1.20.5/6 Snapshots | The NeoForged Project | Page 4

elder swallow Mar 1, 2024, 8:43 AM

#

we can't do that in 1.20.5 since we'd be changing vanilla behavior

true jacinth Mar 1, 2024, 8:43 AM

#

Yeah but fix it for the past™️

elder swallow Mar 1, 2024, 8:43 AM

#

needs to be fixed all the way back to 1.14 kekw

true jacinth Mar 1, 2024, 8:43 AM

#

v good attribute impls we have here

elder swallow Mar 1, 2024, 8:43 AM

#

I don't care that much about the past heh

#

for 1.20.5 I'll remove our patch

true jacinth Mar 1, 2024, 8:43 AM

#

Honestly I wonder how it hasn't come up as a bug yet

elder swallow Mar 1, 2024, 8:43 AM

#

just open a mojira report if you're not happy ioa

elder swallow Mar 1, 2024, 8:44 AM

#

true jacinth Honestly I wonder how it hasn't come up as a bug yet

I suppose that nobody has tried slow falling and very small gravity at the same time

true jacinth Mar 1, 2024, 8:44 AM

#

nobody ever use a slowfall potion on a galacticraft moon?

#

not a single human in the universe

elder swallow Mar 1, 2024, 8:44 AM

#

is GC even ported to 1.14+?

true jacinth Mar 1, 2024, 8:44 AM

#

There's similar mods

#

ad astra or w/e

elder swallow Mar 1, 2024, 8:44 AM

#

the best part is that the slow falling code checks for deltaMovement() <= 0

#

so you're not even going up smoothly

#

one tick you go up, one tick you fall

true jacinth Mar 1, 2024, 8:45 AM

#

wonderful...

elder swallow Mar 1, 2024, 8:45 AM

#

at 0.04 base value you're just floating

#

kekw

elder swallow Mar 1, 2024, 8:46 AM

#

elder swallow for 1.20.5 I'll remove our patch

anyway

fallen igloo Mar 1, 2024, 8:54 AM

#

I'm generating a clean 24w09a from main that we can rebase onto

elder swallow Mar 1, 2024, 8:55 AM

#

wait what

#

can we please squash 24w07a

fallen igloo Mar 1, 2024, 8:57 AM

#

idk if squash and rebase and just rebase make that big of a difference here but sure when you are done make a new branch and squash everything after the snow-lance commits

elder swallow Mar 1, 2024, 8:58 AM

#

the workflow I had in mind was:

squash 24w07a
rebase on top of the latest 1.20.x changes and fix conflicts
update to 24w09a

fallen igloo Mar 1, 2024, 8:58 AM

#

yea that works too I guess

rigid walrus Mar 1, 2024, 8:59 AM

#

and then fix everything broken by item components

fallen igloo Mar 1, 2024, 8:59 AM

#

fun

fallen igloo Mar 1, 2024, 9:00 AM

#

elder swallow the workflow I had in mind was: - squash 24w07a - rebase on top of the latest 1....

but be sure to squash on a different branch so we have the history

elder swallow Mar 1, 2024, 9:01 AM

#

yeah

#

and I'll be sure not to forget the co-authors this time 😔

fallen igloo Mar 1, 2024, 9:04 AM

#

we can rebase the squash onto the 24w09a branch (I killed the workflow before it did the update)

elder swallow Mar 1, 2024, 9:06 AM

#

ok yeah I see

#

still not sure about FBB/RFBB

fallen igloo Mar 1, 2024, 9:08 AM

#

what's to do there?

elder swallow Mar 1, 2024, 9:08 AM

#

in configuration phase, it should be safe to use RFBB from configuration tasks

#

currently the API only allows FBB codecs for config phase

#

this means that you can't send an ItemStack with a config payload, for example

#

I think we'll patch it so that you can always use RFBB for config phase, but with an empty registry access initially, with only the static registries once these are synced, and with all registries once the datapack registries are synced as well

rigid walrus Mar 1, 2024, 9:14 AM

#

why isn't the src entry in the projects/neoforge/.gitignore need to be commented out for command line git for the 24w07a branch?

fallen igloo Mar 1, 2024, 9:15 AM

#

git add --force

#

--force bypasses gitignore

#

and changes are tracked for all checked in files

fallen igloo Mar 1, 2024, 9:24 AM

#

elder swallow in configuration phase, it should be safe to use RFBB from configuration tasks

we'd need to change the protocol after the registry sync to an unbound copy of the protocol and then bind it to the registryAccess

elder swallow Mar 1, 2024, 9:24 AM

#

not really

#

we can do our own RFBB wrapping based on the current state without going through mojang's protocol system

#

an unbound state is dangerous, we risk disconnecting the client if we happen to receive a packet (e.g. keepalive)

#

oh ok, so there is a check that prevents configuration packets from being received outside of the modded configuration phase

fallen igloo Mar 1, 2024, 9:31 AM

#

huh serverbound is always FBB and never RFBB

rigid walrus Mar 1, 2024, 9:31 AM

#

Just found out that by default eclipse tracks changed files that are ignored as long as they are already in the index

fallen igloo Mar 1, 2024, 9:32 AM

#

rigid walrus Just found out that by default eclipse tracks changed files that are ignored as ...

that's not eclipse that is how git works, everything that is checked in is tracked, even if it is ignored

elder swallow Mar 1, 2024, 9:40 AM

#

fallen igloo huh serverbound is always FBB and never RFBB

thonk

#

probably because vanilla didn't need it

#

that's quite problematic

fallen igloo Mar 1, 2024, 9:41 AM

#

so we have to patch that generic type everywhere Thonk

elder swallow Mar 1, 2024, 9:41 AM

#

not everywhere

#

we just need to wrap the bytebuf

rigid walrus Mar 1, 2024, 9:43 AM

#

jgit doesn't seem to support force add

elder swallow Mar 1, 2024, 9:43 AM

#

ok time to squash

#

@misty crow FYI https://github.com/neoforged/Kits/commit/844ea5be58af702392d70f679e16ec6c856d23d8

#

the registry sync task is now before vanilla's datapack sync task

#

it runs before the modded configuration phase hence it needed a special check in the server config listener

elder swallow Mar 1, 2024, 9:54 AM

#

rigid walrus Just found out that by default eclipse tracks changed files that are ignored as ...

oh yeah... the MC sources are in the gitignore lol

rigid walrus Mar 1, 2024, 9:56 AM

#

Should the MC sources be in the gitignore when it's on the kits repo?

elder swallow Mar 1, 2024, 9:57 AM

#

it works fine

#

I like it because it means that we don't have to change the git ignore; only git add --force and git rm -r --cached

fallen igloo Mar 1, 2024, 9:58 AM

#

when you have squashed push the new branch we have to do a bit more than just rebase onto the other branch (we need to update that branch to 24w07a first before we can rebase the fixes)

elder swallow Mar 1, 2024, 10:00 AM

#

I'll rebase on top of 1.20.x now

#

actually I'll join vc for a while

#

ok that's not too bad

elder swallow Mar 1, 2024, 10:34 AM

#

BRUH

#

this suuuucks

wooden tendon Mar 1, 2024, 10:35 AM

#

LOL

#

So sneak prevents you from falling if the height downwards is more then half your step height

#

That is the definition?

elder swallow Mar 1, 2024, 10:36 AM

#

yes

#

it's stupid

wooden tendon Mar 1, 2024, 10:36 AM

#

Seems reasonable

fallen igloo Mar 1, 2024, 10:36 AM

#

@mojang a solution everyone would be happy with: make players not even go down stairs or slabs when sneaking

elder swallow Mar 1, 2024, 10:36 AM

#

if you have a step height of 10 and 1 heart you will die

spiral radish Mar 1, 2024, 10:36 AM

#

concern

white frigate Mar 1, 2024, 10:36 AM

#

fallen igloo @mojang a solution everyone would be happy with: make players not even go down s...

Not sure if you're sarcastic or not.... 😆

fallen igloo Mar 1, 2024, 10:37 AM

#

no that is serious because it is a fix for the reported "bug"

white frigate Mar 1, 2024, 10:37 AM

#

elder swallow if you have a step height of 10 and 1 heart you will die

Seems counterintuitive to me but shrug

elder swallow Mar 1, 2024, 10:37 AM

#

I bet they didn't even try it

#

just got some generic "works as intended"

fallen igloo Mar 1, 2024, 10:38 AM

#

elder swallow if you have a step height of 10 and 1 heart you will die

@azure sparrow do you know if this was considered when debating that ticket?

azure sparrow Mar 1, 2024, 11:04 AM

#

Yes

#

It works as implemented - I personally agree that it this behaviour sucks, but if it’s a change that’s made, it shouldn’t be tracked as a bug

elder swallow Mar 1, 2024, 11:27 AM

#

okay, thanks

kind sable Mar 1, 2024, 12:43 PM

#

The amount of questions we have for the mojank team we should just get a slack invite /s

foggy steeple Mar 1, 2024, 12:46 PM

#

Microsoft owns Mojang

Slack instead of Teams

elder swallow Mar 1, 2024, 12:48 PM

#

@rigid walrus did you forget to add the method content back?

kind sable Mar 1, 2024, 12:48 PM

#

I.... Don't want to work at a place that uses teams....

#

🤣.

elder swallow Mar 1, 2024, 12:49 PM

#

this vineflower bug is unacceptable, I don't understand how it hasn't been fixed yet

dry stumpBOT Mar 1, 2024, 12:51 PM

#

Successfully scheduled reminder on <t:1709333467:F> (<t:1709333467:R>)!

rigid walrus Mar 1, 2024, 12:53 PM

#

elder swallow this vineflower bug is unacceptable, I don't understand how it hasn't been fixed...

I think it has been fixed but the fix hasn't been released yet

elder swallow Mar 1, 2024, 12:54 PM

#

thonk

#

stiff galleon Mar 1, 2024, 1:57 PM

#

harold I push broken code to my dev branches all the time

frail hatch Mar 1, 2024, 2:11 PM

#

tech https://github.com/neoforged/Kits/commit/f0907279443b30314fdd6f3f6296558f70b15078#diff-2ff8a188b91b611d4efa1b1c281f9539c2b740de16801368a1f6a0b51d3548a6R54 make it final, you removed the reject hunk that added that todo

elder swallow Mar 1, 2024, 2:17 PM

#

fuck 😄

blazing valve Mar 1, 2024, 2:17 PM

#

kind sable I.... Don't want to work at a place that uses teams....

Eh, the React based one is "ok"

#

BTW I do wonder what a pain in the ass it was for Mojang to keep that massive branch with the ItemStack component system up to date during its lifespan

kind sable Mar 1, 2024, 2:18 PM

#

blazing valve Eh, the React based one is "ok"

I tried both... As my uni used it... Didn't like either

blazing valve Mar 1, 2024, 2:18 PM

#

kind sable I tried both... As my uni used it... Didn't like either

Well hence the "ok" 😄

#

The previous iteration was practically unusable

#

The new one graduated to "ok"

kind sable Mar 1, 2024, 2:19 PM

#

blazing valve BTW I do wonder what a pain in the ass it was for Mojang to keep that massive br...

Yeah I really wonder how they do these bigger changes...

foggy steeple Mar 1, 2024, 2:20 PM

#

Force push obv smh

kind sable Mar 1, 2024, 2:20 PM

#

What kind of branch management they use....
And when are they gonna merge:

feature/modding-api
feature/new-cool-dimension

frail hatch Mar 1, 2024, 2:22 PM

#

The modding api is datapacks

fallen igloo Mar 1, 2024, 2:24 PM

#

yep the feature/modding-api branch is just datapack blocks and items

restive raft Mar 1, 2024, 2:34 PM

#

https://www.youtube.com/watch?v=iY9OHAd4Aco chonky video

YouTube

slicedlime

News in Data Pack Version 33 (24w09a): Item Components!

Data pack version 33 in snapshot 24w09a introduces a massive change to how item stack data is represented! Here is your ultimate guide to all the changes! #minecraftemployee

slicedlime works as a Tech Lead for Minecraft at Mojang, but the YouTube and Twitch channels are personal projects run entirely in his spare time. This is an unofficial upd...

▶ Play video

kind sable Mar 1, 2024, 2:56 PM

#

fallen igloo yep the `feature/modding-api` branch is just datapack blocks and items

And yet it is not merged in!!
blobpout

lean anchor Mar 1, 2024, 3:53 PM

#

let 'em cook

quiet condor Mar 1, 2024, 4:28 PM

#

kind sable What kind of branch management they use.... And when are they gonna merge: - ``f...

new cool dimension?!

stiff galleon Mar 1, 2024, 4:30 PM

#

the big portal in the ancient city goes to the other end

true jacinth Mar 1, 2024, 4:38 PM

#

elder swallow BRUH

@stray bay Guess you get to open that other bug about it being a fall and not a true step down

#

And I guess I have to write a mixin somewhere, bc that's a stupid ass decision

fallen igloo Mar 1, 2024, 5:05 PM

#

rejects down to 3 hunks in 2 files

#

if someone wants to help me fix KeyBindsList and WalkNodeEvaluator it would be appreciated

frail hatch Mar 1, 2024, 5:16 PM

#

I think looking at the reject vs the file that https://github.com/neoforged/Kits/blob/24w09a/projects/base/src/main/java/net/minecraft/client/gui/screens/controls/KeyBindsList.java#L181C23-L181C33 is where the KeyBindsList reject roughly goes

naive hound Mar 1, 2024, 5:16 PM

#

stiff galleon the big portal in the ancient city goes to the other end

that would be a cool use for that structure.

fallen igloo Mar 1, 2024, 5:26 PM

#

down to 2 hunks in 1 file

#

and I have no idea how to fix WalkNodeEvaluator this time

stray bay Mar 1, 2024, 5:28 PM

#

it occurs to me that they probably could limit fall damage based on the step height, there's already some sort of hook that limits fall damage based on your original height when jumping with wind charges

fallen igloo Mar 1, 2024, 5:29 PM

#

stray bay it occurs to me that they probably could limit fall damage based on the step hei...

yes but that is already pretty hacky

stray bay Mar 1, 2024, 5:34 PM

#

anyway, while a proper step down would be useful, it also seems like it might be difficult to implement

frail hatch Mar 1, 2024, 5:48 PM

#

fallen igloo down to 2 hunks in 1 file

doesn't the second hunk just directly go here: https://github.com/neoforged/Kits/blob/24w09a/projects/base/src/main/java/net/minecraft/world/level/pathfinder/WalkNodeEvaluator.java#L536

#

now that I am back I will look briefly via github to see if I can figure out where the patch stuff for the other hunk should go

#

also ignore the fact apparently I linked base for both instead of from the neoforge project I guess

fallen igloo Mar 1, 2024, 5:52 PM

#

well then it's 1 hunk in 1 file

frail hatch Mar 1, 2024, 6:00 PM

#

I feel like it is a combination of:
https://github.com/neoforged/Kits/blob/24w09a/projects/neoforge/src/main/java/net/minecraft/world/level/pathfinder/PathfindingContext.java#L32
and
https://github.com/neoforged/Kits/blob/24w09a/projects/neoforge/src/main/java/net/minecraft/world/level/pathfinder/PathTypeCache.java#L27

#

for where to patch it

#

though actually

#

maybe not in the cache

#

and just before the first one returns

kind sable Mar 1, 2024, 6:03 PM

#

stiff galleon the big portal in the ancient city goes to the other end

I love that... The negative end...
Tho the opposite of the end would be a mining dimension with no ender man spawning.... Or the local mobs only move if you are not watching them.... Doctor who weeping angels anyone?

stiff galleon Mar 1, 2024, 6:04 PM

#

eye of under

fallen igloo Mar 1, 2024, 6:08 PM

#

hmm maybe it should use the cache

frail hatch Mar 1, 2024, 6:13 PM

#

the only reason I am unsure about the cache part is given if you look in WalkNodeEvaluator#checkNeighbourBlocks vanilla still has the type -> other type thing going on

#

I also don't know if we need an overload as just adjusting the method in the context also will affect things like WalkNodeEvaluator#getBlockPathTypeStatic

fallen igloo Mar 1, 2024, 6:19 PM

#

the problem is that the blockpos now is inside of the PathfindingContext so inaccessible for the calls

#

added a package private getter for the mutable pos

frail hatch Mar 1, 2024, 6:24 PM

#

hence why I was suggesting moving the patch

fallen igloo Mar 1, 2024, 7:55 PM

#

down to 9 classes with compile errors

#

35 errors in 5 files left

#

and now to watching forgecraft

elder swallow Mar 1, 2024, 9:12 PM

#

DataComponents are supposed to be immutable

#

(on stacks at least)

#

Do we want them to be immutable on BEs etc too?

#

Or should we say "Item stack ones are immutable and WILL be copied directly to multiple stacks. Other data component holders will not be copied and may safely be mutated."

#

@frail hatch what do you think? Situational mutability is annoying, might be the lesser evil though.

frail hatch Mar 1, 2024, 9:20 PM

#

idk, mutability is nice, but also for the most part I only really attach stuff to items (and sometimes entities)

lime galleon Mar 1, 2024, 9:27 PM

#

do data components support "serialize on save" semantics (i.e. time dependent serialization)? I'm gathering just from reading they're codec-serialized-only, so probably?

reef parcel Mar 1, 2024, 9:34 PM

#

elder swallow DataComponents are supposed to be immutable

with some notable exceptions, from what I saw thinkies

fallen igloo Mar 1, 2024, 9:35 PM

#

tech lets finish the current stuff and think about expanding it later

#

all compile errors that are left relate to data components

rigid walrus Mar 1, 2024, 10:08 PM

#

fallen igloo all compile errors that are left relate to data components

Are you going to push your changes?

fallen igloo Mar 1, 2024, 10:09 PM

#

oops forgot to push the last one

true jacinth Mar 1, 2024, 10:10 PM

#

fallen igloo tech lets finish the current stuff and think about expanding it later

yeah try to avoid doing feature development during porting

stray bay Mar 1, 2024, 10:18 PM

#

lime galleon do data components support "serialize on save" semantics (i.e. time dependent se...

concern

dry stumpBOT Mar 1, 2024, 10:40 PM

#

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

NeoForm

Build Successful

Version

24w09a-20240301.223831

Build Branch

1.20.5-dev

Minecraft Version

24w09a

rigid walrus Mar 1, 2024, 11:07 PM

#

Fixed some more lambdas #builds message

elder swallow Mar 1, 2024, 11:18 PM

#

true jacinth yeah try to avoid doing feature development during porting

I can think about the feature as I port stuff;)

elder swallow Mar 2, 2024, 12:56 AM

#

btw this sucks

fallen igloo Mar 2, 2024, 7:53 AM

#

yes it does, my idea, just codec and stream codec it and remove INBTSerializable

#

or patch in the lookup provider

elder swallow Mar 2, 2024, 8:17 AM

#

codec it is

#

where did our NBT patches go for the item stack codecs?

fallen igloo Mar 2, 2024, 8:22 AM

#

I did the other one because it actually is easier

fallen igloo Mar 2, 2024, 8:22 AM

#

elder swallow where did our NBT patches go for the item stack codecs?

nuked because ther is no more itemstack nbt

elder swallow Mar 2, 2024, 8:28 AM

#

still, we need to patch in some way to read additional components...........

elder swallow Mar 2, 2024, 8:29 AM

#

fallen igloo I did the other one because it actually is easier

are you porting right now?

#

I have the codec done

fallen igloo Mar 2, 2024, 8:29 AM

#

did you also patch the uses?

rigid walrus Mar 2, 2024, 8:30 AM

#

elder swallow still, we need to patch in some way to read additional components...........

Doesn't vanilla already do that?

elder swallow Mar 2, 2024, 8:30 AM

#

not for recipe outputs, no

elder swallow Mar 2, 2024, 8:30 AM

#

fallen igloo did you also patch the uses?

no since they're all in the test sourceset

#

we don't need test to compile to be able to run the game

#

patching in the holder lookup won't work very nicely with attachments

#

hence the codec

#

looking at NBTIngredient now

fallen igloo Mar 2, 2024, 8:31 AM

#

elder swallow patching in the holder lookup won't work very nicely with attachments

actually there is always a registryAccess available where the methods are used

elder swallow Mar 2, 2024, 8:32 AM

#

yes, it needs propagating through IAttachmentSerializer and AttachmentHolder

fallen igloo Mar 2, 2024, 8:32 AM

#

yep done that

elder swallow Mar 2, 2024, 8:32 AM

#

ok

#

vc?

rigid walrus Mar 2, 2024, 8:36 AM

#

elder swallow we don't need test to compile to be able to run the game

You also don't need all the source code in the running project to compile when using eclipse

elder swallow Mar 2, 2024, 8:36 AM

#

I know I could use ecj but I don't want to 😛

elder swallow Mar 2, 2024, 9:45 AM

#

#

hotbar blur looks disgusting

quiet condor Mar 2, 2024, 9:46 AM

#

you can change the blur in the accessibility settings

#

the amount of*

elder swallow Mar 2, 2024, 9:47 AM

#

I suppose

#

but it looks really bad out of the box

#

looks cool for the game background though - just not for the UI

quiet condor Mar 2, 2024, 9:48 AM

#

will probably be changed in 24w10a

rigid walrus Mar 2, 2024, 9:59 AM

#

javac compiles this::addRenderableWidget to a lambda body

elder swallow Mar 2, 2024, 10:13 AM

#

why though

#

generics?

rigid walrus Mar 2, 2024, 10:15 AM

#

yes

#

almost all decompile errors are caused by generics as well

summer furnace Mar 2, 2024, 11:05 AM

#

they used box blur for performance I guess

#

and it's ... not pretty

elder swallow Mar 2, 2024, 11:14 AM

#

where is my gaussian blur 😭

fallen igloo Mar 2, 2024, 11:21 AM

#

29 tests pass and then we get a critical error

📎 stacktrace.log

dry stumpBOT Mar 2, 2024, 11:32 AM

#

fallen igloo 29 tests pass and then we get a critical error

Created Gist at the request of @slow prism: https://gist.github.com/NeoCamelot/c687db1ca7dd517d0003d733f4b74970

slow prism Mar 2, 2024, 11:32 AM

#

only 29 pass? lol

#

there's 90 tests

elder swallow Mar 2, 2024, 11:36 AM

#

blame the critical error 😛

blazing valve Mar 2, 2024, 12:38 PM

#

It is shader based blur I hope

misty crow Mar 2, 2024, 12:57 PM

#

It's a post-processing shader, yeah

analog ledge Mar 2, 2024, 3:17 PM

#

elder swallow but it looks really bad out of the box

the blur is broken

#

it doesn't calculate the average weights properly compared to the 1.7 version of the shader

#

will probably be fixed

fallen igloo Mar 2, 2024, 3:49 PM

#

elder swallow blame the critical error 😛

Yea it didn't even start the others because of the crash

stray bay Mar 2, 2024, 5:30 PM

#

elder swallow hotbar blur looks disgusting

someone on twitter made a resource pack with a better blur shader iirc

#

https://twitter.com/char_bun/status/1762902650004123972

#

apparently specifically they used a box blur because it's easy to make the radius configurable with a uniform

elder swallow Mar 2, 2024, 5:38 PM

#

well that tweet uses a gaussian blur

#

predictable 😄

analog ledge Mar 2, 2024, 5:41 PM

#

even the linear blur is decent

#

someone fixed the blur to not be bugged

#

stray bay Mar 2, 2024, 6:26 PM

#

yeah box blur has specific high frequency artifacts with certain inputs that make it a bad choice for minecraft

solar zephyr Mar 2, 2024, 6:52 PM

#

my eyes do not like box blur

stiff galleon Mar 2, 2024, 7:05 PM

#

yeah I'm turning that blur off forever when I play 1.20.5+

clear bane Mar 2, 2024, 7:06 PM

#

unless they fix it harold

quiet condor Mar 2, 2024, 7:12 PM

#

which they probably will

analog ledge Mar 2, 2024, 7:13 PM

#

The second image is what it is meant to look like to be clear

#

the first image is unintentional

clear bane Mar 2, 2024, 7:13 PM

#

ah

#

so yeah, I bet they will release it next week then

analog ledge Mar 2, 2024, 7:13 PM

#

it's a one line fix

stiff galleon Mar 2, 2024, 7:36 PM

#

analog ledge the first image is unintentional

I can't tell the difference and they both hurt my eyes 🤷

quiet condor Mar 2, 2024, 7:45 PM

#

stiff galleon I can't tell the difference and they both hurt my eyes 🤷

with the hills in the background you can use them being more square with the box blur

foggy steeple Mar 2, 2024, 8:00 PM

#

#

The blur on one is more boxy and artifact filled

#

On mobile so I can’t easily get images to scale properly

fallen igloo Mar 2, 2024, 8:36 PM

#

were AttributeMap and DefaultAttributes never null safe? the tests currently crash with a NPE because the test entity type in forgeSpawnEggTest does not have attributes

magic sierra Mar 2, 2024, 8:46 PM

#

elder swallow

Are the banners rendered correctly?

#

I wonder if this bug could affect NeoForge

true jacinth Mar 2, 2024, 9:04 PM

#

fallen igloo were AttributeMap and DefaultAttributes never null safe? the tests currently cra...

null safe for what?

fallen igloo Mar 2, 2024, 10:39 PM

#

true jacinth null safe for what?

well for thet test there are no attributes registered for the entity type so the AttributeSupplier is null and throws in AttributeMap#getValue because of that

true jacinth Mar 2, 2024, 11:48 PM

#

That sounds expected

#

there are minimum attribute registration requirements

#

No children of LivingEntity are permitted to have a null attribute registration

fallen igloo Mar 3, 2024, 8:14 AM

#

ok then @slow prism fix the test

fallen igloo Mar 5, 2024, 9:40 AM

#

@elder swallow @true jacinth don't merge the PR on Kits (even if it is done) that should only get merged after the update is pushed to main

elder swallow Mar 5, 2024, 9:40 AM

#

we will have to merge it before

true jacinth Mar 5, 2024, 9:40 AM

#

Retargetting it again will be a headache, i think

elder swallow Mar 5, 2024, 9:41 AM

#

yes

#

because we need to make more networking changes in 1.20.5 anyway

fallen igloo Mar 5, 2024, 9:42 AM

#

still a networking refactor should never happen in the squashed version port

elder swallow Mar 5, 2024, 9:42 AM

#

we don't have the choice

fallen igloo Mar 5, 2024, 9:42 AM

#

we do... don't merge until after the final squash

elder swallow Mar 5, 2024, 9:42 AM

#

that completely freezes networking development on the snapshots

fallen igloo Mar 5, 2024, 9:44 AM

#

we should NEVER do any bigger changes that are not necessary because of MC changes hidden in the update squash

wooden tendon Mar 5, 2024, 9:44 AM

#

Agreed

#

This has been discussed before,

#

Unless it is functionally required

elder swallow Mar 5, 2024, 9:44 AM

#

it is required to fix the current state of the snapshot networking

wooden tendon Mar 5, 2024, 9:45 AM

#

?!?

true jacinth Mar 5, 2024, 9:45 AM

#

Well as it stands networking is broken on the head of 24w09a

wooden tendon Mar 5, 2024, 9:45 AM

#

The flattening is under no circumstances required for snapshot netgworking

#

Neither are all renames

#

Not saying that the PR is bad

#

Just that I agree with schurli

elder swallow Mar 5, 2024, 9:45 AM

#

no, but it is difficult to separate them

fallen igloo Mar 5, 2024, 9:45 AM

#

true jacinth Well as it stands networking *is* broken on the head of 24w09a

then do minimal fixes but the refactor HAS TO wait

wooden tendon Mar 5, 2024, 9:45 AM

#

What?

#

It is not difficult to seperate them for one bit

#

Because he made the PR against the public repo

elder swallow Mar 5, 2024, 9:46 AM

#

I mean that it is very annoying to redo the PR on top of more changes

wooden tendon Mar 5, 2024, 9:46 AM

#

So all changes in that PR are not strictly needed

#

Yes

#

But that is what he has to do

#

He opened the pr unsolicited

true jacinth Mar 5, 2024, 9:46 AM

#

It took as long to rebase it as it did to write it, lol

wooden tendon Mar 5, 2024, 9:46 AM

#

Yes

elder swallow Mar 5, 2024, 9:46 AM

#

wooden tendon He opened the pr unsolicited

that is fine and completely allowed

wooden tendon Mar 5, 2024, 9:46 AM

#

Yes

elder swallow Mar 5, 2024, 9:46 AM

#

I told him to do it in 1.20.5 because the networking already had a number of changes

wooden tendon Mar 5, 2024, 9:47 AM

#

Yeah and there are now 2 maintainers telling you that that was wrong.....

elder swallow Mar 5, 2024, 9:47 AM

#

and there are 2 maintainers telling you that merging the PR might be ok

true jacinth Mar 5, 2024, 9:47 AM

#

What exactly are you concerned about losing if it gets squashed?

fallen igloo Mar 5, 2024, 9:47 AM

#

he can update his PR on the Kits state he just can't merge it there before the final squash

wooden tendon Mar 5, 2024, 9:48 AM

#

fallen igloo he can update his PR on the Kits state he just can't merge it there before the f...

This

true jacinth Mar 5, 2024, 9:48 AM

#

I don't fiddle with the private repo much

wooden tendon Mar 5, 2024, 9:48 AM

#

It is a transparency issue

#

We want these kinds of structural changes public

elder swallow Mar 5, 2024, 9:48 AM

#

it's annoying for people browsing the repo if they get massive changes into one commit

wooden tendon Mar 5, 2024, 9:48 AM

#

On the publicrepo

#

Because as you yourself pointed out, you already missed a review window for the original PR for several reasons

#

And now you are depriving any reviewer outside of the core maintainer team from doing anything!

elder swallow Mar 5, 2024, 9:49 AM

#

it is the core team for a reason, after all 😛

true jacinth Mar 5, 2024, 9:49 AM

#

We can reflect any changes requested on the public facing one

#

(if any are requested... I get the feeling they won't be)

elder swallow Mar 5, 2024, 9:50 AM

#

potentially we can merge the PR right after the final squash, but that means squashing the PR itself to be able to update and apply it reasonably

wooden tendon Mar 5, 2024, 9:51 AM

#

elder swallow it is the core team for a reason, after all 😛

Yes, but the sole argument he has for making these opinionated (there is not a single functional change in his original PR, and that is fine) changes was because he did not get any input with the original team that designed it. Again fine he was busy.

But that in turn means that he can not and should not go in the private repo and get his stuff merged there, were even less (not even all maintainers) can see his changes

fallen igloo Mar 5, 2024, 9:51 AM

#

elder swallow potentially we can merge the PR right after the final squash, but that means squ...

just do a squash merge like on the main repo

elder swallow Mar 5, 2024, 9:51 AM

#

the PR will have to be rebased

wooden tendon Mar 5, 2024, 9:51 AM

#

Then so be it

#

That is the deal with making PRs

elder swallow Mar 5, 2024, 9:51 AM

#

which will be much easier after it is squashed

wooden tendon Mar 5, 2024, 9:52 AM

#

You make PRs you maintain them

#

If you can't be bothered to keep your PR up to date, and rebased

#

Then it does not deserve to get merged

elder swallow Mar 5, 2024, 9:52 AM

#

Orion we work as a team. You can stop with that attitude please

wooden tendon Mar 5, 2024, 9:52 AM

#

WTF

fallen igloo Mar 5, 2024, 9:52 AM

#

elder swallow it is the core team for a reason, after all 😛

this is gatekeeping, we are neoforge and not forge because we wanted no more decisions behind closed doors

wooden tendon Mar 5, 2024, 9:52 AM

#

I am just pointing out that he had a good reason for making this PR

#

Which is that he wanted input on the design

#

And that that is completely fine

true jacinth Mar 5, 2024, 9:53 AM

#

wooden tendon Yes, but the sole argument he has for making these opinionated (there is not a s...

wait why can't all the maintainers see kits

wooden tendon Mar 5, 2024, 9:53 AM

#

And that we will judge his contribution on its merrits

#

But that he can not use the same fucking argument to have his changes merged in the private repo!

elder swallow Mar 5, 2024, 9:53 AM

#

the only reason not to merge the changes in the private repo is because it leads to an obscure final public commit

#

which is fine by me

slow prism Mar 5, 2024, 9:53 AM

#

true jacinth wait why can't all the maintainers see kits

because there's maintainers that haven't said a word in years

wooden tendon Mar 5, 2024, 9:54 AM

#

true jacinth wait why can't all the maintainers see kits

I have no idea, it was at least in the past, I once asked sci and maty to fix it, because I have no idea how the permission system works. I have not heard back. Maybe it has been fixed, maybe not

true jacinth Mar 5, 2024, 9:54 AM

#

Is keeping the public-facing PR up to date with the internal PR not an acceptable solution?

wooden tendon Mar 5, 2024, 9:55 AM

#

elder swallow the only reason not to merge the changes in the private repo is because it leads...

There is two things I would like:

Split the functional required changes to make 1.20.5 work of from the PR. (this can be then merged into kits no problem)
Move the vanity changes to the public PR, back to were it was

fallen igloo Mar 5, 2024, 9:55 AM

#

I think it is the maintainers/neo team that has access

wooden tendon Mar 5, 2024, 9:55 AM

#

true jacinth Is keeping the public-facing PR up to date with the internal PR not an acceptabl...

Sadly that would basically be tripple the work

#

And double the review work

#

But it is doable

true jacinth Mar 5, 2024, 9:55 AM

#

Well reviewers don't need to look at both of them

elder swallow Mar 5, 2024, 9:55 AM

#

well no - we review the private one and assume that it's in line with the public one 😛

fallen igloo Mar 5, 2024, 9:55 AM

#

true jacinth Is keeping the public-facing PR up to date with the internal PR not an acceptabl...

you don't even need that just update it once after the final squash

wooden tendon Mar 5, 2024, 9:56 AM

#

Yeah as long as you do not merge the private one

#

You are basically good I tink

#

Unsure

#

Git is not my strong suite

true jacinth Mar 5, 2024, 9:56 AM

#

That's if the history lines back up for it to work

wooden tendon Mar 5, 2024, 9:56 AM

#

So no idea how this looks

slow prism Mar 5, 2024, 9:56 AM

#

fallen igloo I think it is the maintainers/neo team that has access

anyone in the org can see it. and people not in the org (i.e. the people that haven't said a word in years let alone post split) shouldn't count if they aren't in the org. but that's another matter

wooden tendon Mar 5, 2024, 9:56 AM

#

slow prism anyone in the org can see it. and people not in the org (i.e. the people that ha...

So is it fixed now?

true jacinth Mar 5, 2024, 9:57 AM

#

Basically I don't know what the state of the repo will look like when it is time to apply the changeset

elder swallow Mar 5, 2024, 9:57 AM

#

fallen igloo you don't even need that just update it once after the final squash

it's more likely that the person who will do the final squash will update the PR too

slow prism Mar 5, 2024, 9:58 AM

#

wooden tendon So is it fixed now?

it was always like that

elder swallow Mar 5, 2024, 9:58 AM

#

in any case we have a problem with the current workflow because we don't really have a window to make changes that are 1) in separate commits but 2) are all part of the initial release

#

e.g. the removal of Event.Result will also be one single commit that will need to be applied after the final squash

true jacinth Mar 5, 2024, 9:59 AM

#

Yeah that is a bit of a workflow issue

slow prism Mar 5, 2024, 9:59 AM

#

e.g. removing attachments in favour of components should also be one singular commit with no prior releases

elder swallow Mar 5, 2024, 9:59 AM

#

harold

slow prism Mar 5, 2024, 9:59 AM

#

e.g. renaming mods.toml should also be one singular commit with no prior releases

#

harold

fallen igloo Mar 5, 2024, 10:00 AM

#

elder swallow e.g. the removal of `Event.Result` will also be one single commit that will need...

and that is fine because the main change happened on another (public) repo and it is a separate commit

elder swallow Mar 5, 2024, 10:00 AM

#

it's not fine because it's not scalable

#

and it sucks

#

I had to rush the capability rework PR rebase in the 30 min that I had between the squash was done and I had to leave

#

and that was a single PR

#

we're already talking about 3-4 potential PRs

fallen igloo Mar 5, 2024, 10:01 AM

#

tech these are the technical limitations of our transparency policy and private porting process

elder swallow Mar 5, 2024, 10:02 AM

#

The workflow I would think makes sense is that at some point we squash an intermediate port commit, and then apply larger commits on top of that

#

The only question is how to deal with work that happens in the mainline at the same time. Potentially cherry-pick

true jacinth Mar 5, 2024, 10:02 AM

#

Does the porting process need to be private anymore now that patches are actually like, legible

fallen igloo Mar 5, 2024, 10:03 AM

#

elder swallow I had to rush the capability rework PR rebase in the 30 min that I had between t...

shadows will have from the RC until the release (so at least 24 hours) and then we probably won't release immediately

fallen igloo Mar 5, 2024, 10:03 AM

#

elder swallow The workflow I would think makes sense is that at some point we squash an interm...

that is the first RC

elder swallow Mar 5, 2024, 10:03 AM

#

e.g. you might see in the final update commits:

Port to 24w09a
Networking refactor
Remove Event.Result in favor of more specific enumerations
Port to 1.20.5-pre1
Replace data attachments by vanilla's component system
Port to 1.20.5

#

Schurli you are too focused on the existing workflow

elder swallow Mar 5, 2024, 10:04 AM

#

elder swallow e.g. you might see in the final update commits: - Port to 24w09a - Networking re...

the ONLY problem with this workflow is how we deal with work that happens concurrently in the mainline

fallen igloo Mar 5, 2024, 10:04 AM

#

elder swallow e.g. you might see in the final update commits: - Port to 24w09a - Networking re...

no the network refactor gets merged after on the public repo

elder swallow Mar 5, 2024, 10:04 AM

#

I am discussing a hypothetical workflow that would be better for large changes

fallen igloo Mar 5, 2024, 10:05 AM

#

elder swallow Schurli you are too focused on the existing workflow

we can't change the workflow on our own that is a decision for all maintainers and the council because it requires loosening our transparency policy

elder swallow Mar 5, 2024, 10:06 AM

#

oh come on

#

that doesn't prevent discussing the changes

#

and I don't see how it requires the council, the fuck

fallen igloo Mar 5, 2024, 10:06 AM

#

policy changes are council stuff

slow prism Mar 5, 2024, 10:06 AM

#

next up: article 1, section 2b of the federal constitution of neoforge prevents each state from modifying law d

elder swallow Mar 5, 2024, 10:06 AM

#

it requires maintainer consensus because it is a change in maintainer workflow

#

and that's it

true jacinth Mar 5, 2024, 10:07 AM

#

I don't even think we're losing transparency here, so long as we write down things and make it visible

All three of my staged 20.5 release changes are visible on public rn

slow prism Mar 5, 2024, 10:07 AM

#

(my point is don't overly involve votes and 3 specific people)

elder swallow Mar 5, 2024, 10:07 AM

#

we can btw push the port commits to a public branch with that workflow - hence making the update process more transparent

true jacinth Mar 5, 2024, 10:07 AM

#

If we make major deviations from those public views and do not document it, then we have a problem

fallen igloo Mar 5, 2024, 10:08 AM

#

true jacinth I don't even think we're losing transparency here, so long as we write down thin...

then you can just update them after we squash for the first RC

slow prism Mar 5, 2024, 10:09 AM

#

at this point the private repo can be pushed public just fine

#

and maybe we should

elder swallow Mar 5, 2024, 10:09 AM

#

I think it would make sense to freeze mainline development and publish already (on github, not maven) the port once the prereleases land

#

maybe RCs, pre-releases is a bit early

true jacinth Mar 5, 2024, 10:09 AM

#

isn't the vanilla src in /projects visible atm

slow prism Mar 5, 2024, 10:10 AM

#

push a comit to remove it and squash

true jacinth Mar 5, 2024, 10:10 AM

#

squishy repo

#

anyhow parts of a good discussion here - it is 2 am and I should sleep

Do try to spare my sanity from rebasing that pr if possible :p

slow prism Mar 5, 2024, 10:11 AM

#

sleep is fo rthe weak

#

(go sleep)

elder swallow Mar 5, 2024, 10:12 AM

#

I suppose it could be on a port/1.20.5 branch for example, to avoid triggering releases

#

so the idea is that:

once the first 1.20.5 prerelease or RC lands we freeze 1.20.4 PRs
we remove MC sources, squash and publish kits to a public port/1.20.5 branch (public but not released/published)
impactful PRs can target that branch
once 1.20.5 releases, we push another update commit to port/1.20.5 for that update
we rebase the entire branch into 1.20.x (should be a fast-forward because of the feature freeze)

fallen igloo Mar 5, 2024, 10:18 AM

#

elder swallow so the idea is that: - once the first 1.20.5 prerelease or RC lands we freeze 1....

the second point needs the sources removed first
the last point needs to be a rebase merge not a merge commit

elder swallow Mar 5, 2024, 10:19 AM

#

removing the sources is part of the squash

#

should be a fast-forward merge because of the feature freeze
meaning that commits get added without a merge commit 😛

#

it's equivalent to a rebase merge if the target branch is strictly outdated

#

but yeah on the github UI it would likely be rebase and merge

fallen igloo Mar 5, 2024, 10:20 AM

#

elder swallow > should be a fast-forward merge because of the feature freeze meaning that comm...

it would not be a single commit since squashing again would again hide the changes in the big commit

elder swallow Mar 5, 2024, 10:21 AM

#

yeah that's what I mean

fallen igloo Mar 5, 2024, 10:31 AM

#

so the full new workflow for an update would be:

start kits branch with action
fix rejects
fix compile errors
fix tests
run tests
fix test failures
update to next snapshot (when released)
repeat 2 - 7 until first RC
once the first RC lands freeze PRs on main
remove MC sources
squash
publish the kits branch to a public port/<version> branch
PRs can target that branch (if they are impactful they get merged there)
once MC releases, we push another update commit to the port/<version> branch on main for that update
we rebase-merge the entire branch into the current version branch on main (should be a fast-forward because of the feature freeze)

elder swallow Mar 5, 2024, 10:31 AM

#

yes

#

RC is too late I think, potentially we do it one week before the expected release date

#

for 1.20.4, RC was on the 6th of decement and the release was on the 7th

#

ah but 1.20.4 was a hotfix, right

slow prism Mar 5, 2024, 10:36 AM

#

harold

elder swallow Mar 5, 2024, 10:36 AM

#

late pres or RCs should be fine yeah

#

I wouldn't codify the exact merge freeze time - just check with the rest of the team

fallen igloo Mar 5, 2024, 11:14 AM

#

fallen igloo so the full new workflow for an update would be: 1. start kits branch with actio...

prisma locust Mar 5, 2024, 1:51 PM

#

why can't we just only use kits for the purpose of fixing rejects

#

and as soon as we have something compiling (i.e. what we have now with 24w09a) remove MC sources and move it to a public branch

#

and repeat that for each snapshot we decide to port to (not necessarily all)

#

patch generation is no longer torturous to sit through

elder swallow Mar 5, 2024, 1:56 PM

#

if we are not careful we will mess up the linearity of the git history

#

that is the reason why I propose a mainline freeze

summer furnace Mar 5, 2024, 1:57 PM

#

keeping history pretty should literally not be a priority over keeping the process sane

#

that's the whole point merge commits exist: to indicate the merging of two lines of work that have happened concurrently

#

use them.

#

that doesn't mean we shouldn't hold PRs that will be annoying to merge later, though, just no need to enforce a complete freeze

elder swallow Mar 5, 2024, 2:04 PM

#

I think that a complete freeze of ~1 week makes sense

median wasp Mar 5, 2024, 2:11 PM

#

prisma locust and as soon as we have something compiling (i.e. what we have now with 24w09a) r...

Yeah it would be neat to have porting be semi-open

slow prism Mar 5, 2024, 2:44 PM

#

it would mean that there's more than 20 people that can release mods for new versions early

fallen igloo Mar 5, 2024, 2:47 PM

#

summer furnace that's the whole point merge commits exist: to indicate the merging of two lines...

we use merge commits at work (not me but others) and it makes it very hard to find anything in the tree

clear bane Mar 5, 2024, 2:48 PM

#

I have seen the trees they have here....
yeah, I can se why that would be hard

fallen igloo Mar 5, 2024, 2:50 PM

#

slow prism it would mean that there's more than 20 people that can release mods for new ver...

remembers covers finishing the port of a CB mod befor the version is released

summer furnace Mar 5, 2024, 3:04 PM

#

fallen igloo we use merge commits at work (not me but others) and it makes it very hard to fi...

we are not talking about merging all PRs with merge commits, though. if you see one commit after the porting work is done, named "Merge fixes from 1.20.4 branch", you know exactly where to look, and it wouldn't add that much noise at all, IMO

#

but 🤷‍♂️ ultimately I'm not the one doing the bulk of the work for this, so whatever feels the least headache inducing to you

prisma locust Mar 5, 2024, 3:07 PM

#

ok, this is going to sound like a dumb question, but what is the exact reason, legally, we cannot have the patched source files available publicly?

#

because we already publish a decompiler and the patches necessary to make the decompiled code compile again

summer furnace Mar 5, 2024, 3:09 PM

#

public sources means we are doing the copyright infringement every time someone clones the repo

#

the patches + decompiler is "clean-ish"

#

we provide development tools but not the actual game files

#

which is a very clear line we can avoid crossing

prisma locust Mar 5, 2024, 3:10 PM

#

if I published sources but symmetrically encrypted them using a hash of the Minecraft client JAR, would that also be copyright infringement?

wooden tendon Mar 5, 2024, 3:10 PM

#

Yes

prisma locust Mar 5, 2024, 3:11 PM

#

hm

summer furnace Mar 5, 2024, 3:14 PM

#

if anyone can access it we become a public distributor of copyrighted material. there's a sort of "implicit agreement" that what we do is ok based on the fact mojang has never contacted the forge project (or any other mod loader project) to complain about this kind of development process, but we could not expect that allowance to extend so far as to include the wide public in this development process

fallen igloo Mar 5, 2024, 3:14 PM

#

if you provide the sources in any way it is a violation, if you only provide the tools you are in the clear

median wasp Mar 5, 2024, 3:15 PM

#

Using patches also allows people to easily see the changes

prisma locust Mar 5, 2024, 3:15 PM

#

so providing a reversibly encrypted version of the sources is equivalent to providing the sources, but providing patches is not?

fallen igloo Mar 5, 2024, 3:16 PM

#

summer furnace if anyone can access it we become a public distributor of copyrighted material. ...

under EU law you are explicitly allowed to decompile software to develop against, no license can prevent that

prisma locust Mar 5, 2024, 3:16 PM

#

to be clear, I'm not advocating for breaking the rules, I'm just trying to understand them

summer furnace Mar 5, 2024, 3:16 PM

#

fallen igloo under EU law you are explicitly allowed to decompile software to develop against...

yes but not to SHARE that decompilation with your friends. the private repo is technically in violation of copyright laws, there's just an understanding that mojang knows about it and chooses to allow it

elder swallow Mar 5, 2024, 3:17 PM

#

sharing with close friends is ok

summer furnace Mar 5, 2024, 3:17 PM

#

prisma locust so providing a reversibly encrypted version of the sources is equivalent to prov...

the patches are "minimal" for a purpose, encrypting still contains the data in whole, even if it's not accessible

elder swallow Mar 5, 2024, 3:17 PM

#

depending on how you view it, the repo can be considered legal as long as it remains private - it's on github servers though so idk about that

fallen igloo Mar 5, 2024, 3:17 PM

#

summer furnace yes but not to SHARE that decompilation with your friends. the private repo is t...

it is not public distribution so we are not doing anything stricly illegal (we are in the legal gray area)

summer furnace Mar 5, 2024, 3:18 PM

#

I don't know if this is specific to spanish laws or to the whole of the EU

#

but eg, wrt music

#

legally speaking if your neighbours are hearing you play music

#

you are doing a public performance and are subject to broadcast licensing fees

fallen igloo Mar 5, 2024, 3:19 PM

#

summer furnace legally speaking if your neighbours are hearing you play music

this is a legal gray area here (I think there is a minimum number of people)

summer furnace Mar 5, 2024, 3:19 PM

#

(but in spain copyright infringement is only punishable to the extent of the (intent of) monetary gain -- which is zero in the case of loud music, hence no one would try to go after you legally)

slow prism Mar 5, 2024, 3:39 PM

#

elder swallow depending on how you view it, the repo can be considered legal as long as it rem...

imagine MS protecting their code on their own platform

blazing valve Mar 5, 2024, 4:28 PM

#

elder swallow sharing with close friends is ok

Don't discuss that in too much detail, honestly.

elder swallow Mar 5, 2024, 4:30 PM

#

I mean legally

blazing valve Mar 5, 2024, 4:31 PM

#

That is exactly why you shouldn't. You might not like the answer.

elder swallow Mar 5, 2024, 5:03 PM

#

actually you are right, there is an exception for computer programs under Swiss law

#

#

so you can watch movies together but not play paid games 😛

median wasp Mar 5, 2024, 6:33 PM

#

A benefit of doing porting in the open would be that people get to read PR comments fully

#

Because Discord cuts those off

#

E.g. #neoforge-private message

wooden tendon Mar 5, 2024, 8:18 PM

#

@true jacinth or @elder swallow Why does the Network PR do away with phase specific listeners?

#

The whole point of the phase specific listeners is that during configuration you have a proper way of marking a phase completed, without having to capture a bunch of stuff.

#

On top of that this introduces weird handling related to the sending player (now being sometimes null, instead of an optional)

#

On top, we need to reintroduce versionless channels

true jacinth Mar 5, 2024, 8:20 PM

#

Why do we need versionless channels

wooden tendon Mar 5, 2024, 8:20 PM

#

They are required to be able to communicate with any other platform other then NF

#

Which was the whole point of us fixing the protocol in the first place

#

No other platform other then (Neo)Forge has versioned channels

#

So if you want to support a plugin from a client side mod

true jacinth Mar 5, 2024, 8:20 PM

#

Why is versionless required though? Versioned optional works

stiff galleon Mar 5, 2024, 8:21 PM

#

if you want to version your channels, bump your mod's version harold

wooden tendon Mar 5, 2024, 8:21 PM

#

You need support for version less channels baked in by default

wooden tendon Mar 5, 2024, 8:21 PM

#

true jacinth Why is versionless required though? Versioned optional works

You removed all checks that verified optional channels......

#

Versioned optional as such does not work

#

Because you hard code the check

true jacinth Mar 5, 2024, 8:25 PM

#

Then that's a bug, it should permit versioned optional to connect to versionless

#

Anyway back to the original question

wooden tendon Mar 5, 2024, 8:25 PM

#

I would suggest simply reverting that change

#

The concept to require a version is not great anyway

#

And having versionless channels was explicitly requested by modders and the other loaders

#

I don't see a reason to un do this

true jacinth Mar 5, 2024, 8:26 PM

#

Do you know why it was requested? Is it not for the same goal that optional channels serve?

#

Providing a version should be a good habit instead of an afterthought

wooden tendon Mar 5, 2024, 8:27 PM

#

true jacinth Do you know *why* it was requested? Is it not for the same goal that optional ch...

No. There was a specific request for complete binary and functional compatibility with how Buckit works, additionally the old systems also allowed this (both on the network api level, as well as on the displaytest level)

wooden tendon Mar 5, 2024, 8:27 PM

#

true jacinth Providing a version should be a good habit instead of an afterthought

I agree and it indeed was in the original design that you had to supply one

true jacinth Mar 5, 2024, 8:27 PM

#

wooden tendon The whole point of the phase specific listeners is that during configuration you...

What does having the phase-specific listener do to accomplish this goal? The only difference between the listener types was that they had a different context object (but that is now unified)

wooden tendon Mar 5, 2024, 8:28 PM

#

true jacinth What does having the phase-specific listener do to accomplish this goal? The on...

No actually, if you looked at the API correctly you will see that the configuration handler got given a differnent handler in the first place

#

And that different handler had for example the ability to mark the current running task as complete

true jacinth Mar 5, 2024, 8:28 PM

#

The main IPayloadContext retains this capability

#

See IPayloadContext#finishCurrentTask

wooden tendon Mar 5, 2024, 8:29 PM

#

Yeah but you should not be able to complete a configuration task from a play payload handler

true jacinth Mar 5, 2024, 8:29 PM

#

and you aren't, because it will throw :p

wooden tendon Mar 5, 2024, 8:29 PM

#

And what does that do for play phases? crash?

true jacinth Mar 5, 2024, 8:29 PM

#

it says as much

wooden tendon Mar 5, 2024, 8:29 PM

#

Wtf

#

Why?

#

Just create the compile time safety

#

If you have a payload for the configuration phase

#

Just have it have a specific handler

true jacinth Mar 5, 2024, 8:29 PM

#

It was already absent - the method was present on the clientside for configuration handlers, but threw in that environment

wooden tendon Mar 5, 2024, 8:30 PM

#

Yes because we could not figure out a way to send a consent payload back yet sanely

#

But it is / was planned

#

And it would only cause a crash in a situation where the payload handler was already broken

#

Now a play payload handler can invoke this at any time

true jacinth Mar 5, 2024, 8:31 PM

#

The very slight compile time sanity does not justify the cost of maintaining two separate paths here - you won't have a ConfigurationTask.Type in the context of a play payload anyway

wooden tendon Mar 5, 2024, 8:31 PM

#

true jacinth The very slight compile time sanity does not justify the cost of maintaining two...

?!?, How do you know that

#

The Type is generally a public static final field

#

On the ConfigurationTask it self

#

So in 100% of all cases the modder will simply have access to it

true jacinth Mar 5, 2024, 8:32 PM

#

We should be able to make a reasonable assumption that people can read the javadoc here - @throws UnsupportedOperationException if called on the client, or called on the server outside of the configuration phase.

#

The cost of two paths for one method's worth of segmentation is too high

#

as is evident by the previous implementation

wooden tendon Mar 5, 2024, 8:32 PM

#

true jacinth The cost of two paths for one method's worth of segmentation is too high

IUn your opinion

#

We discussed this with more then 4 people

#

An we agreed that it is worth it

#

To have a compile time safe API

#

If we can do this without too much overhead, and we can.

#

You just need to maintain a decent interface and context implementation

#

Which is really not hard

#

Mojang does it for their listeners

#

And you are telling me that we can't maintain a class hierarchy that is made up out of three types?

#

Seriously?

true jacinth Mar 5, 2024, 8:36 PM

#

wooden tendon If we can do this without too much overhead, and we can.

re-introducing the path is a nontrivial amount of overhead

wooden tendon Mar 5, 2024, 8:36 PM

#

true jacinth re-introducing the path is a nontrivial amount of overhead

?!?

#

I don't understand´

true jacinth Mar 5, 2024, 8:36 PM

#

That's the reason you had so much copypaste in the first place

wooden tendon Mar 5, 2024, 8:36 PM

#

true jacinth That's the reason you had so much copypaste in the first place

Which can be reduced, trivially

#

You can make a proper OOP design without a lot of duplication.

#

I left a review requesting changes for now

#

Some of the PR is really nicely done

#

The staticness of NetworkRegistry is very nice

true jacinth Mar 5, 2024, 8:41 PM

#

Okay, top level here: The two paths require introducing a subclass of IPayloadContext and IPayloadHandler
IPayloadRegistrar#configuration then has to be updated to be generic and take ? super IConfigurationPayloadHandler
#play stays as-is
#common stays as-is, but users of common now need to downcast on their end to access IConfigurationPayloadHandler (instead of just checking the protocol and invoking finishCurrentTask)

NetworkRegistry#handleModdedPayload has to be updated to construct a context based on the listener's concrete type instead of just the context, needing two more context impls

wooden tendon Mar 5, 2024, 8:41 PM

#

I am not sure why I did not do that from the beginning

true jacinth Mar 5, 2024, 8:42 PM

#

It was longstanding for it to be instanced

#

I think you just went with it

wooden tendon Mar 5, 2024, 8:43 PM

#

true jacinth Okay, top level here: The two paths require introducing a subclass of `IPayloadC...

Right now out of the top of my head, you can solve the duplication problem by handling this in a listener specific implementation

#

Which would probably work very nicely

#

If you split up the NetworkRegistry into smaller pieces

#

With each their own functionality and scope

#

The code duplication would then be extremely minimal

#

Maybe even none existent

true jacinth Mar 5, 2024, 8:46 PM

#

Bleh, and I have to de-record-ify ClientPayloadContext and ServerPayloadContext to make them extensible

wooden tendon Mar 5, 2024, 8:47 PM

#

I will point out that you started this without asking anyone from the networking working group that have worked on this topic for nearly half a year now.

#

We have been were you are right now

#

We have not come to this somewhat idiotic API by accident

#

I agree that some stuff (like the subsystem specific context system, where each subsystem is its own type) is very opinionated

#

And might/should be changed

true jacinth Mar 5, 2024, 8:48 PM

#

I still think it is worth trading one method's worth of compile time safety for a single path

#

if there was more code relating to the other path, there would be more reason to have two paths

wooden tendon Mar 5, 2024, 8:48 PM

#

Not from the discussions we had

#

It was already problematic that we threw an exception on the client side

true jacinth Mar 5, 2024, 8:48 PM

#

or if we guaranteed that finishCurrentTask was callable on both sides - but we don't have that either

wooden tendon Mar 5, 2024, 8:48 PM

#

People climbed the barriers and wanted us to introduce a "NeoForge" Ack payload

true jacinth Mar 5, 2024, 8:49 PM

#

So ultimately we still place a burden on the reader in both cases

wooden tendon Mar 5, 2024, 8:49 PM

#

true jacinth *or* if we guaranteed that `finishCurrentTask` was callable on both sides - but ...

Again that was/is the plan

dry stumpBOT Mar 5, 2024, 8:49 PM

#

Quote #1777 added.

plucky brook Mar 5, 2024, 8:50 PM

#

thinkies maty has to fix stuff

true jacinth Mar 5, 2024, 8:50 PM

#

well I am happy to stage things for later, but this looks silly :p https://i.imgur.com/nBoTrvB.png

Imgur

true jacinth Mar 5, 2024, 9:18 PM

#

@wooden tendon this is the minimal changeset needed to support config packets having a special handler class https://github.com/neoforged/Kits/pull/3/commits/859f250d44cbc11695826d3fcb9bbb73d28c7f00

#

it has its own flaws:

Impossible to register a non-configuration handler for a configuration payload without registering it as common [might be a non-issue, should anyone really be doing that?]
NonExtendable override introduced in IConfigurationPayloadHandler
Very long generic declaration for DirectionalPayloadHandlerBuilder due to the introduction of H

#

plus the code volume (which adds a maintainence burden)

elder swallow Mar 5, 2024, 9:25 PM

#

true jacinth Providing a version should be a good habit instead of an afterthought

I agree

#

Most mods probably don't provide a version at all, right now

true jacinth Mar 5, 2024, 9:25 PM

#

I'm pretty sure versioned optional channels can be made to work in whatever method non-versioned channels were working before

#

there is no real basis for them to be unable to

stiff galleon Mar 5, 2024, 9:26 PM

#

why should mods provide versions

#

if I introduce an incompatibility, I bump the mod version

true jacinth Mar 5, 2024, 9:26 PM

#

the mod version is not checked

#

only channel versions are

elder swallow Mar 5, 2024, 9:26 PM

#

true jacinth I still think it is worth trading one method's worth of compile time safety for ...

I agree here as well

stiff galleon Mar 5, 2024, 9:27 PM

#

if we don't have a good way to say "the version of this mod on the client is not compatible with the version of that mod on the server" then should we look into that?

#

rather than having versions on specific channels

true jacinth Mar 5, 2024, 9:27 PM

#

Your network version is not directly comparable to your mod v ersion

elder swallow Mar 5, 2024, 9:27 PM

#

Software design is all about tradeoffs. And here we have a well documented contract. Not everything can be proven at compile time - it is a trade-off between catching bugs earlier and not making the code too hard to read/write

stiff galleon Mar 5, 2024, 9:27 PM

#

Your network version is not directly comparable to your mod v ersion
yours isn't, mine is

true jacinth Mar 5, 2024, 9:28 PM

#

i.e. you may have pushed a fix, but it was only relevant to the client - needing a new version

stiff galleon Mar 5, 2024, 9:28 PM

#

there's a specific digit I bump whenever I break network compat

true jacinth Mar 5, 2024, 9:28 PM

#

Well then you have a network version :p

#

use that as your channel version

restive raft Mar 5, 2024, 9:28 PM

#

you could tie in your mod version to your network version

true jacinth Mar 5, 2024, 9:28 PM

#

done™️

stiff galleon Mar 5, 2024, 9:28 PM

#

I thought channels didn't have versions anymore

true jacinth Mar 5, 2024, 9:28 PM

#

Right now (1.20.4) they optionally have versions

#

in my (1.20.5) refactor, they are required to have versions

In both cases, the channel may still be optional

slow prism Mar 5, 2024, 9:29 PM

#

neo itself uses non strict channel version, so no, channels have to be versioned or not at all

true jacinth Mar 5, 2024, 9:29 PM

#

Yeah, it's effectively always versioned

#

just you may have a blank version

#

how can I even check that cross-loader channels work? Do I need to spin up a fabric mod somehow?

#

If I'm understanding the setup correctly, non-neo channels are stored in ATTRIBUTE_ADHOC_CHANNELS via register payload

#

these do not interact with channel versions whatsoever

#

so the version is irrelevant for any consumers that are not Neo

elder swallow Mar 5, 2024, 9:48 PM

#

wooden tendon Yeah but you should not be able to complete a configuration task from a play pay...

You should not be able to complete one on the client side either, but the API doesn't protect against that. There's only so much we can reasonably protect against

true jacinth Mar 5, 2024, 9:48 PM

#

Thus, version can be mandatory for registration without having any impact on Bukkit/Fabric/whatever cross-connections

true jacinth Mar 5, 2024, 9:50 PM

#

elder swallow You should not be able to complete one on the client side either, but the API do...

The clientside path was mentioned to be for a reply packet that is NYI

elder swallow Mar 5, 2024, 9:50 PM

#

true jacinth how can I even check that cross-loader channels work? Do I need to spin up a fa...

Orion linked a fabric test mod somewhere

#

Might be available in the description of his PR from ~3 weeks ago

elder swallow Mar 5, 2024, 9:51 PM

#

true jacinth The clientside path was mentioned to be for a reply packet that is NYI

Well, that is obscure at best thinkies

true jacinth Mar 5, 2024, 9:53 PM

#

Ye that has come up a couple times now

#

it was also the case with the isConnected naming

#

as the rationale for the name is rooted in future work that only makes sense after that future implementation

#

now I have to figure out how to setup a snapshot fabric env to check this against, screm

#

screm the test mod has to be ported as well bc fabric api changed

elder swallow Mar 5, 2024, 10:09 PM

#

Well yeah, vanilla did change 😛

true jacinth Mar 5, 2024, 10:26 PM

#

well this is what fabric gets me rn https://i.imgur.com/dCkJZ23.png

Imgur

#

so guess that needs some looking at...

#

Also not sure if this is a me artifact or was just broken on 20.5 as-is

#

ah wait nevermind

#

I forgot to do an optional reg on the server

true jacinth Mar 5, 2024, 11:22 PM

#

that does in fact work™️ https://i.imgur.com/PPZWeac.png

Imgur

#

There was one ever so slight bug that was preventing it from working at first, but otherwise it is fine: https://i.imgur.com/8Nzsmu3.png

Imgur

restive raft Mar 5, 2024, 11:26 PM

#

gotta love those off by one errors

blazing valve Mar 6, 2024, 12:56 AM

#

true jacinth Providing a version should be a good habit instead of an afterthought

it will honestly always stay 1.0 for my mod 😛

true jacinth Mar 6, 2024, 12:59 AM

#

stabolb

frail hatch Mar 6, 2024, 12:59 AM

#

I just set it to the artifact version of my mod container

#

xlurk

shut wraith Mar 6, 2024, 1:00 AM

#

true jacinth Providing a version should be a good habit instead of an afterthought

I only change the network format when I make other, incompatible code changes, which I only do at a bump in MC version, so... It will always be an afterthought for me at least

true jacinth Mar 6, 2024, 1:00 AM

#

Well at least I can make you think about it

shut wraith Mar 6, 2024, 1:00 AM

#

stabolb you cannot

#

It we be set to something and promptly forgotten /hj

elder swallow Mar 6, 2024, 9:55 AM

#

how am I supposed to access the Connection from a payload handler?

fallen igloo Mar 6, 2024, 2:07 PM

#

what do you need it for?

stiff galleon Mar 6, 2024, 2:09 PM

#

they want to noop the handler in singleplayer

#

(which begs the question, "why do you want to noop the handler in singleplayer?")

fallen igloo Mar 6, 2024, 2:12 PM

#

also if the player is null and it uses the Minecraft.getInstance().player then the connection in the minecraft instance should also be null since that call is delegated to the player afaik

elder swallow Mar 6, 2024, 2:17 PM

#

good point

#

the problem was that the player shouldn't be cached then

#

see #maintenance-talk

fallen igloo Mar 6, 2024, 2:18 PM

#

on another note, it's snapshot day 🐸

elder swallow Mar 6, 2024, 2:18 PM

#

true

clear bane Mar 6, 2024, 2:19 PM

#

what wonders will they give us this week harold

jagged yoke Mar 6, 2024, 2:19 PM

#

component crafting 😎

clear bane Mar 6, 2024, 2:20 PM

#

remind me...
were you part of Mojang, or not thinkies

jagged yoke Mar 6, 2024, 2:20 PM

#

Nah lol

naive hound Mar 6, 2024, 2:20 PM

#

I think at this point, Misode has them on speed dial lol

magic sierra Mar 6, 2024, 2:21 PM

#

https://twitter.com/slicedlime/status/1765366312019145033

slicedlime 💙💛 (@slicedlime) on X

🤔

elder swallow Mar 6, 2024, 2:21 PM

#

jagged yoke component crafting 😎

would be nice

#

along with a much needed Ingredient refactor

magic sierra Mar 6, 2024, 2:21 PM

#

https://twitter.com/SidedSquare/status/1765368590776058014

EightSidedSquare (@SidedSquare) on X

@slicedlime @gegybeans Item components in recipes? Background looks like the output of a stonecutter.

elder swallow Mar 6, 2024, 2:21 PM

#

wait

#

wtf Misode

jagged yoke Mar 6, 2024, 2:22 PM

#

elder swallow would be nice

*will be nice 🙂

elder swallow Mar 6, 2024, 2:22 PM

#

jeez 😄

#

from the twitter post?

clear bane Mar 6, 2024, 2:22 PM

#

~~super secret mojang discord ofc~~

naive hound Mar 6, 2024, 2:29 PM

#

potion recipe jsons when?

jagged yoke Mar 6, 2024, 2:30 PM

#

Slicedlime's tweet is more than an hour old, we've had time to analyze xD
But also it was one of the options we suspected when he said this in his stream:

I've been working on a massive thing, that's kind of ended up on top of all the item components stuff, that hopefully we'll see soon. It should be pretty exciting for pack makers. And that's a very big new thing.
https://youtube.com/clip/UgkxyWThjY6oNV3FeXLzDt0FbSLVbQCrB2gP?si=jpuFovDP-t80aAJQ

stiff galleon Mar 6, 2024, 2:31 PM

#

probably talking about datapack items/blocks

quiet condor Mar 6, 2024, 2:34 PM

#

jagged yoke component crafting 😎

what's that?

elder swallow Mar 6, 2024, 2:35 PM

#

NBT crafting but with the new component system instead

stiff galleon Mar 6, 2024, 2:35 PM

#

quiet condor what's that?

itemstack capabilities but in vanilla

quiet condor Mar 6, 2024, 2:35 PM

#

oh

#

sounds cool

fallen igloo Mar 6, 2024, 2:35 PM

#

naive hound potion recipe jsons when?

lol we don't even have that in neo (because I'm not done with it)

clear bane Mar 6, 2024, 2:36 PM

#

its Mojang, don't tempt the mojang harold

stiff galleon Mar 6, 2024, 2:36 PM

#

fallen igloo lol we don't even have that in neo (because I'm not done with it)

mojang adding potion recipes immediately after we do would be hilariously in-character

naive hound Mar 6, 2024, 2:38 PM

#

🤔 so we just need to leverage that phenomenon to get what we want in vanilla. foolproof /j lol

white frigate Mar 6, 2024, 2:59 PM

#

clear bane remind me... were you part of Mojang, or not <:thinkies:932605814995292192>

I wish

wild pine Mar 6, 2024, 3:02 PM

#

https://www.minecraft.net/en-us/article/minecraft-snapshot-24w10a snapshot time

stiff galleon Mar 6, 2024, 3:02 PM

#

Added variants of Wolves
Better than Wolves reportedly inconsolable

#

Spotted Wolf - A variant that spawns in a new location for wolves - the Savanna Plateau biome, in larger packs of 4 to 8
Striped Wolf - A variant that spawns in a new location for wolves, the Wooded Badlands biome, in larger packs of 4 to 8
wait those are just hyenas

#

vanilla recipes can now set itemstack data of the result

#

cooking recipes still can't set the count, weirdly

misty crow Mar 6, 2024, 3:09 PM

#

Banner patterns are a datapack registry now

slow prism Mar 6, 2024, 3:09 PM

#

<@&1067092163520909374>

open kindle Mar 6, 2024, 3:14 PM

#

magic sierra https://twitter.com/slicedlime/status/1765366312019145033

if Mojang manages to do single and multiple for just about every command, why can't we have 1 component and %d components as well

white frigate Mar 6, 2024, 3:16 PM

#

not worth it for a debug tooltip ?

open kindle Mar 6, 2024, 3:19 PM

#

eh, if commands are worth it, so are debug tooltips

white frigate Mar 6, 2024, 3:19 PM

#

most commands are meant to be used by players ig

#

🤷

open kindle Mar 6, 2024, 3:20 PM

#

I'd argue the same for advanced tooltips

spiral radish Mar 6, 2024, 3:20 PM

#

https://mdcfe.dev/mc-changes?ver=24w10a
https://www.minecraft.net/en-us/article/minecraft-snapshot-24w10a
https://github.com/neoforged/Snowman/commit/53cedf7a1c8fe1f1aaec46fbda9e87da34307467

blazing valve Mar 6, 2024, 3:24 PM

#

stiff galleon (which begs the question, "why do you want to noop the handler in singleplayer?"...

If it's within the same JVM and the packet is syncing static state, you already have the state

#

Yeah yeah, there are better ways, it just is

stiff galleon Mar 6, 2024, 3:30 PM

#

wolf variants are a datapack registry

#

when a wolf spawns, it iterates over the loaded wolf variants, until it finds one that matches the spawn biome, and uses that

#

so you get at most one wolf type per biome

#

but the wolf variants specify a biome holderset, so you can easily add them for custom biomes

open kindle Mar 6, 2024, 3:31 PM

#

what happens if there's no variant? cancelled spawn?

spiral radish Mar 6, 2024, 3:31 PM

#

So how many mods do you reckon we'll get fixing that

stiff galleon Mar 6, 2024, 3:31 PM

#

if there's no variant for that biome, it uses the default/old/pale wolf

prisma locust Mar 6, 2024, 3:36 PM

#

stiff galleon when a wolf spawns, it iterates over the loaded wolf variants, until it finds on...

can't wait for that to be a problem when modpacks have 200 wolf variants

stiff galleon Mar 6, 2024, 3:36 PM

#

"why are my taiga wolves clown-themed"

analog ledge Mar 6, 2024, 3:37 PM

#

open kindle what happens if there's no variant? cancelled spawn?

        return (Holder<WolfVariant>)lvRegistry2.holders()
            .filter(pHolder$Reference1 -> ((WolfVariant)pHolder$Reference1.value()).biomes().contains(pHolder1))
            .findFirst()
            .orElse(lvRegistry2.getHolderOrThrow(PALE));```

#

just gives pale wolf

short cove Mar 6, 2024, 3:41 PM

#

analog ledge ```java return (Holder<WolfVariant>)lvRegistry2.holders() .f...

Replace findFirst with findAny fr

#

It's too limiting

#

Mojang seems to forget multiple datapacks can be enabled

stiff galleon Mar 6, 2024, 3:47 PM

#

wouldn't findAny just do the same thing

short cove Mar 6, 2024, 3:48 PM

#

I was under the impression from reading the javadoc it finds a random one

stiff galleon Mar 6, 2024, 3:48 PM

#

no

#

it's just not guaranteed to be deterministic

#

the underlying implementation is free to decide which thing to return

short cove Mar 6, 2024, 3:49 PM

#

aw

stiff galleon Mar 6, 2024, 3:49 PM

#

but, if you're non-parallely iterating over a list, which we are, that implementation is still just gonna be "grab the first one"

open kindle Mar 6, 2024, 3:50 PM

#

short cove Mojang seems to forget multiple datapacks can be enabled

datapacks may override things, don't forget that

#

so findFirst does actually make sense

#

since it's grabbing the first match from the first (= top-most) datapack

short cove Mar 6, 2024, 3:50 PM

#

Overriding can still work

stiff galleon Mar 6, 2024, 3:50 PM

#

well, only the topmost datapacks' wolves are going to be in the registry anyway

short cove Mar 6, 2024, 3:51 PM

#

You just override a specific wolf

short cove Mar 6, 2024, 3:51 PM

#

stiff galleon well, only the topmost datapacks' wolves are going to be in the registry anyway

??

#

It loads all of them

stiff galleon Mar 6, 2024, 3:51 PM

#

no I mean

#

if you make a datapack that overrides the vanilla spotted wolf with a clown texture

#

the registry has the clown spotted wolf, and not the vanilla spotted wolf

short cove Mar 6, 2024, 3:51 PM

#

Well yes

#

If overridden

stiff galleon Mar 6, 2024, 3:52 PM

#

okay, good, everyone agrees then

short cove Mar 6, 2024, 3:52 PM

#

But if you want to add the clown spotted wolf and keep the vanilla one, it should pick randomly

#

imo

stiff galleon Mar 6, 2024, 3:52 PM

#

anyway if mojang wanted a random wolf

#

the correct implementation would be to first make a filtered list of valid wolves for the given biome, and then randomly select one from that

median wasp Mar 6, 2024, 4:05 PM

#

Is there a .shuffled() on Java Streams?

spiral radish Mar 6, 2024, 4:06 PM

#

Nope

median wasp Mar 6, 2024, 4:07 PM

#

Util.getRandom on .toList() would work then

stiff galleon Mar 6, 2024, 4:07 PM

#

yeah that's faster than shuffling

median wasp Mar 6, 2024, 4:08 PM

#

Meh

#

It's O(n) either way

#

A bit faster maybe

#

But we're talking much less than a millisecond

fallen igloo Mar 6, 2024, 4:45 PM

#

and now to waiting for NeoForm

quiet condor Mar 6, 2024, 5:29 PM

#

oh yeah all tools have the same damage and attack speed now

stiff galleon Mar 6, 2024, 5:30 PM

#

IIRC that was a bug

quiet condor Mar 6, 2024, 5:31 PM

#

was?

stiff galleon Mar 6, 2024, 5:31 PM

#

from the previous snapshot

#

oh it's listed as fixed for this snapshot https://bugs.mojang.com/browse/MC-268800

foggy steeple Mar 6, 2024, 5:54 PM

#

rigid walrus Mar 6, 2024, 6:47 PM

#

Only 1 new compile error

digital drum Mar 6, 2024, 6:49 PM

#

qq: components essentially replace nbt right?

quiet condor Mar 6, 2024, 6:50 PM

#

stiff galleon from the previous snapshot

oh I thought it was this snapshot

fallen igloo Mar 6, 2024, 6:57 PM

#

digital drum qq: components essentially replace nbt right?

yes

digital drum Mar 6, 2024, 6:59 PM

#

fallen igloo yes

will nbt be removed in the forseeable future?

fallen igloo Mar 6, 2024, 7:00 PM

#

the format probably not but itemstack nbt is "gone" (it is moved to components and the custom_data component)

naive hound Mar 6, 2024, 7:07 PM

#

Since i absolutely love the attachmentTypes that Neo had and this is basically that same system i'm super excited for this vanilla change

fallen igloo Mar 6, 2024, 7:16 PM

#

rigid walrus Only 1 new compile error

let's see how many rejects/errors we get on kits

reef parcel Mar 6, 2024, 7:17 PM

#

love that enchantments don't have to be deserialized every single time they need to be accessed for any reason on itemstacks

plucky brook Mar 6, 2024, 7:23 PM

#

fallen igloo the format probably not but itemstack nbt is "gone" (it is moved to components a...

except for serialization afaict

fallen igloo Mar 6, 2024, 7:24 PM

#

fallen igloo let's see how many rejects/errors we get on kits

10 patch failures

lean anchor Mar 6, 2024, 7:25 PM

#

So, how DO components work now, in code? Something like EnchantedBook.enchantments(stack).level(Enchantment.POWER)?

#

Similar to how data attachments work?

reef parcel Mar 6, 2024, 7:31 PM

#

kinda, yeah. stack.get(DataComponents.ENCHANTMENTS).getLevel(Enchantments.POWER)

median wasp Mar 6, 2024, 7:32 PM

#

And they're stored as the difference between the stack's components and the item's default components

lean anchor Mar 6, 2024, 7:33 PM

#

oh, derp. I crossed two ideas in my head, sec

#

fix'd

#

but question answered

reef parcel Mar 6, 2024, 7:34 PM

#

reef parcel kinda, yeah. `stack.get(DataComponents.ENCHANTMENTS).getLevel(Enchantments.POWER...

not even "like", that is literally what it is, it's such a good design you can basically guess it first try lol

lean anchor Mar 6, 2024, 7:34 PM

#

I see this quickly turning interface-based... like IEnchantableItem having an enchantments method that reads values

#

So your common stuff is abstracted away

reef parcel Mar 6, 2024, 7:35 PM

#

what lol

lean anchor Mar 6, 2024, 7:35 PM

#

That's what I do in CM6 now, anyway

reef parcel Mar 6, 2024, 7:36 PM

#

if you want your item to have enchantments, you just give it the component

lean anchor Mar 6, 2024, 7:36 PM

#

#

like, this but with components.. the common code vanishes and all you need is to add an interface

median wasp Mar 6, 2024, 7:37 PM

#

Here's an example from their code:

public static int getItemEnchantmentLevel(Enchantment enchantment, ItemStack itemStack) {
   ItemEnchantments itemEnchantments = itemStack.getOrDefault(DataComponents.ENCHANTMENTS, ItemEnchantments.EMPTY);
   return itemEnchantments.getLevel(enchantment);
}

lean anchor Mar 6, 2024, 7:38 PM

#

Oh. Yeah, that's 🔥

#

That'll reduce SO much code, it'll be hillarious

reef parcel Mar 6, 2024, 7:39 PM

#

yeah, if they went the interface way, it would be pointless

median wasp Mar 6, 2024, 7:41 PM

#

Well getOrDefault is actually on DataComponentHolder, which is an interface

#

Which means we might get it on other objects 👀

reef parcel Mar 6, 2024, 7:41 PM

#

that's not what I mean lol

median wasp Mar 6, 2024, 7:41 PM

#

It would be neat if we could also have it on entities

#

Yeah, Ik, but still

lean anchor Mar 6, 2024, 7:46 PM

#

It'll come to blocks/block entities long before entities, lol

jagged yoke Mar 6, 2024, 7:46 PM

#

Blocks as well, would be nice if copy_components in a loot table actually did that

median wasp Mar 6, 2024, 7:48 PM

#

Blocks don't really have a way to store components, unless there's a separate map of BlockPos to component inside a chunk, serialized to disk

#

At which point it's just a slightly lighter-weight BE

jagged yoke Mar 6, 2024, 8:04 PM

#

I meant block entities, my bad

median wasp Mar 6, 2024, 8:07 PM

#

Ah, I thought you said blocks and block entities

#

Well

#

I thought Nano did

digital drum Mar 6, 2024, 8:13 PM

#

what are slot ranges?

median wasp Mar 6, 2024, 8:14 PM

#

armor.* is for all five armor slots

digital drum Mar 6, 2024, 8:15 PM

#

ok kinda getting it

lean anchor Mar 6, 2024, 8:34 PM

#

Wonder if we can convince 'em to use range format (hotbar.[1..3])

#

Meaning, any of the first 3 hotbar slots

digital drum Mar 6, 2024, 9:04 PM

#

lean anchor Wonder if we can convince 'em to use range format (`hotbar.[1..3]`)

how?

lean anchor Mar 6, 2024, 9:15 PM

#

Mojang has eyes and ears better_eyes

fallen igloo Mar 6, 2024, 9:16 PM

#

lean anchor Mojang has eyes and ears <:better_eyes:862070075158298635>

in here

reef parcel Mar 6, 2024, 9:17 PM

#

boo!

stiff galleon Mar 6, 2024, 9:18 PM

#

lean anchor Mojang has eyes and ears <:better_eyes:862070075158298635>

they keep them in a jar in the break room

slow prism Mar 6, 2024, 9:39 PM

#

also reminder to sync the component types reg

fallen igloo Mar 6, 2024, 9:55 PM

#

it's a normal registry do we need any additional logic there?

shut wraith Mar 6, 2024, 11:32 PM

#

Whoa, didn't look at today's snapshot yet. Bunch of interesting stuff here

true jacinth Mar 7, 2024, 12:17 AM

#

Hopefully the networking changes don't have that many rejects

solar zephyr Mar 7, 2024, 12:56 AM

#

foggy steeple

this sounds like a fun bug

slow prism Mar 7, 2024, 7:39 AM

#

fallen igloo it's a normal registry do we need any additional logic there?

not what i said. mojang is syncing components with int ids

fallen igloo Mar 7, 2024, 7:54 AM

#

well we do reg id sync for all registries anyway so that is also not a problem

slow prism Mar 7, 2024, 7:55 AM

#

no?

fallen igloo Mar 7, 2024, 7:58 AM

#

wdym no? yes we do have reg id sync

slow prism Mar 7, 2024, 7:58 AM

#

it's not for all registries

elder swallow Mar 7, 2024, 10:15 AM

#

any reason why we don't sync all of them?

fallen igloo Mar 7, 2024, 10:20 AM

#

idk I added the reg to be synced

elder swallow Mar 7, 2024, 10:27 AM

#

@slow prism why not sync all of them?

stray bay Mar 7, 2024, 10:36 AM

#

naive hound Since i absolutely love the attachmentTypes that Neo had and this is basically t...

one big difference is that they're immutable [obviously you can change their value, but if they have a composite type you have to make a new copy for every change]

#

if i understand correctly from what people have been saying

slow prism Mar 7, 2024, 10:48 AM

#

elder swallow <@561254664750891009> why not sync all of them?

not all registries must be on the client

#

and synced registries mandate that the contents of the reg are the same on the client and on the server

elder swallow Mar 7, 2024, 10:52 AM

#

hmmm true, syncing recipe serializers would be pointless for example

foggy steeple Mar 7, 2024, 11:18 AM

#

fallen igloo well we do reg id sync for all registries anyway so that is also not a problem

If you make the worldgen structure registry synced, I’ll start throwing hands as then you made my structure mod no longer serverside only

naive hound Mar 7, 2024, 11:51 AM

#

stray bay one big difference is that they're immutable [obviously you can change their val...

I see. 🤔

stiff galleon Mar 7, 2024, 4:44 PM

#

worldgen structure registry isn't syncable anyway harold

#

or not able to be made syncable

#

not without substantially rewriting how the datapack registries are synced

foggy steeple Mar 7, 2024, 4:47 PM

#

Structure type registry is not a datapacks registry

#

So if one makes that synced, me throw hands

stiff galleon Mar 7, 2024, 4:48 PM

#

ah I thought you were talking about structure jsons nvm

stray bay Mar 8, 2024, 5:37 AM

#

huh they added item component output to the furnaces specifically, did they add it to the crafting table in a previous snapshot?

#

also datapackable banner patterns (and wolf types - kind of makes me wonder if they'll go back and do it to other animal variant types / add more)

quiet condor Mar 8, 2024, 6:14 AM

#

I think they also added it to crafting tables in the snapshot

edgy arrow Mar 9, 2024, 3:11 PM

#

I'm waiting for datapackable paintings iirc

frail hatch Mar 9, 2024, 3:22 PM

#

@elder swallow I realized a potential issue with immutable components. How will say energy handlers etc we try to attach to stacks work given the whole concept of those is that they are mutable

elder swallow Mar 9, 2024, 3:24 PM

#

they just attach a new component on the stack

#

#

I REALLY want to make an immutable subclass of ItemStack and FluidStack

#

components should really be immutable

prisma locust Mar 9, 2024, 3:26 PM

#

elder swallow they just attach a new component on the stack

uh do you mean they attach a new component every time the energy level changes?

elder swallow Mar 9, 2024, 3:26 PM

#

they replace the previous instance yes

prisma locust Mar 9, 2024, 3:26 PM

#

isn't that going to be a complete disaster performance wise

elder swallow Mar 9, 2024, 3:26 PM

#

no

prisma locust Mar 9, 2024, 3:27 PM

#

i haven't actually read 24w09a in detail yet

elder swallow Mar 9, 2024, 3:27 PM

#

potentially it's just allocating a single Integer every time the energy level in an item changes

#

which is completely negligible

prisma locust Mar 9, 2024, 3:27 PM

#

this does not spark joy

elder swallow Mar 9, 2024, 3:27 PM

#

it really doesn't matter

prisma locust Mar 9, 2024, 3:28 PM

#

for that use case no

blazing valve Mar 9, 2024, 3:28 PM

#

It will be less nice for AE2 portable cells

prisma locust Mar 9, 2024, 3:28 PM

#

but in general you're telling me every component has to be copied on write?

elder swallow Mar 9, 2024, 3:28 PM

#

prisma locust but in general you're telling me every component has to be copied on write?

yes

blazing valve Mar 9, 2024, 3:28 PM

#

OTOH those changes would only happen on user interaction

elder swallow Mar 9, 2024, 3:28 PM

#

it's actually quite cool, it makes stack copies very cheap

prisma locust Mar 9, 2024, 3:28 PM

#

oh because you just shallow copy components

#

true

#

hmm

elder swallow Mar 9, 2024, 3:28 PM

#

because they only copy the component map lazily

#

well, lazy copy + shallow copy yes

blazing valve Mar 9, 2024, 3:29 PM

#

yeah and those item stack copies happen... a lot

prisma locust Mar 9, 2024, 3:29 PM

#

yeah I didn't consider the savings there

elder swallow Mar 9, 2024, 3:29 PM

#

it also makes access to the data much cheaper than NBT access

blazing valve Mar 9, 2024, 3:29 PM

#

did you already port fluidstack over to components or is that something for after the minimal port?

elder swallow Mar 9, 2024, 3:29 PM

#

I kinda want to do it in the port

blazing valve Mar 9, 2024, 3:29 PM

#

Aight. BTW. are components guaranteed to have equals/hashCode?

#

Or how do they handle it?

elder swallow Mar 9, 2024, 3:30 PM

#

they should implement equals

#

idk about hashcode

blazing valve Mar 9, 2024, 3:30 PM

#

Okay mods will 100% fuck that up 😄

#

But they'll probably notice if their stuff doesn't stack at all

#

I wonder if we should check in dev that two "default" instances of a component are equal (if there is such a concept, but since they are serializable I'd assume so)

elder swallow Mar 9, 2024, 3:31 PM

#

yeah they should also implement hashcode

#

there is no concept of a default instance for arbitrary components

prisma locust Mar 9, 2024, 3:31 PM

#

in theory one could use reflection to check for that

#

in dev

blazing valve Mar 9, 2024, 3:31 PM

#

But we can copy components... how? Serialization / Deserialization?

elder swallow Mar 9, 2024, 3:31 PM

#

they are immutable

#

you just make a shallow copy of the backing map

blazing valve Mar 9, 2024, 3:32 PM

#

Ah, so no copy to begin with 🤔

elder swallow Mar 9, 2024, 3:32 PM

#

indeed

#

serialization/deserialization happens with codecs

blazing valve Mar 9, 2024, 3:32 PM

#

But they still have serialization/deserialization requirements, so we could make a copy

elder swallow Mar 9, 2024, 3:32 PM

#

there is the concept of a "default" component set that is provided by the Item

blazing valve Mar 9, 2024, 3:32 PM

#

I am just spitballing how we could avoid the footguns

#

or rather, help modders avoid them

#

Didn't Guava have some testing util to test equals/hashCode or am I dreaming that up

elder swallow Mar 9, 2024, 3:34 PM

#

some reflection could work

#

but it sucks 😄

#

could probably check the components directly when they get added to a stack

frail hatch Mar 9, 2024, 3:35 PM

#

elder swallow could probably check the components directly when they get added to a stack

that sounds bad if you have to add to the stack to replace instances

elder swallow Mar 9, 2024, 3:35 PM

#

for example here you can check if the T param implements equals and hashcode

blazing valve Mar 9, 2024, 3:36 PM

#

Well whatever check we add for dev it should not completely kill the perf in dev either 😄

#

but it'd certainly be useful to come into this version with some constraints so we don't get a complete buggy mess of a version

elder swallow Mar 9, 2024, 3:37 PM

#

can only check once per class and in-dev, yeah

#

BTW I introduced a bug in the port

#

FluidStack#equals does not check for the amount lol

#

so it is not even suitable as a component as is

prisma locust Mar 9, 2024, 3:38 PM

#

on the other hand

#

if you don't implement equals and hashCode properly will your component even work

elder swallow Mar 9, 2024, 3:40 PM

#

well, unless you're AE2 or a similar mod I'm not sure you'll care about that

blazing valve Mar 9, 2024, 3:40 PM

#

prisma locust if you don't implement equals and hashCode properly will your component even wor...

Stacking will break for sure

elder swallow Mar 9, 2024, 3:40 PM

#

what if you only forget hashCode

blazing valve Mar 9, 2024, 3:40 PM

#

Isn't there an ItemStack set in Creative Tabs?

#

It would certainly break net.neoforged.neoforge.common.util.ItemStackMap 😄

#

But in real subtle and hard to debug ways...

prisma locust Mar 9, 2024, 3:42 PM

#

blazing valve It would certainly break `net.neoforged.neoforge.common.util.ItemStackMap` 😄

wait that's a thing?

elder swallow Mar 9, 2024, 3:42 PM

#

I just saw some COMPONENT_INTERNER.intern(this.components.build())

prisma locust Mar 9, 2024, 3:42 PM

#

elder swallow what if you only forget hashCode

that one we can check trivially with your idea of instantiating some default components and checking the return values of those methods

#

if(a.equals(b) && (a.hashCode() != b.hashCode())) throw new YouJustLostTheGame();

elder swallow Mar 9, 2024, 3:42 PM

#

there is no default to instantiate

blazing valve Mar 9, 2024, 3:42 PM

#

It would be nice if we could determine if a component class is immutable or not via reflection

#

And if someone really wants to make their component mutable, force them to implement a marker interface or some-such

#

At the very least mods like AE2 could still optimize all cases where immutable components are used then

prisma locust Mar 9, 2024, 3:44 PM

#

if all vanilla components are immutable we should probably try to enforce the same

elder swallow Mar 9, 2024, 3:46 PM

#

we should yes

#

my god

#

btw I'm redoing fluid stack entirely and I'll send a PR with that

prisma locust Mar 9, 2024, 3:58 PM

#

elder swallow my god

why wouldn't they just... implement ItemStack#hashCode....

elder swallow Mar 9, 2024, 3:58 PM

#

well cause you generally don't want the count in there

prisma locust Mar 9, 2024, 3:58 PM

#

oh yeah

#

so, count is the last mutable field on an itemstack?

#

(mutable without some amount of copying, at least)

elder swallow Mar 9, 2024, 4:00 PM

#

😛

prisma locust Mar 9, 2024, 4:00 PM

#

popTime

#

it's literally used for the animation in the GUI and nothing else from what I can tell

elder swallow Mar 9, 2024, 4:03 PM

#

wtf is this now

prisma locust Mar 9, 2024, 4:04 PM

#

concern

#

what executor are they running the accept task on

elder swallow Mar 9, 2024, 4:06 PM

#

mainThreadExecutor

#

could be worse

prisma locust Mar 9, 2024, 4:08 PM

#

elder swallow `mainThreadExecutor`

smells like a potential issue for singleplayer

#

can't wait for someone to do async observer lines from 1.12 but with player heads

blazing valve Mar 9, 2024, 4:09 PM

#

I mean, the skull block entity seems to do some real cursed shit too

prisma locust Mar 9, 2024, 4:10 PM

#

indeed

elder swallow Mar 9, 2024, 4:28 PM

#

oh the 24w10a port wasn't finished...

elder swallow Mar 9, 2024, 4:51 PM

#

https://github.com/neoforged/Kits/pull/4

#

comments welcome

#

here is the full diff for those who can't see the PR: https://gist.github.com/Technici4n/72f3191818ca3710b11a19d5fcb3fa10

Gist

FluidStack update.diff

GitHub Gist: instantly share code, notes, and snippets.

digital drum Mar 9, 2024, 5:02 PM

#

elder swallow here is the full diff for those who can't see the PR: https://gist.github.com/Te...

LGTM

elder swallow Mar 9, 2024, 5:02 PM

#

it is very close to ItemStack with count -> amount and item -> fluid

digital drum Mar 9, 2024, 5:03 PM

#

elder swallow it is very close to `ItemStack` with `count` -> `amount` and `item` -> `fluid`

imo count is much better than amount

#

optically

elder swallow Mar 9, 2024, 5:03 PM

#

I would also prefer if we changed FluidStack amount to count...

#

because having two names for the same thing is annoying and pointless

digital drum Mar 9, 2024, 5:03 PM

#

elder swallow because having two names for the same thing is annoying and pointless

agreed

elder swallow Mar 9, 2024, 5:04 PM

#

and if we want a shared Stack abstraction for transfer later it would be very helpful to have the same name already

digital drum Mar 9, 2024, 5:04 PM

#

elder swallow and if we want a shared `Stack` abstraction for transfer later it would be very ...

it would be IStack if we were to, right?

elder swallow Mar 9, 2024, 5:05 PM

#

maybe, not sure about the name

#

potentially IResourceStack or whatever

digital drum Mar 9, 2024, 5:05 PM

#

actually this sounds nice, because mods like mekanism can simplify their ChemicalStack impls

elder swallow Mar 9, 2024, 5:06 PM

#

that too yeah

#

the question is how much value it really adds to the entire ecosystem if the only mod that benefits from this is mekanism

#

since potentially we could patch getAmount on ItemStack if necessary

digital drum Mar 9, 2024, 5:07 PM

#

elder swallow the question is how much value it really adds to the entire ecosystem if the onl...

it only affects a certain demographic of mods, ones which really emphasise adding new resources like the previously mentioned ChemicalStack from mek

#

not sure how many of those to argue for this to be added

elder swallow Mar 9, 2024, 5:33 PM

#

I think that unifying fluid and item transfer has a lot of value

digital drum Mar 9, 2024, 8:36 PM

#

elder swallow I think that unifying fluid and item transfer has a lot of value

I see the vision here 🔥

lime galleon Mar 9, 2024, 10:48 PM

#

so something like a fluid or item handler item, now has to copy the inventory on all modification because of immutability?

#

or the inventory has to be external and referred to by some instance-specific key

true jacinth Mar 9, 2024, 10:58 PM

#

Something of that sort

#

But the idea is that it makes ItemStack a cheap-to-copy type

#

since you can shallow copy all the Components

#

So with that the burden of copy-on-write for an inventory is much smaller

fallen igloo Mar 9, 2024, 11:29 PM

#

elder swallow oh the 24w10a port wasn't finished...

it wasn't? I did tho... (did you check the new branch?)

prisma locust Mar 9, 2024, 11:51 PM

#

fallen igloo it wasn't? I did tho... (did you check the new branch?)

the 24w10a branch gave me an error when I attempted to compile from source

stray bay Mar 10, 2024, 1:44 AM

#

true jacinth So with that the burden of copy-on-write for an inventory is much smaller

right, i hadn't considered that, you still need to copy a 27 item or however many array but you don't need to copy anything inside it

#

as long as we don't add anything mutable to itemstacks

#

just define a class called ItemStackItemStackHandler that takes care of the copies automatically

elder swallow Mar 10, 2024, 9:56 AM

#

prisma locust the 24w10a branch gave me an error when I attempted to compile from source

Same

elder swallow Mar 10, 2024, 10:47 AM

#

@rigid walrus do you also have compilation errors?

rigid walrus Mar 10, 2024, 10:49 AM

#

only a single compilation error that is exclusive to ecj

fallen igloo Mar 10, 2024, 10:59 AM

#

prisma locust the 24w10a branch gave me an error when I attempted to compile from source

I could run a full assemble without errors

elder swallow Mar 10, 2024, 11:01 AM

#

What the hell

#

Might have been trying to run the tests hmm

fallen igloo Mar 11, 2024, 10:51 AM

#

MC-268727 just got assigned

spiral radish Mar 11, 2024, 11:09 AM

#

Is that the race condition one?

boreal flint Mar 11, 2024, 11:09 AM

#

https://bugs.mojang.com/browse/MC-268727

fallen igloo Mar 11, 2024, 11:10 AM

#

spiral radish Is that the race condition one?

yes

elder swallow Mar 11, 2024, 2:59 PM

#

#

what the hell

#

https://github.com/neoforged/Kits/blob/13abb64ec60ae76086200f15551e8f499ed8e3e0/projects/base/src/main/java/net/minecraft/resources/RegistryDataLoader.java#L40

#

this is imported but doesn't exist

dreamy crystal Mar 11, 2024, 3:01 PM

#

Did it get deleted by accident? Or did it just... not get decompiled and added?

summer furnace Mar 11, 2024, 3:01 PM

#

it's definitely used there

#

stiff galleon Mar 11, 2024, 3:02 PM

#

gradle flub?

summer furnace Mar 11, 2024, 3:02 PM

#

also used in the Wolf class

#

so there's 0 chance it's an oversight by mojang

#

this is a decompilation / repo setup issue

elder swallow Mar 11, 2024, 3:03 PM

#

I think that schurli forgot to add new files under base and neoforge when he genned the sources

summer furnace Mar 11, 2024, 3:03 PM

#

forgetting git add seems likely

elder swallow Mar 11, 2024, 3:03 PM

#

he probably did git add

#

but he should have done git add -f

dreamy crystal Mar 11, 2024, 3:04 PM

#

gitignore only applies to untracked files IIRC, so yeah

summer furnace Mar 11, 2024, 3:04 PM

#

oh there's a gitignore on the folders we have files in? ew.

fallen igloo Mar 11, 2024, 3:04 PM

#

I did git add --force projects/base/src and git add --force projects/neoforge/src

dreamy crystal Mar 11, 2024, 3:04 PM

#

spooky

summer furnace Mar 11, 2024, 3:04 PM

#

do you have a WolfVariant java file?

elder swallow Mar 11, 2024, 3:04 PM

#

no new file in https://github.com/neoforged/Kits/commit/307446045b4debd7982eee7ab750c17a8d9a0049

fallen igloo Mar 11, 2024, 3:05 PM

#

wth git

dreamy crystal Mar 11, 2024, 3:06 PM

#

let me guess: it checked only the immediate children

elder swallow Mar 11, 2024, 3:07 PM

#

fallen igloo I did `git add --force projects/base/src` and `git add --force projects/neoforge...

well I just ran that and it is now adding stuff

#

fallen igloo Mar 11, 2024, 3:09 PM

#

dreamy crystal let me guess: it checked only the immediate children

no I just ran them and it added them (idk why it didn't work before)

#

I blame the intellij console

dreamy crystal Mar 11, 2024, 3:15 PM

#

rip

magic sierra Mar 11, 2024, 7:00 PM

#

MC-172289 fix & MC-92484 fix are still useful?

stiff galleon Mar 11, 2024, 7:09 PM

#

you did the thing where you didn't explain what you're talking about again

#

why do you keep doing that?

magic sierra Mar 11, 2024, 7:14 PM

#

the fix of this

#

and this

#

are still useful?

fallen igloo Mar 11, 2024, 7:14 PM

#

if it is not fixed by mojang then we are gonna leave the fix in

#

but since we are removing our attributes idk if the fixes are even still possible

magic sierra Mar 11, 2024, 7:20 PM

#

the fix: /attribute @s minecraft:player.entity_interaction_range base set 4.5
or
/attribute @s minecraft:player.block_interaction_range base set 3

elder swallow Mar 11, 2024, 7:20 PM

#

the fixes were removed if they were part of our massive GameRenderer reach patch

#

that's a behavior change though

restive raft Mar 12, 2024, 12:26 PM

#

magic sierra MC-172289 fix & MC-92484 fix are still useful?

in the future, please link to the bugs and include why you think they might not be useful (i.e., what prompted you to say that in the first place)

#

it can come off as quite random if you ask questions like that out of the blue

foggy steeple Mar 12, 2024, 1:54 PM

#

restive raft it can come off as quite random if you ask questions like that out of the blue

But why phone?

#1.20.5/6 Snapshots