of mods and updates updating only a few mods at a time... you could end up having to build, optimize, and cache hundreds or perhaps even thousands of such structures

Plus, you can only build that when loading a chunk -- so it gets moved to world load

[fallow horizon]

DFU optimization takes 1 second for vanilla, and I don't think the classloading is that expensive either

It's also only necessary to actually build a new DFU structure if a mod did an update that requires DFU

[magic estuary]

My point is that if you do that, you've lost literally all benefits of DFU. Just use some custom system meant for parallel versioning chains at that rate

[fallow horizon]

I suppose

so your conceptual argument is it's better to make a simpler DFU replacement in Neo itself

[magic estuary]

My argument is that doing this with DFU removes the major benefit DFU gives you and results in having to store a lot of these structures

And doing it outside of DFU makes it a pain to interact nice with vanilla fixers

...so I suppose my actual argument is that modders wanting to both maintain world compat and change their data structure should do it on a vanilla version update so they can just add to the vanilla schemas

Because that's way simpler, it works for 99% of the few cases where people actually want to do this, and there's no worried about parallel versioning woes

[vapid gale]

Found the issue, it's a really sneaky one: the VertexFormat#offsetsByElement array is indexed by the element's ID and not the element's position in the format's element list, so the element offsets in the offset fields in IQuadTransformer were wrong, which caused the baked lighting to be applied incorrectly in VertexConsumer#putBulkData()

[fallow horizon]

good catch

how long did this take you to find

[primal spade]

30 seconds compilation time included

[vapid gale]

About 10 minutes or so 😄

With VertexConsumer not having an endVertex() method anymore, the current implementation of QuadBakingVertexConsumer with the automatic "quad export" to a consumer when the fourth vertex is completed doesn't seem feasible anymore. All current uses of it use the buffered sub class (which is IMO a more sensible approach anyway). I'm considering removing the buffered subclass and making the primary implementation act as "buffered" instead. Any objections?

[primal spade]

We could auto export on vertex begin, however I do not object

[vapid gale]

That would work, but you'd always have a problem on the last quad and starting a new vertex just to force the export seems like an awful solution to me

[primal spade]

Indeed

No objections 😄

[vapid gale]

👌

[gusty flower]

Wait so I predicted that?

[torn notch]

hopefully that becomes a recurring element to updates

[fickle atlas]

In case Mojang sees this, UnbakedModel.getDependencies() and ModelBakery.loadingStack seem entirely unnecessary as for the former, all dependencies are lazily loaded anyway and for the latter, ModelBakery.getModel has no way to call itself recursively, so the stack will always be empty before and after the method is called.

[stable vessel]

I just want to point out, mods can do custom model loading, and do so.
This may also cause getModel to be caused recursively.

[fickle atlas]

getModel does not work recursively for multiple reasons

[red sigil]

neo/forge model loading is meant to work so that you use the provided modelGetter within resolveParents

[fickle atlas]

How is that related? That's a vanilla API.

[red sigil]

the provided modelGetter is literally ModelBakery::getModel

[fickle atlas]

Yes

[red sigil]

I was just pointing it out because custom model loading was mentioned

I think I realize that confusion (a bit late though)

that was primarily in response to what JTK said

it wasn't relatedto getModel being recursive or not

[lone tree]

https://fxtwitter.com/HeartlessMoesh/status/1797567326155690105

https://fxtwitter.com/HeartlessMoesh/status/1797567201240633503

[red sigil]

new prerelease today?

[lone tree]

With 9 bugs fixed?

[red sigil]

sure

1.21 is supposed to release in just over a week

that means we should be seeing development wind-down

less fixes each prerelease

I would expect a prerelease today, maybe another one on friday and RC1 on monday/tuesday, if not RC1 on friday

[light arrow]

I think rc1 friday because thursday next week is release (so they have time for an rc2 on mondey if they need to)

[lone tree]

And the hotfix

[lone tree]

*18 bugs fixed

[unreal timber]

that'll be 1.21.1 on the 14th

[odd flare]

can we please get a 1.21.10?

[light arrow]

1. we are in pre releases so no regular snapshots
2. wednesday is frog day

[red sigil]

https://www.youtube.com/watch?v=nZXRV4MezEw

[red sigil]

I guess that was a no. :P

[lone tree]

Pre-release today or tomorrow

[alpine pelican]

New version detected: 1.21-pre3.

[odd flare]

oh boi

[lone tree]

Nice

[odd flare]

@twilit slate sounce the bell

[alpine pelican]

New version detected: 1.21-pre3.

[half plaza]

And it's only 10pm in my timezone

[weary moon]

1.21 Pre-3

https://www.minecraft.net/en-us/article/minecraft-1-21-pre-release-3
https://mdcfe.dev/mc-changes?ver=1.21-pre3
https://github.com/neoforged/Snowman/commit/9c1f9ae0ca0cb5a8e93ef04aff99595dec25a35e

<@&1067092163520909374>

[median jungle]

Ohno

[jagged stag]

christ tis an early one

[odd flare]

\O/

[lone tree]

Pin this

[red sigil]

oooh

early pre

[alpine pelican]

Minecraft 1.21 Pre-Release 3

The third pre-release for 1.21 is now available and it comes with even more bugfixes!Additionally, we introduced some intentional but undocumented changes to the WindBurst Enchantment in Pre-Release 1 last week. To remedy this, we're including those change...

Jun 5, 2024, 9:59 AM

Minecraft 1.21 Pre-Release 3

The third pre-release for 1.21 is now available and it comes with even more bugfixes!Additionally, we introduced some intentional but undocumented changes to the WindBurst Enchantment in Pre-Release 1 last week. To remedy this, we're including those change...

Jun 5, 2024, 9:59 AM

[obsidian tide]

ater, fix your message, it's 3 not 2

[weary moon]

Snowblower doesn't look like it's currently running
I think someone needs to run the action on the repo now? I'm not sure what the current procedure is since @hard ivy changes

[weary moon]

fixed

[obsidian tide]

https://cdn.discordapp.com/emojis/731993501717495859.gif?size=48

Additionally, we introduced some intentional but undocumented changes to the WindBurst Enchantment in Pre-Release 1 last week. To remedy this, we're including those changes in this changelog instead.

[half plaza]

https://github.com/neoforged/NeoForm/actions/runs/9382118210/job/25833318258#step:4:169
1 compile error

[primal spade]

We were on the verge of greatness

[agile canyon]

Welp, I’m going back to sleep

I’ll deal with this later

[twilit slate]

that was way too early in the morning... unfair

[austere ocean]

guh, I just woke up

[agile canyon]

Minor changes: https://gist.github.com/ChampionAsh5357/d895a7b1a34341e19c80870720f9880f/revisions#diff-d98e96ce400ed97b94b31eb272da1aaf8e5bc302879c32f34168e708477578c6

[crystal umbra]

Ash I have a question to sate my curiosity: how do you make the primers? Do you actually look at the diff in day snowman and go through it all or is there some other strategy you employ

[agile canyon]

I use vanilla gradle to download the sources and then use IDEA's diff comparator. But yeah, basically, I read the entire diff, document any public changes (and private ones if I believe they're highly subject to mixin) and then organize it by importance

And then, because I get annoyed sometimes, I take it out by writing goofy headers

[jagged stag]

what's the state of the kits branch

[vapid gale]

Main compiles and runs, patch gen is broken (spits out the ZIP instead of the unpacked patch files) and the test sourceset is a mess of errors

[jagged stag]

so the usual

[light arrow]

as long as patches are broken we can't push to a port branch... who broke it

[jagged stag]

NG

probably

[half plaza]

generateClientBinaryPatches tries to read the patches directory as a regular file

[light arrow]

unpackPatches should not need the bin patches

[half plaza]

but publishing does

[light arrow]

we don't need publishing on the port branch in the beginning

[red sigil]

aren't ports branches ci-enabled? unless you are ok with the publish failing

[light arrow]

it is fine if publish fails in the beginning

[half plaza]

unpackSourcePatches currently just generates output.jar in the patches directory

[light arrow]

uhm.. @white pond @blissful nebula what happened there?

[blissful nebula]

Yes I am aware

But I have had a problematic week

I need to fix it

Just you know

No time yet

I am unsure what it is

I hope to have some time in an hour

[red sigil]

don't spread yourself too thin, if there's no time there's no time. at the end of the day porting early is just a "nice to have"

[blissful nebula]

Yeah, my best guess is: I fucked up something with the caching

Where it wants to centrally cache that task output

Which only supports files (because it was written for userdev tasks)

And now it derps XD

[half plaza]

https://github.com/neoforged/NeoGradle/blob/602139fd9143709af1a9589dd4f633ab8dc12a87/platform/src/main/java/net/neoforged/gradle/platform/extensions/DynamicProjectExtension.java#L253
Shouldn't this be using zipTree?

[blissful nebula]

Probably

[hard ivy]

anyone willing to placate the Elephant while Orion's away

[hard ivy]

it seems like it, yea

currently, it just visits the zip itself and copies it over

willing to test it and make a PR?

[blissful nebula]

I know why that broke

I blame maty for that one

And his jst rp

Pr

He bricked like 6 of those

So I missed one during fixing

needs to create a mini test setup

[weary moon]

Updated pin with link to Snowman commit

[half plaza]

https://github.com/neoforged/FancyModLoader/commit/4307030113a94ed52246dcbf7e02464cb2be178c
Seems to have broken some of the neogradle tests

[white pond]

in what way

[half plaza]

It currently checks for Caused by: net.neoforged.fml.ModLoadingException: Loading errors encountered: [\n\tfml.modloading.brokenfile\n] in a couple of tests but now neoforge logs this:

Caused by: net.neoforged.fml.ModLoadingException: Loading errors encountered:
    - File /private/var/folders/lr/439_fwvd3m76p9vy50d57kcc0000gn/T/spock_configuring_of_the__0_tempDir15627936413731285732/runs_configuration_after_dependencies/build/resources/main is not a valid mod file

[blissful nebula]

Yes I am aware

I am about to push a PR that fixes those

[half plaza]

generateClientBinaryPatches still seems to be broken with my pr due to: https://github.com/MinecraftForge/BinaryPatcher/blob/master/src/main/java/net/minecraftforge/binarypatcher/Generator.java#L100

[blissful nebula]

Are we still using MFs BinPatcher?

[half plaza]

and it supplies every patch file as a seperate parameter

[half plaza]

Seems like it according to this stacktrace line: at net.minecraftforge.binarypatcher.Generator.lambda$loadPatches$10(Generator.java:122)

[jagged stag]

the installer uses the neo one iirc

[half plaza]

@blissful nebula https://github.com/neoforged/NeoGradle/blob/NG_7.0/common/src/main/java/net/neoforged/gradle/common/caching/CentralCacheService.java#L287
This currently just hashes file inputs without checking if they're a directory.
Should it hash all the files in the directory if the file is a directory?

[blissful nebula]

That is a very very good question, I would say so yeah

[light arrow]

I'd say tomorrow rc1

[unreal timber]

we should bet on when it will be and the closest person gets a "Snapshot Guesser" role until the next guess.

[obsidian tide]

I vote <t:1717755180:f>

[unreal timber]

if no one else guesses, you'll win lol

[obsidian tide]

win by default is boring tho

[vapid gale]

I vote <t:1717765200:F>

[unreal timber]

I vote <t:1717712460:R>

[obsidian tide]

ah yes

I guess if they release before 6 you win

[unreal timber]

6:21 to be specific

[light arrow]

that's in the middle of the night for mojang

[unreal timber]

I don't trust them not to pull something like that

[obsidian tide]

everyone can vote what they thing no matter how likely other people believe it to be ¯_(ツ)_/¯

[austere ocean]

I vote at any point tomorrow :3c

[agile canyon]

Can we use timestamps please? Figuring out these times is gonna be a pain

[austere ocean]

that would be <t:1717768800:F>

[agile canyon]

Let's go with this <t:1717764983>

[light arrow]

I say 5 minutes after that

[austere ocean]

hang on I need to write these down

keeping yall accountable

[red sigil]

<t:1717759200:f>

[austere ocean]

I do not like spreadsheets very much

[unreal timber]

oh no. what have I started lol

[austere ocean]

at least there's not money on the line

[unreal timber]

that can be fixed /j

[light arrow]

@TeamMember we're starting a snapshot lottery, whoever is closest wins the pot /s

[agile lantern]

<t:1717772400:f>

[true totem]

i bet 30 minutes after that ^

[austere ocean]

your bets... have been noted

[torn notch]

37 minutes after Adrian :p

[austere ocean]

aaaagh stop making me do math, I'm a programmer :p

[unreal timber]

which means if adrian wins it will have to be almost exactly on time lol

[uncut river]

how do yall type time zones like this?

[light arrow]

bets are closed and locked in in 1 hour

[digital scaffold]

https://hammertime.cyou/

[lone tree]

<t:1717767600:f>

[mellow nexus]

<t:1717767660:f>

price is right, baybe

[light arrow]

no longer valid you missed the deadline

[red sigil]

hm?

the timestamp shows in 3 hours here?

clock says 12:35, timestamp says 15:41

oh nevermind

deadline for submission

[austere ocean]

today's the day :3

I'm noting all the bets atm

[light arrow]

Caltinor, PlatinPython are already out, currently closest is gigaherz

[austere ocean]

indeed

[odd flare]

wait
we were making a prediction?

I will say <t:1717770615:f>

[austere ocean]

you're too late, betting closed a few hours ago

[odd flare]

ah

then unofficially, I play too, but I will lose anyway

[austere ocean]

alright, as of now platin, caltinor, and gigaherz are out

[obsidian tide]

no more pings

[austere ocean]

how was I meant to know you have a custom ping :p

[light arrow]

huh platin has a custom ping? hmm

[obsidian tide]

yes

[red sigil]

your own fault for having it :p

[obsidian tide]

yes 🙃

[odd flare]

does python also still work

[obsidian tide]

yes

[austere ocean]

made a spreadsheet with conditional formatting stuffs

[remote void]

Quick, start a python convo in #squirrels-🦊

[obsidian tide]

I get enough of those already

[jagged stag]

paithon

[austere ocean]

[obsidian tide]

fuck

[jagged stag]

get rust

[austere ocean]

pythοn

legally distinct o glyph™️

pyth(Ctrl-Shift-u03bf)n

[sinful hazel]

well done for being the closest

👏

[austere ocean]

ope

sound the alarm

[fiery galleon]

https://www.minecraft.net/en-us/article/minecraft-1-21-pre-release-4
https://github.com/neoforged/Snowman/commit/449794a3aa808e04955757813a7ba59c7152a482
<@&1067092163520909374>

[austere ocean]

conglaturations, you done it

[twilit slate]

1.21 - pre 4

https://www.minecraft.net/en-us/article/minecraft-1-21-pre-release-4

https://github.com/neoforged/Snowman/commit/449794a3aa808e04955757813a7ba59c7152a482

[turbid hatch]

delaying the release until the person you want to win is closest 😉

[red sigil]

ooh was only 1 hour late

[austere ocean]

The "Burning Time" reduction effect of Fire Protection and "Knockback Resistance" effect of Blast Protection now stack from wearing several pieces
👀

[red sigil]

nice

[viscid rune]

After a Primed TNT has gone through a Nether Portal, it will no longer be able to destroy Nether Portal blocks
👀

[prisma smelt]

oh no

[alpine pelican]

Minecraft 1.21 Pre-Release 4

Frog day on a Friday? What could go wrong? We're back with 1.21 Pre-Release 4 today, with a couple more bug fixes, and some tweaks to the Blast Resistance and Fire Protection enchantments.ChangesThe "Burning Time" reduction effect of Fire Protection and "K...

Jun 7, 2024, 12:32 PM

[fiery galleon]

silly bot, i already beat you too it

[light arrow]

so you can or can't use a tnt canon to strand someone in the nether

[hard ivy]

👏

[agile canyon]

What’s my prize?

[austere ocean]

a firm handshake

[prisma smelt]

can if you don't shoot it through the portal, can't if you do

[agile canyon]

🤝

[alpine pelican]

Minecraft 1.21 Pre-Release 4

Frog day on a Friday? What could go wrong? We're back with 1.21 Pre-Release 4 today, with a couple more bug fixes, and some tweaks to the Blast Resistance and Fire Protection enchantments.ChangesThe "Burning Time" reduction effect of Fire Protection and "K...

Jun 7, 2024, 12:32 PM

[light arrow]

ah ok so they made it so you can't strand someone else in the nether

[light arrow]

@jagged stag it's triggering twice

[hard ivy]

say hello to the <@&1248617351486902333>

[jagged stag]

sci broke it

[light arrow]

how?

[hard ivy]

you know I didn't have n8n experience

[light arrow]

sci only added the snowman trigger node

[austere ocean]

if this is an ongoing thing I'd better update my spreadsheet

[jagged stag]

it worked until sci touched it

so it's his fault lol

[light arrow]

now I think last one it did 2 too

[austere ocean]

dubious programming project: website that allows the bets to be submitted and checked automatically

[jagged stag]

it didn't detect the new version either

so more reasons to blame him

[hard ivy]

I don't know how that workflow works

[austere ocean]

[turbid hatch]

I wonder how it feels for a mojang dev to release a pre-release and then see people here decompile it and figure out what they did

[hard ivy]

@jagged stag go fix it

[fiery galleon]

1k changed files? what are you looking at
my snowman shows 14 changes files

[austere ocean]

"throw it to the dogs"

[hard ivy]

since I am clearly incompetent in this manner

[turbid hatch]

I know, but the figuring out part isn't 😉

[hard ivy]

this is exactly why I wanted to delegate this to someone who had n8n experience

because I knew I'd mess something up

I don't know how the workflow for the new MC version works

which I've stated multiple times

both here and in the internal channels

and I've made my inexperience known multiple times

so the fact that this broke really upsets me, because I knew it could've broke at my hands

[jagged stag]

[hard ivy]

i'm never touching n8n again

[hard ivy]

now excuse me as I have to manually run the workflows which should've been triggered

[jagged stag]

no

don't

i need to fix this

[hard ivy]

then sure

if anyone needs me, I'm drowing my sorrows in reading

[prisma smelt]

what are you reading

[hard ivy]

things